Date: 20151102  
Docket: T-1156-12  
Citation: 2015 FC 1156  
Ottawa, Ontario, November 2, 2015  
PRESENT: The Honourable Mr. Justice Annis  
BETWEEN:  
GILEAD SCIENCES, INC AND GILEAD  
SCIENCES CANADA, INC  
Plaintiffs  
(Defendants to the Counterclaim)  
and  
IDENIX PHARMACEUTICALS, INC,  
UNIVERSITA DEGLI STUDI DI CAGLIARI,  
L'UNIVERSITÉ MONTPELLIER II AND  
CENTRE NATIONAL DE LA RECHERCHE  
SCIENTIFIQUE  
Defendants  
AND BETWEEN:  
IDENIX PHARMACEUTICALS, INC,  
Plaintiff to the Counterclaim  
and  
GILEAD PHARMASSET LLC, GILEAD  
SCIENCES, INC, AND GILEAD SCIENCES  
CANADA, INC  
Defendants to the Counterclaim  
Page: 2  
AND  
UNIVERSITA DEGLI STUDI DI CAGLIARI,  
L'UNIVERSITE MONTPELLIER II AND  
CENTRE NATIONAL DE LA RECHERCHE  
SCIENTIFIQUE  
Third Parties to the Counterclaim  
PUBLIC JUDGMENT AND REASONS  
(Confidential Judgment and Reasons issued October 9, 2015)  
I.  
Introduction  
[1]  
Gilead Sciences, Inc and Gilead Sciences Canada, Inc, (together with the Defendant by  
counterclaim Gilead Pharmasset LLC, hereinafter referred to collectively as [Gilead], seek a  
declaration that Canadian Patent No 2,490,191 [the ‘191 Patent] is invalid. They initiated the  
claim as interested persons under section 60 (1) of the Patent Act, RSC 1985, c P-4, as amended,  
s 27[the Act].  
[2]  
The Defendant, Idenix Pharmaceuticals Inc [Idenix] and the other defendants in the  
main action are the owners of the ‘191 Patent.  
[3]  
Gilead Sciences, Inc through its subsidiary, Gilead Pharmasset LLC, is the owner of  
Canadian Patent 2,527,657 [the ‘657 Patent] filed on April 21, 2004 and issued June 14, 2011.  
Page: 3  
The ‘657 Patent includes claims to a novel compound, “sofosbuvir”, for the treatment of  
Hepatitis C viral infections.  
[4]  
Gilead claims to be an interested person as having a reasonable basis to believe that the  
manufacture, use or sale of sofosbuvir would be impugned by the Defendants as an infringement  
of the ‘191 Patent. This has proved an accurate assumption, as evidenced in this matter by the  
counterclaim of Idenix.  
[5]  
Gilead Sciences Canada, Inc filed a New Drug Submission for sofosbuvir on or about  
May 21, 2013. Sofosbuvir was approved for sale in Canada through the issuance of a Notice of  
Compliance on December 13, 2013. Gilead Sciences Canada, Inc now sells sofosbuvir in Canada  
under the brand name SOVALDI.  
[6]  
SOVALDI is a “revolutionary” drug. It represents a significant breakthrough in the  
treatment of the Hepatitis C virus [HCV], which is said to infect more than 2.2% of the world’s  
population. It is an oral treatment that is highly effective with few side effects when compared  
with other treatments.  
[7]  
Gilead seeks to invalidate the ‘191 Patent on the grounds of insufficient disclosure, lack  
of demonstrated utility/sound prediction and overbreadth. It does so on the basis that Idenix has  
claimed a huge number of compounds and their use which lack sufficient disclosure, utility on  
any basis in the specification of the’ 191 Patent, or in light of the common general knowledge at  
Page: 4  
the relevant date. Idenix filed its patent application before it had made or tested any compound  
within the ‘191 Patent’s scope, despite more than 18 months of effort to do so.  
[8]  
Conversely, Idenix claims to be the inventor of the novel nucleoside inhibitor that  
provides the anti-viral activity in sofosbuvir. It advances the claim, in part, on the basis that the  
novel nucleoside was an analogue of one of its earlier inventions.  
[9]  
It denies the claim that its inventors did not soundly predict the utility of the novel  
nucleoside in sofosbuvir. It notes that the novel nucleoside is differentiated by only one  
substituent with similar stereochemical characteristics to that disclosed in Idenix’s earlier  
invention, which has demonstrated antiviral activity.  
[10] Idenix also denies that the ‘191 Patent does not disclose how to make the novel  
nucleoside. It claims that the synthesis of the compound was not novel. Based upon the  
disclosure in the ‘191 Patent, it could be made by the chemist of ordinary skill in the art, having  
recourse to the common general knowledge and routine experimentation. It claims the named  
inventor of the ‘657 Patent synthesized the novel compound with little difficulty, whereas the  
Idenix chemist assigned the task, similarly made the compound, but without knowing so because  
he was not acting like a skilled discovery chemist, but rather a process chemist.  
[11] Idenix counterclaims that the novel nucleoside in sofosbuvir was anticipated by the ‘191  
Patent. It seeks a declaration that the ‘657 Patent is invalid and that Gilead has infringed certain  
claims of their patent, along with other remedies of a permanent injunction, damages etc. Idenix  
Page: 5  
has no product, either on the market or pending authorisation by Health Canada, which is  
covered by the ‘191 Patent.  
[12] Idenix argues in the alternative, that to the extent that it does not have priority of a valid  
invention, the ‘657 Patent is invalid under section 53 of the Act for knowingly omitting naming  
Lieven Stuyver as an inventor; a misrepresentation that it claims was made willfully for the  
purpose of misleading.  
[13] After careful consideration of the evidence and submissions of the parties, for the reasons  
that follow, I allow Gilead’s claim. I declare the ‘191 Patent to be invalid and dismiss Idenix’s  
counterclaim in its entirety.  
II.  
A.  
Background Facts  
Idenix and its Patents  
[14] Idenix is a pharmaceutical company founded in the late 1990s. It was interested in  
discovering, developing and commercializing innovative therapeutics in areas of unmet medical  
needs.  
[15] Idenix had collaboration agreements with the Universita Degli Studi Di Cagliari  
[Cagliari], L’Université Montpellier II [Montpellier], and Centre National de la Recherche  
Scientifique [CNRS]. In particular, Idenix established with Montpellier and CNRS a  
Page: 6  
collaborative chemistry laboratory located in Montpellier, France that was overseen by Dr Gilles  
Gosselin (an employee of CNRS) and by Dr Richard Storer, Idenix’s Executive Director of  
Chemistry.  
[16] Idenix also established with Cagliari a collaborative biological testing laboratory located  
in Cagliari, Italy that was overseen by Professor Paolo La Colla of Cagliari.  
[17] On November 29, 2001, WO 01/90121 [US ‘121] was published disclosing the D-ribose  
2’-Methyl (up), 2’-Hydroyl (down) [hereinafter referred to as “2’-C-Me/OH”] nucleoside  
structures and derivatives for antiviral activity against HCV. On December 6, 2001, WO  
01/92282 [US ‘282] was published similarly disclosing the 2’-C-Me/OH structures for antiviral  
activity against flaviviruses and pestiviruses. These compounds were the precursors that led to  
the development of the claimed novel invention, which claims coverage for D-ribose 2’-Methyl  
(up), 2’- Fluorine (down) [hereinafter referred to as “2’-C-Me/F”] nucleoside structures and their  
derivatives.  
[18] On June 28, 2002, Idenix filed US application 60/392,350 [US ‘350] and US Patent  
Application 60/392351 [US ‘351], both of which are referred to in the ‘191 Patent. US ‘350  
contained, among its extensive claimed compounds, 2’-C-Me/F compounds. Conversely, the US  
‘351 omitted reference to any 2’-C-Me/F compounds. US ‘351 is no longer relied upon by Idenix  
as a priority application for the ‘191 Patent.  
Page: 7  
[19] On 31 March 2003, it was announced that Novartis was to acquire a majority stake in  
Idenix, together including the right to jointly develop its 2’-C-Me/OH nucleoside candidate  
NM283 to treat HCV.  
[20] On April 28, 2003, Idenix filed the US application 60/466,194 [US ‘194]. On May 14,  
2003, Idenix filed US application 60/470,949 [US ‘949]. Both are priority applications for the  
‘191 Patent.  
[21] On June 27, 2003, Idenix filed a Patent Cooperation Treaty [PCT] Patent Application  
for its Canadian Patent ‘191, naming the academic institutions as co-proprietors of the Patent.  
They claimed several broad genera of nucleoside alleged analogues of the 2’-C-Me/OH  
nucleoside. However, of its 49 claims, 32 of the claimed analogues were for 2’-C-Me/F  
compounds. The application was based on priority documents: US ‘350, US ‘351 (no longer  
relied upon) US ‘194 and US ‘949.  
[22] On or about September 22, 2008, the Canadian Intellectual Property Office [CIPO]  
issued a Requisition objecting to the ‘191 Patent Application because of; inter alia, a lack of  
unity of invention. Regarding the various genera claimed by Idenix, the examiner identified 15  
different classes of compounds. The examiner requested that Idenix restrict the claims of the  
patent application to one class of compounds.  
[23] On March 23, 2009, Idenix submitted amended claims to the CIPO deleting 17 of the  
claims and narrowing the scope to one class of “Formula IX” compounds, consisting of 32  
Page: 8  
claims of the 2’-C-Me/F genus. These became the claims to the ‘191 Patent as finally issued on  
August 3, 2010 and which are similar to the claims of compounds covered by the ‘657 Patent of  
Gilead.  
[24] In amending its claims, Idenix did not remove the extensive information relevant to the  
2’-C-Me/OH nucleosides that are no longer relevant to the 2’-C-Me/F nucleosides. As a result,  
the ‘191 Patent as issued on its face is a highly confusing document. In addition, the ‘191 Patent  
discloses no information with respect to the fluorination step to create the 2’-C-Me/F nucleoside,  
as Idenix never succeeded in synthesizing the compound, to its knowledge at least, prior to filing  
its application.  
[25] During this litigation, Idenix amended its statement of defence “solely for Canada and for  
the purpose of simplifying the issues to be determined at trial in this proceeding in Canada,  
without making any admission”, to state that “Idenix will not be defending nor asserting a claim  
of infringement in respect of Claim 1 of the ‘191 Patent.” The effect of this amendment is a  
matter of contention in this dispute.  
B.  
Initial Development of the 2’-C-Me/OH Compounds by Idenix  
[26] Up to late 2004, the Montpellier site conducted all of Idenix’s discovery chemistry  
efforts, while the Cagliari site carried out most of the biological testing (viral screening). Starting  
in 2004, Idenix’s facilities in Cambridge, Massachusetts also conducted process chemistry and  
development biology work.  
Page: 9  
[27] In the late 1960’s scientists developed a class of 2’-C-Me/OH nucleosides. In the early  
2000s, Idenix discovered that some known nucleoside analogues, which have a 2’-C-Me/OH  
structure, had activity in in vitro [performed in the laboratory] assays of certain Flaviviridae  
viruses. Idenix filed patent applications US ‘121 and US ‘282 in respect of these compounds.  
One of these compounds (known as NM283) was a prodrug of a nucleoside analog having a 2’-  
C-Me/OH sugar ring and cytosine base.  
[28] Idenix (then Novirio) held a chemistry retreat in Maui in December 2001. Dr Gosselin,  
Professor La Colla, Dr Sommadossi, Dr Standring and Dr Storer attended, along with the  
members of its chemistry discovery team. Among other subjects discussed at that meeting, the  
Idenix chemists identified several nucleoside analogues of the 2’-C-Me/OH compounds that they  
would attempt to synthesize, one of which was the 2’-C-Me/F nucleoside.  
[29] On March 28, 2002, Idenix assigned the synthesis of the 2’-C-Me/F nucleosides to Dr  
Jean François Griffon, a chemist located in Montpellier who had previous experience with  
fluorination of nucleosides. Dr Griffon and his team, which included Ms Audrey Chappe and Ms  
Elodie Pecheux, attempted to make 2’-C-Me/F nucleosides for two years without success.  
[30] Dr Alistair Stewart and Jingyang Wang, Idenix chemists located in Cambridge  
Massachusetts, continued to work on the synthesis of the 2’-C-Me/F compounds in late 2004-  
early 2005. They claim to have succeeded in or around January 2005 at a time when Idenix had  
received information from a former Pharmasset employee about its successful synthesis of the 2’-  
C-Me/F compound. Idenix clearly successfully synthesized the 2’-C-Me/F nucleoside in March  
Page: 10  
2005. Idenix did not test a 2’-C-Me/F nucleoside until March 2005, after the publication date of  
the Gilead ‘657 Patent.  
[31] Idenix contends that Dr Griffon was responsible for its inability to demonstrate that it had  
successfully synthesized the 2’-C-Me/F compound. It argues that he did not meet the standards of  
a skilled discovery chemist in many respects.  
[32] Idenix contends that he succeeded in synthesizing a 2’-C-Me/F nucleoside without  
knowing it. It argues that Albany Molecular Research Inc. [AMRI] synthesized the molecule in  
tests it carried out in 2014 using the same intermediate compound, fluorination reagent, Deoxo-  
Fluor®, and conditions followed by Dr Griffon during his February 2003 experiments.  
[33] Idenix argues that because Dr Griffon did not characterize the products of the reaction he  
failed to realize his success. It adds that because Dr Griffon misrepresented to his superiors that  
only one new compound was formed during the experiment, which was not the target compound,  
it then failed to pursue the most obvious and only known methodology that could successfully  
synthesize the 2’-C-Me/F nucleoside.  
C.  
Gilead Parties and their Patents  
[34] Gilead Sciences Inc is an American biopharmaceutical company that discovers, develops  
and commercializes innovative therapeutics in areas of unmet medical need.  
Page: 11  
[35] Gilead Sciences Canada, Inc [Gilead Canada] is a wholly owned subsidiary of Gilead  
Sciences Inc Gilead Canada sells pharmaceuticals in Canada for use in the treatment of diseases  
such as HIV, cystic fibrosis and Hepatitis.  
[36] Pharmasset Inc, under different changing names and country of origin, was a small  
biopharmaceutical company formed in the late 1990s. Its mandate was to discover and develop  
new compounds for the treatment of HIV, Hepatitis B virus [HBV] and HCV.  
[37] On 30 May 2003, Pharmasset filed US Provisional Patent Application 60/474,368 [US  
‘368] for a 2’-C-Me/F compound, supported by replicon testing data showing anti-viral activity.  
[38] On 21 April 2004, Pharmasset filed the Pharmasset PCT application W0 2005/003147  
claiming priority from US ‘368. It was subsequently published on January 13, 2005 [W0 Clark].  
The US ‘368 application and the Pharmasset PCT disclose, inter alia, the synthesis of a 2’-C-  
Me/F cytosine compound and the activity of this compound against HCV in a replicon assay.  
This synthesis was subject matter of a paper by Clark et al, Design, Synthesis and Antiviral  
Activity of 2'-Deoxy-2'-fluoro-2'-C-methylcytidine, a Potent Inhibitor of Hepatitis C Virus  
Replication, J. Med. Chem., 48, 5504-5508 (2005) [the Clark Paper].  
[39] On June 14, 2011, after extensive testing and trials, the Canadian‘657 Patent was issued.  
Page: 12  
[40] In November 2011, Gilead Sciences Inc purchased Pharmasset. Following the acquisition,  
Pharmasset became known as Gilead Pharmasset LLC. As noted, Gilead Pharmasset LLC is the  
current owner of the ‘657 Patent.  
[41] Gilead filed its statement of claim on June 15, 2012, with Idenix filing its defence and  
counterclaim on November 19, 2012. The parties have since amended their pleadings on several  
occasions.  
D.  
Pharmasset Synthesizes the 2’-C-Me/F Nucleoside  
[42] During the 2003 period when Idenix was attempting to synthesize the 2’-C-Me/F  
compound, Pharmasset was engaged in a similar endeavor. According to Dr Patterson, one of the  
chemists working with Pharmasset at the time, he believed that Pharmasset set about  
synthesizing 2’-C-Me/F compounds after Jeremy Clark, a recently hired chemist, discovered that  
either Idenix’s US ‘121 or US ‘282 Patent omitted them from their coverage. Pharmasset  
chemists thereafter identified the patent with the omission in coverage as the “Idenix” Patent. Dr  
Stuyver claims that he found the omission of coverage for the 2’-C-Me/F compound in an Idenix  
patent which partly qualifies him a co-inventor.  
[43] Dr Patterson testified that, Mr Clark, having found the coverage omission in the Idenix  
patent, was assigned the task of attempting to synthesize the 2’-C-Me/F compound. After a few  
months of effort, in or around May 2003, Mr Clark succeeded in making the 2’-C-Me/F  
Page: 13  
nucleoside with a cytidine base (identified as PSI-6130). Pharmasset biologists, overseen by  
Lieven Stuyver, tested PSI-6130 and found it to have anti-HCV activity.  
[44] On or about May 30, 2003, following the testing of PSI-6130, Pharmasset filed the US  
‘368 Application in respect of 2’-C-Me/F nucleosides. The US ‘368 application included no  
claims. It named Jeremy Clark and Lieven Stuyver as the inventors.  
E.  
Dr Stuyver Contests his Removal from the Pharmasset Patent  
[45] On April 21, 2004, Pharmasset filed the PCT application for the ‘657 Patent naming only  
Jeremy Clark as the inventor. [REDACTED]  
[46] Dr Stuyver contends that he found the omission in the Idenix patent and that he gave Mr  
Clark the idea of synthesizing the 2’-C-Me/F compound based on his work in other patents  
involving fluorine at the 2’ (down) position. He claims that Pharmasset removed his name out of  
spite because he left to move back to Europe and that he signed the declaration under duress,  
applied by Dr Otto.  
Page: 14  
III.  
A.  
The Witnesses  
Expert Witnesses  
(1)  
Gilead’s Experts  
[47] Gilead called three experts to testify at trial: a nucleoside chemist (Dr Wnuk), a virologist  
(Dr Seeger) and a pharmaceutical scientist (Dr Krise).  
(a)  
Dr Stanislaw Wnuk  
[48] Dr Wnuk is a Professor of Chemistry at Florida International University, a position he has  
held since 2002. Dr Wnuk obtained a PhD in organic chemistry from Mickiewicz University in  
Poznan, Poland. From 1985 until 1986, he completed postdoctoral research at the University of  
Alberta, specializing in the synthesis of modified nucleoside analogues and novel fluorination  
reactions. Dr Wnuk has worked in the area of nucleoside chemistry, including fluorination of  
nucleosides, for thirty years. Dr Wnuk is an author of over 180 publications, 80% of which relate  
to nucleosides or nucleotides, and 30% of which relate to fluorination.  
[49] Dr Wnuk is qualified by the Court to provide expert opinion evidence in the following  
areas: organic chemistry, medicinal chemistry and nucleoside chemistry; synthesis of  
nucleosides, nucleotides, fluorothioethers, and analogues thereof; fluorination methodologies to  
synthesize fluorinated nucleosides; developing methods for the fluorination of thioethers,  
alcohols, sugars, nucleosides and nucleotides, including methods for the introduction of fluorine  
Page: 15  
at the 5’ and 2’ positions of purine and pyrimidine nucleosides; and analytical and purification  
methods used in the synthesis of organic compounds.  
(b)  
Dr Christoph Seeger  
[50] Dr Seeger has been a Professor of Virology in the Department of Microbiology at the Fox  
Chase Cancer Center since 1990. Dr Seeger obtained a PhD in Microbiology from the University  
of Basel in 1982, focusing on the study of retroviruses. Dr Seeger’s laboratory has produced cell  
lines that are used to study the lifecycle of Hepatitis B and Hepatitis C viruses and screen for  
antiviral compounds. Dr Seeger’s laboratory was the second in the United States and the third  
worldwide to publish on the HCV replicon assay and its use to evaluate and screen antiviral  
compounds. Dr Seeger’s work has been funded for over thirty years by the National Institues of  
Health and his publications have been cited over 3500 times. Dr Seeger has more than thirty  
years’ experience in research in the field of virology, including extensive experience studying the  
HCV.  
[51] Dr Seeger was qualified to provide expert opinion evidence in the following areas:  
molecular biology and virology with an emphasis on human pathogenic viruses including  
Hepatitis viruses (Hepatitis B virus, Hepatitus C virus) and relevant animal model systems  
belonging to the Hepadnaviridae and Flaviviridae family of viruses, assessment of usefulness of  
nucleosides and nucleotides for the treatment of infections caused by these viruses, model  
systems and cell lines for screening antiviral compounds, including enzyme assays, plaque  
assays, cell-based assays, replicon assays, and high-throughput screens for the discovery and  
Page: 16  
development of antiviral compounds, interpretation of results from in vitro assays and animal  
studies to assess the activity and toxicity of antiviral compounds.  
(c)  
Dr Jeffrey Krise  
[52] Dr Krise is an Associate Professor of Pharmaceutical Chemistry at the University of  
Kansas. His research focus is on the optimization of drug properties in order to maximize the  
drug’s therapeutic usefulness. Dr Krise completed his undergraduate degree in Pharmacy in 1993  
and obtained his PhD in Pharmaceutical Chemistry with Honours in 1998 at the University of  
Kansas under the tutelage of Valentino Stella, a foremost expert in prodrugs. Dr Krise  
synthesized, evaluated, and wrote papers on a novel prodrug approach that employed phosphates  
for improving water solubility barriers of drugs. Dr Krise was qualified to provide expert opinion  
evidence in the following areas: pharmaceutical sciences; pharmacokinetics (absorption,  
distribution, metabolism and elimination); drug formulation; prodrugs, including nucleoside and  
nucleotide prodrugs; metabolism of sofosbuvir; and drug discovery and development. Dr Krise  
was the only prodrug expert to testify in this case.  
(2)  
Idenix’s Experts  
[53] Idenix called three experts to testify at trial: a chemist (Dr Damha); a second chemist (Dr  
Barrett); and a virologist (Dr Lamarre).  
Page: 17  
(a)  
Dr Masad Damha  
[54] Dr Damha is Chair of the Chemistry Department and James McGill Professor of  
Chemistry at McGill University, previously serving as Professor of Chemistry and other  
positions in the Chemistry Department at McGill since 1992. He has been Director, Graduate  
Program, Department of Chemistry, 2000-2004, 2006-2008, and 2010. He holds senior positions  
in various societies including President, Oligonucleotide Therapeutics Society, Board of the  
International Society of Nucleosides, Nucleotides & Nucleic Acids; sits on editorial advisory  
boards; and is the recipient of various scientific awards and honours. He has published  
approximately 155 papers in peer-reviewed journals and book chapters, many of which include  
the synthesis of nucleoside analogues. He has made more than 80 invited presentations at  
national and international meetings in the area of bio-organic chemistry, including the synthesis  
of nucleosides, nucleotides and oligonucleotides. He was an organizer of the 2012 International  
Roundtable on Nucleosides, Nucleotides and Nucleic Acids in Montreal.  
[55] Dr Damha was qualified to provide expert opinion evidence in the following areas:  
medicinal chemistry and bio-organic chemistry, including the design, synthesis and analysis of  
nucleosides and nucleoside analogues for use as antiviral agents including via prodrug strategies.  
(b)  
Dr Anthony Barrett  
[56] Dr Barrett is the Glaxo Professor of Organic Chemistry and the Director of the Wolfson  
Centre for Organic Chemistry in Medical Science in the Department of Chemistry at Imperial  
College of Science, Technology and Medicine [IC] in London, England. He is also the Sir  
Page: 18  
Derek Barton Professor of Synthetic Chemistry and Head of the Synthesis Section in the  
Department of Chemistry at IC. He is a Fellow of the Royal Society and a fellow of the Academy  
of Medical Sciences. He trained 127 PhD graduate students and 192 postdoctoral research  
associates in connection with a variety of research topics in the organic chemistry field, including  
synthetic organic chemistry and medicinal chemistry. He has published 408 peer-reviewed  
articles relating to various aspects of organic chemistry. These include 29 publications on the  
synthesis of carbohydrates and nucleosides, which appear relevant to these proceedings. He is the  
recipient of prizes from various organizations such as the Royal Society of Chemistry and the  
American Chemical Society for his contributions to organic chemistry.  
[57] Dr Barrett was qualified to provide expert opinion evidence in the following areas:  
organic and medicinal chemistry, including the synthesis in nucleosides and carbohydrates,  
analytical chemistry techniques and procedures as used in synthetic chemistry, including the  
techniques used in the interpretation of the results received and the standard laboratory practices  
of synthetic chemists.  
(c)  
Dr Daniel Lamarre  
[58] Dr Lamarre is Professor of Biochemistry and a director of the Molecular Research  
Laboratory lmmunovirology at Université dé Montreal. He previously held a number of senior  
research-based positions at Boehringer Ingelheim, in Laval Québec, with a focus on HCV and  
the discovery of antiviral compounds. He has held numerous fellowships, awards, and research  
chairs. He supervised three separate teams of approximately 25 scientists in a screening for  
biological activity of molecules against HCV via inhibition of the NS3 protease, NS3 helicase,  
Page: 19  
and NS5B polymerase. He has been involved in several important discoveries, including the  
antiviral drug, ciluprevir and the discovery and development of other significant HCV/HIV drug  
candidates.  
[59] Dr Lamarre was qualified to provide expert opinion evidence in the following areas: in  
biochemistry with a particular expertise in virology, industrial drug discovery and development,  
and structure-based rational drug design. His expertise includes a specific focus in HCV and the  
Flaviviridae family of viruses.  
(3)  
Comments on the Witnesses  
[60] As between the nucleoside synthesis witnesses, Gilead submitted that Dr Wnuk is the  
only expert in this case who is qualified to give opinion evidence in the area of fluorine  
chemistry. Conversely, they argued that neither Dr Damha nor Dr Barrett is qualified by the  
Court to provide expert opinion evidence in relation to fluorine chemistry or the fluorination of  
nucleosides.  
[61] I would disagree that neither Dr Damha nor Dr Barrett is qualified to speak to the topic of  
chemical fluorination of nucleosides. It is apparent however, that Dr Wnuk focused his career on  
the synthesis of nucleosides, including an extensive amount of time dealing with the fluorination  
of nucleosides. I weigh this as a factor in the evidence of the contending experts that favours  
Gilead with respect to the fluorination issues.  
Page: 20  
[62] Gilead further submits that Dr Wnuk is the only expert who had personal experience with  
fluorination during the relevant period and is therefore qualified to express an opinion on this  
topic from first-hand experience in the 2003-4 period. This appears to be an accurate assessment.  
Dr Barrett acknowledged that the chemistry at issue in this case was not of interest to him at the  
relevant time. Similarly, Dr Damha did not have experience with the fluorination of nucleosides  
in the 2003-4 period.  
[63] This is a fair submission, as there is an advantage to having an expertise in fluorination of  
nucleosides at the relevant time. This is particularly germane in being able to describe the limits  
of common general knowledge on the subject matter and the nature of experimentation at the  
time, as opposed to providing opinions based on ex post facto reviews of written materials  
provided to the witnesses or obtained through searches conducted a decade after the events.  
[64] Gilead was also critical of both Dr Damha and Dr Barrett concerning the superior  
qualities they attributed to the skilled chemist. Dr Damha testified that the skilled person had  
problem-solving and reasoning skills that would not be uncreative. He attributed this to the fact  
that by the time they graduate these persons would have made a contribution to some aspect of  
the knowledge in the field.  
[65] I agree with Gilead’s submission that this view could lead to a greater attribution of  
knowledge and skills than the imaginary skilled chemist would possess. This requires restraint in  
some areas when relying on Dr Damha’s opinions about the skilled chemist’s achievements in  
the synthesis of new compounds.  
Page: 21  
[66] I find that Dr Barrett similarly overreaches in describing the basic skills of the person  
skilled in drug discovery, testifying that he or she was “brilliant at reactions, but not inventive”.  
The superior laboratory skills attributed to the person of ordinary skill in the art [POSITA/ skilled  
chemist] by Dr Barrett diminishes somewhat the weight I attribute to his opinion of the skilled  
person’s ability to synthesize compounds by trial and error experimentation. I also take this into  
account in tempering the high degree of obviousness that Dr Barrett portrays in synthesizing the  
2’-C-Me/F compound and the contrasting harsh criticisms he has of Dr Griffon’s work in his  
failed attempts to synthesize the target compound.  
[67] I make a few comments where I find that the witnesses are overreaching. Overall, I did  
not find that their cross-examination played much of a role in the trial. I preferred different  
witnesses on different issues, based on the content and depth of support for their opinions, rather  
than how they testified. I also tended to give more weight to the evidence of the witnesses who  
were involved in the events in 2003-4, when in conflict with the opinions of the experts.  
B.  
Fact Witnesses  
(1)  
Gilead’s Fact Witnesses  
[68] Gilead did not lead evidence from any fact witnesses in respect of the main claim, but  
relied on admissions and read-ins from the discovery process. I also gave leave to Gilead to file  
foreign testimony from a former Idenix chemist, Dr Alistair Stewart. Gilead called one fact  
witness (Dr Otto) in respect of Idenix’s counterclaim contesting inventorship.  
Page: 22  
(a)  
Dr Alistair Stewart  
[69] Dr Stewart was the Director of Chemistry, Manufacturing and Controls at Idenix in  
Cambridge, Massachusetts. Dr Stewart received a PhD in Organic Chemistry from the University  
of Oxford in 2003, having studied under Professor George Fleet. Dr Stewart joined Idenix in  
September 2003 as a Research Scientist 1 in the Process Chemistry group. Dr Stewart worked on  
a project to make 2’-C-Me/F nucleosides from mid-2004 to early 2005, and oversaw the work of  
Ms Jingyang Wang. Dr Stewart was not called to testify at trial. However, after hearing parties  
on a motion in writing, I gave leave to Gilead to file Dr Stewart’s witness statement and related  
cross-examination transcript from the U.K. Proceeding as evidence in this trial. I provide my  
reasons for doing so when considering Idenix’s efforts to synthesize the 2’-C-Me/F compounds  
and the related evidence of Ms Wang from the UK proceedings, which was admitted on consent.  
(b)  
Dr Michael J. Otto  
[70] Dr Otto was the Chief Scientific Officer at Pharmasset Inc from 1998 to 2012. His role  
was to direct the chemistry and biology program, coordinate research efforts, and run the  
development program for an HIV drug. He holds a PhD in Microbiology from the Medical  
College of Wisconsin. Dr Otto testified as to the discovery of PSI-6130, the filing of patent  
applications for PSI-6130, and his recollection of events in respect to allegations made by Dr  
Lieven Stuyver concerning Dr Stuyver’s alleged role in the discovery of PSI-6130 as one of its  
inventors.  
Page: 23  
(2)  
Idenix’s Fact Witnesses  
[71] Idenix called six fact witnesses to testify at trial (Drs Standring, Griffon, Patterson,  
Clemens, Stuyver, and Professor La Colla). The parties agreed to the admission of foreign  
testimony from a seventh witness, Ms Wang in lieu of viva voce evidence at trial.  
(a)  
Dr David Standring  
[72] Dr Standring holds a PhD in Bio-organic Chemistry from Harvard University. He was  
employed by Idenix Pharmaceuticals Inc. in Cambridge, MA in a variety of senior management  
positions from 2000 until his departure from the company in 2013. Dr Standring is not a listed  
inventor on the ‘191 Patent. He was not involved in the chemistry in respect of the compounds at  
issue.  
(b)  
Professor Paolo La Colla  
[73] Professor La Colla is a Professor of Microbiology at Cagliari. Professor La Colla acted as  
the Director of the Department of Biomedical Science and Technology at the University of  
Cagliari from 2002 to 2008. He oversaw the collaborative work done at the University of  
Cagliari between the University of Cagliari and Idenix Pharmaceuticals Inc in respect of testing  
anti-HIV, anti-HBV and anti-HCV compounds. Professor La Colla did not make any decisions in  
respect of which compounds to test and had no involvement with the synthesis of the  
compounds. Professor La Colla was the only named inventor of the ‘191 Patent (out of four  
named inventors) called as a witness at this trial.  
Page: 24  
(c)  
Dr Steven Patterson  
[74] Dr Patterson has a PhD in Organic Chemistry with a minor in Biochemistry from Georgia  
State University. He was employed at Pharmasset between February 2000 and June/July 2004.  
He became the head of the analytical chemistry group in or around February 2003. Dr Patterson  
was called to testify by Idenix and spoke about the idea to make 2’-fluoro (down)- 2’-methyl(up)  
compounds at Pharmasset and how Jeremy Clark was the first person to make such a compound.  
(d)  
Dr Jean François Griffon  
[75] Dr Griffon is a Senior Research Scientist at Idenix in Montpellier in the nucleoside  
analogues group. Dr Griffon obtained a PhD in Organic Chemistry in 1998 from Montpellier in  
the laboratory of Professor Jean-Louis Imbach under the supervision of Dr Gilles Gosselin. Dr  
Griffon synthesized a number of 2’- and 3’-fluoro substituted nucleosides during the course of  
his PhD Dr Griffon joined Idenix in 2001. He was assigned the synthesis of 2’-fluoro (down)-2’-  
methyl (up) nucleosides on March 28, 2002 and he worked on this project until mid-2004.  
(e)  
Dr Lieven Stuyver  
[76] Dr Stuyver received a PhD in Human Genetics in 1992. He was the head of the biology  
group at Pharmasset from 1998 to 2004. Dr Stuyver was named as an inventor on provisional US  
‘368, which is cited as the priority application to the ‘657 Patent, but was not named as an  
inventor on the PCT patent application that led to the Canadian ‘657 Patent, or the ‘657 Patent  
itself.  
Page: 25  
(f)  
Ms Jingyang Wang  
[77] Ms Jingyang Wang is a Principal Research Scientist at Idenix in Cambridge,  
Massachusetts. Ms Wang received a Bachelor of Science degree in chemistry in 1989 from  
Nankai University, Tianjin, China; a Master’s of Science in Organic Chemistry in 1995 from  
University of Manchester, United Kingdom; and a Master’s of Science degree in Organic  
Chemistry in 1998 from the University of Maine. Ms Wang joined Idenix Pharmaceuticals, Inc in  
2002 as Process Chemist and was promoted to the position of Research Scientist in May 2004.  
Ms Wang worked on a project to make 2’-fluoro(down)-2’-methyl(up) nucleosides from late  
2004 until early 2005 under the direction of Dr Alistair Stewart. The parties agreed to have Ms  
Wang’s witness statement and related cross-examination from the U.K. proceeding entered into  
evidence in this trial in lieu of viva voce testimony.  
(g)  
Dr Alexander Clemens  
[78] Dr Clemens is a process chemist who works for AMRI. Dr Clemens performed  
experiments requested by Idenix in June-July 2014, as well as a repeat of those experiments in  
August 2014 intended to demonstrate that Dr Griffon had synthesized a 2’-C-Me/F nucleoside,  
without realizing it. Dr Clemens was not involved in the development of the protocols for the  
experiments.  
Page: 26  
C.  
The Failure of Jeremy Clark to Testify  
[79] Neither party called Mr Clark to testify. Idenix submitted at one point that the Court  
should draw an inference against Gilead by its failure to call him. Dr Otto provided evidence of a  
strained relationship between Mr Clark and Pharmasset (now Gilead). He felt that he had not  
been appropriately compensated for his invention He sent letters to the company, communicated  
with the company’s lawyers and filed lawsuits against the company and against Dr Schinazi.  
While he is a former employee, he is not under Gilead’s control. He did not testify in the United  
Kingdom case. It is not a surprise that he did not testify in this matter.  
[80] Gilead for its part, suggests that Idenix should have called Mr Clark. Idenix conducted a  
discovery of Mr Clark and asked a full day of questions relating to his work. It also points out  
that Idenix made allegations in respect of Clark’s work, and therefore, Idenix assumed the burden  
of proof on such facts. Gilead asks the Court to infer that Mr Clark’s evidence would not have  
assisted Idenix. I do not see why Gilead would want to or needed to call Mr Clark because of  
these facts. Idenix had his evidence from the discovery of Mr Clark and could have called him. In  
addition, Dr Patterson described Mr Clark as someone with a somewhat idiosyncratic  
temperament. I attribute no adverse inference to either party for the failure to call Mr Clark.  
IV.  
Scientific Principles and their Relation to the Issues  
[81] The following is a brief introduction to the requisite chemistry and biology principles that  
inform this decision along with some discussion of their relation to the issues under  
Page: 27  
consideration. The scientific background information is excerpted from the expert reports and  
parties’ submissions.  
A.  
Chemical Notation  
[82] The majority of the chemistry at issue is organic chemistry. Organic chemistry is carbon-  
based chemistry. Each carbon atom in an organic molecule is most commonly capable of forming  
four bonds to other atoms. Carbon atoms can be characterized according to the number of other  
carbons they are attached to. A “primary carbon” is a carbon bonded to a single carbon atom.  
Carbons that are bonded to two or three carbon atoms can be described as secondary or tertiary  
carbons. When attached to alcohols, they denote a secondary or tertiary alcohol.  
B.  
Sugar Rings, Nucleosides, Nucleotides and Nucleic Acids  
[83] The ‘191 and ‘657 Patents both relate primarily to nucleosides and nucleotides that are  
modified at what is described as the 2’ carbon position of their sugar ring (also referred to as a  
carbohydrate ring or in this matter ribose). Nucleosides and nucleotides are compounds to which  
a heterocyclic base (also referred to as a “Base” or “Nucleobase” discussed below) is attached to  
a sugar ring.  
[84] Nucleosides and nucleotides are the building blocks of nucleic acids. Nucleic acid is a  
general term used to describe a polymer (a large molecule composed of many repeated subunits)  
of nucleotides linked together by bonds. Deoxyribonucleic acid [DNA] and ribonucleic acid  
Page: 28  
[RNA] are the primary nucleic acids in cells that serve to encode and carry out genetic  
information in living organisms.  
[85] The HCV genome consists of a single strand of positive-sense RNA. The modified  
nucleosides or nucleotides are designed to mimic the natural nucleosides that are incorporated by  
the Hepatitis C virus to form RNA. When this occurs, analogues nucleotides are useful in the  
treatment of HCV by their disruption of the replication process that is required for new viruses to  
be formed.  
(1)  
The Sugar Ring  
[86] The Sugar Ring can take the structure of a five membered ring containing an oxygen  
atom. The sugar rings found in RNA and DNA are “ribose” and “deoxyribose” respectively.  
They are identical except at the 2’ “down” position. The RNA Ribose contains a hydroxyl group  
at the 2’-C (down) position, while DNA Deoxyribose contains a hydrogen atom at that position.  
They may be depicted using a “Haworth projection” as follows:  
Page: 29  
[87] A Haworth projection is a common way of representing the three-dimensional perspective  
of sugar rings, nucleosides, nucleotides and prodrugs. The Haworth projection provides  
information about whether substituents are attached above [(up)] or below [(down)] the plane of  
the sugar ring. Hydrogen atoms on the sugar ring are not depicted. Carbon atoms are represented  
by either a vertex (a point where two or more straight lines meet), or a line without a carbon  
symbol. The relative positions of the carbons around the sugar ring are denoted using the prime  
(‘) symbol along with the number of the carbon.  
(2)  
Nucleosides  
[88] A “nucleoside” is a chemical compound that consists of a base and a sugar ring. The  
sugar ring and the base are connected to each other by a chemical bond, known as a “glycosidic  
bond.” “Glycosylation”, or sugar-base condensation, is the process of coupling the sugar and  
base. An example of a nucleoside is depicted below. The spatial arrangement of the components  
on the drawing is illustrated using a wedge and dash drawing with the following bonds: a straight  
line is a bond in the plane of the paper; a bold wedge is a bond coming out of the page (in the  
(up) position); a hatched line is a bond coming into the paper (in the (down) position).  
Page: 30  
[89] Both DNA and RNA use combinations of bases to perform their coding function. In  
DNA, these bases are thymine, adenine, cytosine and guanine. In RNA, thymine is replaced by  
uracil. Adenine and guanine are classified as “purines” and contain a double-fused ring.  
Cytosine, thymine, and uracil only contain one ring and are classified as “pyrimidines”. These  
bases are often represented by a single letter (A, C, G, T or U).  
[90] The numbering on the atoms in the bases follows standard convention as illustrated  
below:  
C.  
Stereochemistry  
[91] “Stereochemistry” involves the study of the relative three-dimensional spatial  
arrangement of atoms that form the structure of molecules and their manipulation. Stereoisomers  
are isomers that differ in spatial arrangement of atoms, rather than order of atomic connectivity.  
An important branch of stereochemistry is the study of chiral molecules. A chiral molecule exists  
in two different forms, which are mirror images of each other.  
[92] A chiral molecule and its mirror image are called “enantiomers” of each other.  
Page: 31  
[93] Enantiomers have identical chemical and physical properties such as boiling point,  
melting point, density and refractive index. However, they differ in how they interact with  
polarized light, and may have different biological properties. A mixture of enantiomers in equal  
proportions is referred to as a “racemic mixture.”  
[94] Diastereoisomers (diastereomers) are stereoisomers of a compound having two or more  
chiral centers that are not a mirror image of another stereoisomer of the same compound  
[95] The distinction between these terms may be seen from the following drawing  
[96] The process of applying a Grignard reagent in the schemes in the ‘191 Patent describing  
how to synthesize the 2’-C-Me/OH compounds would result in a racemic mixture of  
diastereomers of both the 2’-C-Me/OH and 2’-C-OH/Me compounds. These can be separated to  
obtain the desired 2’-C-OH/Me enantiomer as the intermediate used to synthesize the 2’-C-Me/F  
target compound.  
Page: 32  
D.  
Fluorination  
[97] The term “fluorination” refers to the addition of at least one fluorine atom to a compound.  
By January 2004, fluorination reactions were understood to generally proceed via two different  
mechanisms: “electrophilic fluorination” and “nucleophilic fluorination”. Different starting  
materials and fluorinating reagents were required for electrophilic fluorination reactions as  
compared to nucleophilic fluorination reactions.  
[98] Electrophilic fluorinating agents that were known as of January 8, 2004 include: F2;  
SelectFluor; N-flurobenzenesulfonimide; and ClO3F. Nucleophilic fluorinating agents that were  
known as of January 8, 2004 include: HF and HF-based reagents (e.g. HF-pyridine, HF-pyridine/  
AlF3 , anydrous HF, HF/Fe (AcAc)3; AgF and AgF-based reagents (AgF/NH4F); Et3N●3HF; KF  
and KF-based reagents (e.g. KHF2); DAST; Deoxo-Fluor®; tetrabutylammonium fluoride  
(TBAF); tris(dimethylamino)sulfonium difluorotrymethylsilicate (TASF); Bu4NH2F3;  
Bu4NHF2/Fe(AcAc)3; and perfluoro-1-butanesulfonyl fluoride (PBSF).  
[99] The 2’-C-Me/F compounds were first synthesized using the nucleophilic fluorination  
agent of DAST, and thereafter using a more stable version of it sold under the trademark of  
Deoxo-Fluor®.  
E.  
Nucleophilic substitution reactions  
[100] A “nucleophile” (nucleophilic fluorinating agents such as DAST) is an electron rich  
species that reacts with electron deficient species, termed “electrophiles,” to generate new  
Page: 33  
molecule. A reaction between a nucleophile and an electrophile is called a “nucleophilic  
substitution reaction.” In such reactions, the nucleophile reacts with the electrophile, while  
another moiety, termed a leaving group, is released.  
(1)  
Mechanisms of nucleophilic substitution reactions  
[101] Nucleophilic substitution reactions can proceed by two different reaction mechanisms.  
These are called SN1 and SN2 mechanisms, and have different stereochemical consequences.  
[102] A “SN2 mechanism” proceeds with complete inversion of stereochemistry and gives rise  
to a single product. This is illustrated in the scheme below. The starting material with a hydrogen  
atom above the plane and a hydroxyl group below the plane; the product has a hydrogen atom  
below the plane and a fluorine atom above the plane: an inversion.  
[103] In contrast, a “SN1 mechanism” gives a mixture of stereoisomers. This is illustrated in the  
scheme below:  
Page: 34  
[104] Whether a nucleophilic substitution reaction will occur via a SN1 or SN2 mechanism  
depends on a number of factors, such as the degree and type of substitution, the steric hindrance  
of the substrate, the strength of the nucleophile, the stability of the leaving group, and the solvent  
used, among others.  
[105] The synthesis of the 2’-C-Me/F nucleoside by Gilead was carried out on the 2’-C-OH/Me  
compound by a SN2 fluorination of the tertiary alcohol in that compound with inversion of the  
stereochemistry resulting to synthesize the 2’-C-Me/F compound.  
F.  
Protecting Groups  
[106] A nucleoside (and sugar) has multiple potential reaction sites. During a synthetic  
sequence, some of these reaction sites may need to be blocked from chemical reactions so that a  
selective chemical transformation can occur at the targeted reaction site. A “protecting group” is  
a functional group that is temporarily placed on a potentially reactive site of the nucleoside to  
block it from undergoing unwanted reactions. A protecting group can be removed once the  
desired transformation is achieved. Certain protecting groups are more compatible with certain  
reaction conditions than others.  
Page: 35  
G.  
Isolating Molecules  
[107] After performing reactions, chemists must isolate and purify the products of their reaction  
for analysis and identification. Most commonly, chemists use chromatography to separate and  
purify reactions.  
[108] Chromatography involves the separation of compounds based on different partitions  
between the mobile phase and the stationary phase. The mobile phase is usually a solvent or  
mixture of solvents and the stationary phase is usually an inorganic solid like silica or alumina.  
Chromatography generally proceeds by allowing a solution containing the mixture of products to  
flow through the stationary phase. Due to the differences in partition, the products of the mixture  
progress (elute) at different rates through the stationary phase and are collected separately and  
evaporated.  
[109] Three main forms of chromatography bear consideration for the purposes of this case:  
thin layer chromatography [TLC], column chromatography, and high-performance liquid  
chromatography [HPLC].  
(1)  
Thin Layer Chromatography  
[110] TLC is commonly used to monitor progress of a reaction. TLC provides a chemist with a  
simple, inexpensive, efficient and relatively fast determination to analyze a reaction mixture and  
to monitor the progress of a reaction and the extent of conversion of starting material into  
Page: 36  
product. It also provides an empirical and approximate measure of the purity of the product or  
products.  
[111] In TLC, usually a glass or alumina plate is covered with a stationary phase, usually silica  
or alumina. The solutions of interest are spotted on the plate at a set distance from the bottom.  
The reaction mixture is spotted on the plate, and the bottom of the plate is placed in a small  
amount of solvent or solvent mixture which migrates up the plate by capillary action. The  
compounds in the reaction mixture are dragged along with the mobile phase at different speeds  
depending on how avidly they interact with the silica or alumina. The compounds in the reaction  
mixture are separated based on differences in solubility, polarity and absorption of compounds.  
[112] While monitoring an organic synthesis, a chemist can place the starting material in the  
left-hand lane on a TLC plate, and the reaction mixture in the right hand lane. If both lanes show  
exactly the same spots on the TLC plate, then the chemist would usually assume that no reaction  
has occurred. If there was a different spot on the TLC, when compared to the original lane, this  
would imply that there is one reaction product. If there were multiple additional spots on the new  
lane, the inference would be that multiple reaction products were formed in the reaction.  
[113] A TLC plate may be employed after the separation of the reaction products to assist in  
determining whether it contains a sugar, such as a 2’-C-Me/F nucleoside. The TLC plate is  
stained with sulphuric acid, which preferentially charcoals carbohydrates, such as sugars. Spots  
containing sugars char darkly under UV light. Compounds that do not contain carbohydrates  
(such as a base that was not successfully coupled to a sugar) will not char as darkly.  
Page: 37  
(2)  
Purification by Chromatography  
[114] Various forms of chromatography are used to separate mixtures of different compounds  
for analytical purposes. These include: column chromatography, HPLC and reverse phase HPLC.  
[115] In column chromatography, the mobile phase – the reaction mixture – is poured on top of  
a stationary phase – a column that is typically made of silica gel. Additional solvent is  
continuously added and fractions are eluted from the base of the column. Different compounds  
from the reaction mixture flow through the stationary phase at different speeds. The reaction  
product is collected manually in vails that are combined into similar reaction products, which can  
be further analyzed.  
[116] HPLC is similar to column chromatography. It is a more sophisticated chromatographic  
technique in which the solution mobile phase is pumped under pressure through a finely divided  
stationary phase and the eluted solution monitored for composition change by a detector. HPLC  
is usually much more efficient in separations than column chromatography.  
H.  
Characterizing Molecules  
[117] Chromatography is used to separate reaction products, but not characterize them. Many  
methods of characterization of molecules exist. The two relevant methods in this matter are mass  
spectrometry [MS] and nuclear magnetic resonance [NMR].  
Page: 38  
(1)  
Mass Spectrometry [MS]  
[118] MS analysis provides a mass spectrum that may provide the molecular composition of the  
compound as well as providing some information on its structural features. MS can also be  
coupled with HPLC allowing for separation and characterization using one procedure.  
(2)  
Nuclear Magnetic Resonance [NMR]  
[119] NMR spectroscopy is a frequently employed technique that is used for the structural  
characterization of compounds. It determines either completely or partially, the structure of a  
compound by giving information on the types of groups of atoms present and their connectivities.  
[120] NMR can also be used to authenticate a sample of a compound by comparisons of the  
NMR spectra with those from an authentic sample. NMR spectroscopy is the primary method  
used by chemists to characterize unknown compounds and to authenticate other samples.  
(3)  
High-pressure Liquid Chromatography Attached to a Mass Spectrometer [LC/MS]  
[121] HPLC attached to a mass spectrometer is a more recent technology to separate and  
characterize reaction compounds. It operates by a small amount of reaction mixture being  
introduced into to an HPLC column. After leaving the column, the eleunt immediately passes  
into a mass spectrometer, allowing the chemist to monitor, in real time on an on-going basis, the  
molecular weights of the compounds leaving the column.  
Page: 39  
I.  
Nucleotides and the Formation of Ribonucleic acid [RNA] and Deoxyribonucleic acid  
[DNA]  
[122] Nucleosides are precursors of nucleotides. Nucleotides are nucleosides that contain a  
phosphate bonded to the oxygen atom of the alcohol (hydroxy) unit of the sugar. A nucleotide is  
defined as a nucleoside having a mono-, di-, or triphosphate group attached to the sugar ring at  
the 2’, 3’ and/or 5’ position, with the 5-carbon site the most common as is the situation in this  
matter. The nucleotide in sofosbuvir, i.e. a prodrug of a 2’-C-Me/F is a nucleotide with a  
monophosphate at the 5’ position.  
[123] An example of the three forms of a nucleotide is depicted below.  
[124] The conversion of a nucleoside into a nucleotide by the attachment of the phosphate (for  
5-  
example PO3 ) to the C-5 alcohol unit is a chemical process that can occur naturally in vivo  
(inside the cell). Their formation starts after the nucleoside or monophosphate nucleotide enters  
the cell. The entry into the cell is facilitated by the use of “prodrugs,” which are discussed below.  
Page: 40  
[125] Once inside the cell, the nucleoside, or a mono or di-phosphate nucleotide, undergo up to  
three consecutive phosphorylation steps at the hydroxyl group attached to the 5’-carbon to form  
5’-nucleotide mono-, di and triphosphates respectively. “Phosphorylation” is a metabolization  
process that may occur in vivo whereby a phosphate group is naturally added to a nucleoside or  
to a mono or di-phosphate nucleotide to become a nucleotide. It is catalyzed by enzymes called  
“kinases”.  
[126] A polymerase is an enzyme which can take the nucleotide building blocks and couple  
them together to make strands of DNA or RNA. For a polymerase to link one nucleotide to the  
next, it is necessary for the nucleotide to be in a triphosphate form (having three phosphate  
groups at the 5’ position of the sugar ring). For ribonucleosides, a hydroxyl group at the 3’  
position (the 3’ OH group) of one nucleotide will be coupled with a 5’ phosphate group at the  
next nucleotide to result in the phosphate bridge. By this process the triphosphate is converted  
back to a monophosphate, which forms the phosphate bridge to the next nucleotide.  
[127] In summary, as part of the DNA and RNA replication machinery of cells, a nucleoside  
triphosphate is the required starting material for forming a phosphate bridge to an adjoining  
nucleotide to form a growing strand of RNA or DNA. Kinase enzymes are responsible for  
phosphorylating nucleosides and mono and diphosphate nucleotides. Polymerases are enzymes  
that are responsible for synthesizing the growing strands of RNA or DNA.  
Page: 41  
J.  
Viruses  
[128] “Viruses” are the smallest of all self-replicating organisms. While self-replicating, they  
have no metabolism of their own, but rather are obliged to invade cells and direct subcellular  
machinery to produce more viruses. All viruses carry a genome composed of viral nucleic acid  
(either RNA or DNA) enveloped in a protein coat, called a capsid. Viruses infect living things  
and make use of their host’s cellular reproduction mechanisms to reproduce themselves. Like  
normal cells, nucleic acids are the genetic material of viruses. Viruses carry their own  
polymerase called viral polymerase. Viruses can be classified based upon whether they contain  
RNA or DNA as their genetic material.  
[129] Viruses are categorized into families. A family of viruses can include 10, 20 or 30  
viruses. The Flaviviridae family encompasses numerous viruses of significant global concern that  
affect both humans and animals.  
[130] The Flaviviridae virus family includes three types: flavivirus, hepacivirus and pestivirus.  
Some examples of these types include:  
Page: 42  
[131] Flaviviridae show similarity in their genome (their genetic code, which provides the  
blueprint for the proteins required to replicate the virus), which includes the viral RNA  
polymerase. All Flaviviridae viruses have viral RNA polymerase enzymes (flavivirus - NS5,  
pestivirus - NS5B, hepacivirus - NS5B) that replicate RNA in the same way by building RNA  
using nucleoside triphosphates.  
(1)  
Hepatitis C  
[132] HCV is one of the most important viruses of the Flaviviridae family because of its  
significant impact on human health. HCV is a disease of the liver. It is transmitted by blood-to-  
blood contact, including through contaminated blood and blood products. Once an individual is  
infected, HCV continuously replicates and spreads within hepatocytes (liver cells). About 185  
million people, roughly 2.2% of the human population, are infected with HCV, and virus  
transmission remains a significant public health concern.  
[133] The most notable feature of HCV infections is that they typically persist, often for  
decades, with more than 70% of cases with acute Hepatitis C progressing to chronic Hepatitis.  
Patients with chronic Hepatitis are predisposed to developing chronic active Hepatitis, cirrhosis  
of the liver, and hepatocellular carcinoma, all of which are responsible for hundreds of thousands  
of deaths each year. The death rate from HCV will continue to climb for at least 10 years,  
because of the decades-long lag time between acute infection and liver failure.  
Page: 43  
(2)  
The HCV genome  
[134] The HCV genome consists of a single strand of positive-sense RNA. The organization of  
the HCV genome is illustrated below:  
[135] The non-structural [NS] proteins that are encoded by this genome are involved in viral  
replication and protein synthesis. For example:  
(a)  
NS3 - a protease which helps process the polyprotein to generate mature proteins  
by cleaving it at specific target sequences; and  
(b)  
NS5B - a polymerase that is responsible for the multiplication of the viral  
genome. It is the enzyme that connects the nucleotides together to make a chain of viral  
RNA.  
Page: 44  
(3)  
HCV replication  
[136] An understanding of the HCV lifecycle is an important aspect of antiviral drug discovery  
since antiviral drugs often inhibit virus-specific functions that are essential for replication. The  
HCV lifecycle includes the following steps, as illustrated in the figure below:  
[137] The steps are as follows:  
(a) The HCV particle attaches itself to (Step 1), and enters, the host liver cell (i.e.  
hepatocyte) (Step 2);  
Page: 45  
(b) Once inside the cell, the outer shell of the HCV particle falls apart (uncoating) to  
expose the RNA strand that the virus carries and will use to make copies of itself  
(Steps 3 and 4);  
(c) The virus uses the host cell’s components to read the information present on the  
RNA strand and produce proteins such as NS5B (the viral protein responsible for  
making new copies of the HCV RNA strand, also called the HCV polymerase)  
(Step 5);  
(d) NS5B recognizes and binds to specific nucleotide triphosphates that are found in the  
cell, and incorporates them into the new strand of RNA (RNA replication) (Step 6).  
[138] To synthesize RNA (either the replicative negative strand or the positive strand viral  
RNA), nucleotides triphosphates are specifically required, which are used by the NS5/NS5B  
polymerase to incorporate ribonucleotides to a growing nascent chain based on the template  
RNA.  
K.  
Nucleoside analogues to treat viral infections  
[139] When developing nucleoside analogues to treat viral infections, cytotoxic effects must be  
avoided. The aim is to kill the virus, not the host cell. Thus, the requirements known in 2003 for  
direct acting nucleoside and nucleotide analogues for HCV infections should have the following  
properties:  
Page: 46  
be able to enter infected cells without harming them;  
be able to convert into nucleotide triphosphates;  
be recognized in their triphosphate form by the HCV NS5B polymerase protein;  
be incorporated into the new growing HCV RNA strand in place of a naturally-occurring  
nucleotide; and  
possess some property that, once incorporated into the new RNA strand, inhibits the  
complete replication of the growing HCV RNA strand.  
L.  
Evaluating HCV Treatments  
[140] The study of HCV and the development of anti-HCV therapies have been hindered by  
difficulties in developing appropriate model systems. This is because HCV is a species-selective  
virus that infects only humans and chimpanzees. Instead, various in vitro model systems were  
used in 2003. These are discussed below.  
(1)  
Phosphorylation assay  
[141] The phosphorylation assay is not used to measure direct antiviral activity. As set out  
above, nucleoside analogue inhibitors need to be converted to their triphosphate form in vivo to  
be recognized by a polymerase. The phosphorylation assay simply determines whether a  
nucleoside analogue inhibitor can be phosphorylated in vitro. This assay uses normal cells that  
Page: 47  
are not infected with a virus, and therefore, this assay must be done in conjunction with  
additional assays to fully test antiviral activity.  
(2)  
Polymerase assay  
[142] As discussed above, a nucleoside analogue used in the treatment of HCV should inhibit  
the activity of NS5B. The HCV polymerase assay may be used by the skilled virologist to assess  
anti-HCV activity (i.e. it can be used to determine the ability of different test compounds to  
inhibit the HCV polymerase’s activity).  
[143] However, like the phosphorylation assay, the polymerase assay is not ideal. First, it  
involves the large-scale production of soluble viral polymerase in bacteria or insect cells. Second,  
it can only be used if the test compound is in its tri-phosphorylated form. Third, it does not  
provide data about cytotoxicity. In 2003, a skilled virologist would have been aware of the  
drawbacks of the polymerase assay.  
(3)  
BVDV, (Bovine Viral Diarrhea Virus) surrogate model  
[144] In 2003, there was no HCV cell culture system or a convenient small animal model in  
which to screen potential anti-HCV compounds. Therefore, prior to the development and  
adoption of the HCV replicon assay, some scientists used an in vitro assay system based on the  
pestivirus BVDV as a surrogate model for testing nucleoside analogues for anti-HCV activity.  
Page: 48  
[145] Madin-Darby bovine kidney [MDBK] cells are first seeded onto a monolayer of cultured  
cells and then infected with BVDV, followed by the addition of serial dilutions of test  
compounds. After several rounds of viral replication, the initial infection gives rise to visible  
structures called plaques through the diffusion of the virus from the original site of infection to  
new sites. The number of plaques on each plate is compared to a control plate that was not tested  
with any test compound. If a compound inhibits BVDV RNA replication, it will reduce the  
number of plaques formed as compared to the control.  
(4)  
The replicon assay  
[146] The replicon assay was first reported in 1999 by Lohmann et al in the preeminent  
“Science” journal. A replicon is a portion of the HCV genome that is able to mimic HCV RNA  
replication in human hepatocytes. A replicon cell culture system can be used to directly measure  
the ability of a compound to prevent successful replication of the HSV RNA. After adding a test  
compound to the cell culture, the cells are incubated for a given period of time to allow for viral  
replication, and then the amount of replicon RNA in the cells is measured. By comparing the  
amount of replicon RNA in the cells that received the test compound to the amount of replicon  
RNA in control cells, it is possible to quantify the anti-HCV effect of the test compound.  
(5)  
Measures of antiviral activity and toxicity  
[147] In order to be an effective antiviral compound for HCV treatment, a nucleoside analogue  
must act selectively against the HCV virus, but also be non-toxic to HCV-infected or uninfected  
cells. A compound’s antiviral activity is expressed in the following ways.  
Page: 49  
[148] The “Effective Concentration 50%” [EC50] is a measure of a compound's antiviral  
activity, i.e. the amount of a given compound required to reduce virus titer or RNA or protein in  
virus infected cells by 50%. EC50 is often used interchangeably with a measure of Inhibition  
Concentration 50% [IC50], which refers to the concentration of a drug that is required for 50%  
inhibition of the RNA polymerase activity in vitro. By determining the amount of test compound  
required to reduce viral replication (i.e. plaque number) by 50%, the test compound's EC50 can  
be determined. The lower the EC 50 value, the more potent the compound.  
[149] The “Cytotoxic Concentration 50%” [CC 50] on the other hand, is a measure of a  
compound's cellular toxicity, and refers to the concentration of the compound required to reduce  
the number of cells by 50%. The larger the CC 50 value, the less toxic the compound.  
[150] The “Therapeutic index” is used to assess EC 50 values in relation to their CC 50 values  
through a ratio [CC 50:EC 50]. Since a high CC 50 value indicates low toxicity, and a low EC  
50 value indicates high anti-viral activity, a preferred compound would have a significantly  
higher CC 50 value than an EC 50 value. A higher therapeutic index is preferable over a lower  
one.  
[151] In order to assess a therapeutic index, the EC and CC values must be obtained for the  
same cell line as that in which the virus is tested. This is because different cell lines show  
different sensitivities to different nucleoside analogues.  
Page: 50  
M.  
Prodrugs  
[152] A “prodrug” is a biologically inactive derivative of a drug that upon administration to the  
human body is converted into its active form by some chemical or enzymatic pathway.  
Generally, a prodrug is a compound that metabolizes to the active compound at some point in  
vivo.  
[153] The negative charge of an unmasked nucleoside mono, di, or triphosphate, i.e. a  
nucleotide, presents a barrier to cellular uptake, thus preventing drugs from reaching their desired  
target. The negatively charged mono-, di, or triphosphate has difficulties passing through the cell  
membrane because the lipid bilayer of the cell membrane resists charged molecules. To  
overcome this issue, nucleotide prodrugs can be prepared to mask the negative charge on the  
phosphate groups to increase the passage through a cell’s membrane.  
[154] Once inside a cell, the prodrug components of the nucleoside/nucleotide are removed.  
This may involve the cleaving of the leaving group in one, or a number of steps, in the  
metabolization process. In the case of a nucleotide monophosphate prodrug, the nucleotide  
monophosphate is then inside the cell and, as it is already monophosphorylated, it is in an  
excellent position to be converted to its di and triphosphate forms.  
Page: 51  
V.  
A.  
The Person Skilled in the Art and Common General Knowledge  
The Skilled Person  
[155] Patent specifications are addressed to a hypothetical person possessing the ordinary skill  
and knowledge to which the patent relates and a mind willing to understand the specification  
(Apotex Inc v Sanofi-Synthelabo Canada Inc, 2008 SCC 61 para 25).  
[156] The Federal Court of Appeal has compared the notional person of ordinary skill in the art  
to the “reasonable person” in the context of negligence law. This “man in the Clapham omnibus  
of patent law” has been described in Beloit Canada Ltd v Valmet OY (1986), 8 CPR (3d) 289 at  
294 (FCA) as:  
The technician skilled in the art but having no scintilla of  
inventiveness or imagination; a paragon of deduction and dexterity,  
wholly devoid of intuition; a triumph of the left hemisphere over  
the right.  
[157] I also find apt the description of the skilled person set out in Valensi v British Radio  
Corporation [1973] RPC 337 at 377, as follows:  
The hypothetical addressee is not a person of exceptional skill and  
knowledge, that he is not to be expected to exercise any invention,  
nor any prolonged research, inquiry or experiment. He must,  
however be prepared to display a reasonable degree of skill in  
common knowledge of the art in making trials and correcting  
obvious errors in the specification as a means of correcting them  
can readily be found.  
Page: 52  
[158] The skilled person has the same capabilities whether the issue is construction of the  
patent, utility and sound prediction, sufficiency of disclosure, or overbreadth.  
[159] The ‘191 Patent is to be construed as of its publication date (January 8, 2004) from the  
perspective of the skilled person.  
[160] The parties are in agreement that the skilled person would comprise a team of persons  
having at least a PhD with about three years of practical work experience, or a Master’s degree  
with a corresponding increase in work experience.  
[161] There is also agreement for the most part, that the skilled team members would comprise:  
(a) a chemist with knowledge or experience in the chemical aspects of drug discovery  
including the synthesis of nucleosides and standard laboratory techniques for synthesising  
and characterising nucleoside compounds;  
(b) a biologist or virologist familiar with the biological aspects of drug discovery particularly  
as it relates to Flaviviridae infections and particularly HCV with experience conducting  
relevant in vitro assays, cell culture systems, and interpreting data from in vitro assays,  
cell culture systems and animal models, and in assessing the activity and toxicity of other  
models for biological activity and toxicity.  
Page: 53  
[162] Idenix’s experts contend that the patent does not address a skilled person with a  
background in pharmacology with knowledge and experience relevant to bioavailability,  
pharmacokinetics, drug delivery and metabolism. Their opinion is that the patent is addressed to  
skilled persons in the art of novel drug discovery comprised in the two groups described above,  
as opposed to those who work on the after-development of the drug, once invented, which is the  
field of the pharmacologist.  
[163] I agree with this submission and as a result, I give less weight to some of Dr Krise’s  
evidence, particularly concerning the meaning of “a leaving group” in the ‘191 Patent. I address  
this issue below.  
[164] I further accept Dr Wnuk’s opinion that the notional person skilled in the art may still  
need direction and would only be starting to get into independent work. This is corroborated by  
Dr Krise’s evidence that while recent graduates may have an in-depth knowledge of a very  
narrow field of study, their knowledge across the remainder of their field remains basic.  
[165] I also accept Dr Wnuk’s opinion that the skilled person would have little experience in  
fluorination synthesis. His opinion was that in 2004 only a small percentage of chemists were  
trained in this art. Dr Wnuk had the advantage of being the only expert to testify with personal  
experience in fluorination in 2004.  
[166] I also agree with Idenix’s experts that skilled persons performing the tasks of a discovery  
chemist must have an awareness of the requirement to characterize compounds resulting from  
Page: 54  
their synthesis experiments. This would require the skilled discovery chemist to employ  
reasonable means at his or her disposal to characterize the compounds synthesized. To the extent  
that there may be constraints on access to technology, these limitations are attributable to the  
organization responsible for drug discovery and not the competence of the skilled person.  
B.  
Common General Knowledge  
[167] Common general knowledge is “knowledge generally known by persons skilled in the  
relevant art at the relevant time”, Apotex v Sanofi Synthelabo Canada Inc, [2008] 3 SCR 265,  
2008 SCC 61 at para 37 (2) [Plavix (SCC)].  
[168] In Eli Lilly and Company v Apotex Inc, 2009 FC 991 at para 97 [Cefaclor], Madam  
Justice Gauthier, adopted with approval the comprehensive description of common general  
knowledge from General Tire & Rubber Co v Firestone Tyre & Rubber Co Ltd (1971), [1972]  
RPC 457 at 480 to 481 (UKCA) which I have summarized without citations as follows:  
The common general knowledge imputed to such an addressee  
must be carefully distinguished from what in patent law is regarded  
as public knowledge.  
Common general knowledge is derived from a common sense  
approach to the practical question of what would in fact be known  
to an appropriately skilled addressee - the sort of man, good at his  
job, that could be found in real life.  
Individual patent specifications and their contents do not normally  
form part of the relevant common general knowledge, though there  
may be some exceptions.  
As regards scientific papers generally,  
Page: 55  
•
it is not sufficient to prove common general knowledge that  
a particular disclosure is made in an article, or series of articles, in  
a scientific journal, no matter how wide the circulation of that  
journal may be, in the absence of any evidence that the disclosure  
is accepted generally by those who are engaged in the art to which  
the disclosure relates.  
•
A piece of particular knowledge as disclosed in a scientific  
paper does not become common general knowledge merely  
because it is widely read, and still less because it is widely  
circulated.  
•
Such a piece of knowledge only becomes general  
knowledge when it is generally regarded as a good basis for further  
action [as opposed to generally known and accepted without  
question] by the bulk of those who are engaged in the particular  
art; in other words, when it becomes part of their common stock of  
knowledge relating to the art.  
•
It is difficult to appreciate how the use of something which  
has in fact never been used in a particular art can ever be held to be  
common general knowledge in the art.  
[169] A party that asserts that a particular piece of information was part of the common general  
knowledge at the relevant time must prove this fact with evidence.  
[170] Idenix relies on the ability of the skilled person to locate the common general knowledge  
by the use of search engines commonly available in the chemical discovery and education world.  
It makes reference to Justice Gauthier’s decision in Cefaclor from para 104 as follows:  
The distinction between common general knowledge and prior art  
which is part of the state of the art for the purpose of assessing  
anticipation and obviousness tends to diminish in modern times  
because of the sophistication of search engines and the availability  
of electronic publications and databases.  
Page: 56  
[171] This does not mean however, that scientific papers found by searches or otherwise  
produced at trial can be used to buttress disclosure deficiencies in the ‘191 Patent unless  
established as meeting the requirements of common general knowledge.  
[172] In addition to expert evidence, the Court has heard from several individuals who were  
actually working on the synthesis of 2’-C-Me/F compounds during the relevant period. I agree  
with Gilead’s submission that where the experience of individuals actually working on the  
problem differs from the theoretical opinions of experts, the contemporaneous factual  
information should be preferred.  
VI.  
A.  
Construction of the ‘191 Patent  
Principles of Construction  
[173] The claims of the patent are to be construed prior to any assessment of validity or  
infringement. It is a fundamental rule that “the claims receive one and the same interpretation for  
all purposes” Whirlpool Corp v Camco Inc, 2000 SCC 67, [2000] 2 SCR 1067 at para 49  
[Whirlpool].  
[174] The modern principles of claim construction were developed by the Supreme Court of  
Canada in the companion cases of FreeWorld Trust v Électro Santé Inc, [2000] 2 SCR 1024 at  
para 13, 2000 SCC 66 [Freeworld] and Whirlpool. These principles of claim construction were  
Page: 57  
later summarized by the Federal Court in Biovail Pharmaceuticals Inc v Canada (Minister of  
National Health & Welfare), 2005 FC 9 at para 15 as follows:  
1.  
A patent is construed as a bargain between the inventor and  
the public. In consideration of disclosing the invention, the  
inventor is given a temporary monopoly to exploit it.  
2.  
It is a statutory requirement that the patent contain a  
specification and end with a claim or claims "defining distinctly  
and in explicit terms the subject-matter of the invention for which  
an exclusive privilege or property is claimed". The specification  
must be sufficiently full, clear, concise and exact "as to enable any  
person skilled in the art or science to which it pertains, or to which  
it is most closely connected, to make, construct, compound or use  
it". (Patent Act, R.S.C. 1985, c. P-4, as amended, s. 27)  
3.  
The patent is notionally addressed to a person skilled in the  
art or science of the subject-matter and is to be read as such a  
person would have read it when it first became public…  
4.  
The claims are to be read in an informed and purposive way  
to permit fairness and predictability and to define the limits of the  
monopoly "[I]ngenuity of the patent lies not in the identification of  
the desired result but in teaching one particular means to achieve it.  
The claims cannot be stretched to allow the patentee to monopolize  
anything that achieves the desired result" (Free World Trust, paras  
31, 32).  
5.  
The claim portion of the patent specification takes  
precedence over the disclosure portion in the sense that the  
disclosure is read to understand what was meant by a word in the  
claims "but not to enlarge or contract the scope of the claim as  
written and thus understood" (Whirlpool, para 52).  
6.  
It is only such novel features that the inventor claims to be  
essential that constitute the "pith and marrow" of the claim. "The  
key to purposive construction is therefore the identification by the  
Court with the assistance of the skilled reader, of the particular  
words or phrases in the claims that describe what the inventor  
considered to be the "essential" elements of his invention"  
(Whirlpool, para 45).  
7.  
Some elements of the claimed invention are essential and  
others are not, based either on common knowledge when the patent  
Page: 58  
was published or according to the intent of the inventor, expressed  
or inferred from the claims…  
8.  
To overclaim is to lose everything. If the inventor  
underclaims, the court will not broaden the monopoly in the  
interests of the "spirit" thereof. This often, as in this case, results in  
layers of claims, each limitation serving as a potential safety net so  
that if the broadest claims fall, the monopoly may be saved in part  
by the more modest claims.  
9.  
Yet a patent is not an ordinary writing. It meets the  
definition of a "regulation" in the Interpretation Act, and must be  
read to assure the attainment of its objects. "Claims construction is  
a matter of law for the judge, and he was quite entitled to adopt a  
construction of the claims that differed from that put forward by  
the parties." (Whirlpool, para 52)  
[175] Claims are to be interpreted in a purposive manner in order to “achieve fairness and  
predictability and to define the limits of the monopoly” (Dimplex North America Ltd v CFM  
Corp, 2006 FC 586, 54 CPR (4th) 435 at para 49, aff’d 2007 FCA 278, 60 CPR (4th) 277  
[Dimplex]).  
[176] A patent is to be given a “purposive construction”, not a technical or literal construction,  
taking into account the entire context of the specification, including the disclosure and the  
claims, as they would be understood by the skilled person reading the patent with the object of  
understanding what the inventor claims to have invented and the scope of the monopoly the  
inventor wishes to claim (Freeworld paras 39-40, 44, 50).  
[177] Although a patent is to be given a purposive construction, the Supreme Court has held  
that competitors and the public are entitled to clear and definite rules as to the extent of the  
monopoly conferred. The discretionary elements of claims interpretation (e.g. the “spirit of the  
Page: 59  
invention”) must be kept to a minimum, consistent with giving the inventor protection for that  
which he has actually in good faith invented. Predictability is achieved by tying the patentee to  
its claims; fairness is achieved by interpreting those claims in an informed and purposive way  
(Freeworld para 43).  
B.  
Construction of Claims of the ‘191 Patent  
(1)  
Claim 1  
[178] Idenix is not defending or asserting a claim of infringement in respect of Claim 1. On that  
basis it only refers to Claim 1 for the purpose of defining the substituents of the other claims that  
are relevant to those claims.  
[179] Gilead argues that Idenix cannot resile from its representation to the CIPO that Claims 1-  
3 are one invention. Gilead claims that this was the implied representation made by Idenix when  
it was required to narrow its claims to describe a Patent Application for only one invention  
wherein Claims 1-3 represent the genus claims of the Patent. Gilead submits that the invention  
therefore, is described by Claims 1-3 and the derivatives thereof.  
[180] Gilead wishes to include Claim 1 as describing the invention particularly because the 2’  
(up) position of Claim 1 includes 45 different substituents. Claims 2 and 3 have this substituent  
as a methyl group, CH3. All but the methyl substituent have never been made, tested or their  
synthesis disclosed. At every turn in the case therefore, Gilead claims that the invention cannot  
be predicted, its synthesis has not been disclosed, or is overly broad.  
Page: 60  
[181] I reject Gilead’s submission, which is unsupported by any jurisprudence of similar facts.  
Nor do I understand that Idenix is prevented from abandoning the defence of a too broadly stated  
claim. The Supreme Court in Teva Canada Ltd v Pfizer Canada Inc, 2012 SCC 60, [2012] 3 SCR  
625 [Viagra] at para 80 sanctioned this practice as being acceptable:  
[80] I would not make too much of the fact that Claim 1  
included over 260 quintillion compounds. The practice of  
cascading claims — although it may, as in this case, result in  
claims that are overly broad — is a common one that does not  
necessarily interfere in every case with the public’s right to  
disclosure. The skilled reader knows that, when a patent contains  
cascading claims, the useful claim will usually be the one at the  
end concerning an individual compound. The compounds that do  
not work are simply deemed invalid. In accordance with s. 58, any  
valid claim — in this case, Claim 7 — survives despite the  
existence of invalid claims.  
[Emphasis added]  
[182] Therefore, Claim 1 in the ‘191 Patent is considered only for the purpose of the  
descriptions of its substituents as they apply to Claims 2 and 3. Claim 1 covers a large number of  
nucleoside analogues (or their pharmaceutically acceptable salts thereof) which are compounds  
of Formula (IX). It is depicted as follows:  
Page: 61  
(a)  
Substituents of Claim 1  
(i)  
Substituent “X” in the sugar ring component  
[183] This substituent can be oxygen (O), sulfur (S), SO2, or CH2. This encompasses both  
naturally occurring sugar rings (X = O or CH2) and modified sugar rings (X = S or SO2).  
However, in Claims 2 and 3, it is defined only as an oxygen atom.  
(ii)  
Substituent “C(Y3)3 ” at the 2’-up position in the sugar ring  
component  
[184] This substituent is a carbon [C] bonded to three atoms [Y3]. Each Y3 atom is either a  
hydrogen [H] or a halogen (F, Cl, Br, or I). The three Y3 atoms can be the same or different (e.g.  
CF3 , CF2Cl, etc.). The evidence of Dr Wnuk is that forty-four different C(Y3)3 groups are  
encompassed by this substituent in Claim 1, in addition to the methyl group C(H3)3. This  
substituent is a methyl group (CH3) in both Claims 2 and 3.  
(iii)  
Substituents for the “Base” at the 1’ position  
[185] The Base substituent is repeated in Claims 2 and 3 and therefore, is relevant to this  
litigation. It is one of a large number of different natural and non-natural purine and pyrimidine  
bases, that includes at least all of the bases listed in the definition of purine and pyrimidine set  
out on p. 104 of the ‘191 Patent, as follow:  
Page: 62  
[186] The term “purine” or “pyrimidine” base includes, but is not limited to, adenine, N6-  
alkylpurines, N6-acylpurines (wherein acyl is C(O)(alkyl, aryl, alkylaryl, or arylalkyl), N6-  
benzylpurine, N6-halopurine, N6-vinylpurine, N6-acetylenic purine, N6-acyl purine, N6-  
hydroxyalkyl purine, N6-alkylaminopurine, N6-thioalkyl purine, N2 -alkylpurines, N2 -alkyl-6-  
thiopurines, thymine, cytosine, 5-fluorocytosine, 5-methylcytosine, 6-azapyrimidine, including 6-  
azacytosine, 2- and/or 4-mercaptopyrmidine, uracil, 5-halouracil, including 5-fluorouracil, C5-  
alkylpyrimidines, C5-benzylpyrimidines, C5-halopyrimidines, C5-vinylpyrimidine, C5-acetylenic  
pyrimidine, C5-acyl pyrimidine, C5-hydroxyalkyl purine, C5-amidopyrimidine, C5-  
cyanopyrimidine, C5-iodopyrimidine, C6-iodo-pyrimidine, C5-Br-vinyl pyrimidine, C6 -Br-vinyl  
pyrimidine, C5-nitropyrimidine, C5-amino-pyrimidine, N2-alkylpurines, N2-alkyl-6-thiopurines,  
5-azacytidinyl, 5-azauracilyl, triazolopyridinyl, imidazolopyridinyl, pyrrolopyrimidinyl, and  
pyrazolopyrimidinyl. Purine bases include, but are not limited to, guanine, adenine,  
hypoxanthine, 2,6-diaminopurine, and 6-chloropurine. Functional oxygen and nitrogen groups on  
the base can be protected as necessary or desired. Suitable protecting groups were well known to  
those skilled in the art, and include trimethylsilyl, dimethylhexylsilyl, t-butyldimethylsilyl, and t-  
butyldiphenylsilyl, trityl, alkyl groups, and acyl groups such as acetyl and propionyl,  
methanesulfonyl, and p-toluenesulfonyl.  
[187] Idenix experts acknowledged that the bases covered by this definition may be linked to  
the sugar ring by either a nitrogen atom (N-linked) or a carbon atom (C-linked). Throughout the  
patent there are also references to schemes and compounds with non-natural and C-linked bases,  
none of which would work with the invention, and indeed would most likely be harmful if  
ingested by a human or other host.  
Page: 63  
[188] Dr Wnuk was of the opinion that the skilled person would have understood the Formulae  
of the ‘191 Patent to include a vast number of non-natural bases, thereby referring to an  
enormous number of nucleosides.  
[189] Dr Barrett stated his opinion that in the context of the ‘191 Patent, the POSITA would  
understand the term “Base” to encompass purine or pyrimidine bases that could be used in the  
treatment of Flaviviridae (viral) infections. The claimed class of compounds and their use as  
antiviral agents is the focus of the ‘191 Patent. He similarly opined that the skilled person reading  
the ‘191 Patent would understand that the novel modifications to the claimed compounds do not  
occur on the base, but rather occur at the 2’ position of the sugar ring. Thus, the person skilled in  
the art would read the term “Base” narrowly to refer to the natural bases.  
[190] This opinion is consistent with the conclusions of the Supreme Court in Burton Parsons  
Chemicals v Hewlett-Packard, [1976] 1 SCR 555 at page 563 [Burton]. It stands for the  
proposition that where it is obvious to the skilled person that within the specification there are  
substituents that are incompatible with the use of the invention, those compounds will not be  
selected.  
[191] Dr Damha testified to the same effect as Dr Barrett. He indicated that a nucleoside with a  
C-linked base would be unnatural. The person skilled in the art would understand N-linked bases  
to be the normal reference for natural nucleosides. Therefore to a skilled person, if one simply  
refers to a “base” or a “purine or pyrimidine base” this refers to the N-linked base. Unless  
specified, the normal understanding is that bases are N-linked bases.  
Page: 64  
[192] The terms “purine” and “pyrimidine” bases are specifically defined on page 104 of the  
‘191 Patent to include the four commonly known naturally-occurring bases used to make RNA:  
Adenine (A), Guanine (G), Cytosine (C), and Uracil (U)  
[193] While I understand that the definition of Base found in the Patent refers to a large number  
of natural and unnatural bases, I prefer the evidence of doctors Drs Barrett and Damha that the  
skilled person would understand that bases refer to those used in the treatment of Flaviviridae  
that are N-linked, thereby specifying the four commonly known naturally-occurring bases used to  
make RNA.  
(iv)  
Substituents “R1” and “R2” at the 5’ (up) and 3’ (down) position in  
the sugar ring component:  
[194] These substituents are one of a large number of different substituents that are specifically  
listed beneath Claim 1 of the ‘191 Patent. The large number of possible substituents can be  
broadly categorised as either an atom (e.g. H), a chemical group (e.g. a phosphate), or groups of  
substituents (e.g. CO-alkyl and lipid). This description of the R1 and R2 substituents also applies  
Page: 65  
to Claim 2. It does not apply to Claim 3 where the R1 and R2 substituents are narrowed to a  
hydroxyl.  
(2)  
“Phosphate” and “Pharmaceutically Acceptable Leaving Group”  
[195] There is a construction issue for Claim 2 in respect of the terms “phosphate” and  
“pharmaceutically acceptable leaving group”. The construction of these terms is determinative of  
portions of issues agreed upon by the parties regarding Idenix’s infringement claim. The relevant  
portion of the description of these substituents, in particular as they apply to the 5’ position on  
the sugar ring, is as follows:  
phosphate … or a pharmaceutically acceptable leaving group  
which when administered in vivo is capable of providing a  
compound wherein R1 is … phosphate  
[196] Gilead is seeking a narrow interpretation of this wording such that the prodrug  
sofosbuvir, which is a phosphoramidate at the 5 position, is not captured by this wording because  
it is neither a “phosphate” nor a “leaving group”. Gilead submits that “phosphate” can only be  
construed as a monophosphate – PO4. It further argues that a “leaving group” is interpreted by  
the skilled person as one that is cleaved off in its entirety, whereas the prodrug used in sofosbuvir  
and Sovaldi is cleaved off in stages.  
[197] I find that the construction of these terms is best left to the section on infringement where  
they have a practical application. I may state here however, that I interpret “phosphate” to  
include di- and tri-phosphates, while, “a pharmaceutically acceptable leaving group which when  
Page: 66  
administered in vivo is capable of providing a compound wherein R1 is … phosphate,” I  
conclude refers to prodrugs generally, without regard to how they are metabolized in vivo.  
(3)  
Claim 2  
[198] I have already pointed out that Claim 1 in the ‘191 Patent is relevant for the purpose of  
the descriptions of its substituents which apply to Claims 2 and 3. Claim 2 has the following  
structure:  
[199] Claim 2 is the compound of Claim 1, wherein hydrogen is the substituent in the 2’ (up)  
position on the sugar ring making it a methyl, while oxygen is the sole “X” substituent on the  
sugar ring making it a furanose. The other substituents remain the same. The Court’s conclusions  
on the construction of the terms “Base”, “phosphate” and “leaving group” made with respect to  
Claim 1, apply to Claim 2.  
Page: 67  
(4)  
Claim 3  
[200] Claim 3 is the same compound as Claim 2, except that R1 and R2 are H, meaning that the  
3ꞌ and 5ꞌ groups are hydroxyl (OH). The issues concerning the interpretation of “phosphate” and  
“leaving group” do not arise in this Claim. The Court’s conclusions on the construction of the  
term “Base” made with respect to Claim 1, apply to Claim 3.  
(5)  
Claims 4 to 32  
[201] Claims 4 to 32 incorporate “the compounds of claims 1 to 3”. It is understood that these  
Claims incorporate within their scope all of the compounds claimed in each of Claims 1-3. Idenix  
however is not defending any claims, or claiming infringement of claims based on Claim 1.  
(6)  
Claims 4 to 5 and Claims 17 to 18  
[202] Claims 4 to 5 cover the use of the compounds of Claims 1 to 3 (or a pharmaceutically  
acceptable salt thereof) in the treatment of a host infected with a Flaviviridae virus (Claim 4), or  
more specifically, a host infected with HCV (Claim 5). To the extent that these and other Claims  
are based on Claim 1, they are not valid.  
(7)  
Claims 6 to 9  
[203] Claims 6 to 9 (which depend on Claim 4) cover the use of the compounds of Claims 1 to  
3 (or a pharmaceutically acceptable salt thereof) in combination or alternation with a second anti-  
Page: 68  
viral agent in the treatment of a host infected with a Flaviviridae virus. Claim 7 provides a list of  
groups of the second anti-virals and Claims 8 and 9 narrow the second anti-viral to types of  
interferons.  
(8)  
Claims 10 to 12  
[204] Claims 10 to 12 (which depend on Claim 4) cover the use of the compounds of Claims 1  
to 3 (or a pharmaceutically acceptable salt thereof) in the treatment of a Flaviviridae infection  
wherein the compound or pharmaceutically acceptable salt is in the form of a “dosage unit”.  
[205] The term “dosage unit” refers to a defined amount of compound that is administered to a  
whole organism (e.g. a human or animal). Such amounts can be provided in the form of a tablet,  
capsule, injection, suspension, syrup, patch, or the like.  
(9)  
Claims 11 and 12  
[206] Claim 11 limits the amount of active ingredient in the dosage unit (50 to 1000 mg). Claim  
12 specifies that the dosage unit should be a tablet or capsule.  
(10) Claim 13  
[207] Claim 13 (which depends on Claim 4) covers the use of the compounds of Claims 1 to 3  
(or a pharmaceutically acceptable salt thereof) in the treatment of a human infected with the  
Flaviviridae virus.  
Page: 69  
(11) Claims 14 to 16  
[208] Claims 14 to 16 (which depend on Claim 4) cover the use of the compounds of Claims 1  
to 3 (or a pharmaceutically acceptable salt thereof) in the treatment of a host infected with a  
Flaviviridae virus, wherein the compound of Claim 1 to 3 (or a pharmaceutically acceptable salt  
thereof) is in its “substantially pure form”. The term “substantially” means that, on a weight  
basis, the β-D-isomer is at least 85-90% of the product (or a pharmaceutically acceptable salt  
thereof), the remainder of which could comprise other chemical species or enantiomers.  
(12) Claims 15 and 16  
[209] Claims 15 and 16 relate to such uses wherein the compound is at least 90% (Claim 15) or  
95% (Claim 16) by weight of the β-D-isomer.  
[210] To be “at least 90%” or “at least 95%” “by weight of the β-D-isomer” means that on a  
weight basis, at least 90% or 95% of the compound (or its pharmaceutically acceptable salt) is  
present in the β-D-isomeric form, the remainder of which could comprise some impurities or  
other compounds.  
(13) Claims 19 to 32  
[211] Claims 19 to 32 cover the use of the compounds of Claims 1 to 3 (or a pharmaceutically  
acceptable salt thereof) for the preparation of a medicament for the treatment of a host infected  
with a Flaviviridae virus.  
Page: 70  
[212] The term “medicament” means a medicine or drug in a specified formula with an  
acceptable therapeutic index.  
(14) Claims 20 to 32  
[213] Claims 20 to 23 cover the manufacture of medicaments which contain a claimed  
compound with one or more additional anti-viral agents.  
[214] Claim 21 provides a particular list of groups of second anti-virals and Claims 22 and 23  
narrow the second anti-viral to types of interferon.  
[215] Claims 24 to 26 cover the manufacture of medicaments which contain a claimed  
compound in the form of a dosage unit.  
[216] Claim 25 limits the amount of the active ingredient in the dosage unit (50 to 1000 mg).  
[217] Claim 26 specifies that the dosage unit should be a tablet or capsule.  
[218] Claim 27 covers the manufacture of a medicament wherein the compound of Claim 1 to 3  
(or a pharmaceutically acceptable salt thereof) is in substantially pure form.  
[219] Claims 28 and 29 cover the manufacture of a medicament wherein the compound is at  
least 90% by weight of the β-D-isomer (Claim 28) or at least 95% by weight of the β-D-isomer  
(Claim 29).  
Page: 71  
[220] Claim 32 covers the use of the compounds of Claims 1 to 3 in the manufacture of a  
medicament for the treatment of a host infected with HCV.  
VII. Utility  
A.  
General Principles  
(1)  
Requirement for Utility  
[221] It is common ground that there was no demonstrated utility as a June 27, 2003. Idenix has  
admitted that it did not test any compound falling within the scope of the claims of the ‘191  
Patent until March 2005. Therefore, the only argument for establishing utility in the 2’-C-Me/F  
compounds is based on sound prediction from its testing of the 2’-C-Me/OH compounds.  
[222] Gilead alleges that the ‘191 Patent is invalid because at the filing date of June 27, 2003,  
Idenix failed to soundly predict the promised utility.  
[223] Section 2 of the Patent Act defines “invention” as “any new and useful art, process,  
machine, manufacture or composition of matter, or any new and useful improvement in any art,  
process, machine, manufacture or composition of matter.”  
[224] The Supreme Court of Canada [SCC] in Viagra, at paras 37 to 40, has recently  
summarized many of the requirements needed to establish that the invention is “useful” or has  
“utility”:  
Page: 72  
All that is required to meet the utility requirement in s 2 is that the invention  
described in the patent do what the patent says it will do, that is, that the promise of  
the invention be fulfilled.  
When an invention's utility cannot actually be demonstrated by way of tests or  
experiments, its utility can nevertheless be established by means of a sound  
prediction.  
The failure to disclose information on testing compounds goes to the issue of  
disclosure of the invention, not to that of disclosure of the invention’s utility.  
The invention must be useful as of the date of the claim (June 27, 2013) based on  
the information and expertise then available.  
If a patent sought to be supported on the basis of sound prediction is subsequently  
challenged, the challenge will succeed if the prediction at the date of application  
was not sound, or, irrespective of the soundness of the prediction, there is evidence  
of lack of utility in respect of some of the area covered.  
[225] Where the specification does not promise a specific result, no particular level of utility is  
required; a “mere scintilla” of utility will suffice. However, where the specification sets out an  
explicit “promise”, utility will be measured against that promise: Consolboard Inc v MacMilan  
Bloedel (Saskatchewan Limited), [1981] 1 SCR 504 at 525 [Consolboard] ; Pfizer Canada Inc v  
Canada (Minister of Health), 2008 FCA 108, [2009] 1 F.C.R. 53. The question is whether the  
invention does what the patent promises it will do.  
Page: 73  
[226] Before considering whether the inventors had demonstrated utility, the Court must  
consider what the specification says the invention will do. The claims at issue are to be read and  
considered in light of the specification through the eyes and understanding of the person skilled  
in the art, in relation to the science and information available at the time of filing: Apotex Inc v  
Wellcome Foundation Ltd, 2002 SCC 77, [2002] 4 SCR 153 [Wellcome/AZT] at para 51;  
Consolboard, supra at 521; Freeworld, supra at para 44; Eli Lilly Canada Inc v Novopharm Ltd,  
2010 FCA 197 at para 80 [Eli Lilly/Zyprexa].  
[227] Whether the patent contains a promise of utility is to be ascertained as a first step in the  
utility analysis. Because the promise doctrine holds an inventor to an elevated standard, it will  
only apply where a clear and unambiguous promise has been made. Where the validity of a  
patent is challenged on the basis of an alleged unfulfilled promise, the patent will be construed in  
favour of the patentee where it can reasonably be read by the skilled person as excluding this  
promise: Apotex Inc v Pfizer Canada Inc, 2014 FCA 250 at para 66.  
(2)  
The Promise of the ‘191 Patent  
(a)  
The Compounds for which Utility is Claimed  
[228] Gilead asserts that the stated utility of the ‘191 Patent is for all of the claimed compounds,  
which it argues includes those covered under Claim 1 of the Patent. On this basis, it has led  
evidence from its experts that the description of the ‘191 Patent would be invalid, as it does not  
identify which compound has activity and in which Flaviviridae virus.  
Page: 74  
[229] As indicated above, I find that the compounds under Claim 1 are not at issue. The  
relevant compounds therefore are the 2’-C-Me/F nucleosides in Claims 2 and 3 and their  
nucleotides and other derivatives thereof. In my view, these compounds are sufficiently  
identified for the purposes of utility. It is the 2’-C-Me/F sugar ring that is the foundation of the  
‘191 Patent and recognized by both parties as the basis for its antiviral effect, as well as that of  
the ‘657 Patent.  
[230] If the utility is demonstrated or soundly predicted in a nucleoside of the ‘191 Patent, that  
is sufficient for the purposes of meeting the requirements of the Act.  
(b)  
Stated Utility  
[231] Once the compounds are narrowed to the genus of those described in Claims 2 and 3, the  
promise of the ‘191 Patent is not an issue of undue contention. Utility is only to be considered for  
the claims in issue in light of the specification. Gilead and Idenix agree that the Patent contains a  
promise that the compounds are useful in the treatment of Flaviviridae infections, including HCV  
infections, in humans and other hosts.  
[232] Dr Seeger, who testified on behalf of Gilead on this issue, excerpted numerous references  
from the ‘191 patent, opining that the claimed compounds were all stated to be useful in the  
treatment of Flaviviridae infections, including HCV infections, in humans and other hosts.  
[233] The primary reference was to the objects clause at p. 12, stating as follows:  
Page: 75  
Further, given the rising threat of other flaviviridae infections,  
there remains a strong need to provide new effective  
pharmaceutical agents that have low toxicity to the host.  
Therefore, it is an object of the present invention to provide a  
compound, method and composition for the treatment of a host  
infected with hepatitis C virus.  
It is another object of the present invention to provide a method  
and composition generally for the treatment of patients infected  
with pestiviruses, flaviviruses, or hepaciviruses.  
[234] These promises are supported by similar references elsewhere in the specification and  
claims repeating the promise including: The “Field of Invention” at p. 1; the “Summary of  
Invention” at p. 12; the “Detailed Description of the Invention” section at pp.21-22; the “Features  
of the invention” at pp. 23-28; the “Active Compounds” section at p. 100; the “Pharmaceutical  
Composition” section at p. 117; and repeated again in the claims of the ‘191 Patent, for example  
at Claim 4 for the “Use of a therapeutically effective amount of a compound as claimed in  
anyone of Claims 1 to 3 (now considered as Claims 2 and 3), or a pharmaceutically acceptable  
salt thereof in the treatment of a host infected with a Flaviviridae virus” and related Claims.  
[235] Dr Seeger was not contradicted by Idenix. Dr Lamarre, who was Idenix’s only witness  
speaking to utility. He stated in his report at para 69 as follows:  
[69] With respect to what the claimed invention is useful for,  
both Dr Seeger and I agree that the claimed compounds, or  
pharmaceutically acceptable salts or prodrugs thereof, “are useful  
in the treatment of Flaviviridae infections, including HCV  
infections, in humans and other hosts”.  
Page: 76  
(c)  
Toxicity  
[236] Idenix appears to make an overly broad submission with respect to the ‘191 Patent not  
containing any promise of being free of toxicity. I cite para 188 of its written submissions to this  
effect:  
[188] Finally, there is nothing in the ‘191 Patent’s claims or  
disclosure (or for that matter the expert reports of either party) to  
suggest a promise that the claimed compounds are: free of toxicity,  
have reduced side effects, or would necessarily satisfy clinically or  
regulatory approval standards.  
[Emphasis added]  
[237] I would agree with this submission with respect to any promise regarding reduced side  
effects or necessarily satisfying clinically or regulatory approval standards. For example in Eli  
Lilly Canada Inc v Mylan, 2015 FC 17 at paras 87-92 the Court concluded that references to  
toxicity would not include side effects arising from clinical trials.  
[238] Where I disagree with Idenix, if this is the position it takes, is that the Patent, in its  
demonstration of antiviral activity, makes no promise of demonstrating a satisfactory therapeutic  
index. It is to be noted that this index is used to assess a compound’s antiviral activity in relation  
to its cellular toxicity. The evidence of Dr Lamarre on cross-examination was that antiviral  
activity that is toxic is not useful (p. 3264), while Professor La Colla stated that “for an antiviral  
compound, you want to inhibit the viral multiplication without killing the cell.”  
[239] Similarly, Dr Seeger testified as follows:  
Page: 77  
An effective antiviral compound must act selectively against the  
virus and be non-toxic to the same virus-infected or uninfected  
cells. A compound is not useful for treating of viral infection if it is  
also toxic to the host more specially, a skilled virologist would not  
consider a compound to be useful for treating a Flaviviridae if it  
had high cytotoxicity.  
[240] I am of the view that the skilled chemist would interpret treatment for a host infected with  
Hepatitis C virus as requiring effectiveness in combination with low toxicity as measured in  
accordance with the therapeutic index.  
(d)  
The Level of Utility and How Utility is Demonstrated or Predicted  
[241] I find that the ‘191 Patent makes no promise of any specific result or level of treatment. I  
conclude that Gilead must therefore, prove that Idenix has not demonstrated or soundly predicted  
a scintilla of utility in the 2’-C-Me/F compounds for the treatment of Flaviviridae to succeed in  
its argument.  
[242] However, I disagree with one of Idenix’s submissions of how the skilled person  
determines a scintilla of utility is demonstrated or predicted. It advances an argument based on  
Dr Lamarre’s opinion in his November report under the title “How a skilled person can measure  
the utility of the invention?” which concludes at para 86 with the following statement:  
In summary, from reading the ‘191 Patent, the skilled biochemist  
would understand that a compound is useful against Flaviviridae if  
it can exhibit a positive antiviral result in at least one assay in  
respect of at least one virus from the Flaviviridae family.  
[Emphasis added]  
Page: 78  
[243] There are number of issues that appear to be addressed by Dr Lamarre in the foregoing  
statement. First, he appears to be saying that the wording of a patent determines how utility is to  
be demonstrated, apparently as part of its promise. Second, he states that one assay of one virus  
in one type ( e.g. HBV) of the Flaviviridae family is sufficient to demonstrate utility in all the  
other viruses of that type, as well as all of the other viruses in the other two types (e.g. HAV and  
HCV) making up the family of Flaviviridae.  
[244] I will leave the issue of one assay demonstrating utility across the spectrum of  
Flaviviridae until Idenix’s claims of sound prediction are considered. Idenix needs to be able to  
establish this as part of its factual basis to be able to soundly predict the utility of the 2’-C-Me/F  
compounds based on assays of the 2’-C-Me/OH compound.  
[245] Returning to the issue of the promise of the Patent determining what constitutes sufficient  
evidence to demonstrate a scintilla of utility, it would appear that Dr Lamarre is attempting to  
make up for the dearth of assaying data found in the ‘191 Patent. By his measure, claiming that  
testing data available from only one assay is sufficient to demonstrate utility based on the skilled  
person’s reading of the Patent.  
[246] I agree with Gilead that no such interpretation can be made from the ‘191 Patent. More to  
the point however, the sufficiency of evidence to support a finding of utility is not a conclusion  
that will be dictated to by the claims or promises of the patent. It is determined as a function of  
the soundness of the factual basis, in this case for the prediction and the reasonableness of the  
line of reasoning for the prediction.  
Page: 79  
[247] Moreover, I understand this position is intended to confront the shortage of testing data  
for the 2’-C-Me/OH compounds in the Patent. This is only relevant if Idenix is required to  
disclose the basis for the sound prediction in the Patent. Idenix’s position is that it is not required  
to do so. Otherwise, Idenix possesses abundant evidence of extensive testing by Idenix of its 2’-  
C-Me/OH nucleosides in different forms of assays to argue that these results should apply across  
the full spectrum of the Flaviviridae family.  
(3)  
Sound Prediction  
(a)  
General Principles  
[248] The doctrine of sound prediction can be relied upon by an inventor to justify patent  
claims whose utility has not actually been demonstrated, most often because the invention has  
not been made, but can be soundly predicted based upon the information and expertise available.  
If a patent sought to be supported on the basis of sound prediction is subsequently challenged, the  
challenge will succeed if, per Justice Pigeon in Monsanto Co v Commissioner of Patents, 1979  
244 (SCC), [1979] 2 SCR 1108, at p. 1117, the prediction at the date of application was  
not sound, or, irrespective of the soundness of the prediction, “[t]here is evidence of lack of  
utility in respect of some of the area covered,” (Wellcome/AZT at para 56).  
[249] In order to establish that a sound prediction has been made by the inventor, the Supreme  
Court of Canada set out a three-part test in Wellcome/AZT at para 70, as follows with my  
underlining:  
1.  
There must be a factual basis for the prediction;  
Page: 80  
2.  
3.  
The inventor must have an articulable and “sound” line of  
reasoning from which the desired result can be inferred  
from the factual basis; and  
There must be proper disclosure (contested by Idenix).  
[250] The party challenging the patent for lack of utility has the legal burden of proof.  
However, since the relevant documents and information are usually or often in the possession or  
control of the patent owner, the challenger can meet its evidentiary burden by offering expert  
evidence on the relevant issues. Bell Helicopter Textron Canada Limitée v Euorcopter, 2013  
FCA 219 at paras 154 & 161-162.  
(b)  
The Standard for a Sound Prediction  
[251] Idenix cites the decisions of the Federal Court of Appeal in Eli Lilly Zyprexa supra at 76  
and in this Court Astrazeneca Canada Inc v Apotex Inc, 2014 FC 638 at para 182  
[AstraZeneca/Pregabalin] with respect to the appropriate standard for a sound prediction being a  
“prima facie reasonable inference”. In the latter case, the Court states that “[t]he situating of a  
prima facie reasonable inference between mere speculation and certainty provides the clearest  
guidance to the proper approach to be taken in this case.”  
(c)  
The Sound Prediction is an Inferred Fact Not Proven in This Matter  
[252] As noted, the Supreme Court has stated in paras 70 and 71 of Wellcome/AZT that a sound  
prediction is an inferred (“sound” line of reasoning from which the desired result can be inferred  
Page: 81  
from the factual basis) fact (“It bears repetition that the soundness (or otherwise) of the  
prediction is a question of fact”).  
[253] The predicted inferred fact that Idenix says arises from its primary facts based on the  
antiviral activity demonstrated in the 2’-C-Me/OH compounds and sound line of reasoning is that  
in 2003, the skilled person could soundly predict that, when synthesized, the 2’-C-Me/F  
nucleoside will demonstrate therapeutic antiviral activity such as is found in its 2’-C-Me/OH  
compound.  
[254] It is common ground that to prove a fact, the evidence must establish the fact on a level of  
proof on a balance of probabilities, usually described as a likelihood or probability. Anything  
below that evidentiary standard is not an established fact. Anything found not to be a fact on the  
balance of probabilities from the evidence is a mere possibility, or if dealing with inferences,  
mere speculation.  
[255] By the Court’s analysis that follows, I find that Idenix has not established on the balance  
of probabilities from the relevant evidence presented to the Court that the skilled chemists on the  
date of filing of the ‘191 patent application could soundly predict any antiviral activity in a 2’-C-  
Me/F nucleoside prior to its synthesis. Accordingly, the ‘191 Patent is invalid for lack of utility,  
demonstrated or soundly predicted, as a finding of fact.  
Page: 82  
B.  
Idenix’s Claim of Sound Prediction  
(1)  
Overview  
[256] In broad strokes, Idenix claims that the inventors’ sound prediction was based on the  
following three items:  
1. a vast body of knowledge regarding antiviral activity of 2’-C-Me/OH nucleoside  
analogues;  
2. an expectation of activity of 2’-C-Me/OH nucleoside analogues against the family  
of Flaviviridae viruses because they have  
a. similar genomic organization,  
b. conserved sequence/structure motifs and  
c. conserved mechanism of action; in respect of their (NS5/NS5B) viral RNA  
polymerase enzymes; and because  
d. predictive surrogate and representative Flaviviridae virus models were  
known; and  
3. wide acceptance in the nucleoside field that fluorine is an isostere for a hydroxyl  
group.  
[257] Items 1 and 2 represent the first factual basis for establishing that the 2’-C-Me/OH  
nucleosides exhibit antiviral activity against HCV. This conclusion is also somewhat a form of a  
Page: 83  
prediction, inasmuch as Idenix’s testing data for its 2’-C-Me/OH compounds, apart from tests on  
chimpanzees, was on flaviviruses and pestiviruses, but not the Hepatitis C virus using the  
Replicon assays then available. Idenix’s argument is that its tests for antiviral activity in  
flaviviruses and pestiviruses are nevertheless surrogates to predict the antiviral effectiveness for  
HCV. Idenix relies on the literature to support its conclusion that the 2’-C-Me/OH compound  
was tested across the spectrum of Flaviviridae, including using Replicon assays. It also relies on  
its chimp data, which is an equivalent of the human cell line.  
[258] Idenix claims that Item 3 provides the factual basis for the sound of the line of reasoning  
to predict antiviral activity in the 2’-C-Me/F nucleosides based upon that demonstrated in the 2’-  
C-Me/OH nucleosides. It submits that the similarity of the stereochemical structures of the  
fluorine and hydroxyl group as isosteres means that replacing the hydroxyl group in the 2’-C-  
Me/OH nucleoside with a fluorine would result in a likelihood of some activity in the 2’-C-Me/F  
nucleoside, because the fluorine “mimics” the hydroxyl.  
[259] My analysis that follows will first consider the factual basis relating to the 2’-C-Me/OH  
nucleoside as described in Idenix’ a submissions in Items 1 and 2 above. I find that there is a  
satisfactory factual basis to conclude that in 2003 the 2’-C-Me/OH nucleosides demonstrated  
therapeutic antiviral activity against the HCV.  
[260] The more detailed analysis concerns Item 3 pertaining to the factual basis and line of  
reasoning to soundly predict antiviral activity in the 2’-C-Me/F nucleosides based on that in the  
Page: 84  
2’-C-Me/OH nucleosides . I find that Gilead has met its onus of demonstrating that the sound  
prediction of utility in the 2’-C-Me/F nucleoside is a mere possibility and is speculative.  
[261] For the purpose of making factual findings that could become relevant on appeal of this  
matter should that occur, I complete my analysis by considering whether Gilead has  
demonstrated that Idenix is required to sufficiently disclose the sound prediction of utility in the  
2’-C-Me/F compounds in the ‘191 patent, and if so being assumed, whether it has demonstrated  
that it failed to do so. I conclude that Idenix is not required to disclose the basis for its sound  
prediction, but had that been the requirement, the ‘191 Patent fails to do so  
(2)  
Analysis of Idenix’s Prediction of Activity across the Flaviviridae Family (Items 1  
and 2)  
(a)  
A Vast Body of Knowledge regarding Antiviral Activity  
[262] As indicated, I am prepared to accept Items 1 and 2 of Idenix’s submissions above that  
there was a factual foundation to predict antiviral activity in the 2’-C-Me/OH nucleosides across  
the Flaviviridae family. With respect to Item 1, I conclude that Idenix, as the inventor of the use  
of the 2’-C-Me/OH nucleoside as an antiviral compound, had been working on the invention for  
approximately 4 years at the time of the application. It had acquired considerable knowledge  
about the compound’s antiviral activity. I do not believe that this proposition is seriously  
contested by Gilead.  
Page: 85  
[263] Professor La Colla began testing an adenine compound in July 1999. Thereafter, between  
October, 2000 and May, 2001, he tested a number of 2’-C-Me/OH nucleosides using different  
bases, including the other three naturally occurring bases (cytosine, guanine, and uracil).  
[264] On November 29, 2001, WO 01/90121 application [Novirio 121] was published  
disclosing claims of the 2’-C-Me/OH structures for antiviral activity against HCV with the four  
natural bases (adenine, cytosine, guanine, and uracil). On December 6, 2001, WO 01/92282  
[Novirio 282] published the disclosure of the 2’-C-Me/OH structures for antiviral activity  
against flaviviruses and pestiviruses.  
(b)  
An Expectation of Activity of the 2’-C-Me/OH Nucleoside Analogues against  
the Family of Flaviviridae Viruses  
[265] With respect to Item 2 above, I also do not believe that the Items 2(a), (b) and (c) are  
controversial in stating that the family of Flaviviridae viruses have similar genomic organization  
with conserved sequence/structure motifs and conserved mechanisms of action in respect of their  
(NS5/NS5B) viral RNA polymerase enzymes. The evidence of Dr Damha on this point was not  
seriously challenged.  
[266] Only Item 2(d) claiming that there were known predictive surrogate and representative  
flaviviridae virus models is challenged. This relates mostly to Idenix’s failures to test the 2’-C-  
Me/OH nucleosides using the Replicon assays. This assay tests directly on HCV cell line.  
Page: 86  
[267] Instead, Idenix’s tests were conducted in what were described as “HCV mimic” or  
surrogate models. The surrogate tests used by Idenix include those for: BVDV, Yellow Fever  
Virus, West Nile Virus, and Dengue Virus. They are described as a surrogate models because  
they do not test HCV cell lines, which are limited to humans or chimpanzees.  
[268] I am satisfied that the assays conducted by Idenix and reported by other in the  
documentation known to Idenix supports a conclusion that the inventors had reasonable  
expectation of antiviral activity from the 2’-C-Me/OH nucleoside across the spectrum of  
Flaviviridae viruses.  
(i)  
Chimp Data  
[269] Idenix conducted tests on Chimpanzees on its NM283 - 2’-C-Me/OH adenosine  
nucleoside, being the most analogous surrogate to human hosts. Idenix pointed out that the  
Federal Drug Administration [FDA] authorized Idenix to conduct Investigational New Drug  
[IND] human clinical trials using the 2’-C-Me/OH nucleoside (NM107/NM283 -adenosine) in  
December 2002, which it argues it would not have occurred if it were not satisfied with Idenix’s  
testing data. No HCV Replicon data was required before these trials were allowed. Dr Seeger  
acknowledged that this authorization was likely based on the Chimp testing data.  
(ii)  
Carroll et al Article  
[270] In addition, Idenix relies upon a study in The Journal of Biological Chemistry published  
in January, 2003 by Steven Carroll and other scientists from Merck describing data from tests  
Page: 87  
conducted independently using Idenix’s 2’-C-Me/OH adenine compound. I accept Idenix’s claim  
that this paper is significant because it is said to confirm the therapeutic activity of the Idenix  
compounds and its prodrug forms, including its mechanism of action of chain termination  
activity across the Flaviviridae family based on the conserved critical sequences of their RNA-  
dependent RNA polymerase.  
[271] The Merck conclusions used comparative testing, including HCV replicon assays. The  
tests confirmed antiviral activity at concentrations resembling those that Professor La Colla  
obtained in his cell-based surrogate assays. I find that these results largely nullify Gilead’s  
complaint about the lack of replicon assays to support Idenix’s sound prediction.  
[272] Gilead objected to Professor La Colla testifying about this article. It was not mentioned in  
the ‘191 Patent and it argued that this was the first time Gilead was aware that Professor La Colla  
was relying on it. Professor La Colla claims to have learned of the article by way of an email  
circulated amongst Idenix discovery personnel. The email was never introduced into evidence.  
Professor La Colla acknowledges that no one requested his documents and therefore, none were  
produced.  
[273] After hearing argument on this issue, I indicated that I would allow the document to be  
reviewed, but would have to consider the weight attributed to Dr La Colla’s claim to having  
recollected receiving the article and being “happy” because it confirmed his prediction that the  
2’-C-Me/OH compounds would prove to have antiviral activity.  
Page: 88  
[274] It is difficult to understand how an article of such singular importance could not have  
been referenced in the ‘191 Patent. This suggests that none of the inventors would have been  
aware of its existence. The absence of the production of documents in Professor La Colla’s  
possession is a further concern.  
[275] Nevertheless, it may well be that documents known to the inventors did not make it into  
the Patent. I do not believe that Professor La Colla would intentionally mislead the Court. The  
article, which was peer-reviewed and published in a well-known scientific journal, should have  
been of high interest to Idenix. It is the sort of document that would be circulated to all members  
of the discovery team given its relevance and corroboration of their work. Moreover, Professor  
La Colla should have been a priority recipient in such circumstances, given the important role he  
played in testing Idenix’s compounds, the validation of which was the focus of the Carroll  
article.  
(iii)  
Eldrup et al, Bhat et al and Olsen et al Studies  
[276] The ‘191 Patent referenced three studies that were discussed in oral sessions on Hepatitis  
C Virus Flaviviridae at the 16th International Conference on Antiviral Research (April 27, 2003,  
Savannah, Ga.). These include studies by (1) Eldrup et al describing the structure activity  
relationship of 2'-modified nucleosides for inhibition of HCV, (2) by Bhat et al describing the  
synthesis and pharmacokinetic properties of nucleoside analogues as possible inhibitors of HCV  
RNA replication, and (3) Olsen et al. describing the effects of the 2'-modified nucleosides on  
HCV RNA replication. The authors reported that 2'-modified nucleosides demonstrated  
inhibitory activity in cell-based replicon assays. These are references only to the published  
Page: 89  
abstracts of the presentations at the Savannah conference. However, Dr Storer and others from  
Idenix attended and reported on the presentations and were definitely within the knowledge of  
the inventors.  
(c)  
Gilead Submissions  
[277] I accept Dr Seeger’s evidence that normally without the Replicon assay data, there would  
not have been a sufficient basis for a sound prediction of antiviral activity across the Flaviviridae  
family based on the 2’-C-Me/OH nucleoside testing data. The evidence on the development of  
assaying technology is convincing that the advent of the replicon assay commencing in 1999 was  
immediately recognized as a significant breakthrough in the common general knowledge of the  
skilled virologist to test for HCV.  
[278] Inventors of patent applications filed in relation to Flaviviridae antivirals at the beginning  
of the 20th century invariably adopted it to validate their conclusions. This includes those of Dr  
Damha filed in that period and most of the patents referred to by witnesses in the relevant time  
period throughout the trial. Replicon assays were clearly a requirement for validating HCV utility  
by the filing date of the ‘191 Patent.  
[279] Regarding other criticisms by Gilead, I would also accept Dr Seeger’s statement that  
Idenix had concentrated its efforts on certain nucleosides querying to some extent these  
limitations. However, I do not see this impacting on the reliability or validation of its data set  
underlying its prediction of utility, in so far as the sound prediction is based upon the antiviral  
activity of the 2’-C-Me/OH nucleosides. I also find that Idenix’s testing responds to Dr Seeger’s  
Page: 90  
contention that the literature recognized that all nucleoside analogues are not necessarily active,  
or are active with different potency across the family of Flaviviridae viruses.  
[280] The sound prediction of Idenix focuses on the similarity of the antiviral activity at the 2’  
position on the core sugar ring of the 2’-C-Me/OH and 2’-C-Me/F nucleosides. Gilead has only  
patented two nucleosides, which confirms the essential issue for the sound prediction is with  
respect to the substituents at the 2’ position on the ring.  
[281] Gilead also argues that the testing conducted by Idenix was too limited to establish the  
therapeutic antiviral activity across the breadth of Idenix’s claims. This complaint refers once  
again to the particularly large number of compounds covered by Claim 1, mostly in reference to  
the 2’(up) position or the number of nucleosides bases in Claims 2 and 3. I have already  
discussed and dismissed this form of overbreadth argument, based upon the skilled person’s  
more selective interpretation of the ‘191 Patent. With respect to the large number of bases  
claimed, it was pointed out to Dr Seeger in cross examination that bases were similarly described  
in Gilead’s ‘657 Patent.  
[282] In summary therefore, apart from issue of the soundness of the prediction based on steric  
mimicry which is addressed below, I find that Idenix’s reliance on the activity demonstrated in  
the 2’-C-Me/OH cytidine nucleoside to establish therapeutic antiviral activity against the  
Flaviviridae family, including HCV is sufficient as one aspect of its factual basis for its sound  
prediction of antiviral activity of HCV by a similar 2’-C-Me/F nucleoside compound.  
Page: 91  
(3)  
Sound Prediction of Utility Based on Isosteric Mimicry of Fluorine for a  
Hydroxyl Group  
[283] I find that Gilead has demonstrated that Idenix had no sound line of reasoning to predict  
the antiviral activity of the 2’-C-ME/F compound based on its hypothesis of “steric mimicry.”  
Professor La Colla, as the only one of four inventors that testified, did not demonstrate that the  
inventors considered that their prediction based on steric mimicry was anything more than a  
possibility. The contemporaneous documentation also demonstrates that the inventors believed  
that the hydroxyl group at the 2’ (down) position on the sugar ring was essential. The Court  
similarly finds Dr Lamarre’s opinion of sufficient steric mimicry such as to predict the antiviral  
activity of the 2’-C-Me/OH compound in the 2’-C-Me/F nucleoside at best, to be an educated  
guess, not supported by the common general knowledge.  
[284] Rather, the Court accepts the testimony of Dr Seeger and others that the 2’ down position  
on the D-Ribose sugar ring was recognized as a highly conserved and selective a position and  
that fluoride is too unpredictable a substituent regarding potential toxicity to provide a sound  
basis to conclude that the prediction of antiviral activity from the substitution of a fluorine for the  
hydroxyl was anything but mere speculation.  
(a)  
The Inventors’ Factual Foundation for the Sound Prediction of Utility  
[285] The ‘191 Patent lists four inventors, two of whom are chemists (Drs Storer and Gosselin),  
and two of whom are pharmacologists (Dr Sommadossi and Professor La Colla). Dr Storer and  
Dr Gosselin were both highly experienced research chemists. Neither was called to testify at trial,  
Page: 92  
even though Dr Gosselin is still affiliated with Idenix and Dr Storer was a key witness in the  
affairs of Idenix at the time. Idenix called one inventor to testify, Professor La Colla.  
[286] Gilead argues that Professor La Colla had been rehearsed significantly for his testimony  
at trial and was led extensively during his direct examination. Gilead submits that it is not clear  
whether Professor La Colla’s testimony at trial is truly reflective of what he believed and knew at  
the relevant time. I agree to some extent with this submission, but conclude its effect is  
somewhat attenuated by the fact that Professor La Colla was testifying in his second language,  
aided only from time to time by an interpreter.  
[287] Nevertheless, the Court is required to give somewhat diminished weight to Professor La  
Colla’s testimony upon being advised of the singular lack of confidence that Idenix displayed  
towards him by denying him the opportunity to answer fundamental questions relevant to his role  
in the invention during the discoveries. In a portion of the discovery transcripts running over  
several pages, the Court noted that every single question that Gilead asked in an attempt to  
understand what Professor La Colla considered was his contribution to the invention was refused.  
Idenix eventually provided answers to these questions; some during the trial and further  
discovery was contemplated but avoided by answering a “stack” of written questions. When the  
Court learns from references to a discovery transcript that a party’s witness has been kept on a  
very short leash, it cannot but conclude that it should not have more confidence in the witness’s  
testimony than the party does.  
Page: 93  
[288] More importantly however, the Court has a number of substantive concerns with  
Professor La Colla’s testimony regarding his sound prediction of antiviral activity in the 2’-C-  
Me/F nucleosides. Overall, I find that he expressed his sound prediction in terms of a possibility,  
as opposed to a likelihood or probability. He also seem to confuse the strength of his personal  
belief that his prediction would be sound, with the necessity of demonstrating by objective  
evidence that it was a probability.  
[289] Specifically, Professor La Colla repeatedly relied on the patent application of Emory  
University, PCT Publication No. WO 99/43691, entitled “2'-Fluoronucleosides” [“the Emory  
Patent application”], as did Idenix, as its rationale to support his conclusions of the soundness of  
his predictions based on steric mimicry. However, the Emory Patent application was speculative  
in its statements, and more importantly, did not contain any data upon which a prediction of  
utility for a fluoride substitution of the 2’ (down) hydroxyl on the sugar ring could be based.  
[290] Professor La Colla summarized his reasoning for his sound prediction of utility of the 2’-  
C-Me/F compound in the passage as follows:  
THE WITNESS: … And my position was that the compound to be  
synthesized with the high priority was 2'-methyl (up) and 2'-fluoro  
(down). The reasons that I had to propose this priority were based  
on the knowledge of the Emory patent --  
…
Emory, Emory University, the Emory patent.  
…
THE WITNESS: Where a huge number of 2'-fluoro derivatives  
were described.  
Page: 94  
I had two reasons for my personal choice of making 2'-methyl (up),  
fluoro (down). The first was that in this same patent, it was  
recognized that fluorine could be the best substituent for an  
hydroxy group based on the fact that the CO bond is totally similar  
to the CF bond, that OH and F are acceptor, are hydrogen acceptor  
groups, and also based on my more general knowledge working  
with so many chemists, during which we often had the problem of  
trying the synthesis of new compounds by changing an OH with  
the fluorine or vice versa. And what was remembered by me was  
that in most of those cases, the choice was successful because OH  
and fluoro are so-called bioisosteres, so changing one for the other  
doesn’t change the biological activity of the compound.  
The third reason was that in the Emory patent, the antiviral  
activities claimed for 2'-fluoro compounds were the following:  
HIV, HBV, HCV, and also they were claimed to have all  
anti-tumor activity.  
So the reason I was interested in this spectrum of activity and in the  
difference that the 2'-fluoro compounds had in spectrum of activity  
with respect to our 2'-methyl compounds was the simple question,  
will substitution of OH (down) with a fluorine give 2'-methyl (up)  
compounds an additional or a totally new spectrum of activity, that  
was the question.  
But, of course, I had a privileged answer, this was just a scientific  
curiosity and could be solved only by making that compound.  
But the prediction that I could be able to make was that according  
to my knowledge and reasoning, 2'-methyl (up), 2'-fluoro (down)  
were going to show the same activity and thus the same spectrum  
of activity of 2'-methyl (up), 2'-OH (down). This because of said  
reasons of the bioisostere. So I could not believe that the  
bioisostere could lead to a total change in the behaviour of the  
molecule as an inhibitor becoming inhibitor not only of  
RNA-dependent RNA polymerase but also of reverse  
transcriptases, also, which are the enzymes HIV and HBV. I  
couldn't believe that. And I preferred to stick to the hypothesis that  
due to the bioisosteres and due to the strong, strong conviction that  
the methyl group could be prevalent in dictating the broad  
spectrum activity, I could make the predizione, the prediction -- .  
[Emphasis added]  
Page: 95  
[291] Professor La Colla indicates that it was a “scientific curiosity that could just be solved by  
making the compound.” The term “just” in this sense I interpret to mean “only.” The contingent  
character of a successful outcome is furthermore consistent with the speculative nature of the  
language he employs to describe his prediction, i.e. a belief, a question to be answered, a  
preference or a hypothesis.  
[292] Later in direct examination, when attempting to explain his prediction of utility in the  
face of doubts expressed by others on the Idenix discovery team of chemists at the Maui meeting,  
he stated:  
And I stood for the fluorine, the fluorine substitution of OH  
because of the Emory patent that was extremely clear in that  
respect and also guided other inventors in doing the 2' substitution  
of an OH with a fluorine.  
[Emphasis added]  
[293] In fact, the passage in the Emory Patent application with respect to the ability to  
substitute a fluorine for a hydroxyl, which Idenix argues provides the rationale for its sound  
prediction, is expressed as a possibility, at best. It reads follows:  
In designing new biologically active nucleosides, there have been a  
number of attempts to incorporate a fluoro substituent into the  
carbohydrate ring of the nucleoside. Fluorine has been suggested as  
a substituent because it might serve as an isopolar and isosteric  
mimic of a hydroxyl group as the C-F bond length (1.35 A) is so  
similar to the c-o bond length (1.43 A) and because fluorine is a  
hydrogen bond acceptor. Fluorine is capable of producing  
significant electronic changes in a molecule with minimal steric  
perturbation The substitution of fluorine for another group in a  
molecule can cause changes in substrate metabolism because of the  
Page: 96  
high strength of the C-F bond (116 kcal/mol vs. C-H = 100  
kcaUmol).  
[Emphasis added]  
[294] Later, when confronted with the proof that the Emory Patent application contains no  
claims of antiviral activity, and indeed no data on anti-HCV or Flaviviridae viruses, Professor La  
Colla resiled from his “extremely clear” reliance on the Patent, indicating that the Emory Patent  
application was only useful to establish “a simple chemical fact” that the hydroxyl group and  
fluorine were isosteres.  
[295] He also describes his conclusion as a “guess”. The following excerpt is taken from the  
transcript of his cross-examination:  
Q.  
Okay. And so this Emory patent has no data on anti HCV or  
Flaviviridae viruses; correct?  
A.  
Q.  
A.  
Q.  
Yes. But, the reason I …  
Professor  
The reason I mention this patent  
Professor, it was a simple question, you answered it, there  
is no data in this patent; correct?  
A. Yes, but it is not because of data that I mentioned this  
patent. It is because it is very clearly said that the CO bond is  
bioisostere to the CF bond, that is enough. I mention this Emory  
patent just to say that other people already established a simple  
chemical fact, the bioisostere between OH and fluoro, I didn't  
apply that. I started from the knowledge of this patent to make any  
guess on any new and good anti HCV compounds. [pp. 1893-4]  
[Emphasis added]  
Page: 97  
[296] I also note that in reference to his own work as a foundation for his prediction, no  
examples or description of his involvement in fluorine substitutions or compounds containing a  
fluorine was provided. This contrasts with the very specific evidence of Dr Stuyver, who claims  
to have played an important role in Gilead’s synthesis of the 2’-C-Me/F compound. Dr Stuyver  
supports his claim however, by reference to his own work with fluorines, including his role in a  
patent application by Gilead with a fluorine substitution in the 2’ (down) position on the sugar  
ring. Despite this, his view is that he could not predict antiviral activity in the 2’-C-Me/F  
compound, which could only be determined by testing.  
(b)  
Contemporaneous Evidence Relevant to the Prediction of Utility of the 2’-C-  
Me/F Compound  
[297] The parties presented evidence about a number of events or documents which were  
relevant to the soundness of the prediction of utility in a 2’-C-Me/F compound, were it  
synthesized. The Court’s analysis will follow these events as they were presented in the  
arguments of Idenix.  
(i)  
The Discussion at the December 18, 2001 Maui Chemistry Meeting  
to Synthesize the 2’-C-Me/F Compound  
[298] The evidence regarding the Maui chemistry meeting is relevant both to Professor La  
Colla’s prediction and that of the Idenix chemistry group concerning the utility of the 2’-C-Me/F  
compounds. It was attended by the ‘191 Patent inventors (Dr Gosselin, Professor La Colla, Dr  
Sommadossi, Dr Storer), along with Dr Standring and other Idenix discovery chemistry  
Page: 98  
personnel. A preparatory canvassing the issues and providing some thoughts on them was  
apparently prepared by either Dr Storer or Dr Gosselin, neither of whom was called to testify.  
[299] Idenix refers to passages in the meeting notes that confirm that “[t] he 2'Me substituent is  
clearly important and appears to make an essential contribution to activity.” Dr Lamarre states  
that the importance of this finding cannot be stressed enough. Dr Seeger disputes this point,  
indicating antiviral activity is not necessarily an addition of A and B molecules. On the other  
hand, Idenix does not really dispute that changes of a substituent of any compound can radically  
change its characteristics. The more relevant evidence on this issue is how the steric similarities  
of fluorine and the hydroxyl group can form the basis for a sound prediction that the substitution  
of the fluorine will likely demonstrate antiviral activity.  
[300] On this item, the meeting note contains predictions of the likely success of other 2’-C-  
Me/OH analogues that Idenix was planning to synthesize by replacing the 2’ hydroxyl with a  
range of other substituents. The note clearly reflects a negative prediction of utility for any of the  
suggested nucleosides to be synthesized with various substituents replacing the 2 hydroxyl group  
on the sugar ring.  
The 2’-Hydroxyl  
For a RNA virus we anticipate that this would be essential but it is  
worth checking with a few key compounds in the 2’beta-Me C and  
G series what the exact requirements are particularly since the  
compounds of interest have a tertiary rather than the usual  
secondary OH. Analogs lacking the OH will likely confirm its  
essential nature but should be done. Can the OH be replaced by  
other groups which will allow H-bond acceptance or donation. A  
wide range of groups could be of interest including NH3, NHAc,  
CN, F (although often toxic), CO2H,CONH2, OMe. A priority  
Page: 99  
should be to investigate whether the OH can be moved by making  
CH2OH. If this were possible, it would open up a wide variety of  
other 5m sugar rings such as those with additional heteroatoms. It  
is also worth seeing if the hydroxyl position could be accessed  
from 1’ or 3’ by CH2OH.  
[Emphasis added]  
[301] The note presents a pessimistic prediction of any compounds substituting a fluoride for  
the hydroxyl group, anticipating (1) that hydroxyl “would be essential” and (2) that analogs  
fabricated “will likely confirm its essential nature”. It also described (3) a number of substituents  
to be considered as replacements for the 2’ hydroxyl, of which the fluoride was not the primary  
compound. The fluoride was however, singled out (4) by being described as “often toxic.”  
[302] The views expressed in the note appear to be consistent with Gilead’s position that in  
June 2003, the likelihood of utility could not be expected from the 2’-C-Me/F compound, if  
synthesized.  
[303] Professor La Colla testified about the 2 hydroxyl position being “essential” as follows:  
Q.  
“For a RNA virus, we anticipate that this would be  
essential.”[as read]  
What is this referring to?  
A.  
So as the methyl (up) was recognized as fundamental for  
the anti RNA plus activity, also the 2' hydroxy (down) position was  
believed to be important by definition because that position makes  
the nucleoside, the physiologic nucleoside or the ribo analogues,  
usable for RNA synthesis, the nucleoside, the physiologic  
nucleosides and the inhibitory nucleoside analogues, in particular  
those with 2' methyl (up).  
Page: 100  
So, perhaps in a short sentence, the point was there will be some  
possibility to substitute OH with another group allowing hydrogen  
bonds. And the answer was widely discussed, and everyone came  
out with its own proposal or way of viewing this problem on the  
basis of the literature that all of us knew.  
So if I may, I can give you which was my position.  
And my position was that the compound to be synthesized with the  
high priority was 2' methyl (up) and 2' fluoro (down).  
The reasons that I had to propose this priority were based on the  
knowledge of the Emory patent.  
[Emphasis added]  
[304] Professor La Colla’s evidence confirms that other Idenix chemists thought that the  
hydroxyl group would likely prove essential because of its selective role in the RNA. Professor  
La Colla again cites the Emory Patent application as the basis for his sound prediction.  
Moreover, in his areas of specialty, Professor La Colla appears to support the conclusions of Dr  
Seeger that the 2’(down) position in the DNA/RNA ring is unique because of the profound role it  
plays in physiological biology.  
And there are two types of nucleosides, one with the OH in this  
position [2’ down], physiological nucleosides, and another kind of  
physiological nucleosides with an H in this position. Because of the  
lack of the oxygen, these later compounds are called deoxy. The  
ribo are used to build up RNAs, the deoxy ribo are used to build up  
DNAs. So between the two compounds, there is a profound  
difference in terms of what they are useful for. La Colla p. 1460.  
[Emphasis added]  
(ii)  
Toxicity - March 7, 2002 – Gemcitabine - Roche Patent Publication -  
WO 02/18404 [404 Patent]  
Page: 101  
[305] With respect to the comment that fluoride was often toxic, Professor La Colla  
understandably could not recall the specific discussions at the Maui meeting. He testified  
nevertheless, to his disagreement with the comment based on the conclusions he drew from the  
Emory application described above.  
[306] On the matter of toxicity, the Roche 404 Patent, which was referred to in the ‘191 Patent,  
was the subject of comment by Dr Lamarre in his reply report. The 404 Patent includes a claim  
for a di-fluoro nucleoside analogue with the structure of a 2’-F(up), 2’-F(down).  
[307] These compounds are similar to the drug Gemcitabine, not to be confused with the  
gemcitabine process as a means of synthesizing the 2’-C-Me/F compound discussed in the  
disclosure of synthesis section below. Gemcitabine has a long history in treating cancer, i.e. a  
drug that kills cells. Dr Seeger explained somewhat this background, as follows:  
[…]originally investigated as a potential antiviral drug -- I  
expanded on that [in my report]  
-- was determined to be too toxic to be a useful  
antiviral agent at the time and was developed  
instead as an anti-cancer agent. It used today against  
pancreatic cancer and other cancers. (p. 856)  
[308] Based on the 404 Patent, Idenix argues that “it was known to the skilled person and to the  
inventors that fluorine substitutions at the 2’-position were not toxic.”  
[309] I find this overstates Dr Lamarre’s evidence. It was only intended to mitigate the  
contention that fluorides can be toxic, which he openly acknowledges, as follows:  
Page: 102  
[229]The inventors contemplated using a fluorine substitutions  
despite that it was known that some, but not all compounds  
containing a fluorine may exhibit toxicity in a host.  
[230]Indeed, the inventors were aware of, and included on page  
10 of the '191 Patent, the International Publications from Roche  
(WO 02/18404) and Emory (WO 99/43691), which both disclose  
that 2'-F(down) can be used to provide antiviral activity in the  
treatment of hepatitis C.  
[Emphasis added]  
[310] As noted, the Emory Patent application only discusses the possibility of the Fluorine  
providing activity against HCV. With respect to the 404 gemcitabine patent, Dr Lamarre does not  
rely upon the patent as a basis for sound prediction, but only in respect of refuting concerns about  
toxicity always occurring.  
[311] Moreover, Dr Seeger refers to gemcitabine and unsuccessful attempts to place a fluorine  
at the 2’ (down) position on the sugar ring, with a hydroxyl in the (up) position. I cite his  
evidence at page 863 of his examination in chief, where he notes that the sugar ring with the  
fluorine (down) and hydroxyl (up) and at the 2’ position kills cells:  
Q.  
At paragraph 224(c), you discuss nucleoside analogues and  
cytotoxicity. Can you share with the Court the Skilled Virologist's  
understanding?  
A.  
I alluded yesterday to the historical aspect of drug  
discovery in reference to nucleoside analogues. I tried to point out  
that, historically, this field started out in the field of cancer therapy.  
The goal was to kill cancer cells because they would divide  
rapidly. Only subsequently was it taken up by virologists. The  
challenge again is to find the ones that are non toxic. I refer here to  
a cancer drug. It is called gemcitabine. What is interesting about  
that drug is that it has a fluoro (down), in the (down) position. It  
also has a fluoro in the (up) position. What is particularly  
Page: 103  
interesting is, in the 1980s, chemists at Eli Lilly said, "If I have the  
fluoro(up) and the hydroxyl(down), let's put the fluoro into the  
(down) position. I might have a great antiviral, a new antiviral." He  
produced it, and then the virologist told him, "Your drug actually  
kills cells."  
[Emphasis added]  
[312] It is not clear that the Roche 404 Patent, even though referenced in the ‘191 Patent was  
within the knowledge of Professor La Colla. It was not put to him for comment in an otherwise  
detailed walk-through of the evidence that supported his prediction of utility of the 2’-C-Me/F  
compound. Nonetheless, conceding that it is difficult to refute knowledge of the inventors when  
specifically referenced in the patent, I accept that the inventors were aware of the contents of the  
404 patent. This is in addition to considering Dr Seeger’s evidence on the toxicological history of  
modifications at the 2’ position described above.  
(iii)  
October 2, 2002, Idenix’s Strategy and Progress Meeting  
[313] Some 10 months after the Maui meeting, a strategy and progress meeting was held by  
Idenix. Efforts continued during the intervening period working on analogues of the 2’-C-Me/OH  
nucleosides without any progress being made in the synthesis of the 2’-C-Me/F compound. The  
slides prove that not only was the 2’-C fluoride (down) compound not the “lead” or prime target  
at the time, but experimentation was underway at both the 2’ and 3’ positions of the sugar ring.  
[314] The lack of progress generally being made in synthesizing new 2’-C-Me/OH analogues is  
evident in the slides used at the meeting, with the slide below showing what little had been  
accomplished in the interim:  
Page: 104  
[315] The further portion of the slide at page 10 demonstrates that at this time the 2’-C-Me/F  
compound was still not the only target composition under development.  
[316] Professor La Colla commented on the middle figure [the methyl is inferred at the 2’ (up)  
position of the ribose ring in these figures] which are as follows:  
This compound was my preferred compound at the time, it contains  
a 2'-methyl (up) and a 2'-fluoro (down). And the fact that the  
chemists accepted to put it allow the compounds to be synthesized  
and that they say currently in progress made me very happy but  
this means that they accepted the whole prediction and the basis for  
that prediction.  
Page: 105  
[317] Professor La Colla’s statement of a compound being preferred over others is about a  
triage. In my view, it is not indicative that, if fabricated and tested, the compound would likely  
demonstrate some utility.  
[318] I also do not accept Professor La Colla’s conclusion that Idenix’s decision to proceed to  
attempt to synthesize the compound is indicative of it accepting his prediction, or that it is a basis  
to conclude a likelihood of utility from the compound’s possible synthesis.  
[319] The evidence in this case suggests strongly that the drug discovery industry adopts a  
procedure that I would describe as “make-and-test” all candidate compounds that demonstrate  
some reasonable possibility of success. In other words, drug developers do not require a likely or  
probable successful outcome for a decision to go forward to synthesize a new compound for  
testing.  
[320] The make-and-test a range of analogue candidates was certainly reflected in the approach  
outlined in the minutes of the December 2001 Maui meeting. A list of possible substituents at the  
2’ HO position were targeted for development, even though the minutes evidenced little hope for  
success.  
[321] This methodology was confirmed independently by the evidence of Dr Patterson. He  
described his reluctance to test for analogues using the Idenix patent application as follows:  
Dr. Schinazi came to us with the news that the compound we called  
the Idenix compound was active against hepatitis C, he wanted us  
to consider making analogues. Dr. Watanabe presented some  
Page: 106  
analogues of interest, nobody was particularly interested in making  
that molecule [the fluorine substituent for HO], certainly I wasn't,  
because we knew that that was made at Idenix and we are certain  
that Idenix would have covered all the reasonable analogues.  
[Emphasis added]  
[322] Dr Patterson’s description of the certainty of Idenix’s predictable behaviour to make and  
test all reasonable analogues applies to what transpired at the December 2001 meeting in Maui.  
Fluoride was a reasonable analogue based on a possibility of activity, as were a number of other  
substituents.  
[323] Moreover, if going forward with attempting to synthesize a new drug is evidence of a  
sound prediction, the test would become circular and tautological. The decision to attempt  
synthesis would predict utility, as opposed to whether or not the compound can be predicted to be  
useful, if synthesized.  
(iv)  
November 28, 2002, - Roche Patent Application - No. WO  
02/094289 [289]  
[324] The further Roche ‘289 Patent Application describes antiviral activity in a compound  
described as 2'-deoxy-2'-fluorocytidine with a single fluoro at the 2’ (down) position on the sugar  
ring. The first and only reference to the Roche ‘289 Patent Application occurred during  
examination in chief of Dr Stuyver. Dr Stuyver pointed out that the Roche publication predated  
his unpublished Pharmasset patent application US 60357411. It described a similar Pharmasset  
compound [PSI-0262], which established antiviral activity for the same compound by testing  
using the HCV replicon.  
Page: 107  
[325] Although published in 2002, Patent Application ‘289 was not referenced in the ‘191  
Patent. None of the experts or other witnesses made mention of it in their reports or testimony. It  
was not brought to the attention of Professor La Colla, nor discussed during his testimony. I am  
not satisfied that the 289 Patent was known to the inventors of the ‘191 Patent on the filing date.  
Therefore, it is not considered as forming part of their knowledge for the purpose of determining  
the utility of the ‘191 Patent.  
(v)  
April 1-4, 2003 - Summary of an Idenix internal Scientific Update  
Course on Fluorination  
[326] Idenix claims that it was clear that the prediction for using fluorine in place of the  
hydroxyl group was well known within Idenix based upon information received from a Scientific  
Update Course on Fluorination. Dr Griffon and Audrey Chappe attended this course for the  
purpose of assisting in the synthesis of the 2’-C-Me/F nucleoside. They prepared a report that  
was later referenced to have been received by email-dated more than two months later on June 9,  
2003 from Dr Storer to Moussa Chadhari attaching the Report. According to Dr Lamarre this  
report shows that the prediction for using fluorine in place of the hydroxyl group was well known  
within Idenix, when at page 3 of the Summary, it discusses that replacing the hydroxyl group  
with a fluorine would act similarly with the hydrogen-bond donor.  
[327] The report was not referred to by Professor La Colla, concerning information received  
two weeks prior to the filing of the Patent application, while Dr Storer did not testify. It appears  
to report on an opinion of one of the instructors at the program regarding synthesis, not the  
pharmaceutical properties of a 2’-C-Me/F nucleosides. In that regard, it appears to be no more  
Page: 108  
than the statement in the Emory Patent application, only more generalized to physical properties  
and of even less certain and authoritative a source. It obviously does not deal with other issues  
such as toxicity problems associated with fluorine. I cannot attribute much weight to this  
evidence given its source and the manner by which the evidence has been presented to the Court.  
(c)  
The Evidence of the Experts  
(i)  
Dr Lamarre’s Opinion on Steric Mimicry  
[328] Dr Lamarre was Idenix’s expert who it was pointed out was the only expert qualified to  
opine on both virology and isosteric propreties of atoms and compounds. His expert opinion in  
this regard was not challenged on cross-examination.  
[329] Dr Lamarre’s general statement on the predictive outcome of fluorine as a substitute for a  
hydroxyl group is contained at para 172 of his report as follows:  
By June 27, 2003, the skilled person would also have known that  
when designing new biologically active nucleosides by a substrate-  
based Rational Drug Design, often hydroxyl (OH) groups of the  
nucleoside can be replaced by fluorine (F) atoms.  
[Emphasis added]  
[330] With respect, the statement is not as strong as was needed to predict a likelihood of  
antiviral activity in the 2’-C-Me/F nucleosides. The term “often” is not a scalable measure above  
or below a probability of expectation. It appears to be the type of compromise term sometimes  
reached where there may be unwillingness to state a proposition in as clear a fashion as is  
Page: 109  
required, but is as far as the expert is willing to go. In any event, the opinion is at best ambiguous  
as a backhand sort of support of predictability, which is not what would be expected from an  
expert supporting a sound prediction of steric mimicry.  
[331] The statement is also not specific to the issue of utility, instead being generalized to a  
situation of “when designing new biologically active nucleosides.” This issue here is not about  
designing new nucleosides. There was no rationale design process followed by Idenix. The  
evidence demonstrates that Idenix proceeded exactly as Dr Patterson said they would, namely  
trying to synthesize any number of 2’-C-Me/OH analogues with the view to testing them for  
antiviral activity, with no likelihood of success as they set off to do so.  
[332] The heart of Dr Lamarre’s opinion on steric mimicry is contained in his para 173 of his  
report citing a portion of the introduction of a 1998 article by Professor Jeffrey McAtee and  
others entitled A Completely Diastereoselective Electrophilic Fluorination of a Chiral,  
Noncarbohydrate Sugar Ring Precursor: Application to the Synthesis of Several Novel 2′-  
Fluoronucleosides in the Journal of Organic Chemistry, volume 63, pp. 2161-2167.  
[333] As the Emory Patent application relied on the McAtee article for the propositions it cites  
regarding steric mimicry, the McAtee article appears central to Idenix’s case on this issue. Apart  
from this article, no other literature was cited for fluorine substitutions for the hydroxyl group at  
the 2 position on the sugar ring.  
Page: 110  
[334] Dr Lamarre cited the following passage from the article, with his emphasis and footnotes  
in square brackets, in support of his conclusions:  
Fluorine may also serve as an isopolar and isosteric mimic of a  
hydroxyl group since the C - F bond length (1.35 Å) is so similar to  
the C - O bond length (1.43 Å) and because fluorine is a hydrogen-  
bond acceptor. The ability of fluorine to mimic a hydroxyl group  
makes this atom uniquely suited to nucleoside analogues as a  
replacement of OH in the sugar portion of a nucleoside. In addition  
to our long standing interest in the synthesis of novel nucleoside  
analogues, [3] we were interested in incorporating an R -fluorine  
substituent at the 2′ position of the sugar ring for several reasons.  
First, the electronegativity of fluorine should stabilize the anomeric  
bond and suppress a significant pathway of in vivo decomposition,  
[4] thereby improving the acid stability of the nucleoside (Scheme  
1).  
Second, hydroxyl groups often serve as “handles” for the first step  
in oxidative degradation of biomolecules in vivo. [2c] By replacing  
OH with F, it is possible to create a ribo-like sugar that has a  
substituent at the 2′ position sterically and electronically similar to  
a hydroxyl group, but which cannot undergo oxidative catabolism.  
[Emphasis of Dr Lamarre]  
[335] While Gilead did not cross examine Dr Lamarre on this evidence, the Court is entitled to  
make a few observations on some of the obvious aspects of the McAtee article. First, the  
proposition of substituting a fluorine for a hydroxyl in the article is described in terms of  
possibilities: “Fluorine may also serve as an isopolar and isosteric mimic of a hydroxyl group”  
and “[b] y replacing OH with F, it is possible to create a ribo-like sugar.”  
[336] Second, the passage from the introduction to the article cited by Dr Lamarre bears no  
relation to the purpose of the study. The study was reporting on unsuccessful attempts to  
Page: 111  
synthesize 2’ fluoro (down) nucleosides, because they were lacking in availability, as was  
described in the passage just after that quoted by Dr Lamarre, as follows:  
Finally, few of the ribo or (down) 2′-fluoro nucleoside derivatives  
were known. Therefore, we endeavored to find a simple and  
efficient synthesis of these compounds in order to generate novel  
analogues for biological evaluation.  
[337] The authors’ goal was not met in terms of synthesizing fluorinated nucleosides for  
biological evaluation, inasmuch as only 2 of the 10 compounds synthesized proved active and  
they turned out to be toxic. (“[u] nfortunately, only compounds 15b and 18b showed any antiviral  
activity against HIV, but they exhibited toxicity as well, indicating that the compounds are  
nonselective inhibitors of HIV” p. 2163). This failed experiment would seem to speak to the  
opposite of any sound prediction of antiviral activity, besides negating the common general  
knowledge in support of the predictability of either activity or non-toxicity of 2’ fluoro (down)  
nucleosides.  
[338] More importantly, there are also no footnotes supporting the portion of the article cited by  
Dr Lamarre in his report, although other statements in the comment are footnoted. From that  
point of apparently no supporting literature, one would have thought that between 1998 and 2003  
the literature and texts would have included examples of fluorines “mimicking” the 2’ (down)  
hydroxyl. I can find no basis therefore, to consider sufficient evidence was provided to support  
any the propositions on steric mimicry of fluorine likely being able to replace a hydroxyl as  
representing the common general knowledge or seriously supported in the literature.  
Page: 112  
[339] The Court also does not need expert opinions to advise it that if the McAtee article is the  
best that Dr Lamarre could come up with as the basis for his sound prediction, this speaks legions  
for the hypothetical nature of his proposition. It also would explain why, as a scientist, he appears  
reluctant to make a stronger statement on the probability of the prediction of utility for the 2’-C-  
Me/F nucleosides.  
[340] Finally, I should comment on Dr Lamarre’s testimony that introduced a factor, namely  
described as the “Structure Activity Relationship” or SAR, which appears relevant to the  
mimicry argument, but is not discussed in his report. I quote a fairly long passage from the  
transcript of Dr Lamarre’s introduction of this concept into the evidence when discussing steric  
mimicry:  
Q. So paragraph 171, you talk about “given this foundation”, what  
is your view and, of what the skilled person would gather, starting  
with the foundation that was known by the inventors?  
A. So the inventors were, I guess, quite knowledgeable about this  
2' position, the methyl “up” being a major part of the invention,  
and the OH (down), which is in the natural substrate, they knew  
that this, the enzyme is quite -- you know, the enzyme is able to  
make a difference at that position when there is a hydrogen at that  
position versus OH. The hydrogen being in a deoxyribose  
nucleotides and being used for DNA, for example.  
So clearly, the OH is an important group at that position.  
MR. JUSTICE ANNIS: I am not following that, I would like to  
understand that better. I haven't heard anybody talk about the H  
and the OH yet -- H is the DNA, right, OH is for the --  
THE WITNESS: Correct. We have discussed the ribonucleotides is  
the OH, at that position. And in the deoxy --  
MR. JUSTICE ANNIS: It is H.  
Page: 113  
THE WITNESS: It is an H, yes. And these substrates are used for  
DNA synthesis, and now we are talking virus that has an RNA  
genome. So even if there is in the cell a lot of deoxy nucleotide pull  
MR. JUSTICE ANNIS: But would methyl H, did it work, did it  
have antiviral activity?  
THE WITNESS: The H?  
MR. JUSTICE ANNIS: Methyl H, yes.  
THE WITNESS: No.  
MR. JUSTICE ANNIS: It doesn't work at all?  
THE WITNESS: No, that is what I was getting at, the enzyme is  
really able to discriminate in between the OH and hydrogens just  
because it has an RNA genome, and it will not use the H (down),  
which is, you know, there is a lot of pull in the cells, so it is able to  
make that discrimination between the two.  
So the 2' position is very, you know, there is what we call limited  
SAR in that position.  
MR. JUSTICE ANNIS: What is SAR?  
THE WITNESS: Structure activity relationship. When the  
chemists are exploring --  
MR. JUSTICE ANNIS: Structure?  
THE WITNESS: Structure activity relationship. That is a term that  
we use in medicinal chemistry.  
So at that position, the fluorine, that is the importance for this  
patent, the fluorine is considered as a very good isosteric mimic  
because of the nature of the fluorine, because of the size of that  
atom which is close to the, has a similar size than the OH, and to  
the fact that this fluorine atom, when attached to the carbons at that  
2' position, the length of this C-F is similar to the length of a COH  
or the CO, hydrogens.  
So you have an atom of similar size at the proper place, at this  
place, which is a good mimic for that type of inhibitor.  
[Emphasis added]  
Page: 114  
[341] The analytical concept of the structure activity relationship appears to be relevant to this  
case, or at least Idenix’s argument based on mimicry of the structure so as to produce activity by  
the analog nucleoside.  
[342] It is from this passage that I made the above conclusion that Dr Lamarre agreed with Dr  
Seeger’s general proposition that the 2’ (down) position on the RNA nucleoside is highly  
selective. I believe that this statement was also supported by Dr Lamarre’s references to their  
being “a lot of pull” and that “this is what we call limited SAR in that position.”  
[343] In other words, his testimony appears to be that the 2’ (down) position on the ribose ring  
is highly selective to structure, which explains why fluorine would have the best chance of  
replacing the hydroxyl and still showing activity - despite the limited SAR and pull in that  
position.  
[344] I found Dr Lamarre to be a very fair witness. I am satisfied that he did his best in giving  
his opinion, but was careful not to overreach. He knew he could not go any further and left it at  
that. That said, I do not find in the circumstances however, based on his own carefully nuanced  
evidence, that he makes the case of antiviral activity being a probability from a fluorine  
substitution of the 2’ (down) hydroxyl on the ribose ring.  
[345] To convince the Court more evidence was needed from the literature and perhaps more  
evidence regarding the structure activity relationship concept. If it is common ground that the 2’  
(down) position on the ribose ring has “limited SAR”, it suggests that there are comparative  
Page: 115  
measures of “pull” that may be used to measure activity. In any event, that is speculation, but the  
lack of probative evidence to support Idenix’s claims is not.  
[346] In addition, the McAtee article appears to support the conclusion that one still has to  
worry about toxicity, even if the nucleoside analogue proves active, as was the comment to the  
same effect made at the Idenix Maui meeting. The fact that Idenix had to rely upon what appears  
frankly to be an unhelpful article in 2009 to make an argument based on predictable steric  
mimicry in 2003, is indicative of the absence of any solid documentation, and therefore  
experimental data underpinning a claim that utility would be probable as a result of a fluorine  
substitution of the hydroxyl in the 2’-C-Me/OH nucleoside.  
[347] That is not unexpected for a compound, yet to be invented, that Idenix is predicting can  
fool a highly selective and essential chemical position in medical science, to achieve therapeutic  
activity against a most pernicious virus.  
(ii)  
Dr Seeger’s Opinion on the Conserved Nature of the 2’ (down)  
Hydroxyl Group on the Sugar Ring  
[348] Where I find Dr Lamarre’s evidence to fall short of the requirement to establish a sound  
prediction, (which is not Idenix’s legal onus), I find that Dr Seeger’s opinions, already presaged  
by the passages cited from Dr Lamarre’s testimony, discharge Gilead’s onus in referencing the  
selectivity of the 2’ (down) hydroxyl group in distinguishing between RNA and DNA analogues.  
Page: 116  
[349] Dr Seeger’s evidence basically comes down to predicting any change in the hydroxyl  
group as “a big change at an important place on the RNA.” His evidence is probably best  
highlighted in the cross-examination conducted challenging some of his views, which I find did  
not succeed in attaining its objective, but not for lack of trying.  
[350] Dr Seeger points out that any change to any compound will change its properties. Starting  
with a methyl (up) and the hydroxyl (down) that provide the activity in the 2’-C-Me/OH  
nucleoside does not change that.  
Q.  
They have advanced their knowledge with respect to that  
position in a compound that has 2' methyl (up)?  
A.  
No. The hydroxyl is there. They leave the hydroxyl natural.  
The expectation is in fact, we discussed it this morning a change  
there will probably be dramatic or it will be profound.  
Nevertheless, we can try. Why not? Let's figure it out. Really, the  
methyl is methyl(up). That is the methyl(up), and we are talking  
about the hydroxyl (down) here. The hydroxyl is still here. It still  
has the structure of the RNA, if you wish with that methyl(up).  
Q.  
That is ground zero for them. It is not any given compound.  
You starting with methyl(up) and OH(down). You have a body of  
knowledge with respect to that?  
A.  
Here it is not a matter of addition. In other words, you are  
not saying, “Now I am having A here, and then I go from A plus B  
equals this.” Every molecule you make has now its own new  
property. If you change one atom, it has an influence on the whole  
molecule; right? That is why everyone is new whether you start  
with a methyl or with the natural base or whatever you do. You are  
creating something new.  
[Emphasis added]  
Page: 117  
[351] Dr Damha said much the same thing, including the variability that fluorine can introduce  
into compounds (page 2026-7):  
THE WITNESS: So when you replace an OH group with fluorine,  
OH is a polar functional group. Fluorine is less polar, yet is  
electronegative like oxygens. So for that reason, when you do a  
replacement of an OH to a fluoro, you are kind of keeping the same  
type of electronegativity of the group, but you are altering it  
slightly to endow the molecules with different properties.  
And so I give references to the fact that many fluorinated organic  
compounds exhibit interesting properties, as has been shown the  
case for many type of molecule, steroids, nucleoside.  
[352] Justice Michael Phelan noted in Alcon Canada Inc v Cobalt Pharmaceuticals Company,  
2014 FC 462 [Alcon/moxifloxacin] at para 160 that even small changes can have profound  
effects on activity.  
[160] The importance of the stereochemistry of the Claim 12  
compounds was discussed by a number of experts. It appears to be  
common ground that even small changes to a molecule can have  
profound effects on activity.  
[Emphasis added]  
[353] Dr Seeger’s view is that that the change at the hydroxyl group is a “big change at an  
important site” [Emphasis added], the point being that you just do not know the nature of its  
impact:  
A.  
The polymerase as a family, they have this distinction.  
There are structural indications of why it is so, and I could expand  
on it. I won't because it will take some time. That is exactly the  
point, this distinction. That is why, when you change this residue,  
Page: 118  
you don't know whether it works at all, whether it goes into RNA,  
or whether it goes into DNA. It must have an impact.  
…
Q.  
This is your conclusion reading this cite?  
A.  
No, that is not the conclusion from me. Again, I am not  
manipulating the building blocks of nature; right? There are just  
two kinds. The only distinction is that hydroxyl. If you go and  
change that chemically, this must have an impact.  
…
Q.  
Can we just look at the next paragraph of your affidavit  
where you say it can have a significant impact. You are saying,  
"modifications in the 2’ position can have a significant impact," not  
must?  
A.  
I can say "can." I can say "must." I just submit, it is the  
most conserved region. It is an important distinction. You change  
it, the expectation is it must have an impact. It can. It must. You  
name it. This is a big change at the very important site.  
[Emphasis added]  
[354] Dr Seeger’s summary opinion was that the results of a change at the 2 hydroxyl position  
is unpredictable, the only way to find out is by testing:  
Q.  
Thank you, Dr Seeger. Given the sensitivity that exists at  
the sugar ring and in particular, as you have mentioned, at the 2'  
position, what would the Skilled Virologist's understanding be as to  
whether a factual basis or a Sound Prediction could be made if a  
modification is made to the sugar ring?  
A.  
As I said, there is no way you can make any prediction.  
You have to do the experiment and find out what the activity of  
that particular compound is.  
Page: 119  
(iii)  
Dr Patterson’s Opinion on Predictability of Activity  
[355] Idenix did not hesitate to seek opinions on the issue of the prediction of activity from Dr  
Patterson, although not called as an expert and without notice in its witness statement. Gilead  
objected to this evidence, which I nevertheless allowed. I concluded that Dr Patterson’s evidence,  
like that of Dr Stuyver, objected to later by Idenix, was relevant coming from someone who was  
an active participant in the historical events, as opposed to experts opining on those events a  
decade later.  
[356] The important passages of Dr Patterson’s evidence are, as follows:  
Q.  
A.  
And why were you reasonably optimistic?  
Well, we have replaced a hydroxyl with the fluorine. They  
are similar in size, shape, they have similar electronic structure, so  
it’s, you know, if, you know, I think it’s reasonable to expect that it  
would be active.  
[…]  
Q.  
A.  
Q.  
And other people substituted hydroxyl for fluoro?  
Yes. Dr Watanabe was pretty well known for doing that.  
And why would that have been done?  
MR. MARKWELL: Objection [already ruled upon].  
THE WITNESS: Well, we like to talk about isosteres in medicinal  
chemistry. I prefer the term “bioisosteres”, but it’s a similar  
concept. So if you have a biological molecule, let’s say a peptide or  
an amide, we have these structures that are similar, though not  
necessarily identical, but similar in size, shape, and electronics, and  
very often, not always but very often, they are accepted by the  
biological systems. [p. 1780]  
Page: 120  
[357] Idenix also relies upon Dr Patterson’s testimony that fluorine in the 2’ (down) position  
would be his “number one pick” for developing an analogue of the Idenix Compound, which was  
how Pharmasset described the starting material. I point out that a number one pick of a chemical  
reaction is not evidence of a sound prediction, but I also consider this aspect of his testimony,  
which is similar to that of Dr Damha when discussing the sufficiency of disclosure of synthesis.  
[358] Dr Patterson’s evidence on the prediction of utility was more nuanced than simply argued  
by Idenix. He presented a long discourse on the problems with predicting outcomes of nucleoside  
analogues generally.  
THE WITNESS: So in that context, I was optimistic, but I did not  
share his absolute certainty.  
MR. JUSTICE ANNIS: Because there is some problem with  
nucleosides?  
THE WITNESS: Yes, sir, a nucleoside has to be a substrate of  
multiple enzyme systems, so it’s very difficult to predict whether a  
particular nucleoside is active. They are, in effect, prodrugs, so  
they have to be transported, usually transported into cell and then,  
you know, converted to the monophosphate, then the diphosphate,  
then the triphosphate, with some exceptions. And it is that  
triphosphate that is the active metabolite. So because of all these  
enzyme systems, it is not like you just hit a single system where  
it’s much easier to predict, right.  
…
THE WITNESS: Right, so one of the ways that we medicinal  
chemists suspect whether a molecule, an active molecule works  
against a single target or multiple targets is we ask, does that  
particular molecule follow a discernible structure activity  
relationship? Can we, with a few actives, predict the activity of,  
you know, analogues, will it be more or less active?  
…
Page: 121  
THE WITNESS: Nucleosides rarely do that. And that is because  
they have to act, they have to be activated by no fewer than three  
enzymes to get the active metabolite, and that active metabolite  
then has to be a substrate for the viral polymerase. So in that  
context I was optimistic but I did not share his absolute certainty.  
[Emphasis added]  
[359] Dr Patterson also makes an admission of sort of the medical chemist’s blinkered thinking  
in his testimony as follows:  
I frequently tell people what organic chemists understand best is  
structure, and some people say organic chemists only understand  
structure, and I think that is what we need are some structures on a  
board or something.  
[360] There was very little evidence relating to the 2’-C-Me/OH molecule that one could say  
with confidence that there was a discernible structure activity relationship, which subject was  
hardly mentioned in any of the witness’s testimony. Nor were there any “actives” to speak of to  
predict the utility of the 2’-C-Me/F compound.  
[361] However, the main difficulty I have with Dr Patterson’s opinion evidence was that it  
misses the point of Dr Seeger’s evidence about the selectivity of the 2’ (down) position on the  
RNA and DNA and related issue of the toxicity of fluorines. Dr Patterson spoke of the  
difficulties of predicting outcomes in respect of nucleosides generally. This fails to take into  
consideration to what extent his optimism should have been attenuated when applied to a  
substitution at the most conserved and selective position in chemical biology. It also fails to  
Page: 122  
consider the issue of fluorine often being toxic. In fairness, he was not called to offer these  
opinions.  
(iv)  
Dr Stuyver on the Unpredictability of Toxicity and Activity of the  
Fluorine Compound  
[362] Dr Stuyver, called by Idenix also provided an opinion in cross-examination on the  
predictability of antiviral activity of the new compound with the fluorine at the 2’ (down)  
position before its synthesis, as follows:  
Q.  
Would you agree that if you take the structure that is  
Structure Number 3 on this page here, when you set off on a  
synthesis route, who can tell from looking toward this structure that  
this is going to act as an antimetabolite [that may have a toxic  
effect on cells]?  
A.  
Let me make sure I understand your question. So you ask  
me if from looking to this molecule, this structure, whether you can  
make a prediction whether it is an antimetabolite; is that your  
question?  
Q.  
A.  
Q.  
Yes.  
You can't.  
Is it your view that in the beginning of 2003, looking at this  
structure, you can tell that the drug is going to be a candidate direct  
antiviral?  
A.  
But in the beginning of 2003, this was not a drug, right.  
And in 2003 you make, and in 2002, you make plans, hypothesis,  
you synthesis, and you test. But you cannot say this is -- at that  
moment, you cannot say this is going to be antiviral or  
antimetabolite. [PP 3998 – 3999]  
[Emphasis added]  
Page: 123  
[363] Idenix quotes Dr Stuyver from an article in 2004, after the 2’-C-Me/F nucleosides had  
been synthesized, noting that “fluorine is isosteric with a hydroxyl group” and that antiviral  
activity from the substitution is anticipated as follows:  
Since FdC [the invention] is an analogue of cytidine and fluorine is  
isosteric with a hydroxyl group, it is anticipated that its  
5’triphosphate would inhibit the viral NS5B RNA-dependent RNA  
polymerase.  
[364] If this statement is intended to show that Dr Stuyver has resiled from the inability to  
predict the antiviral outcome prior to testing on the new 2’-C-Me/F nucleoside, I conclude that it  
proves the foresight of hindsight when speaking publicly about an invention claimed to have  
been invented by the speaker, and who also claims to have directed the chemist inventor to make  
the compound. I prefer to attribute the meaning of “educated guess” to the term “anticipated”  
from the passage quoted. This article was not put to Dr Stuyver for confirmation purposes,  
although called as Idenix’s witness and having provided the opinion above on the  
unpredictability of the outcome of the utility of the new compounds without testing.  
(d)  
The Cytotoxicty of Fluorines  
[365] I also find that there was no distinction in the inability to predict whether a new fluorine  
compound would be toxic at either the 2’ (up) or (down) position on the RNA nucleoside  
analogue. In this respect, Gilead also points out that the inventors were aware in a companion  
patent filed also on the same date as the ‘191 Patent of June 23, 2003 that there had been a recent  
death in trials of a compound described as FIAU. It was compound with a fluoro in the 2’ of  
Page: 124  
position. This fact was not contained in the ‘191 Patent. Professor La Colla was aware of toxicity  
concerns with FIAU and other 2’-fluoro compounds.  
[366] Given the evidence of the unpredictability of fluoro compounds exhibiting toxicity, I  
conclude that this was another factor making it impossible to predict whether a new fluoro  
compound will exhibit toxicity without testing.  
(e)  
Alcon/moxifloxacin  
[367] As a final issue, I am required to consider the recent decision of Justice Phelan in  
Alcon/moxifloxacin which Idenix claims applies to the facts in this case and should be followed. I  
disagree. The facts in Alcon/moxifloxacin are highly distinguishable from the novel  
circumstances presented in this matter.  
[368] The 114 Patent claimed a class of quinolone compounds, which includes moxifloxacin,  
characterized by a fused pyrrolidine bicycle at the C-7 position of the quinolone ring.  
[369] The core structure of the quinolones is:  
Page: 125  
[370] The Applicant claimed that the utility of a compound substituting a methyl for a fluorine  
compound at the C-8 position could be soundly predicted, although the compound had not been  
synthesized at the time of filing.  
[371] The Applicant relied on the common general knowledge that compounds with a methoxy  
group at the C-8 position were active. This included European Patent Application No 0 241 206  
[Sankyo] and European Patent 0 230 295 A2 [Kyorin] as examples of prior art teaching that  
compounds with a methoxy group at the C-8 position had good activity. The Court found that the  
inventors therefore, had a sound basis to predict that substituting a methoxy group at that position  
for the fluorine in the example cited in the Patent would either enhance or not hinder the high  
antibacterial activity demonstrated in the example.  
[372] The most obvious distinction between Alcon/moxifloxacin and this matter is that  
compounds with methyl at the C-8 position had previously been synthesized and shown active by  
testing. Obviously, neither Idenix, nor any other drug discovery organization possessed testing  
data on the 2’-C-Me/F nucleosides or a comparable compound with activity of a fluorine at the 2’  
(down) position at the time of filing, except Gilead, which was not otherwise known at the time.  
[373] Another important distinction between the two cases is that there is no comparing the  
predictive outcome from a highly substituted core compound in existence since 1960, with the  
first ever substitution of an element known for its toxicity as a replacement for the most selective  
and highly conserved group in a biological compound.  
Page: 126  
[374] In Alcon/moxifloxacin the Court was dealing with a core compound whose usefulness had  
been demonstrated in administration to humans since 1960. It was not an entirely new compound  
like the 2’-C-Me/OH nucleoside that was just entering the first level of clinical trials, whose  
utility remained to be fully established (and which was eventually withdrawn). This case is quite  
distinguishable from that in Alcon/moxifloxacin.  
(f)  
The Last Word to Wellcome/AZT  
[375] I find Justice Binnie’s comments in Wellcome/AZT at paras 63 and 64 highly apposite in  
noting the distinction between the predictability of chemical reactions and the predictability of  
the pharmacological effects, and thus of the pharmacological utility of new substances:  
63  
Our Federal Court of Appeal subsequently applied the  
doctrine of “sound prediction” in the context of a patent for a  
pharmaceutical product in Ciba-Geigy AG v. Commissioner of  
Patents (1982), 65 C.P.R. (2d) 73. In that case, Thurlow C.J.  
upheld product and process claims in relation to certain “new  
amines” useful in cardiac treatment, but added the qualification  
that what is predictable chemically may not be predictable  
pharmacologically, at p. 77:  
The predictability of a particular result seems to me  
to be essentially a question of fact, though in some  
situations it may be a matter of common knowledge.  
With respect to chemical reactions it is apparent  
from the foregoing that knowledge in the chemical  
art as to the predictability of chemical reactions has  
advanced considerably in the 50 years since  
Chipman Chemicals Ltd. v. Fairview Chemical Co.  
Ltd., [1932] Ex. C.R. 107, was decided. The  
predictability of chemical reactions should not,  
however, be confused with predictability of the  
pharmacological effects and thus of the  
pharmacological utility of new substances.  
Page: 127  
[Justice Binnie’s emphasis]  
64  
Thurlow C.J. was not laying down as a matter of law that  
pharmacological utility cannot be predicted because, as he said,  
predictability is “essentially a question of fact”. It will depend on  
the evidence. In Beecham Group Ltd. v. Bristol Laboratories  
International S.A., [1978] R.P.C. 521 (H.L.), for example, claims  
in respect of a semi-synthetic penicillin were invalidated as being  
little more than an announcement of a research project (p. 570). In  
that case, on the facts, Lord Diplock stated at p. 579:  
The evidence in the instant case is overwhelming  
that it is not yet possible to predict in advance what,  
if any, special therapeutic advantages will be  
possessed by a penicillin made to a particular  
formula. The only way to find out is to make it and  
discover what its therapeutic characteristics are by  
conducting extensive tests upon it in vitro and in  
vivo.  
[376] I find the evidence bordering on overwhelming that it would not have been possible to  
soundly predict in advance the therapeutic advantages of substituting a fluoride for the hydroxyl  
on the 2’-C-Me/OH nucleoside. I agree with Drs Seeger and Stuyver, and implicitly, the Idenix  
chemists at their Maui meeting that the only way to find the 2’-C-Me/F nucleoside’s therapeutic  
value was to make it and discover what its therapeutic characteristics are by reliable and valid  
testing. This is what Gilead did, and Idenix did not.  
(4)  
Conclusion on the Sound Prediction of Utility of the 2’-C-Me/F Compound  
[377] For all the reasons above, I find that Gilead has demonstrated that Idenix could not  
demonstrate, nor soundly predict as an inferred fact or prima facie reasonable inference, the utility  
of the 2’-C-Me/F nucleoside on June 27, 2003 as an effective treatment of HCV.  
Page: 128  
C.  
Disclosure of Sound Predictions of Utility in the Patent  
(1)  
The Requirement to Disclose Utility for New Use Inventions  
[378] The parties disagree on the requirement stated in the jurisprudence to disclose the factual  
basis and line of reasoning for the sound prediction in the patent. Idenix, relying on the analysis  
of Justice Rennie, as he then was, in AstraZeneca/Pregabalin limits the disclosure requirement of  
soundly predicted utility to “new use” inventions.  
[379] Gilead, argues that there remains some wiggle room in AstraZeneca/Pregabalin and the  
subsequent Court of Appeal decision of Bell Helicopter Textron Canada Limitée v Eurocopter  
Société, 2013 FCA 219 [Bell Helicopter]. I disagree that Justice Rennies’s reasoning leaves much  
to the imagination on the necessity to disclose the utility for a “new compound” invention. He  
states at para 158 of AstraZeneca/Pregabalin that he was “compelled [my emphasis] to follow  
the Supreme Court’s remarks in Teva sildenafil and the interpretation of Wellcome/AZT endorsed  
by Justice Gauthier in Apotex Inc v Sanofi-Aventis, 2013 FCA 186 (), para 42 [Sanofi-  
Aventis Plavix]).”  
[380] I adopt Justice Rennie’s reasoning, which I find properly reflects the binding  
jurisprudence on our Court. Accordingly, as the ‘191 Patent relates to a sound prediction of a  
new composition, there is no utility disclosure requirement. I assume that this exemption from  
disclosure applies both to the factual basis and the line of reasoning.  
Page: 129  
[381] Despite my conclusion that Idenix is not required to disclose the factual basis for its  
sound prediction of utility, I nevertheless, am required to determine whether its disclosure was  
sufficient should the law on this issue change in the future, which admittedly is contentious:  
Pfizer Canada Inc v Pharmascience, 2013 FC 120 at para 157; AstraZeneca/Pregabalin (which  
is under appeal).  
(2)  
Common General Knowledge in Disclosure of Utility  
[382] The jurisprudence has recognized a dispensation from disclosure of all or elements of a  
sound prediction if it is a matter of common general knowledge. This applies to both disclosure  
of the factual basis and the line of reasoning. The law was described by Justice Noel, as he then  
was, in Bell Helicopter at paras 153-155, as follows:  
[153] Where the factual basis can be found in scientifically  
accepted laws or principles or in information forming part of the  
common general knowledge of the skilled person, then no  
disclosure of such factual basis may be required in the  
specification. On the other hand, where the factual basis is reliant  
on data which does not form part of the common general  
knowledge, then disclosure in the specification may indeed be  
required to support a sound prediction.  
[154] As noted in the Manual of Patent Office Practice issued by  
the Canadian Patent Office (at paras 12.08.04b and 12.08.04c),  
since a sound line of reasoning is directed to a skilled person, those  
elements of the doctrine of sound prediction that would be self-  
evident to that person in view of the common general knowledge  
need not be explicitly disclosed in the specification. The soundness  
of a line of reasoning can also be effectively assessed by asking  
whether the skilled person would accept the logic presented in the  
specification and derive from the sound prediction as a whole an  
expectation that the invention will provide the promised utility.  
Page: 130  
[155] As a result, where the sound prediction is based on  
knowledge forming part of the common general knowledge and on  
a line of reasoning which would be apparent to the skilled person  
(which is often the case in mechanical inventions), the  
requirements of disclosure may readily be met by simply  
describing the invention in sufficient detail such that it can be  
practiced. A contextual approach is thus appropriate in each case.  
(3)  
Is the Disclosure by Idenix Adequate?  
[383] Apart from any reliance on the common general knowledge and assuming disclosure is  
required, the issue here to be considered is whether the ‘191 Patent describes a factual basis and  
sound line of reasoning to soundly predict antiviral activity in the 2’-C-Me/F nucleoside. As  
Gilead concentrated on the disclosure of the factual basis of testing the 2’-C-Me/OH nucleosides  
being surrogates of HCV, I will limit my comments to this issue as well. It is further narrowed to  
the question of Idenix disclosing its testing data on the 2’-C-Me/OH nucleoside as part of its  
factual basis, as well its factual basis to claim that the BVDV and other assays may be used as  
surrogates for demonstrating antiviral activity against HCV.  
(a)  
Disclosure of Factual Basis of Idenix’s Testing Data on the Me/OH  
Compound  
[384] Idenix claims it disclosed testing data in the ‘191 patent in Examples 25 and 26 and by  
reference to its prior PCT applications Nos. WO 01/09121 [Application ‘121] and WO 01/92282  
[Application ‘282].  
Page: 131  
(i)  
Examples 25 & 26  
[385] Example 25: Example 25, while it describes various tests including those that could be  
used to assess the antiviral activity of the compound against a specific virus, contains no data.  
[386] Example 26: Example 26 discloses experimental data in respect of Compound F. Gilead  
argues that Compound F is not within the scope of the claims and that the data is flawed.  
[387] The structure of Compound F is not disclosed in the ‘191 Patent. Moreover, Idenix’s  
claims are based upon Compounds NM107, NM283, which were not publicly disclosed.  
Compound F is also not the same compound used by Idenix as the basis for its claims in the ‘191  
Patent. Compound F has an unnatural and highly modified tricyclic base. The structures of  
Compound F in comparison with that of Idenix’s NM 107 compound are depicted below. The  
data from Compound F is not properly disclosed and is not relevant therefore to the 2’-C-Me/F  
nucleoside.  
[388] The data in Example 26 on Compound F was displayed in two tables, one above the  
other, as follows:  
Page: 132  
[389] There are obvious problems with the tables. While the chemical formula is provided,  
there is no structure for the compound necessary to understand it stereochemical properties  
around which structure activity relationships apply.  
[390] The EC50 values in the bottom row of the first table are said to reflect assays done in  
quadruplicate or triplicate. They are expressed in single numbers when there should be averages  
or a standard deviation. Moreover, the table does not describe which assay was used. The  
specification describes two different assays by which a compound’s EC 50 could be calculated  
that measure different aspects of antiviral activity. The skilled virologist would want to know  
which assay was conducted in order to be able to properly interpret the data presented. Dr Seeger  
testified that in the absence of experimental details, there was little useful information provided  
in the first table. I accept his testimony on these points.  
Page: 133  
[391] The second table contains no headings in the top row. EC 50 values are repeated in the  
table already referred to in the first table, where the second table is said to relate to CC 50 values  
on toxicity. Cell line information is absent, which is significant as values differ depending upon  
cell lines. Dr Seeger concludes that the tables in Example 26 do not clearly show any  
corresponding cytotoxicity data in respect of the cell lines that were used in the antiviral activity  
assays. It is therefore, not possible to determine the therapeutic index for Compound F.  
[392] Dr Lamarre’s opinions expressed in his report on the subject and his brief testimony do  
not deal with the majority of issues raised by Dr Seeger. Specifically, Dr Lamarre did not  
challenge Dr Seeger’s conclusion that it was not possible to determine the therapeutic index for  
Compound F.  
(ii)  
Patent Application ‘949  
[393] Dr Lamarre relied on data from Patent Application ‘949 which disclosed Idenix’s testing  
data to demonstrate activity in its 2’-C-Me/OH compounds. However, Idenix amended its ‘191  
Patent Application, including striking the introductory paragraph at page 1 by which specific  
reference to Application ‘949 was removed. As far as the Court may determine by removal of  
this paragraph, no further reference can be found in Patent ‘191 to Patent Application ‘949. This  
is assumed to be the reason that Idenix did not attempt pursue reliance on the data from this  
Application during argument.  
Page: 134  
(iii)  
Applications ‘121 and ‘282  
[394] Idenix claims these applications disclose antiviral activity data for a number of 2’-C-  
Me/OH nucleosides against both BVDV and YFV. Gilead did not make submissions with respect  
to the testing data in these applications.  
Idenix Pharmaceuticals discloses the use of branched nucleosides  
in the treatment of flaviviruses (including HCV) and pestiviruses in  
International Publication Nos. WO 01/90121 and WO 01/92282.  
Specifically, a method for the treatment of hepatitis C infection  
(and flaviviruses and pestiviruses) in humans and other host  
animals is disclosed in the Idenix publications that includes  
administering an effective amount of a biologically active, 1', 2', 3'  
or 4'-branched ß-D or ß-L nucleosides or a pharmaceutically  
acceptable salt or derivative thereof, administered either alone or in  
combination with another antiviral agent, optionally in a  
pharmaceutically acceptable carrier.  
[395] In its argument challenging Idenix’s reliance on Patent Application ‘949, Gilead  
submitted that Idenix is not entitled to incorporate data by reference because this is explicitly  
prohibited by the Regulations made under the Act. As it turns out, Idenix did not rely on  
application ‘949.  
[396] However, Application ‘282 discloses anti-BVDV in anti-YFV activity for the 2’-C-  
Me/OH nucleosides. Dr Lamarre describes this testing evidence in his November 2014 report as  
follows:  
In particular, the 2'-Me(up), 2'-OH(down) compounds were tested  
using all four natural bases against the flavivirus (yellow fever  
Page: 135  
virus) and pestivirus (bovine viral diarrhea virus) prototypes for  
their genera. A spectrum of antiviral activity is shown on pages  
186, 190, and 191 of WO 01/92282. From the data and structures  
described in WO 01/92282, the skilled person would understand  
that the key aspect of this nucleoside analogue is the 2'-Me(up)  
substituent because all other substituents are the same as a  
naturally occurring ribonucleoside. Further, based on the data  
provided, the skilled person would expect similar antiviral activity  
for other Flaviviridae. The antiviral data for the flavivirus and  
pestivirus prototypes is tabulated below [which is not reproduced],  
which clearly demonstrates the antiviral activity of 2'-Me(up), 2'-  
OH(down) compounds:  
[397] I find that the ‘191 Patent, by reference to the ‘282 Application and its purpose for doing  
so, has sufficiently disclosed the antiviral testing data of 2’-C-Me/OH nucleosides for the  
flavivirus (yellow fever virus) and pestivirus (bovine viral diarrhea virus). I do not agree  
however, with Dr Lamarre’s comment that “based on the data provided, the skilled person would  
expect similar antiviral activity for other Flaviviridae”, which is intended to encompass the HCV.  
(iv)  
Disclosure of Factual Basis to Claim that BVDV May Be Used As a  
Surrogate for Demonstrating Antiviral Activity against HCV  
[398] Idenix has not disclosed its testing data on chimpanzees in the ‘191 Patent. This would  
have satisfied the requirement to disclose its testing data regarding HCV. It has already been  
noted that none of Idenix’s data was based upon the replicon assay, which the skilled virologist  
would have expected in 2003 to demonstrate antiviral activity against HCV.  
[399] The ‘191 Patent discloses genetic similarities and protein function similarities amongst  
members the Flaviviridae family and at page 2 specifically provides that “bovine viral diarrhea  
virus (BVDV) is often used as a surrogate to study the HCV virus” [my emphasis]. At page 3 of  
Page: 136  
the ‘191 Patent references are made to the Koonin and Dolja (1993) article discussed by  
Professor La Colla in relation to common motifs characteristic of RNA directed RNA  
polymerases.  
[400] I do not find this information sufficient however, to disclose a factual basis to claim that  
BVDV may be used as a surrogate for demonstrating antiviral activity against HCV.  
(v)  
Common General Knowledge  
[401] The question here is whether the failure to disclose the utility in the Patent can be  
compensated for by the common general knowledge. Idenix advances the following materials as  
demonstrating that it was common general knowledge in June 2003 that BVDV was an adequate  
surrogate to test for HCV.  
1. Slide Presentations of Bhat et al., Eldrup et al., and Olsen et al.  
from the 16th International Conference on Antiviral Research  
(April 27, 2003, Savannah, Ga.)  
[402] These documents, which were disclosed on page 10 of the ‘191 Patent and described  
above, consist of slide presentation decks used as visual backups for the oral presentations. The  
presentations precede the application date of June 27, 2003 by just over two months.  
[403] Dr Seeger states in his December Report that presentation slides from scientific  
conferences would typically not be published and the skilled virologist would typically not have  
been provided with copies of them. Based solely on the information provided in the ‘191 Patent,  
Page: 137  
a skilled virologist would not have known whether the abstracts had been published. A skilled  
virologist would have had to perform a literature search to find this out. As well, Dr Seeger  
points out that the published abstracts themselves do not provide any data or structure or  
information about the nature of the nucleosides tested. Although some presentation abstracts can  
be published in conference newsletters or conference abstract books, the information that can be  
gleaned from an abstract is limited and not peer reviewed. In this case the only information  
mentioned was that “nucleoside analogs modified at the 2’-position” had shown antiviral activity  
using the replicon HCV assay.  
[404] The abstracts attached to the conference program were of no assistance as they only made  
reference to undisclosed nucleoside analogs modified at the 2’-position being found to inhibit  
synthesis of viral RNA in a cell-based replicon assay in the absence of cytotoxicity, and nothing  
else of relevance.  
[405] Idenix is unable to provide evidence of any weight as to the availability of the documents  
in the form of details as to even when or how it received a copy of the presentation slides, or  
even their general availability at this time a decade later. There is no indication that the abstracts  
were made available to Idenix, other than that they are now in its possession. The person who  
attended the conference, Dr Storer, was not called. He sent an email to Dr Standring, but without  
attachments. I suspect the only person who could have testified on receipt or public availability  
of these documents was Dr Storer.  
Page: 138  
[406] Accepting Dr Seeger’s evidence that the information orally presented based on the  
Savannah slide presentations which were not distributed, but referred to in the ‘191 Patent, I  
conclude that this information does not meet the requirement of section 81 (2) of the documents  
being available to the public on June 27, 2003. I also conclude that they do not represent  
common general knowledge either in content, timeliness or evidence of general acceptance.  
[407] Because there are no publicly available documents disclosing the factual basis that  
BVDV assays may be used as a surrogate for demonstrating antiviral activity against HCV, the  
‘191 Patent does not disclose a key element of its factual basis for a sound prediction of utility.  
Idenix has made no attempt to claim that BVDV testing as surrogates for testing for HCV is  
common general knowledge, nor is there any evidence that it is.  
2. Carroll et al Article  
[408] This article was not mentioned in the ‘191 Patent. There is also no evidence that the  
content of the paper by Carroll et al was “generally known and accepted” by a majority of those  
working in the field. Rather, the Carroll Paper was published on April 4, 2003, less than three  
months before the filing of the ‘191Patent. Further, the mere fact that something was published is  
not sufficient to establish that it formed part of the common general knowledge. While I have  
admitted the document in respect of Professor La Colla having knowledge of it, the fact remains  
that there is no contemporaneous documents emanating from the inventors to establish that this  
paper was known and accepted by the person skilled in the art as part of the common general  
knowledge. I conclude that it was not part of the common general knowledge.  
Page: 139  
[409] The Court concludes that it was not common general knowledge in June 2003 that BVDV  
was an adequate surrogate to test for HCV.  
(b)  
Disclosure of Sound Line of Prediction  
[410] Gilead has not challenged the lack of disclosure of the sound line of reasoning claimed by  
Idenix based on steric mimicry of the fluorine of the hydroxyl group in a substitution.  
Accordingly, I will leave it for another occasion whether a reference as brief as that at page 10 of  
the ‘191 Patent to the Emory Patent application could constitute sufficient disclosure of a line of  
reasoning of steric mimicry.  
PCT Publication No. WO 99/43691 to Emory University, entitled  
“2'-Fluoronucleosides” discloses the use of certain 2' -  
fluoronucleosides to treat HCV.  
[411] The passage referred to in the Emory Patent application is as follows:  
In designing new biologically active nucleosides, there have been a  
number of attempts to incorporate a fluoro substituent into the  
carbohydrate rmg of the nucleoside. Fluorine has been suggested as  
a substituent because it might serve as an isopolar and 15 isosteric  
mimic of a hydroxyl group as the C-F bond length (1.35 A) is so  
similar to the c-o bond length (1.43 A) and because fluorine is a  
hydrogen bond acceptor. Fluorine is capable of producing  
significant electronic changes in a molecule with minimal steric  
perturbation The substitution of fluorine for another group in a  
molecule can cause changes in substrate metabolism because of the  
high strength of the C-F bond (116 kcal/mol vs. C-H = 100 20  
kcaUmol).  
Page: 140  
(c)  
Conclusion  
[412] The Court concludes that Idenix has not disclosed a sound prediction of utility in the 2’-  
C-Me/F nucleoside as it had not disclosed its factual basis for its position in June 2003 that the  
2’-C-Me/OH nucleosides had antiviral activity against HCV.  
VIII. Sufficiency of Disclosure of the Synthesis of the Invention  
A.  
Introduction and Legal Principles  
(1)  
Introduction  
[413] The principal, issue, which takes up almost all of the discussion in this section is whether  
the ‘191 Patent sufficiently discloses how to make the 2’-C-Me/F nucleoside. I conclude that  
Gilead has established that the ‘191 Patent as supplemented by the common general knowledge  
does not sufficiently disclose how to synthesize the 2’-C-Me/F compound.  
[414] This determination takes all of Idenix’s arguments into consideration and is based on my  
factual findings, including preferring the evidence of Gilead’s experts to that of Idenix’s. As a  
result, it is not necessary for the Court to consider Idenix’s contention that the enabling  
disclosure of the synthesis of the 2’-C-Me/OH nucleosides for the test of anticipation under  
section 28.1 is the same as the disclosure of how to make the compound under section 27(3)b) of  
the Act.  
Page: 141  
[415] Gilead also raises a preliminary argument that based upon Claim 1 of the ‘191 Patent, and  
treating all claims as one invention, there was a failure to disclose the invention as it was  
impossible to identify any single invention contrary to the requirements set out in the Supreme  
Court decision of Viagra. I reject this argument, concluding that Viagra has no application to the  
facts in this matter.  
(2)  
Legal Principles  
(a)  
Disclosure Requirements  
[416] Disclosure is an important element of the bargain between the inventor and the public. It  
is the quid pro quo for valuable proprietary rights to exclusivity which are entirely a statutory  
creature of the Patent Act. In return for the grant of the monopoly, the inventor must provide the  
public with an adequate description of the invention with sufficiently complete and accurate  
details so as to allow a workman skilled in the art to which the invention relates, to construct or  
use it Wellcome/AZT, at para 37.  
[417] The Supreme Court described the essentials of disclosure in Pioneer Hi-Bred Ltd v  
Canada (Commissioner of Patents), [1989] 1 SCR 1623 at p. 1637-38 and 1641, [Pioneer Hi-  
Bred] as follows with citations omitted:  
Canadian courts have stated in a number of cases the test to be  
applied in determining whether disclosure is complete. The  
applicant must disclose everything that is essential for the  
invention to function properly. To be complete, it must meet two  
conditions: it must describe the invention and define the way it is  
produced or built. The applicant must define the nature of the  
invention and describe how it is put into operation. A failure to  
Page: 142  
meet the first condition would invalidate the application for  
ambiguity, while a failure to meet the second invalidates it for  
insufficiency. The description must be such as to enable a person  
skilled in the art or the field of the invention to produce it using  
only the instructions contained in the disclosure…and once the  
monopoly period is over, to use the invention as successfully as the  
inventor could at the time of his application.  
[…]  
It is nonetheless clear that apart from steps which appear to be  
obvious and common knowledge for an experimenter skilled in the  
art, a person to whom the disclosure is addressed "is not required to  
exercise or to be possessed of more, and, if the specification  
contains something that necessitates the working out of a problem,  
the patent cannot be supported".  
[Emphasis added]  
[418] In determining whether Idenix complied with s 27(3), the Supreme Court in Viagra at  
paras 50 & 51provides a framework for the analysis of the sufficiency of disclosure in the form  
of three questions that must be answered:  
(a) What is your invention?  
(b) How does it work?  
(c) Having only the specification, can the person of ordinary skill in the art produce  
the invention using only the instructions contained in the disclosure?  
[419] This framework reflects the language of section 27(3)(a) and (b) of the Patent Act which  
is as follows:  
Page: 143  
27. (3) The specification of an 27. (3) Le mémoire descriptif  
invention must doit :  
(a) correctly and fully describe a) décrire d’une façon exacte et  
the invention and its operation  
or use as contemplated by the  
inventor;  
complète l’invention et son  
application ou exploitation,  
telles que les a conçues son  
inventeur;  
(b) set out clearly the various  
steps in a process, or the  
method of constructing,  
b) exposer clairement les  
diverses phases d’un procédé,  
ou le mode de construction, de  
making, compounding or using confection, de composition ou  
a machine, manufacture or d’utilisation d’une machine,  
composition of matter, in such d’un objet manufacturé ou  
full, clear, concise and exact  
terms as to enable any person  
skilled in the art or science to  
which it pertains, or with  
which it is most closely  
d’un composé de matières,  
dans des termes complets,  
clairs, concis et exacts qui  
permettent à toute personne  
versée dans l’art ou la science  
dont relève l’invention, ou  
dans l’art ou la science qui s’en  
rapproche le plus, de  
connected, to make, construct,  
compound or use it;  
confectionner, construire,  
composer ou utiliser  
l’invention;  
[420] Neither party has raised any issue about how the invention works. “What” is the invention  
shall be briefly discussed when considering Gilead’s interpretation of the Viagra decision. The  
sufficiency of the disclosure of whether the skilled chemist can make the compound based using  
only the instructions in the disclosure is obviously the primary concern of this section.  
(b)  
Common General Knowledge and Routine Experimentation  
[421] I have already addressed some aspects of the law with respect to common general  
knowledge and routine experimentation. It is common ground that both are permissible to assist  
Page: 144  
the skilled person’s task of making the invention work, (Pioneer Hi-Bred, p. 1641 supra; Teva  
Canada Limited v Novartis AG, 2013 FC 141 at para 384.  
[422] Idenix’s position on recourse to these two principles in this case is perhaps best summed  
up at para 371 of its submissions as follows:  
[371]… With respect to the ‘191 Patent, the skilled person is  
taught how to make the precursor molecule for the fluorination.  
With respect to the fluorination step, the skilled person would  
know of various possible fluorinating agents that he or she could  
try, and would be most drawn to the most widely used reagent for  
converting an alcohol to a fluoride, namely DAST or an equivalent  
reagent such as Deoxofluor. While exact reaction conditions were  
not known (given that the claimed compound was novel), finding  
appropriate conditions would be exactly within the realm of routine  
trial and error experimentation expected of a skilled chemist.  
[Emphasis added]  
[423] I have already commented on the broadening of what constitutes common general  
knowledge in the age of computers citing Justice Gauthier in Cefaclor at para 104, which I repeat  
here:  
The distinction between common general knowledge and prior art  
which is part of the state of the art for the purpose of assessing  
anticipation and obviousness tends to diminish in modern times  
because of the sophistication of search engines and the availability  
of electronic publications and databases.  
[424] Idenix’s position on the limit to common general knowledge found in scientific databases  
is described as follows from its written final submissions.  
Page: 145  
It is not Idenix’s position that the common general knowledge  
encompasses all information contained in the Scifinder or Beilstein  
Crossfire databases. The skilled person would need to know what  
to search for before these tools could be of any use. However, if a  
specific transformation was of interest to a skilled chemist, the  
results of the proper search for this specific transformation would  
become of primary importance to the skilled person.  
[425] That is not so much the issue in this matter where the circumstances describe a series of  
searches and the need for experimentation with each one as sub-steps to go forward. This case is  
about whether what is found from the searches can be said to be generally accepted as common  
general knowledge, or conversely whether a series of searches requiring experimentation  
constitutes an onerous burden. It is also about the evidence of the chemists’ work at synthesizing  
the compound in the period of 2002 to 2005.  
(c)  
No issue of Inventive Steps  
[426] The parties have not attempted to advance any serious issue about an inventive step being  
a factor in the reliance on common general knowledge or undue experimentation in delimiting  
common general knowledge or experimentation. It is obvious by definition that something cannot  
be common knowledge if it incorporates an inventive step. Similarly, the Supreme Court in  
Plavix (SCC) specifically excluded trial and error experimentation in aid of an inventive step.  
Justice Rothstein, speaking to the role of experimentation involving an inventive step, stated as  
follows at paragraph 33:  
[33] What amount of trial and error or experimentation is  
permitted before a prior disclosure will not constitute enabling  
disclosure? Certainly, if the applications judge finds that an  
Page: 146  
inventive step was required to get to the invention of the second  
patent, the specification of the first patent will not have provided  
enabling disclosure. But even if no inventive step is required, the  
skilled person must still be able to perform or make the invention  
of the second patent without undue burden.  
[Emphasis added]  
[427] The experts of both parties made general contradictory statements as to the inventiveness,  
or not, in making the compound, but went no further. Dr Barrett claims that no inventiveness was  
involved in the fluorination step. His opinion on this issue is found at para128 of his November  
report:  
[128]I acknowledge that in some cases a combination of known  
reactions may not be routine, as there may be inventiveness in  
determining the sequence of steps to follow. This is not such a  
case. In this case, the ‘191 Patent explicitly sets out two alternative  
methods to synthesize appropriate precursor molecules. Only one  
additional step to what is described in the ‘191 Patent need be  
taken, and this step, the substitution of an alcohol to provide a  
fluoride, would have been straightforward for the person skilled in  
the art, especially since it was amply described in the  
contemporaneous art.  
[Emphasis added]  
[428] Dr Damha is of the same opinion at para 106 of his September report:  
[106]It is my opinion that a PSIA would have been able to  
prepare compounds within the scope of claims of the ‘657 Patent  
using the disclosure of the ‘191 Patent without inventive ingenuity  
and without undue experimentation...  
[429] Dr Wnuk, at para 198 of his November report disagrees with the statement of Dr Damha  
that the preparation of the 2’-C-Me/F compound lacked inventive ingenuity, (he did not respond  
Page: 147  
to Dr Barrett’s comment, as the two reports were served on the same day). However, the issue of  
inventiveness of the 2’-C-Me/F synthesis is not developed in the remainder of the evidence. Dr  
Wnuk simply disagrees with the broad general statement of the Idenix experts, stating that  
neither the common general knowledge nor trial and error experimentation was sufficient to  
advise the skilled person how to make the target compound.  
[430] Gilead prefers to argue that the literature and knowledge of the skilled chemist at the time  
does not establish any common general knowledge with respect to the fluorination step. In  
addition, it submits that the steps and procedures Idenix outlines to arrive at the synthesis of the  
2’-C-Me/F compound would require an undue burden of experimentation. It implicitly relies  
upon Justice Rothstein’s comment in Plavix (SCC) at para 33 above that “even if no inventive  
step is required, the skilled person must still be able to perform or make the invention of the  
second patent without undue burden”. [Emphasis added].  
[431] In this regard, it is noted that in Plavix (SCC), where inventiveness was an issue, the  
Supreme Court at para 78 included “methods for obtaining that compound” as an aspect of “the  
inventive concept”, which I equate to an inventive step.  
[78] In the present case, it is apparent that the inventive concept of  
the claims in the ‘777 patent is a compound useful in inhibiting  
platelet aggregation which has greater therapeutic effect and less  
toxicity than the other compounds of the ‘875 patent and the  
methods for obtaining that compound.  
[Emphasis added]  
[432] Dr Damha was apparently of the same view stating as follows:  
Page: 148  
There is no magic. Really, chemistry, if I, and I say this with due  
respect, is, the inventiveness is the compound, its use, the  
discovery, making it.  
[Emphasis added]  
[433] The concept of inventive steps could have been made relevant to the outcome of this  
decision, if set out as an issue for the Court to decide. It represents a first means to delimit the  
application of common general knowledge and experimentation. It also could have been raised in  
infringement issues to fix the essential features under the “Catnic principle” (Catnic Components  
Ltd v Hill & Smith Ltd, [1982] RPC 183 at pages 242-43 [Catnic] , as endorsed by the Supreme  
Court in the Free World and Whirlpool decisions). Inventive steps similarly are material in  
deciding who the inventor is or are where there may be two steps of inventiveness in terms of  
coming up with the idea of the compound and the making of the invention.  
[434] Both parties appear to have their own reasons on different issues to avoid raising any  
consideration of an inventive step, which they have clearly chosen not to do. I agree that it is not  
necessary here where the case may be decided on alternative findings of fact. But to the Court it  
seems an unusual strategy, where in most litigation normally one advances all submissions that  
might contribute to a successful final judgment.  
(d)  
Gilead’s Submission on the Insufficiency of Disclosure based on Viagra  
[435] Gilead advances two novel submissions at two different places in its submissions that  
unsuccessfully attempt to apply the Viagra decision to have the ‘191 Patent declared invalid. In  
this part, the issue appears to be about the insufficiency of disclosure in light of Viagra, while  
Page: 149  
further below, under the rubric of “what is the invention”, Gilead argues there is no invention  
because Idenix could not synthesize all the compounds in Claims 1 to 3.  
[436] The foundation of both arguments is described at paras 172 to 174 of Gilead’s  
submissions, which I set out below, without footnotes:  
[172]The Court held that while it is possible for different claims  
in a patent to disclose separate inventions, “individual patents must  
be considered on a case-by-case basis” and that the Court must  
look to the whole of the disclosure and the claims to ascertain the  
nature of the invention and methods of its performance.  
[173]The Court looked to the specification as a whole in which  
the patentee described “the invention” as a class of compounds.  
The Court held there was nothing to support the view that the use  
of sildenafil (claim 7) was a separate invention from the use of any  
of the other claimed compounds. No specific attributes or  
characteristics were ascribed to sildenafil that would have set it  
apart from the other compounds.  
[174]Finally, the Court considered the statutory requirement that  
each patent must be in respect of only one invention and the lack of  
any divisional applications filed by Pfizer. The Court held that: “It  
would be disingenuous for Pfizer to imply that there is one  
invention in the patent application for the purpose of complying  
with s 36(1) and then to submit that each claim concerns a distinct  
invention for the purposes of this appeal.  
[437] Gilead’s first raises its argument on the sufficiency of disclosure at paras 175 to 177,  
citing para 75 of the Viagra decision, as follows (without footnotes):  
[175]Sufficient Disclosure – In concluding that the patent did not  
sufficiently describe the invention, the Supreme Court made the  
following important comments of more general application:  
[70] As I noted above, this Court made it clear in  
Consolboard that the specification, which includes  
Page: 150  
the claims and the disclosure, must define the  
“precise and exact extent” of the privilege being  
claimed so as to ensure that the public can, having  
only the specification, make the same use of the  
invention as the inventor (p. 520). In my view, the  
courts below misread Consolboard when they stated  
that the only questions that must be answered are  
“What is your invention?” and “How does it work?”  
Dickson J. did not state that those were the only  
relevant questions. In fact, quoting Minerals  
Separation, he went on to say, at p. 520:  
With respect to each question the description must  
be correct and full in order that, [...]  
when the period of monopoly has expired the public  
will be able, having only the specification, to make  
the same successful use of the invention as the  
inventor could at the time of his application.  
[Emphasis by Gilead]  
[71] The Court reiterated this in Pioneer Hi-Bred:  
“The description must be such as to enable a person  
skilled in the art or the field of the invention to  
produce it using only the instructions contained in  
the disclosure” (p. 1638).  
[…]  
[74] The disclosure in the specification would not  
have enabled the public “to make the same  
successful use of the invention as the inventor could  
at the time of his application”, because even if a  
skilled reader could have narrowed the effective  
compound down to the ones in Claim 6 and Claim  
7, further testing would have been required to  
determine which of those two compounds was  
actually effective in treating ED  
[176]The Court specifically rejected Pfizer’s argument that the  
disclosure was sufficient because third parties were eventually  
given enough information to be able to use the invention:  
The fact that Teva carried out this minor research  
project is irrelevant to Pfizer’s obligation to fully  
disclose the invention. More importantly, what must  
Page: 151  
be considered is whether a skilled reader having  
only the specification would have been able to put  
the invention into practice. The trial judge clearly  
found that the skilled reader would have had to  
undertake a minor research project to determine  
what the true invention was.  
[Emphasis added by Gilead]  
[177]The Supreme Court held that the patent was insufficient  
because “a skilled reader having only the specification would not  
be able to put the invention into operation.” As a remedy, the Court  
held that: “the logical consequence of a failure to properly disclose  
the invention and how it works would be to deem the patent in  
question invalid. This flows from the quid pro quo principle  
underpinning the Act. If there is no quid — proper disclosure —  
then there can be no quo — exclusive monopoly rights.”  
[My emphasis]  
[438] I admit to difficulty in following this argument. The reference to the “failure to properly  
disclose the invention and how it works” seems in reference to the first two questions posed in  
Viagra. Citing the passage referring to “a skilled reader having only the specification would not  
be able to put the invention into operation” seems to be about disclosure of the synthesis of  
compound. Idenix argues that Viagra is basically a case about the disclosure of the invention and  
not about the sufficiency of disclosure of how to make the invention. I am not able to make the  
distinction between the two lines of argument in Viagra, inasmuch as the Supreme Court  
appeared never to have made reference to which of the subparagraphs of section 27(3)(a) or (b) it  
was relying on in the decision.  
[439] I do not think that this is important in any event. I find that Viagra has insufficient  
parallels to be applied to this matter. Viagra is a case about a situation where the invention is  
known and the inventor is quite able to describe it, but leaves the addressee up in the air as to  
Page: 152  
which of two claims contains the accurate disclosure, thereby requiring a minor research project  
to find out.  
[440] The facts of this case bear no relation to those in Viagra where it is clear from Claims 1-3  
and the other dependent claims that the fluorine (down), and a methyl (up) at the 2’ position of  
the carbon ring is the invention, even if Claim 1 consists of millions or more compounds. The  
issue remains whether disclosure of the synthesis of the nucleoside has been disclosed in the  
Patent and by the common general knowledge and routine experimentation. Indeed, Idenix  
argues that a skilled reader having only the specification in hand would be able to synthesize the  
invention in the Patent and by the common general knowledge supplemented by routine  
experimentation.  
[441] The Supreme Court disagreed with the lower courts in Viagra that there were separate  
inventions in the claims. Gilead argues that Idenix has resiled from its representation to the CIPO  
that Claims 1-3 are one invention by not defending Claim 1. I cannot understand on what basis  
this conclusion is founded, or why Idenix is prevented from abandoning the defence of a too  
broadly stated claim. I have already referred to Viagra at paragraph 80 of the decision as  
authority that this practice is acceptable.  
[442] I also do not accept that Viagra stands for the principle that a “minor research project”  
properly describes the test for what constitutes an “undue burden” in terms of making the  
invention based on routine experimentation. The need for the testing in Viagra was only to show  
Page: 153  
that the disclosure of the invention was ambiguous between two possibilities, when the active  
agent was known and should have been disclosed in the first place.  
B.  
What is the Invention?  
(1)  
Quasi-Agreement on the Description of the Invention  
[443] In Viagra, the Supreme Court states at paragraph 53 that the first step in determining  
whether disclosure requirements have been met is to define the nature of the invention:  
[53] In determining whether the disclosure requirements have  
been met in this case, the first step is to define the nature of the  
invention in Patent ’446. This must be done in order to comply  
with s. 27(3) of the Act, which requires, among other things, that  
the specification “correctly and fully describe the invention”.  
Therefore, we must ask: What is the invention in Patent ’446?  
[444] To a point, the parties have generally agreed that the invention in the ‘191 Patent is a  
class of 2’-C-Me/F compounds that have anti-Flaviviridae activity. Idenix’s written submission  
on what the invention is, as follows:  
[2] What is the Invention? By the filing date of June 23, 2003,  
the inventors of the Canadian Patent 2,490,191 had soundly  
predicted that a novel class of compounds would have anti-  
Flaviviridae activity. The compounds of the invention belong to a  
class of nucleoside analogues having a particular modification at  
the 2’ position of the sugar ring, namely a fluorine atom in the  
(down) position and a methyl in the (up) position.  
Page: 154  
[445] Gilead’s counsel described the invention in its oral submissions as follows: “The  
invention here would be the group H compounds together with, what I showed you a minute ago  
from Christiani v Rice, [1930] SCR 443 a means of making that which is invented with utility.”  
[446] Idenix appears to misapprehend Gilead’s position when it states that it accepts that there  
is no requirement for an inventor to have actually made a compound falling within the claims.  
Both parties agree that under Canadian law there is no requirement  
for an inventor to have actually made a compound falling within  
the claims. In other words, the requirements of an “invention”  
and/or inventive concept should not be conflated with an actual  
physical making of a compound of the invention.  
[Emphasis added]  
[447] The confusion may arise as to what Gilead intended by “a means of making that which is  
invented”. Gilead does not limit this to being able to disclose how to make the invention.  
[448] This is apparent from Gilead’s submissions under the heading of the “Nature of the  
Invention”. Gilead again relies on the Viagra decision, but this time it is in support of the  
proposition that the Court may consider extrinsic evidence to ascertain the nature of the  
invention. Where I find this submission difficult to follow however, is that the failure to make the  
2’-C-Me/F compounds despite two years of effort does not appear to be the issue. Rather  
Gilead’s concern is Idenix’s abandonment of Claim 1. Gilead’s submission is as follows at  
paragraphs 188-9 of its written argument:  
Page: 155  
[188]Applying the reasoning from Viagra, there is no reason to  
treat the compounds claimed in claims 1-3 as separate inventions.  
It would be disingenuous for Idenix to say that there is one  
invention in order to overcome an examiner’s office action and  
then submit that each claim concerns a distinct invention for the  
purpose of this action. This is a direct application of the reasoning  
in paragraph 68 in Viagra.  
[189]In Viagra, the Court looked to the actual conduct of the  
inventors to ascertain what their invention was. In this case, the  
inventors’ conduct clearly demonstrates that they did not have an  
invention at the time they filed, since they did not have a means of  
making the compounds claimed. Therefore, the ‘191 Patent fails  
for the fundamental reason that the named inventors had not  
actually made any invention, contrary to the long-standing  
principles of law articulated by the Supreme Court of Canada in the  
1930 decision of Christiani v. Rice.  
[Emphasis added]  
[449] My problem with Gilead’s submission is that I do not find that Idenix is arguing that  
Claims 1-3 are separate inventions. This appears to be Gilead’s conclusion, because Idenix  
subsequently chose not to defend Claim 1. I say this because the reference in paragraph 189  
above to Idenix not having had the means to make the compounds claimed, seems to relate back  
to its submission in paragraph 188. This in turn, appears to relate to the abandoned Claim 1 and  
Idenix’s inability to make all the compounds with the C(Y3)3 group in the 2’ (up) position. I have  
already referred to paragraph 80 of Viagra that acknowledges the propriety of abandoning claims  
that are too broadly framed. I therefore, reject this submission.  
[450] On the other hand, if the submission concerns Idenix’s failure to make the 2’-C-Me/F  
nucleosides by the substitution of the fluorine at the 2’ down position, I agree with Idenix that  
this is not relevant to the definition of the invention. In this matter the invention is the 2’-C-Me/F  
nucleoside with predictable utility. I understand that “What is the invention” is an independent  
Page: 156  
consideration of the requirement of being able to describe how to make the invention. In either  
case, there does not appear to be any requirement to have actually been able to make the  
invention. Otherwise, there would be no basis to predict utility for inventions that had not been  
made; or at least, it would add another factor to soundly predicted inventions.  
[451] If the related issue for soundly predicted inventions is that the prediction of utility is  
hypothetical based on the inventor’s knowledge, when the inventor’s actual knowledge is that the  
compound cannot be made despite attempts to make it, that issue has not been pleaded, or at least  
argued, and is not before me.  
C.  
Does the ‘191 Patent Disclose how to Synthesize the 2’-C-Me/F Compound?  
[452] In the analysis that follows I first consider whether and to what extent the synthesis of the  
2’-C-Me/F nucleoside is expressly disclosed in the ‘191 Patent, bearing in mind the  
acknowledgment that the Patent contains no description of the fluorination step that ultimately  
synthesizes the 2’-C-Me/F nucleoside.  
[453] Thereafter, I consider whether common general knowledge and routine experimentation  
are sufficient to disclose the synthesis of the 2’-C-Me/F compounds. This analysis is based upon  
the testimony of the experts and the contemporaneous evidence from the Idenix and Gilead  
chemists involved in efforts to make the compound during the relevant 2002 to 2004 period.  
Page: 157  
[454] With respect to the Court’s analysis of the work of Dr Griffon, it also considers  
allegations by Idenix that he did not perform to the level of a skilled chemist. This includes  
allegations that he failed to recognize that he had synthesized the 2’-C-Me/F nucleoside, as was  
allegedly demonstrated by experiments conducted for Idenix by AMRI in 2014 simulating his  
February 2003 experiment with Deoxo-Fluor® on a carbohydrate nucleoside.  
(1)  
Is the Synthesis of the 2’-C-Me/F Compound Disclosed in the ‘191 Patent?  
[455] The Court is here considering the express written disclosure contained in the ‘191 Patent  
without the added disclosure of synthesis from the common general knowledge.  
[456] It is common ground that as of the relevant dates, the synthesis of a 2’-C-Me/F nucleoside  
was novel, as Dr Barrett implied by it never having been made. It is also common ground that the  
‘191 Patent contains no information with respect to the fluorination step. On this basis alone, the  
Patent fails to teach how to make the claimed compounds. It relies on common general  
knowledge and experimentation in order to meet the disclosure requirements to carry out the  
essential fluorination step.  
[457] The best Idenix can achieve therefore, is to demonstrate that the intermediate used to  
synthesize the 2’-C-Me/F nucleoside, which is a 2’-C-OH/Me sugar ring or nucleoside required  
for a fluorination by DAST, was disclosed in the ‘191 Patent. No prior step in the synthesis  
process is at issue. I find Idenix does not teach how to make the intermediate compounds in the  
‘191 Patent. Essentially, the few relevant pages found in the ‘191 Patent are directed towards the  
Page: 158  
synthesis of the 2’-C-Me/OH nucleosides of which Idenix held patents, and not the intermediate  
to be fluorinated.  
[458] The ‘191 Patent claims 2’-C-Me/F nucleosides without describing any synthetic scheme  
by which any such compound can be made. As for its precursor the 2’-C-OH/Me sugar ring or  
nucleoside compounds, the synthetic schemes in the ‘191 Patent are directed to the 2’-C-Me/OH  
class of compounds having 2’-OH(down) substituents, with only inferred instruction on how to  
make the precursor 2’-C-OH/Me compounds necessary for the fluorination step.  
(a)  
Gilead’s General Submission on the Disclosure Due to Idenix’s  
Abandonment of Claims in the Application  
[459] As indicated, Idenix abandoned a number of claims and 22 of its 23 formulae in the  
Patent Application. These were not contained in the ‘191 Patent as issued. Both parties directed  
the Court to the file history of the ‘191 Patent which contains evidence on this downsizing step  
by Idenix.  
[460] Idenix has stated that the effect of the downsizing of its Patent Application is that,  
… some of the language used in the ‘191 Patent may relate to some  
of the formulas disclosed in the ‘191 Patent but which are not  
claimed in the ‘191 Patent as issued. As such, some of the  
specification that is directed to formulas other than Formula (IX)  
may not be relevant in construing the claims of the ‘191 Patent.  
[461] Gilead’s witnesses, in particular Dr Wnuk, provided an opinion with respect to the  
irrelevancies in the ‘191 Patent remaining after the removal of claims and formula, along with  
Page: 159  
Claim remaining in issue. As an example, I set out a summary of his opinions in this regard from  
his first report as follows:  
(b)  
It is my opinion the skilled person would have understood  
from the title of the ‘191 Patent and initial pages of the  
specification that the patent is directed to 2’ and 3’ prodrugs of 1’,  
2’, 3’ or 4’ branched nucleosides for use in the treatment of  
Flaviviridae infections, such as HCV infections. The ‘191 Patent  
discusses twenty-three Formulae that each appear to encompass a  
different broad class of substituted nucleosides. The skilled person  
would have understood that, depending on the Formulae, different  
substituents are available at each of the positions on the sugar ring,  
including changes to the sugar ring itself. The skilled person would  
have also understood the Formulae of the ‘191 Patent to include a  
vast number of non-natural bases. As a whole, the skilled person  
would have understood that the ‘191 Patent refers to an enormous  
number of nucleosides with various substituents possible on the  
sugar ring as well as on the base. The skilled person would have  
understood that the ‘191 Patent provides some synthetic schemes  
and examples for making some of the nucleosides discussed in the  
specification or body of the ‘191 Patent (which I understand is that  
part of the patent that precedes the claims), but certainly not all. In  
particular, the skilled person would have understood that the ‘191  
Patent provides no schemes or examples which describe how to  
make nucleosides with a substituent other than OH or H in the 2’  
(down) position or the ‘3 (down) position. The skilled person  
would have also understood that the ‘191 Patent provides examples  
that relate to the preparation of starting materials and intermediates  
that can be used in the synthesis of particular nucleosides with a  
substitution at the 4’ position. The ‘191 Patent also provides  
information regarding some biological assays that it says can be  
used to assess the compounds referred to in the ‘191 Patent, as  
well as biological assay results in respect of a single compound,  
that the ‘191 Patent refers to as Compound F. The skilled person  
would have noted that no chemical structure or drawing for  
Compound F is provided in the ‘191 Patent. Further, the skilled  
person would have noted that the nomenclature provided for  
Compound F is incomplete and thus the exact structure of the  
compound would not have been understood by the skilled person.  
(c)  
The skilled person would have understood that the ‘191  
Patent ends with 32 claims. Claims 1 to 3 relate to particular  
chemical structures. Claims 4 to 32 relate to using the compounds  
of Claims 1 to 3 in the treatment of flaviviridae infections,  
including HCV. Rather than being directed to “modified 2’ and 3’-  
Page: 160  
prodrugs of 1’, 2’, 3’ or 4’-branched nucleosides” or the twenty-  
three Formulae contained in the specification, Claims 1 to 3 are  
directed only to one Formula, Formula (IX). Further, Claims 2 and  
3, while still very broad, limit the nucleosides covered to those  
with a methyl in the 2’ (up) position and a fluoro in the 2’ (down)  
position. The skilled person would have found the focus on this  
particular substitution pattern unexpected, as the ‘191 Patent does  
not highlight or provide any example or scheme directed to a  
nucleoside with these two substituents. Further, the skilled person  
would have been confused by the scope of the claims, as the focus  
in the Examples was clearly on nucleosides with a methyl group at  
the 4’ position and no such nucleosides appear to be covered by the  
Claims.  
[462] Dr Wnuk’s comments on the contents of the ‘191 Patent taken without interpretation by a  
skilled chemist seeking to understand the Patent are irrefutable. Some were confirmed by Dr  
Damha in his cross-examination. This did not, however, extend to the claims’ portion of the  
specification, or to Dr Damha’s opinions of the extent of misdirection.  
[463] Dr Damha, in his reply report to Dr Wnuk’s interpretation of the ‘191 Patent, as it read  
without the benefit of the information from the application file, gave a similar explanation as that  
contained in Idenix’s submissions to the Court, as follows:  
Dr Wnuk, however, did not mention that during the prosecution of  
the ‘191 Patent, the Canadian Intellectual Property Office issued an  
office action requiring the ‘191 Patent be restricted to one  
invention. As such, the claims of the ‘191 Patent were restricted to  
compounds of Formula (IX) and their use. Dr Wnuk’s comments  
about aspects of the disclosure relating to compounds other than  
Formula (IX) are therefore misplaced.  
[464] The true interpretation of the ‘191 Patent is to be arrived at by consideration of what a  
competent person skilled in the art reading the specification at its date of issue would have  
Page: 161  
understood it to have disclosed and claimed, (Freeworld at para 52). As such, the ‘191 Patent  
must be construed without reliance upon extrinsic evidence, which I conclude includes the  
information contained in the CIPO’s files.  
[465] However, it is also my opinion that the skilled chemist would have figured out that there  
was a disconnect between the Claims and much of the content of the specification. This would  
have led to identifying the only relevant formula based on Formula (IX) with the fluorine in the  
‘2 (down) position. In this respect I, therefore, agree with Dr Damha’s opinion that Dr Wnuk’s  
interpretation of the Patent specification is too literal. While reflecting a first impression of the  
document, it does not state what the skilled person would have eventually taken from it.  
[466] On the other hand, very little can be taken from much of the 170 pages, and often the  
specification misdirects. This occurs in the schemes and discussions on synthesis that were only  
stated to be relevant to the disclosure of how to make 2’-C-Me/OH nucleosides, and not to its 2’-  
C- OH/Me diastereomer used as the intermediate to synthesize the 2’-C-Me/F nucleosides.  
[467] On the other hand, I find that the Court was not prevented from construing the Claims,  
which all were clearly directed at compounds with the methyl (up) and fluorine (down) structure  
at 2’ position on the sugar ring. Claim 1, which was abandoned by the Idenix, was similarly  
related to the other compounds claimed with the fluorine at the 2’ (down) position.  
[468] Idenix’s expert witnesses provided assistance to the Court in the construction of the  
Claims, as well as referring the Court to those portions of the ‘191 Patent that they considered  
Page: 162  
relevant to the issues at hand, such as the definition section of Bases. Gilead directed the Court’s  
attention to inconsistencies in the specification that affected the interpretation of the claims, for  
example the various and different references to phosphates. Gilead advanced no general  
argument that the Claims were unintelligible on these issues because of the contents of the  
specifications.  
[469] However, I do give some weight to Dr Wnuk’s opinions that the disclosure of the  
synthesis of the 2’-C-Me/F nucleoside was adversely affected by the fact that so much of the  
specification was irrelevant, making the relevant information more difficult to determine. This in  
addition to the facts that the ‘191 Patent contained no information on the essential fluorination  
step to make the invention and that the schemes in the Patent directed the skilled person to the 2’-  
C-Me/OH as a final product; thereby away from its diastereomer 2’-C-OH/Me compounds as the  
appropriate intermediate to synthesize the 2’-C-Me/F nucleoside, add to my conclusion that the  
express disclosure on the synthesis of the intermediate was also lacking.  
(b)  
Schemes of Synthesizing the Intermediate 2’-C-OH/Me Compound  
[470] It was acknowledged that there were three approaches to synthesizing the 2’-C-Me/F  
nucleoside, two of which use the 2’-C-OH/Me compounds as intermediates. Schemes 11 and 5  
from Dr Barrett’s first report shown below depict the first two approaches. They are the  
Nucleoside Modification and the N-Glycosidation (sugar ring) approaches. The Nucleoside  
Modification approach uses a nucleoside starting material, while the N-Glycosidation approach is  
carried out by modifying a sugar D-ribose ring derivative starting material, which after  
fluorination of the ring, is coupled with a uracil base by glycosidation. For the purpose of these  
Page: 163  
discussions on disclosure of the synthesis of the intermediate, only the methylation step (e.g. the  
step between molecules 64 and 65 in scheme 11 below that adds the CH3 group) that creates the  
2’-C-OH/Me precursors are at issue.  
(i)  
The Nucleoside Modification Approach  
[471] Scheme 11 of Dr Barrett’s drawings starts with a uracil nucleoside compound 62, which  
is protected at the 3'- and 5'-hydroxyl groups on the sugar resulting in compound 63. It in turn is  
oxidated with a reagent to change the 2'-alcohol (CH3) to the 2'-ketone (O) resulting in  
nucleoside 64. These two steps are described in the ‘191 Patent and are not at issue.  
[472] Nucleoside 64 is methylated with methyl lithium in diethyl ether at -78 C following the  
method of A. Matsuda et al. to create nucleoside 65, being the 2’-C-OH/Me nucleoside. This is  
the sought-after intermediate nucleoside that will be fluorinated to produce the 2’-C-Me/F  
Page: 164  
nucleoside. This methlation step (Nuceloside 64 to 65) is to be compared with the schemes found  
in the ‘191 Patent (further below at Scheme 3 under the title of Synthesis of 2 '-C-Branched  
Nucleosides) which do not show or make reference to the intermediate, but only Idenix’s  
compound of the 2’-C-Me/OH nucleoside.  
[473] In Dr Barrett’s drawings, Nucleoside 65 after undergoing further deprotection and  
protection steps is fluorinated with DAST in toluene solution at -20 F resulting in the protected  
target 2’-C-Me/F nucleoside 66, along with elimination products not shown on the scheme. The  
‘191 Patent has no information on this step. Compound 66 is deprotected, resulting in compound  
61 being the target 2’-C-Me/F nucleoside. The entire fluorination step is absent from the ‘191  
Patent.  
(ii)  
The N-Glycosidation Approach  
[474] The N-Glycosidation, or sugar ring approach, follows much of that of the Nucleoside  
approach, except that the starting material is a sugar D-ribose ring and the DAST reagent is not  
in a toluene solution. After the fluorination step, a Vorbrüggen reaction followed by deprotection  
is employed to complete the synthesis to add the base. This approach is depicted in scheme 5  
below from Dr Barrett’s report.  
Page: 165  
(iii)  
The Gemcitabine Approach  
[475] The Gemcitabine “Small Molecule Strategy” approach is so named as it was originally  
used to produce the anti-cancer drug bearing its name. The procedure described in Scheme 25  
below, involves using an aldol or Reformatsky reaction followed by a cyclization to form a sugar  
with the “2’ methyl (up)”- 2’ F (down), substituents already established in their correct positions  
to which a base is coupled. There is no mention or depiction of this scheme in the ‘191 Patent.  
Page: 166  
(iv)  
The Schemes in the ‘191 Patent  
[476] The relevant schemes in the ‘191 Patent all show the target compound with the hydroxyl  
remaining in the 2’ (down) position, being the Idenix patented 2’-C-Me/OH nucleosides. In other  
words, they do not depict either the inverted diastereomer intermediate 2’-C-OH/Me or the target  
2’-C-Me/F compounds. Typical of these schemes is Scheme 3 from the Patent depicting a N-  
Glycosidation approach starting from a sugar ring with a 2’-C-Me/OH nucleoside as a product  
prior to the optional protection and reduction. It is found at page 123 in the Patent in the section  
entitled “B. General Synthesis of 2 '-C-Branched Nucleosides”, 1. Glycosylation of the  
nucleobase with an appropriately modified sugar. The scheme reads left to right across the top  
and then down.  
Page: 167  
[477] Compound 1 is the starting material being the D-ribose sugar ring. Compound 2 is the  
result of the oxidation process of the starting material, while Compound 3 is the result of the  
methylation process being the 2’-C-Me/OH sugar ring compound (the methyl is the substituent  
R6) before coupling with the base. The fourth Compound on the line below shows the coupling of  
the base to the sugar ring making the target 2’-C-Me/OH nucleoside. The fifth and sixth  
compounds, being the protection and optional reduction steps on the scheme, are not relevant.  
[478] Scheme 4 at page 125 of the ‘191 Patent depicts the Nucleoside Modification approach,  
which is similar to Scheme 3, only with the starting material being a nucleoside. Scheme 9, a  
more generalized version, also shows the target compound with the OH in the 2’ (down) position,  
as well as having a purine base. The 2’-C-Me/F nucleoside is only active with the pyrimidine  
bases of uracil or cytidine. These drawings all show the synthesis of the 2’-C-Me/OH nucleoside,  
Page: 168  
while Scheme 9 has a base that would not provide anti-viral activity if the 2’-C-Me/OH  
diastereomer (i.e. the 2’-C-OH/Me compound) was used to make a 2’-C-Me/F nucleoside.  
(c)  
Idenix’s Experts Evidence on the Disclosure in the ‘191 Patent  
[479] Despite the schemes directing the skilled person to the synthesis of the 2’-C-Me/OH  
nucleoside, Idenix’s experts insist that the ‘191 Patent provides the teaching to make the  
intermediate 2’-C-OH/Me compounds. Dr Barrett summarizes Idenix’s case as follows at para  
128 of his November report  
[128]In this case, the ‘191 Patent explicitly sets out two  
alternative methods to synthesize appropriate precursor molecules.  
Only one additional step to what is described in the ‘191 Patent  
need be taken, and this step, the substitution of an alcohol to  
provide a fluoride, would have been straightforward for the person  
skilled in the art, especially since it was amply described in the  
contemporaneous art.  
[Emphasis added]  
[480] Dr Barrett’s explanation of the explicit description of how to make the two alternative  
methods starts with a retrosynthetic analysis of the skilled person working backwards from the  
chemical properties of the 2’-C-Me/F compound. The skilled person, after viewing the target  
compound, is led to the conclusion that the synthetic route must use the DAST fluorination  
method to make the molecule. That research also leads the skilled chemist to recognize that the  
target intermediate molecule must have the methyl (down) and hydroxyl (up) inverted at the 2’  
position from all the compounds depicted in the schemes in the ‘191 Patent.  
Page: 169  
[481] With respect, a retrosynthetic analysis to make a compound that is not described or shown  
in the Patent is not what the skilled person would consider as an explicit description of how to  
make the compound.  
[482] From this undescribed starting point of a retrosynthetic analysis, being the essential first  
step in the novel synthesis, the skilled chemist is said to be directed to the intermediate  
compound via the O’kuru article. It is to be found by the skilled chemists among the few relevant  
passages in the ‘191 Patent. Dr Barrett states as follows:  
At the least, Harry-O’kuru and Example 2 teaches a person skilled  
in the art how to use all of the necessary reagents, and perform all  
of the necessary reactions, for the synthesis of nucleoside 41 other  
than the step in which the alcohol is converted into the fluoride.  
[483] I think it would be better to state that “at the most” a skilled person might find his way to  
the synthesis of the intermediate in the ‘191 Patent, except that I would agree with Gilead’s  
argument that the passage in question quoting the O’kuru article equally directs the POSITA  
away from the 2’-C-OH/Me intermediate as follows:  
EXAMPLE2: REPARATION OF2'-C-METHYLRIBO-8-  
METHYLADENINE  
The title compound was prepared according to a published  
procedure (R.E. Harry O'kuru, J.M. Smith, and M.S. Wolfe, “A  
short, flexible route toward 2'-C-branched ribonucleosides", J. Org.  
Chem. 1997, 62 1754-1759) (Scheme 9).  
[Emphasis added]  
Page: 170  
[484] Example 2, refers explicitly to Scheme 9 which is the 2’-C-Me/OH compound, also with  
the purine rather than the pyrimidine base. The title compound in Scheme 9 is also the 2’-C-  
Me/OH compound, not the intermediate 2’-C-OH/Me compound.  
[485] As was further pointed out by Gilead, the O’kuru article does not explicitly explain or  
point to the methylation step with methyl lithium in diethyl to create the intermediate 2’-C-  
OH/Me compound. Rather it refers to the Matsuda article used by Jeremy Clark, A. Matsuda et  
al, “Alkyl Addition Reaction of Pyrimidinme 2’-Ketonucleosides: Synthesis of 2’-Branched-  
Chain Sugar Pyrimidine Nucleosides”, Chemical Pharamaceutical Bulletin, 1988, 36(3), 945-  
953. The Matsuda article provides the methodology to synthesize the 2’-C-OH/Me precursor  
compound. The O’kuru article only references the Matsuda article in a footnote. The passage  
referring to the footnote is not about antivirals, but cancer compounds. It states:  
In the past 10 years, a number of 2’C- branched nucleosides have  
displayed promising anticancer (Footnote 1) …  
[486] Footnote 1 contains a long list of reference articles, nearly all of which are written by or  
contributed to by Dr Matsuda, none of which are titled, as follows:  
Page: 171  
[487] The Matsuda document, which turns out to be the article at sub-footnote 1(b), is equally  
imprecise in that the reaction type must first be selected from the various examples it provides.  
Selecting the one that provides stereo selectivity, in turn leads to 10 different examples from  
experiments, from which entry 6 is the appropriate one, as follows:  
[488] I agree with Gilead that this is not express disclosure of how too make the intermediate  
compounds needed to synthesize the 2’-C-Me/F nucleoside.  
Page: 172  
[489] Idenix also argues that the discussion on stereochemistry at pages 102-3 of the Patent is  
sufficient to disclose how to synthesize the 2’-C-Me/OH precursor. Idenix here refers to the  
useful properties of nucleosides in the methylation process used to form the 2’-C-Me/OH  
nucleoside. It argues that this “is sufficient to disclose how to synthesize the 2’-C-Me/OH  
precursor.”  
[490] In this respect, Dr Wnuk acknowledged that it would be common general knowledge that  
the process of applying a form of the well-known Grignard reagent referred to in the Patent  
would result in a racemic mixture of diastereomers of both the 2’-C-Me/OH and 2’-C-OH/Me  
compounds that could be separated to obtain the desired precursor enantiomer.  
[491] However, a general reference to a form of a reagent that results in a by-product  
compound of the 2’-C-Me/OH compound that is irrelevant to the teaching instructions that focus  
on the making that compound, cannot be said to sufficiently disclose how to make the  
intermediate compound. In effect, I find that Idenix is relying on common general knowledge to  
advise the skilled reader, both that the 2’-C-OH/Me is the intermediate required to make the 2’-  
C-Me/F nucleosides, as well as how to make that intermediate, with at best snipits of useful  
information well buried in the ‘191 Patent.  
(d)  
Comparison with the Description in the ‘657 Patent  
[492] The ‘191 Patent can be compared with the ‘657 Patent which sets out six schemes with all  
of the details of the steps and the particulars applying to each step to enable synthesis of the 2’-  
C-Me/F nucleoside. An example follows using the sugar ring approach. The scheme starts after  
Page: 173  
the oxidation step described earlier whereby ketone was applied to a protected sugar ring to  
produce compound 3-1. Step 1 is the methylation step to produce the 2’-C-OH/Me sugar ring.  
Step 2 is the fluorination step to produce the 2’-C-Me/F sugar ring. Steps 3, 4 and 5 are the  
Glycosidation steps to produce the nucleoside with the uracil base as compound 3-6.  
Example 1: Synthesis of (2'R)-2'-Deoxy-2'-Fluoro-2'-C-  
Methylcytidine Starting from a Carbohydrate [sugar ring without  
the base]  
Step 1: Compound 3-1 (7.7 g, 0.022 mmol) was dissolved in  
anhydrous diethyl ether and cooled to -78°C. To this solution was  
added MeLi (30 mL, 1.6 M in diethyl ether). After the reaction was  
complete, the mixture was treated with ammonium chloride (l M,  
65 mL) and the organic phase was separated, dried (Na2S04),  
Page: 174  
filtered, and concentrated to dryness. Silica gel chromatography  
followed by crystallization from diethyl ether-hexanes afforded  
1
pure compound 3-2 (6.31 g). H NMR (400 MHz, CDC13): δ 1.40  
(s, 3H), 3.41 (s, 3H), 3.49 (dd, 1H, J= 10.3, 6.89 Hz), 3.57 (dd, 1H,  
J = 10.3, 3.88 Hz). 3.84 (d, 1H, J = 7.3 Hz), 4.03 (m, 1H), 4.48  
(s,lH), 4.58 (m, 3H), 4.83 (d, 1H, J= 11.6 Hz), 7.31-7.36 (m, 10H);  
13C NMR (100 MHz, CDCl3): δ18.4, 55.4, 72.2, 73.4, 79.5, 80.2,  
84.7, 107.4, 127.7, 127.8, 127.83, 128.5, 138.2, 138.3.  
Step 2: Compound 3-2 was dissolved in CH2C2 and was treated  
with DAST (4.0 mL, 30.3 mmol) at room temperature. The  
solution was stirred at room temp overnight. The so-obtained  
mixture was poured into sat NaHC03 (100 mL) and washed with  
sat NaHC03 (l x 15 mL). The organic layer was further worked up  
in the usual manner. Silica gel chromatography (1:5 EtOAc-  
hexanes) gave crude compound 3-3 (0.671 g) that was sufficiently  
1
pure for the next step. H NMR (400 MHz, CDC3): δ 1.43 (d, 3H,  
J= 22.8 Hz), 3.35 (s, 3H), 3.49 (dd, 1H, J= 10.5, 5.4 Hz), 3.55 (dd,  
1H, J= 10.5, 4.1 Hz), 3.87 (dd, 1H, J=23.5, 7.5 Hz), 4.26 (m, 1H),  
4.56 (d, 2H, J= 6.9 Hz), 4.66 (d, 2H, J= 8.2 Hz), 4.72 (d, 1H, J=  
10.8 Hz), 7.29- 7.36 (m, 10H); .13C NMR (100 MHz, CDCI3): δ  
17.0 (d, J= 24.4 Hz), 55.2, 77.1, 73.4,73.8,77.3,80.3,81.2 (d, J= 16  
Hz), 99.7 (d, J= 178.9 Hz), 106.8 (d, J= 32.0 Hz), 127.7, 127.8,  
128.1, 128.3, 128.5, 128.6, 137.8, 138.3; 19F NMR (100 MHz,  
CDC13): δ -8.2 (m, IF).  
Step 3: Compound 3-3 (0.39 g, 1.1 mmol) was dissolved in 1:2  
EtOH-EtOAc and treated with Pd/C (~0.l g) and cyc10hexene (~1  
mL). The mixture was heated to reflux ovemight and then filtered  
through celite*.·The solvent was removed in vacuo and the residue  
was dissolved in pyridine (~5 mL). To this solution was added  
benzoyl chloride (0.22 mL, 1.83 mmol) and the mixture was stirred  
at room temp ovemight. The pyridine was removed in vacuo and  
the residue was partitioned between CH2Cl2 and sat NaHCO3)  
(10.0 mL). The organic phase was dried (Na2SO4), filtered, and the  
solution was concentrated to dryness. Column chromatography  
provided 0.350 g of pure compound 3-4. 1H NMR (400 MHz,  
CDCl3): δ 1.53 (d, 3H, J= 22.4 Hz), 3.39 (s, 3H), 4.46 (dd, JH, J=  
11.6, 4.7 Hz), 4.58 (m, lH), 4.65 (dd, lH, J= 11.6, 3.9 Hz), 4.87 (d,  
1H, J= 9.9 Hz), 5.64 (dd, 2H, J= 24.1, 7.8 Hz), 7.29-7.36 (m, 10H);  
19F NMR (l00 MHz, CDCl): δ-7.5 (m, IF).  
Step 4: A solution of bis(trimethylsilyl)-N-benzoylcytosine (0.28 g,  
0.77 mmol) and compound 3-4 (0.20 g, 0.5 mmol) in 1,2  
dichloroethane (2 mL) and toluene (2 mL) was treated with  
TMSOTf (0.I5 mL, 0.77 mmol). After most of the starting material  
disappeared as judged by TLC, the solution was cooled to room  
Page: 175  
temp, washed with water (1 x 5 mL), brine (1 x 5 mL), dried  
(Na2S04), filtered, and concentrated to dryness. Flash  
chromatography followed by crystallization from CH2Cl2-hexanes  
afforded compound 3-5 (68 mg). mp 241°C; 1H NMR (400 MHz,  
CDCl3): δ 1.49 (d, 3H, J= 22.4 Hz), 4.64 (dd, 1H, J= 12.9, 3.4 Hz),  
4.73 (app d, 1H, J= 9.5 Hz), 4.89 (dd, 1H, J= 12.7, 2.2 Hz), 5.56  
(dd, 1H, J= 20.7,8.6 Hz), 6.52 (d, 1H, J= 15.9 Hz), 7.38-7.67 (m,  
19  
10H), 7.89 (d, 2H, J= 6.9 Hz), 8.07-8.11(m, 5H), 8.67 (s, 1H); F  
NMR (100 MHz, CDCl3): δ 2.85 (m, IF).  
Step 5: Compound 3-5 (40 mg, 0.05 mmol) was dissolved in  
methanolic ammonia and stirred at room temp for 48 h. The  
solution was concentrated to dryness and chromatographed (SiO2)  
eluting with 1:4 EtOH-CH2Cl2. The yield was about 12 mg of pure  
{2'R)-2'-deoxy-2'-fluoro-2'-C-methylcytidine, 3-6. IH NMR (400  
MHz, DMSO-d6): δ 1.16 (d, 3H, J= 22.0 Hz), 3.61 (dd, 1H, J= 11.6,  
5.2 Hz), 3.60-3.83 (m, 3H, J = 10.5, 5.4 Hz), 5.24 (S, 1H,  
exchangeable with D2O), 5.59 (s, 1H, exchangeable with D2O),  
5.71 (d, lH, J = 7.3 Hz), 6.08 (d, 1H, J = 19.0 Hz), 7.24 (d, 1H, J=  
19  
17.7 Hz, exchangeable with D20), 7.87 Cd, lH); F NMR (100  
MHz, DMSO-d6): δ 4.13 (m, IF).  
[493] In light of the details provided in the ‘657 Patent, which is acknowledged by Dr Damha, I  
find it incomprehensible that he could testify that the ‘191 Patent provided the same level of  
specificity of synthesis. This is his evidence from the passage cited below at pages 2517-18 of  
the transcript:  
And do you agree with me that at pages 70 to 88, the Pharmasset  
inventor, Mr Clark, provides a very specific synthetic protocol for  
making 2'-methyl-fluoro nucleosides using both the nucleoside  
approach and a sugar approach?  
A. Yes.  
Q. And he specifies the reagents, the concentrations, temperatures,  
all of the reaction conditions the person reading this patent would  
need to put these compounds into practice; right?  
A. Yes, he provides experimental details to do that.  
Page: 176  
Q. And this same level of information is not found in the ‘191  
Patent; will you agree with that?  
A. I disagree with that.  
…
THE WITNESS: Okay. With the exception, Your Honour, of the  
DAST reaction, everything else, it is in the ‘191 Patent, the  
oxidation --  
(e)  
Conclusion of Disclosure of the Synthesis of the 2’-C-Me/F nucleosides in  
the ‘191 Patent  
[494] Idenix’s experts insist that the schemes that are contained in the description of the ‘191  
Patent are directed to precursors of the compounds claimed. I agree with Gilead’s submission  
that a review of these schemes, as well as other patents filed by Idenix, demonstrates that this  
argument is little more than an attempt by Idenix, with the benefit of hindsight; to read something  
into the ‘191 Patent that is simply not there.  
[495] Moreover, the very same schemes in the ‘191 Patent and identical instructions to make  
the compounds found in the ‘191 Patent are duplicated in prior Idenix patents that make no claim  
for any fluorinated compounds [i.e. the predecessor Patent application 351].  
[496] More importantly, it is indisputable that none of the synthetic schemes or examples  
provided in the ‘191 Patent contain any instructions on the fluorination of nucleosides or sugars  
rings. I agree with Dr Wnuk’s observation, which is indisputable, that the ‘191 Patent does not  
Page: 177  
even suggest what starting materials, reagents, or reaction conditions could be used to make such  
compounds, nor does the ‘191 Patent mention fluorination or fluorination reactions whatsoever.  
[497] In my view, the absence of any meaningful disclosure in the ‘191 Patent that both  
misdirects and confuses the reader as described by Dr Wnuk places the skilled person in  
generally the same circumstances as Dr Griffon, Mr Jeremy Clark and Dr Stewart and Ms Yang  
at the beginning of their research projects to make the compound. They would have no teaching  
apart from the compound formula itself, and some of the generally well known initial and  
completion steps, like that of protection, oxidation and coupling, with nothing whatsoever on the  
Gemcitabine approach.  
[498] The skilled chemist was starting basically with a blank piece of paper, turning to the  
search engines with the hard work ahead required to make the novel 2’-C-Me/F compound, and  
not even aware of all the different unsuccessful attempts by Idenix, when it applied for the ‘657  
Patent.  
[499] The evidence that follows from Idenix’s experts substantiates this conclusion. They rely  
on a retrosynthetic analysis to synthesize the target Compound. A disclosure said to start at the  
end of the process with the synthesized compound is an indication that the significant synthesis  
steps must be worked out by the skilled reader, with little or no help from the ‘191 Patent.  
[500] In conclusion, I find that the ‘191 Patent discloses neither the synthesis of the 2’-C-Me/F  
compound, nor its precursor 2’-C-OH/Me compounds.  
Page: 178  
D.  
Is the Synthesis of the 2’-C-Me/F Compound a matter of the Application of the Common  
General Knowledge and Routine Experimentation?  
(1)  
Retrosynthetic Analysis  
[501] It is common ground that Pharmasset’s patent application WO 2005/003147, published  
on January 13, 2005 [WO Clark Publication] was the first disclosure of how to make any 2’-C-  
Me/F compounds. Anyone endeavoring to make the nucleoside prior to that time would have had  
to work backwards from the formula, with no published literature on how to make the compound.  
[502] The experts also agree that the synthesis of the 2’-C-Me/F nucleoside required a  
“retrosynthetic” form of analysis working backwards from its formula and structure. After the  
Court mused at one point whether a retrosynthetic analysis can be employed as enabling  
disclosure, it disappeared as a description by Idenix of the synthetic process the skilled chemist  
would follow. Dr Barrett for example, resiled somewhat from the requirement to carry out a  
retrosynthetic form of analysis, suggesting that the route to follow was clearly marked out for the  
skilled person simply by looking at the compound, which I think is the same thing. Nevertheless,  
his analysis in his November report refers to the steps in a retrosynthetic analysis (paras 132,  
135, 162, 164 and 181).  
[503] Dr Damha provided the Court with the most fulsome description of what was entailed in a  
retrosynthetic analysis at para 113 of his September 7, 2014 report as follows, with my breakout  
of salient points in square numbered brackets:  
Page: 179  
[113]… In the most general terms, the PSIA will look at a target  
compound to be synthesized. They will do a retrosynthetic or  
partial retrosynthetic analysis [1] working backward from the  
target compound to a known compound. For some target  
compounds a full retrosynthetic analysis which examines all  
potential precursors [2] may not be necessary where the successful  
synthetic pathway is more likely than other less likely routes. This  
is the case with the Claimed Compound which is really a straight  
forward synthesis. From this retrosynthetic analysis, a synthetic  
scheme is outlined which often [3] will require multiple synthetic  
steps through intermediate compounds including the use of  
protecting groups to prevent unwanted reactions from occurring on  
intermediate compounds. A literature search is performed to  
support the reaction conditions [4] for the individual steps in the  
synthetic scheme. [5] Experimentation is then done to carry out the  
synthesis of the target compound. Purification and analytical  
techniques may be employed on [6] intermediates and [7] the target  
compound. This routine approach is the common and accepted  
practice for the PSIA not only in the early 2000s but throughout the  
history of chemistry  
[504] The Court was not presented with any documentation that demonstrated a retrosynthetic  
analysis could be used to teach how to make a compound made in this fashion. This is surprising  
given that it is clearly a common tool of the trade. Dr Stuyver mentioned retrosynthetic analysis  
as a process of discovery of new compounds, not teaching how to make a compound already  
invented. Dr Patterson provided his definition of a retrosynthetic analysis as follows:  
So you will start with the idea of your target molecule in mind, and  
you will work backwards step by step until you find some  
commonly available, commercially available hopefully, starting  
point.  
[505] A retrosynthetic plan is at best a contingent roadmap for a proposed route through a  
labyrinth of research and trial and error experimentation that, when it confronts a dead end, goes  
back to the last step, and then again maybe the last step before that, and tries another route. There  
Page: 180  
is no common general knowledge to any plan, as every plan is different and depends upon  
context, including the abilities of the person making the plan, which for a skilled chemist, is not  
someone with much experience in the field. A retrosynthetic analysis is used precisely because  
the synthesis path is unknown. It represents the chemist’s best estimate of how to synthesize the  
target compound.  
[506] Dr Wnuk disagreed that a retrosynthetic analysis could be employed in this fashion to  
describe common general knowledge of synthesis of a novel compound that would not lead to a  
wide variety of pathways and sub pathways all involving more research and more  
experimentation that would not sufficiently disclose how to make the target compound. I accept  
Dr Wnuk’s evidence as reflecting common sense.  
(2)  
Steps in Idenix’s Retrosynthetic Analysis  
(a)  
A Literature Search  
[507] There was no literature on how to make the 2’-C-Me/F compound. As a result, Drs  
Griffon, Stewart and Jeremy Clark all had to conduct literature searches as the first step  
undertaken in their approach to synthesizing the 2’-C-Me/F nucleoside. Dr Griffon appeared to  
have spent two months on a research project before proposing on June 27, 2002 to undertake his  
first attempted synthesis. His notes show that he was in constant continued research as outcomes  
proved unsuccessful. Dr Stewart spent on average 2 hours a day over 6 months researching the  
synthesis of the 2’-C-Me/F compound. Because Jeremy Clark did not keep good notes, it is not  
clear what research he undertook. It appears that he started with the Gemcitabine approach before  
Page: 181  
switching to the nucleoside approach, and then finishing off with the sugar ring approach.  
According to Dr Patterson, he had considerable help from the experienced chemists at Pharmaset,  
although he apparently rejected advice from the senior chemists to use a nucleoside approach. He  
successfully synthesized the 2’-C-Me/F nucleoside using the nucleoside approach a few months  
later.  
[508] The resort to literature as a starting point in making the compound to determine which  
synthetic pathway to follow is very different from Dr Damha’s understated description above of  
a literature search being “performed to support the reaction conditions for the individual steps in  
the synthetic scheme”. Besides demonstrating that there was no common general knowledge  
supporting the synthesis of the 2’-C-Me/F compound, it is not conceivable that the disclosure of  
enablement is sufficient if it requires a series of literature searches and experimentation at each  
step, bearing in mind that the lead-off route offers three pathways.  
(b)  
Which of the Three Known Initial Pathways for Synthesis?  
[509] Dr Barrett indicated that three choices confronted the skilled person as the opening  
pathway to follow to synthesize the 2’-C-Me/F compound. Dr Wnuk agreed, describing what was  
entailed by these choices, which I describe in a somewhat précised form:  
1.  
Decide which strategy of three to follow to attempt to make  
the molecule; a small molecule strategy [Gemcitabine approach]  
by making an open chain carbohydrate with the desired  
substituents, then formed a sugar ring from that open chain  
carbohydrate, and then attempted to attach the sugar ring to the  
desired base; a sugar strategy wherein the skilled person could  
have begun with an available sugar ring, attempted to install the  
Page: 182  
desired substituents, then attempted to attach the sugar ring to the  
desired base; and a nucleoside strategy wherein the skilled person  
could have started with an available nucleoside and attempted to  
install the desired substituents therein.  
2.  
In each instance, the skilled person would have had  
multiple different starting materials to choose from, multiple  
reagents to choose from, multiple possible routes that could have  
been tried, multiple protecting groups available, and multiple  
possible reaction conditions that could have been attempted for  
each step in the chosen route.  
[510] Idenix’s experts suggest the appropriate pathway would have been evident to the skilled  
person, being either the sugar or the nucleoside route. The evidence of Mr Clark’s efforts is not  
consistent with Idenix’s theory. Mr Clark apparently synthesised the compound some seven  
months after he first discussed the idea with Pharmasset management, and then after first trying  
the Gemcitabine route that proved unsuccessful. Dr Watanabe apparently counselled him to  
follow the nucleoside approach. This was contrary to Dr Coe’s advice, again demonstrating that  
no one knew the most appropriate starting route. The Gemcitabine route was also tried without  
success by Dr Griffon, who of course, tried all approaches, without success.  
[511] Dr Coe opined as follows:  
In our experience and indeed in that of manner other [sic]  
particularly the de Clerc group the most viable routes to fluoro  
nucleosides are by sugar/base condensation methods the anomer  
problem notwithstanding, for the very reasons you have  
discovered, in that the leaving groups generated in situ e.g. in  
DAST reactions are readily attacked by the pyrimidine ring  
nucleophiles or elimination and/or participation of blocking  
groups. Further migrations of groups can readily occur: see our  
papers in JFC 1993 62 145 and 1993 60 239. Having said this some  
of the route [sic] you have tried are OK except that I think you are  
using the wrong reagents, leaving groups and reaction conditions.  
Page: 183  
[512] It turned out to be a good wrong guess in the sense that eventually all three routes have  
proved achievable. It appears that Mr Clark first synthesized the compound via the sugar route.  
But he also did it using the nucleoside route, with better yields, as had been a factor driving Dr  
Griffon to use that route with Deoxy-Fluor, which is a more stable form of DAST. Dr Stewart  
and Ms Wang only succeeded after they had heard that Pharmasset had successfully synthesized  
the compound via the nucleoside route. The point is that until successful, no one had any idea of  
which route to follow. Idenix was in no position, based on the common general knowledge or  
otherwise, to disclose any particular synthesis route to follow in 2004.  
(c)  
Which Fluorination Pathway?  
[513] Dr Wnuk, the only expert with fluorination experience at the relevant time, describes the  
challenges that faced the discovery chemist in determining which fluorination pathway as  
follows:  
i)  
Further, different fluorination reactions, involving multiple  
distinct fluorinating agents, were available to the skilled person in  
January 2004, but there was no understanding of how to make a  
nucleoside with a 2’-fluoro(down) and a 2’-methyl(up). At that  
time, fluorination reactions were understood to generally proceed  
via two different mechanisms: electrophilic fluorination and  
nucleophilic fluorination. Different starting materials and  
fluorinating reagents were required for electrophilic fluorination  
reactions as compared to nucleophilic fluorination reactions.  
[514] He stated that the different fluorinating agents known as of January 8, 2004, included, but  
were not limited to:  
(a)  
Electrophilic: F2; SelectFluor; N-flurobenzenesulfonimide; and ClO3F; and  
Page: 184  
Nucleophilic: HF and HF-based reagents (e.g. HF-pyridine, HF-pyridine/ AlF3,  
(b)  
anydrous HF, HF/Fe(AcAc)3); AgF and AgF-based reagents (AgF/NH4F);  
Et3N●3HF; KF and KF  
[515] The choice of fluorinating reaction (electrophilic or nucleophilic) and, ultimately, of  
fluorinating agent would have influenced the kind of starting material used in a particular route.  
It would add complexity to the particular route chosen, as the skilled person would have had to  
consider using different protecting groups to address the different interactions between the  
fluorinating agents, their solutions and conditions of activity, and the various interactions of the  
precursor compounds.  
[516] I find the evidence as of January 8, 2004 supports Dr Wnuk’s opinion that there was no  
teaching as to how to fluorinate a tertiary carbon at a nucleoside’s 2’ position in a stereoselective  
manner. I also accept his evidence that it was not common general knowledge that using a  
nucleophilic fluorination approach, or indeed any fluorination approach, would successfully  
synthesize a 2’-C-Me/F compound.  
[517] There was no statement of common general knowledge from any text-or peer reviewed  
articles supporting the Idenix experts’ opinion that DAST would be the choice to follow in 2004  
for this particular form of synthesis. The Idenix opinions are largely based upon the number of  
articles from searches conducted regarding fluorination processes generally, none of which  
describe DAST being used to fluorinate a tertiary carbon in a ribose ring or nucleoside.  
Page: 185  
[518] With respect to the articles found from the searches, Dr Wnuk raises concerns for  
numerous reasons about the power of searches conducted a decade after the fact: “based on my  
own experience, [I] know that its search functions are more powerful today than they were even  
5 years ago”. This seems to be demonstrated by Dr Barrett’s testimony about the searches he  
carried out, as stated at para 236 of his November report:  
For instance, as can be seen from Griffon’s initial search, the 3’  
position in the nucleoside 41, for example, contains a –CH(OH)-  
group. I have on October 26, 2014 repeated this search in SciFinder  
with a 2002 or earlier filter and found this gave 426 references but  
none of these were nucleosides. Of these references, 15 were  
reactions using DAST (51) or Deoxo-Fluor® (52). Reaxys  
(Beilstein Crossfire in 2002) is a far better database for searching  
reactions than SciFinder. I have on October 26, 2014 repeated this  
search in Reaxys with a 2002 or earlier filter and found this gave  
44 references for the use of DAST (51) or Deoxo-Fluor® (52) or a  
related sulfur reagent for the conversion of a tertiary alcohol into a  
tertiary fluoride. It is interesting that Griffon only found 48  
references rather than the 426 that I found.  
[Emphasis added]  
[519] Dr Wnuk also challenges the claims of Dr Barrett with the results of his own search. He  
located 32 reactions using DAST as the fluorinating reagent contained in 7 references (5 papers  
and 2 patents). His search also located 25 reactions using HF as the fluorinating reagent  
contained in 10 references (3 papers and 7 patents). Consequently, he concluded that HF appears  
in more references than DAST.  
[520] Dr Griffon used HF as a fluorinating reagent before trying a scheme involving Deoxo-  
Fluor®. I find that the literature of the time did not sufficiently draw the skilled person to the  
Page: 186  
DAST family of reagents when embarking on a new synthesis with no teaching and only  
common general knowledge.  
[521] Dr Coe, the expert in the field consulted by Idenix on the very problem at the time in  
question, proposed four different methods for the synthesis of a 2’-C-Me/F nucleoside, none of  
which involved the nucleophilic fluorination of a tertiary alcohol on a sugar with DAST or an  
equivalent reagent. He mentioned DAST in the cover portion of his letter in passing as an  
example of his advice to adopt the sugar ring approach, as opposed to the nucleoside starting  
material.  
[522] It is significant that he did not recommend a DAST or Deoxo-Fluor® experiment, given  
that there was no mention of Dr Griffon’s February 13, 2003 experiment in Dr Storer’s letter of  
February 9, 2003 in respect of the target compound which was Target 9 as indicated:  
Target 9 is an attempt to replace the tertiary OH of the ribo  
analogue with fluorine. We’ve tried a variety of procedures from  
the exocyclic methylene analogue in attempt to effectively add HF  
across the double bond. We had no success with that. We’re now  
looking at attempting to take the 2’-α methyl anhydro compound  
and open that with fluoride. I’m not too hopeful for success with  
that. Appendix 3 shows a summary of this. Your thoughts on how  
to introduce the tertiary fluoro substituent in compound 1 would be  
appreciated.  
[523] If Idenix is correct that the most obvious first choice route would have been via DAST or  
Deoxo-Fluor®, Dr Coe’s letter should have laid out the scheme involving the need to invert the  
substituents at the 2’ position on the sugar ring etc. Instead, Dr Coe refers to DAST only in  
support of using the sugar route approach “for the very reasons you have discovered in that the  
Page: 187  
leaving groups generated in situ (e.g. in DAST reactions) are readily attacked by the pyrimidine  
ring nucleophiles or elimination and/or participation of blocking groups”. According to Idenix’s  
experts, DAST was the obvious route of any retrosynthetic analysis, yet it is not even stated as  
one of the suggested schemes by Dr Coe. This is not to cast aspersions at Dr Coe. It reflects that  
no one knew how to synthesize the 2’-C-Me/F nucleoside prior to Jeremy Clark finding a route,  
apparently with lots of advice from the senior chemists he worked with.  
[524] I have already stated that I accept the accuracy of Dr Wnuk’s statement that he is not  
aware of any publication prior to 2005 that describes the use of DAST or Deoxo-Fluor® to  
fluorinate a tertiary alcohol on a nucleoside in a stereoselective manner, particularly at the 2’  
position on the Deoxy sugar ring. At that time, it was not known nor could it have been predicted  
with any degree of confidence that DAST or Deoxo-Fluor® could be used to fluorinate a sugar  
(or sugar moiety) bearing a methyl at the 2’ (up) position or at the 2’ (down) position.  
[525] Dr Damha takes issue with this statement, but the examples of the stereospecific  
structures he provides of fluorinating are largely of a secondary and those for tertiary carbons  
bear no similarity with those of the sugar ring with the methyl in the (up) placement at the 2’  
position.  
[526] Dr Wnuk states at para 129 et seq. of his first report that the outcome of using DAST type  
agents depended on the structure, as follows (footnotes omitted):  
Of importance, the behavior of the fluorinating agent  
DAST/Deoxo-Fluor® had been shown to depend on the structure  
of compounds on which it is being used. More specifically, the  
Page: 188  
ability of DAST/Deoxo-Fluor® to generate a fluorinated  
compound was known to be dependent on the structure of the  
starting material (e.g, primary, secondary vs. tertiary alcohol; linear  
vs. cyclic compound).  
[527] An example of the tertiary fluorination relied on by the Idenix experts is found in the  
article of Van Robays M. et al. (1986), J. Chem. Soc. Perkin Trans 1: 251-254, which is also  
cited by Dr Wnuk regarding the unpredictability of fluorination outcomes:  
[528] There was also a problem with the unpredictability of DAST reactions. In January 2004,  
the skilled person would have been aware that fluorination reactions can have unpredictable  
outcomes. This was especially true when one was trying to synthesize a new nucleoside with a  
novel substitution pattern on the furanose sugar ring. Such would have been the case for the  
skilled person, as of January 8, 2004, trying to make a 2’-C-Me/F nucleoside. This was  
confirmed by Mr Clark and the team of authors from Pharmasset in their 2005 article who wrote  
as follows:  
Page: 189  
The fluorination of tertiary alcohols using DAST has been  
reported, but the stereochemistry of such trans-formations is  
substrate-specific and often unpredictable. For instance, Yang et al.  
reported that the DAST fluorination of a tertiary alcohol in 2-  
bromomethyl-DL-myo-inositol proceeds with retention of  
configuration. Wachtmeister et al. obtained a 4-fluoro-1-  
cyclopentanol containing a tertiary fluorine in 25% yield using  
DAST as a fluorinating reagent, and this transformation proceeded  
with inversion of configuration. Further-more, dehydrations or  
eliminations, rearrangements, and ring contractions are often  
pervading problems in the DAST fluorination of highly  
functionalized molecules.  
[529] I agree with Idenix that there were articles that discussed the use of DAST or Deoxo-  
Fluor® in the synthesis of carbohydrates. For example, both parties cited the 2002 Singh article  
Recent Advances in Nucleophilic Fluorination Reactions of Organic Compounds Using  
Deoxofluor and DAST [Singh] Exhibit 35b. Its purpose was “to summarize the most recent  
advances in the introduction of fluoro- and fluorinated groups into organic compounds using  
Deoxo-Fluor® and DAST as nucleophilic fluorinating reagents”. When questioned on the  
comment from the article that “nucleophilic fluorination reactions of organic compounds using  
fluorinating reagents are one of the most widely used methodologies in the field of fluorine  
chemistry”, Dr Wnuk responded that the same would be said about any article reviewing a  
specific reagent.  
A.  
Yes. But I agree with this of how the review articles, review  
articles is, of course, written on DAST, so if review article would  
be written on different fluorinating reagent, like HF, most probably  
the author would start “HF reagent is widely used reagent”, that is  
introductory sentence; yes? Which is very common how we write  
review article. But that is true in terms of DAST.  
Page: 190  
[530] In fact, “one of the most widely” means there must have been other widely used  
methodologies, which was Dr Wnuk’s evidence.  
[531] I do not find that the cross examination of Dr Wnuk established anything but that the use  
of DAST type agents to fluorinate was not common general knowledge to produce a 2’-C-Me/F  
compound in 2004. The cross examination led Dr Wnuk through the process of SN1 and SN2  
reactions with DAST-type agents. He agreed that it was generally known that they could produce  
a substitution and a fluorination in one step using the 2’-C-OH/Me intermediate compound that  
could result in a mixture containing the target compound.  
[532] In doing so, he was referred to the Singh paper in respect of the Wachtmeister paper  
[Exhibit 222] cited at reference 22 from the Table of 13 examples of Table 2 described as  
“Fluorination of Hydroxyl Group Located on a Pyranoside Ring”, which involved a fluorination  
of a tertiary alcohol as follows:  
[533] Dr Wnuk pointed out that the Wachtmeister and related papers are not in respect of a  
fluorination at the 2’ position of the sugar ring or nucleoside. He stated at para 237 of his first  
report that in contrast, these papers discuss fluorination at the 4’ position, which has different  
reactivity and (in a nucleoside) is not as sterically hindered as the 2’ position (due to the  
proximity of the base to the 2’ position of the sugar). In addition, these papers deal with  
Page: 191  
carbocyclic nucleosides or sugars which have CH 2 and not oxygen (O) in the ribose ring.  
Presence of oxygen at the adjacent carbon often affects the fluorination.  
[534] I also could not help but note that in Singh, in the paragraph including the table citing the  
Wachtmeister article, when describing DAST or Deoxo-Fluor® used as the reagent for a  
carbohydrate ring, the authors pointed out at pages 2565-66 the challenges those reactions pose,  
as follows:  
There are only a very small number of naturally occurring  
fluorinated carbohydrate compounds, an example of which is  
nuc1eocidin (53), an antibiotic isolated from streptomyces clavus.  
Because of their low natural abundance, but widespread utility and  
acidity, the syntheses of fluorinated carbohydrates are of great  
importance. However, because numerous protection and  
deprotection steps are required to set up the desired hydroxyl group  
for substitution with fluorine, introduction of fluorine into a  
carbohydrate moiety is difficult. Deoxofluor (1) and DAST (2)  
have been utilized for such molecules (Table 2)  
[Emphasis added]  
[535] During his cross-examination Dr Wnuk, as he was expected to do as an expert assisting  
the Court, acknowledged that the piecing together of various references pulled from articles  
treating a wide variety of situations could produce what at the time could only have been a  
theoretical route to synthesize the 2’-C-Me/F compound. It in no way convinced the Court that  
Dr Wnuk, the only expert testifying with an expertise in fluorinating these compounds at the time  
in question, was not correct in concluding that there existed no common general knowledge that  
a DAST approach should be followed to synthesize the 2’-C-Me/F compound in 2004.  
Page: 192  
[536] Finally, in the alternative, I conclude that being drawn to a reagent, or it being a first  
choice, does not constitute common general knowledge or meet the requirements of section  
27(3)(b). It cannot be otherwise, because routine experimentation cannot serve the purpose of  
establishing the correctness of the choice. To be common general knowledge in relation to the  
synthesis of a compound, the statement must possess some fore-ordained degree of near-  
certainty, to be confirmed, not proven, by routine experimentation. Nor would a first choice of  
the key synthesis reagent be a sufficient basis for going forward.  
[537] Similarly, section 27(3)(b) requires that the description of how to make the compound be  
almost certain, with some scope for routine experimentation to make the invention work.  
Otherwise, this would not be “set(ting) out clearly the various steps in … making … a  
…composition of matter, in such full, clear, concise and exact terms” as to enable its synthesis.  
The statement from the common general knowledge cannot be a description of a likelihood, or a  
sound prediction of a choice of a route, or key reagent to follow to synthesize a compound. That  
would not be sufficient hard coinage for a monopoly.  
(d)  
Toluene as the solvent for DAST reactions  
[538] As described in the excerpt above, Dr Damha is of the view that once the choice of DAST  
is made, the rest is “routine trial and error experimentation expected of a skilled chemist.” This  
view extends to both the choice of toluene as a catalyst and the conditions for its employment.  
[539] In my mind, the resort to toluene as a reagent raises an issue of where common general  
knowledge ends and trial and error experimentation begins. The conditions for use of toluene, if  
Page: 193  
accepted as the common general knowledge, to be employed with the DAST reactions in the  
circumstances, would nevertheless require consideration of the issue of whether the  
experimentation for its use would be undue.  
[540] In either event, if resort to toluene is a matter of trial and error experimentation, I find on  
the evidence of the experts, that its employment in the successful synthesis of the 2’-C-Me/F  
compound would be undue.  
[541] I am in agreement with Dr Wnuk that it was not common general knowledge that the  
synthesis of the 2’-C-Me/F nucleoside using DAST or Deoxo-Fluor® required the critical steps  
of the use of toluene as the solvent, or that the temperature of the solution should be -20 degrees  
as applied in Jeremy Clark’s successful synthesis.  
[542] The evidence of Idenix on the use of toluene was provided by Dr Damha at para 247 and  
48 of his reply report as follows:  
[247]Another thing that Dr Wnuk says is that toluene is not a  
solvent typically chosen for DAST reactions (para 212 of his  
report). However, this is contrary to two papers that he cited in his  
report. In Singh, [supra], the authors discuss the applications of  
Deoxo-Fluor and DAST as the most widely used methodologies to  
introduce fluorine in organic compounds. The authors state at page  
2576:  
Fluorinations with 1 [Deoxofluor] and 2 [DAST]  
were most frequently conducted in anhydrous  
solvents such as CH2 Cl2 , CHCl3 , CCl3 F, hexane,  
isooctane, toluene, H2O and glyme.  
[248]Yang S.S. et al. (1993) Carbohydrate Research 249:259-  
263 was also cited in Wnuk’s report, and also discusses the use of  
Page: 194  
DAST in toluene. These references show that toluene was known  
as a solvent that could be used for DAST fluorinations in the early  
2000s.  
In any event, other solvents including dichloromethane (DCM;  
CH2Cl2) also work for DAST fluorination. Regardless, the choice  
of solvent is another distraction, as it is a factor that can be  
routinely modified by a skilled person, and therefore would not be  
determinative of whether a skilled person would be able to make a  
DAST fluorination work in a routine manner.  
[Emphasis added]  
[543] Dr Wnuk points to the article by Singh, reviewing the literature to the effect that all the  
DAST or Deoxo-Fluor® examples used dichloromethane as the solvent listed, not toluene. The  
Singh article listed toluene as the fifth or sixth option in the review based on papers of many  
other scientists where DAST or Deoxo-Fluor® was used for the conversion. Dr Wnuk notes that  
“nobody is talking on (sic) the toluene, and those were the papers of many other scientists which  
were using DAST for the conversion”. The only article cited in the literature where toluene was  
used with DAST appears to be that of Dr Yang. The Singh article provides no other references,  
which explains perhaps why it is found at the end of the solvents used with DAST type reagents.  
[544] Dr Wnuk also points out the critical nature of the selection of toluene as the solution for  
DAST and the temperature for the reaction. The Singh article states on page 2562 that “For most  
of the compounds, fluorination proceeds below room temperature, sometimes as low as -78°C”.  
This is a range of almost 100°C. The Yang article reported that the toluene used with DAST was  
heated for 10-15 minutes at 70-75°C.  
Page: 195  
[545] The Court also relies upon Dr Wnuk’s observations at paras 190 to 194 of his first report  
that the use of toluene at a temperature of 20°C, along with the cytosine base (instead of uracil),  
was critical for the success of Dr Stewart and Ms Wang in ultimately synthesizing the 2’-C-Me/F  
nucleoside. Idenix appears to make no attempt to contradict this evidence, other than objecting to  
the introduction of Dr Stewart’s evidence, but not that of Ms Wang. I set out Dr Wnuk’s  
evidence in full because Idenix is relying on its employees eventually synthesizing the 2’-C-  
Me/F compound, even though it refused to admit the transcripts of Dr Stewart’s testimony in the  
United Kingdom case.  
[190]Of importance, the fluorination reaction discussed by Dr  
Stewart on page 135 of his laboratory notebook 081 was conducted  
in a toluene solvent. As was known in the art, toluene is not  
commonly used as a solvent for DAST and would not have been  
the solvent of choice by a skilled person for conducting  
fluorination reactions with DAST. It is worth noting that toluene  
was not even used by Dr Stewart in his previous attempts when  
using DAST. However, toluene was the solvent used and described  
by the WO Clark Publication to obtain 2’-fluoro(down)2’-  
methyl(up) nucleoside compounds using DAST. Although I do not  
know for certain, given the reaction conditions used by Dr.  
Stewart, I believe he used the teaching provided in the WO Clark  
Publication to help him with his attempt to synthesize the 2’-  
fluoro(down)-2’-methyl(up) nucleoside.  
[191]In reviewing Ms Wang’s notebook, it appears that during  
the month of December 2004, Ms Wang attempted to synthesize a  
fluorinated nucleoside having a 2’-methyl (up) and a 2’ fluorine  
(down) using a nucleoside compound containing uracil as a starting  
material. Ms Wang attempted to fluorinate the starting nucleoside  
with DAST (in a DCM solvent) but failed due to reorganization  
which is, as indicated above, a common occurrence when using  
this reagent.  
[192]In January 2005, Ms Wang started using another nucleoside  
compound bearing cytosine (instead of uracil) as starting material  
for the fluorination reaction. I note that, as of January 13, 2005, the  
WO Clark Publication had been published and suggested to use, as  
a starting material, nucleoside compounds bearing a protected  
cytosine as a base. Ms Wang protected the nucleoside compound  
Page: 196  
bearing a cytosine (e.g., at the 3’ and 5’ positions of the sugar  
moiety as well as the base) using the same protecting groups as  
those taught by the WO Clark Publication. Ms Wang then made an  
intermediate containing a ketone at the 2’ position of the sugar of  
the protected nucleoside to introduce a methyl substituent at the 2’  
(down) position of the sugar moiety using the methodology  
described in the WO Clark Publication. Following the introduction  
of the methyl substituent at the 2’ (down) position, Ms Wang  
changed the protecting groups at the 3’ and 5’ positions of the  
sugar moiety to those suggested in the WO Clark Publication.  
Finally, in March 2005, Ms Wang attempted to fluorinate a  
protected nucleoside compound at the 2’ position using DAST  
under the conditions taught in the WO Clark Publication.  
[193]I note that the fluorination reaction discussed by Ms Wang  
in her lab book entry of March 9, 2005 was conducted in toluene at  
a temperature of -20 o C. As indicated above, toluene would not  
have been considered to be the solvent of choice for conducting  
fluorination reactions with DAST. However, toluene was the  
solvent used and described by the WO Clark Publication to obtain  
2’-fluoro(down)2’-methyl(up) nucleoside compounds.  
[194]It is interesting to note that in her laboratory notebook 073,  
Ms Wang refers to the process of the “patent” or the “Pharmasset  
patent” without providing a specific reference to identify such  
document. However, it seems that this reference is likely to the  
process and methodology described in the WO Clark Publication  
since Ms Wang appears to have used the experimental conditions  
provided in the WO Clark Publication.  
(3)  
Excess of Choice  
[546] Dr Damha, recognizing the significant amount of choice facing someone setting off to  
synthesize the 2’-C-Me/F compound, states as follows at para 250 of his reply report:  
[250]Despite the wealth of literature available to the skilled  
person, Dr Wnuk also provides sweeping comments that many  
possible routes need to be tried, many different fluorinating agents  
or other reagents available throughout the process with no  
expectation of success (paras 112-5). I do not think that excess  
Page: 197  
choice would mean that a given transformation would not be  
routine. Rather, choice provides a safety net to the chemist in the  
event that one of the choices unexpectedly does not work as  
planned. More importantly, this ignores the schemes in the ‘191  
Patent relating to 2’nucleosides that can be used as logical starting  
points. To the extent Wnuk is suggesting that there are other ways  
to accomplish the conversion of an alcohol on a nucleoside to a  
fluorine, I agree. Unless he proves, which he has not, that these  
don’t work his point is meaningless.  
[Emphasis added]  
[547] The evidence does not support Dr Damha’s views. A wealth of literature available to the  
skilled person setting out choices is not any indication of whether the synthesis of the compound  
was common general knowledge. Certainly, routine trial and error experimentation does not  
occur where there is an excess of choice, not to mention that an excess of choice entirely  
excludes the possibility that the disclosure of the 2’-C-Me/F compound could be made by  
common general knowledge. In addition, I have already pointed out that the minimal inferred  
information contained in the ‘191 Patent provides no logical starting point to synthesize the 2’-C-  
Me/F compound, contrary to Dr Damha’s assertions.  
[548] I also disagree that Dr Wnuk’s evidence can be characterized as demonstrating other  
ways to accomplish the conversion of an alcohol on a nucleoside to a fluorine. His evidence  
demonstrated that Idenix did not meet its obligation to disclose how to synthesize the 2’-C-Me/F  
compound. He concluded that the fact that, in hindsight, individual steps in a chemical synthesis  
have some precedent in the literature does not mean that the overall sequence of steps for making  
a new compound was easy to determine. Most new compounds are made using reagents and  
methods that have been reported previously in the literature. However, determining the proper  
sequence of reactions to conduct, and the appropriate reagents for carrying out each step, often  
Page: 198  
requires significant creativity and/or an extensive amount of experimentation. I agree with these  
views, without making any finding on whether making the 2’-C-Me/F compound was an  
inventive step.  
(4)  
Reliance on Pharmasset Work  
[549] I agree with Gilead, that Idenix cannot rely on what Pharmasset had accomplished, which  
was not known to them when the Application was published. As there was no literature  
describing or available information explaining how the 2’-C-Me/F compound should be  
synthesized, there would be no common general knowledge that there was any synthetic route for  
its synthesis.  
E.  
Dr Griffon’s Work  
[550] Dr Griffon is a central figure in this litigation for a number of reasons. First, he attempted  
to synthesize the 2’-C-Me/F compound using a wide variety of fluorination agents for a period up  
towards two years without success. His failures therefore, represent real-time evidence that  
synthesis of the compound by Idenix could not be completed in reliance upon common general  
knowledge and routine trial and error experimentation.  
[551] Second, it is for this very reason, that, everything else aside, if Idenix cannot provide an  
explanation for Dr Griffon’s lack of success, its claim must fail. Idenix’s argument is that Dr  
Page: 199  
Griffon did not conduct himself as would a skilled chemist, and moreover misstated the results of  
his experiments, thus preventing his supervisors from correcting his alleged inadequacies.  
[552] Third, and a somewhat inconvenient corollary to his alleged incompetence, Idenix is also  
attempting to demonstrate that Dr Griffon successfully synthesized the compound without  
knowing it. Idenix attempts to prove this by a simulation said to be carried out by AMRI of Dr  
Griffon’s Deoxo-Fluor® experiment on synthesizing the 2’-C-Me/F nucleoside in February  
2003. I will deal with all of these contentions below.  
(1)  
Dr Griffon’s Attempts to Synthesize the 2’-C-Me/F Compound  
(a)  
Dr Griffon and Idenix’s Nucleosides Analogues Group  
[553] Dr Griffon is an experienced synthetic chemist who had synthesised hundreds of  
molecules, most of them nucleosides, over the course of his training and time at Idenix. Between  
1994 and 1998, Dr Griffon obtained his PhD in Dr Jean-Louis Imbach’s lab at Montpellier where  
his PhD supervisor was Dr Gilles Gosselin. Both are recognized as highly competent senior  
chemists in their fields. During the course of his PhD thesis, Dr Griffon made a number of 2’ and  
3’ fluoro substituted nucleosides. Following his PhD, Dr Griffon did a two year postdoctoral  
fellowship at the Southern Research Institute in Birmingham Alabama in a lab headed by a well-  
known nucleoside expert named Dr Jack Secrist.  
[554] Dr Griffon joined Idenix in February 2001 in Montpellier France, in the Nucleosides  
Analogues Group. At the time, the same Dr Gosselin who had supervised Dr Griffon’s PhD was  
Page: 200  
the director of research at Idenix. In and around 2004, Dr Griffon’s job responsibilities included  
conducting literature searches, developing novel synthetic strategies, carrying out those  
strategies, and analyzing reaction products.  
[555] Dr Griffon carried out his own TLC analysis and had access to nuclear magnetic  
resonance NMR and MS at Montpellier. Dr Griffon did not have access to liquid  
chromatography-mass spectrometry LC/MS (the process that combines the separation and  
characterization of reaction products in one step) until 2006.  
[556] Dr Griffon reported directly to Dr Gosselin and to Dr Storer, the latter being the senior  
vice president of chemistry at Idenix who split his time between Montpellier, France and  
Cambridge, Massachusetts.  
[557] Drs Gosselin and Storer directed the chemistry efforts at Idenix. Part of Dr Griffon’s job  
was to work with them on novel compounds and to see if new compounds could be synthesized.  
This involved reviewing literature and, if necessary, adapting literature to new applications.  
[558] Dr Griffon was assigned the task of synthesising a 2’-C-Me/F nucleoside in March 2002.  
Initially, the 2’-C-Me/F compound was not Dr Griffon’s priority, but by July 2002, this  
compound had been designated as a “high priority”. Dr Griffon was assisted by Audrey Chappe,  
a technician in the lab, and Elodie Pecheux, a trainee student.  
Page: 201  
[559] Dr Griffon and those assisting him, planned and executed a number of different general  
strategies for synthesizing a 2’-C-Me/F compound between March 2002 and the summer of  
2004. In the period of 2002 to 2004, the Nucleoside Analogues Group at Idenix Montpellier,  
included Drs Griffon, Storer and Gosselin who met every two months or so to review progress  
and identify target molecules and timelines. These meetings lasted approximately 2-4 hours on  
average and were summarised in a series of reports that were then circulated to the chemists.  
[560] In addition to these group meetings, Dr Griffon had other meetings directly with Dr  
Storer and often discussed chemistry problems or issues with the other PhD level chemists at  
Idenix including Drs David Dukhan, Frederic Leroy, and Jean-Christophe Meillon.  
[561] Dr Griffon’s work was summarized in monthly reports. Copies of these reports were sent  
to Drs Storer and Gosselin on a monthly basis, with copies sometimes to Dr Imbach. The purpose  
of these reports was to explain Dr Griffon’s synthetic efforts and achievements. Dr Griffon  
believes that his lab notebooks and monthly reports are an accurate record of his synthetic  
attempts. On cross-examination, Dr Griffon acknowledged that he does not have clear memories  
of his work between 2002 and 2004 separate and apart from looking at the documents.  
[562] In addition to his monthly reports, Dr Griffon summarised his synthetic efforts and  
literature searches in a report prepared at the end of the project. In this final report, Dr Griffon  
reported on his experiments, many of which were tried a number of times, using a number of  
different reaction conditions and involving a mix of nucleophilic and electrophilic fluorinating  
agents.  
Page: 202  
[563] On March 28, 2002, at an Idenix Montpellier Chemistry Meeting attended by Drs Storer,  
Gosselin and Imbach among others, Dr Griffon was given the task of synthesizing a 2’-C-Me/F  
nucleoside. At the time, Dr Griffon's priority was to make a 2'-methoxy (also described as OCH3)  
(down) nucleoside and he was also asked to make a number of other nucleosides with 2'-methyl  
(up) and various groups at the 2' (down) position. Between 2002 and 2004, Dr Griffon worked on  
other projects as well.  
(b)  
Literature Searches  
[564] Dr Griffon commenced his efforts with literature searches. These appear to have been  
discussed at the May 2002 Idenix Montpellier Chemistry Meeting. Idenix has criticized the  
manner in which Dr Griffon carried out the literature searches, indicating that they were too  
restrictive in the substituents entered into the Scifinder search engine and consequently did not  
turn up as many relevant articles as could have been found if properly done. This will be  
discussed below.  
[565] The results of Dr Griffon's literature searches were presented in an Idenix report on June  
27, 2002. His initial search using a general formula for 2'-methyl-2'-substituted nucleosides (with  
a methyl group in the 2' (up) position and various groups in the 2' (down) position, including a  
tertiary fluorine), yielded no results. Consequently, Dr Griffon understood at the time that he was  
being asked to carry out a novel synthesis.  
[566] Dr Griffon also searched using a general formula for a tertiary fluoride structure. Dr  
Griffon explained that he did so because he could not find any references to a sugar or nucleoside  
Page: 203  
with a fluoro methyl group (where a methyl group and fluoro atom attached as substituents to the  
same carbon) so he extended the search to more general structures. This search gave 48 results  
for a tertiary fluorine.  
[567] Dr Griffon refined his search to look for the fluorination of a tertiary alcohol, and this  
search yielded two results. He carried out another refinement using a search for the  
transformation of an epoxide to a linear molecule containing tertiary fluorine, and this yielded  
two results. Dr Griffon conducted these searches in order to explore different potential strategies  
for introducing a fluorine atom at the 2' position.  
[568] Based on all of the searches that he did, Dr Griffon proposed a synthetic route to 2’-C-  
Me/F nucleosides in his June 2002 report. This route was based on synthetic strategies for  
installing fluorine at the 4' position. In his report, Dr Griffon set out scientific references relating  
to 4’ fluorinated nucleoside derivatives before setting out his proposal, which was a 2'-ethenyl  
nucleoside strategy.  
[569] By July 2002, the 2’-C-Me/F project was described as a “high priority” in a summary of a  
presentation given by Dr Griffon at an Idenix Montpellier Chemistry Meeting at that time. Dr  
Griffon noted in his summary that the proposed synthetic strategy had to be refined in order to be  
started as soon as possible.  
[570] Dr Griffon explained on cross-examination that the “high priority” designation meant that  
“the main chemistry effort either by me or Audrey [Chappe]” had to be the 2’-C-Me/F  
Page: 204  
nucleoside. This is reflected in a summary of an Idenix Montpellier Chemistry Meeting held a  
little later in July 2002 where the 2’-C-Me/F target was described as a “new synthetic priority for  
Jean-François Griffon” and matched Dr Griffon's recollection of the project at that time.  
[571] Starting in September 2002, he proposed a series of strategies using different fluorination  
agents to synthesize the 2’-C-Me/F compound. Strategies 1 to 6 were nucleoside strategies, while  
the remaining strategies were sugar ring approaches.  
(c)  
First Attempts at the Synthesis of the 2’-C-Me/F Nucleoside  
[572] He first proposed a 2'-ethenyl nucleoside strategy with silver fluoride (AgF) as the  
fluorinating reagent. He determined that only two compounds had been synthesised, neither of  
which was the desired fluorinated nucleoside.  
[573] In light of this, Dr Griffon made two further nucleoside proposals in his September 2002  
report. One proposal was another 2'-ethenyl nucleoside strategy, using HF/pyridine AlF3 as the  
fluorinating reagent. The other involved the reduction of a protected 2'-iodomethyl (down)  
anhydro nucleoside (which was an undesired product of the failed first 2'-ethenyl strategy)  
followed by fluorination of the protected 2'-methyl(down) anhydro nucleoside with HF/pyridine,  
AlF3. Neither was successful as reported in the November and December 2002 reports.  
[574] Dr Griffon met with Dr Storer, Dr Gosselin and Professor Fleet in December 2002. They  
discussed Dr Griffon's work on the 2'-anhydro nucleoside strategy, and Professor Fleet proposed  
a synthetic strategy which involved the use of a positive fluorine source (an electrophilic  
Page: 205  
fluorinating reagent) reacting with a carbon-carbon double bond in a sugar. Gilead points out that  
Professor Fleet, recognized as a brilliant carbohydrate chemist, did not propose the fluorination  
of a tertiary alcohol with DAST or Deoxo-Fluor®.  
[575] In January 2003, Dr Griffon continued his work on the synthesis of a 2’-C-Me/F  
nucleoside. In his Progress Report he noted that “starting from the 2'-ethenyl derivative [i.e.  
Strategies 1 and 2] … all the attempted experimental conditions failed”. Dr Griffon also noted  
that he was unable to complete the reduction, being the first step of his third Strategy.  
[576] In the January report, Dr Griffon indicated that he would be trying an unprotected  
nucleoside under several experimental conditions. The various attempts involved fluorination  
with: i) HF-pyridine, AlF3 at 80° C and 120° C (Strategy 3 above, but without 3' and 5'  
protection); ii) KF, Kryptofix 2.2.2, pTsOH, DMF, reflux (Strategy 4); and iii) KHF2 refluxed  
in either ethylene glycol or 2-methoxyethanol (Strategy 5). The February 2003 Progress report  
stated that none was successful.  
(d)  
Letter Seeking Assistance from Dr Coe on the Fluorination Step  
[577] In February 2003, Dr Storer of Idenix wrote Dr Coe, a consultant who the Court  
recognizes was an expert in fluorine chemistry. In his letter, Dr Storer described a number of  
targets of interest to Idenix and asked for Dr Coe's suggestions on how they might go about  
synthesizing them. In the first paragraph of his letter Dr Storer noted that “[w] e are OK with the  
nucleoside chemistry, it’s the fluorine chemistry we are struggling with and where your help will  
Page: 206  
be valuable.” In the section of his letter discussing nucleoside analogues, Dr Storer noted that  
“we have a number of nucleoside targets which contain fluorine most of which we are struggling  
with. The targets are all ribonucleosides of one sort or another and have an additional substituent,  
usually a methyl group, at the 2'-position.”  
[578] Dr Storer went on to discuss the 2’-C-Me/F nucleoside, describing it as “an attempt to  
replace the tertiary OH of the ribo analogue with fluorine”. Dr Storer stated that:  
We've tried a variety of procedures from the exocyclic methylene  
analogue in attempt to effectively add HF across the double bond.  
We had no success with that. We're now looking at attempting to  
take the 2'-α methyl anhydro compound and open that with  
fluoride. I'm not too hopeful for success with that. Appendix 3  
shows a summary of this. Your thoughts on how to introduce the  
tertiary fluoro substituent in compound 1 would be appreciated.  
(e)  
The Deoxy-Fluor Fluorination of a D-Ribose Nucleoside with a Uracil Base  
[579] Dr Griffon's February 2003 progress report also sets out a summary of making a 3'-5'-  
protected 2’-C-OH/Me nucleoside and then fluorinating it using Deoxo-Fluor® (which is also  
described in the report as bis-(2-methoxyethylamino) sulphur trifluoride). This is the experiment  
that AMRI attempted to simulate in 2014. Dr Damha testified that Deoxo-Fluor® is very similar  
to DAST and is thermally more stable over a range of temperatures than DAST. Dr Griffon’s  
proposed synthesis was based upon the Wachtmeister article, wherein the substitution was of the  
tertiary C-4 hydroxyl group with a fluorine using Deoxo-Fluor® on the proper intermediate 2’-C-  
OH/Me nucleoside (molecule 5). The fluorination step in the strategy is highlighted in the  
diagram below.  
Page: 207  
[580] The February 2003 progress report records Dr Griffon's view in relation to the Deoxo-  
Fluor® experiment that “one new compound was formed during the reaction”, and he identified  
this as a compound with a carbon-carbon double bond at the 2' position (a 2'-ethenyl derivative).  
His view at the time was that the reaction had failed. Dr Griffon repeated the reaction on 19  
February 2003 using slightly different reaction conditions. Idenix argues that the 2’-C-Me/F  
nucleoside was in fact synthesized during this experiment. Idenix is highly critical of Dr Griffon  
in respect of his failure to separate and analyze the results, which is considered in detail below.  
(f)  
March 2003 LiF and TBAF as the fluorinating reagents  
[581] In March 2003, Dr Griffon made two new proposals for the 2’-C-Me/F project. One was a  
2'-anhydro unprotected nucleoside strategy using LiF as the fluorinating reagent. The other was a  
2'-methyl-arabino nucleoside strategy using (CF3SO2) 2 O to make an intermediate and TBAF as  
a fluorinating reagent to displace the intermediate on a protected nucleoside.  
[582] Dr Griffon's April 2003 progress report indicates that the reaction to fluorinate an  
unprotected 2'-anhydro nucleoside using LiF which he had proposed in March 2003 had failed.  
Page: 208  
(g)  
Dr Griffon Attends a Fluorination Course  
[583] From April 1-4, 2003, Dr Griffon attended a fluorination course in Stratford-upon-Avon  
with one of his colleagues from Idenix, Dr Claire Pierra. They summarized what they had learned  
at the course in a report which they sent to Dr Storer who subsequently forwarded it to other  
Idenix chemists. Their email states that they “focused on the electrophilic and nucleophilic  
fluorine sources because we thought it was the most interesting part for our current chemistry.”  
The course was called “Making and Using Fluoroorganic Molecules” and was given by Professor  
Jonathan Percy and Dr Alison Stuart. The report discussed numerous electrophilic fluorinating  
reagents and nucleophilic fluorinating reagents, and set out potential applications of some of  
those reagents for the synthesis of Idenix targets. There does not appear to have been any special  
consideration of DAST type reagents.  
(h)  
Dr Coe Replies  
[584] On April 9, 2003, Dr Coe replied to the letter which Dr Storer had sent him in February  
2003. Dr Coe discussed the “more interesting nucleoside problems” and stated that:  
[i]n our experience and indeed in that of manner other particularly  
the de Clerc group the most viable routes to fluoro nucleosides are  
by base/sugar condensation methods the anomer problem not  
withstanding, for the very reasons you have discovered, in that the  
leaving groups generated in situ e.g. in DAST reactions are readily  
attacked by the pyrimidine ring nucleophiles or elimination and/or  
participation of blocking groups. Further migrations of groups can  
readily occur. See our papers in JFC 1993 62 145 and 1993 60 239.  
Having said this some of the routes you have tried are OK except  
that I think you are using the wrong reagents, leaving groups and  
reaction conditions.  
Page: 209  
[585] Dr Coe went on to propose four different routes to the 2’-C-Me/F target compound. None  
of the methods proposed by Dr Coe involved the fluorination of a tertiary alcohol (in particular, a  
2’-C-Me/F nucleoside or sugar) with DAST or Deoxo-Fluor®.  
a. Method 1 involved the fluorination of a 2'-methyl (down) 2'-hydroxyl (up) sugar via  
an imidazole sulfonyl intermediate using Et3N●3HFE as the fluorinating reagent.  
b. Method 2 involved a similar approach to Method 1, via an imidazole sulfonyl  
intermediate, but starting with a nucleoside instead of a sugar.  
c. Method 3 involved the fluorination of a 2'-anhydro nucleoside with AHF or  
Bu4NH2F.  
d. Method 4 involved the reaction of a tertiary alcohol with pyridine/HF.  
(i)  
Dr Griffon’s Work from May 2003 to July 2004  
[586] A summary of the Idenix Montpellier Chemistry Meeting in May 2003 indicates that the  
2’-C-Me/F project was “still a high priority” at that time.  
[587] Dr Griffon's May 2003 progress report indicates that he had made “attempts for  
introducing the fluorine atom at the 2'-(down) position following Paul Coe's report”.  
Page: 210  
[588] One of the strategies that Dr Griffon described in his May 2003 progress report is a 2'-  
methyl-arabino nucleoside strategy using an imidazole sulfonyl intermediate and Et3N●3HF as  
the fluorinating reagent adapted from the recommendation of Dr Coe. Dr Griffon determined that  
this step was unsuccessful.  
[589] Dr Griffon's May 2003 report indicated that when he attempted the fluorination step of  
this strategy, “no reaction occurred; the starting material was mainly recovered.”  
[590] The other strategy which Dr Griffon described in his May 2003 progress report in the  
section referencing Dr Coe's report was a 2'-anhydro nucleoside strategy using Bu4 NH2F as the  
fluorinating reagent. Dr Coe proposed the reaction of a 2'-anhydro nucleoside with Bu4NH2F in  
his April 2003 letter in Method 3.  
[591] In or around July 2003, Dr Griffon prepared a summary of his work on the 2’-C-Me/F  
project between September 2002 and May 2003. In that report, he also proposed an electrophilic  
fluorination sugar strategy involving fluorination of a 1,2-double bond with an electrophilic  
reagent, Selectfluor.  
[592] The summary of the July 2003 Idenix Montpellier Chemistry Meeting indicates that the  
2’-C-Me/F project was “still a high priority” at that time. The summary also notes that “a strategy  
starting from the corresponding fluorinated sugar might be the solution” and that “up to now, all  
procedures starting from a nucleoside were unsuccessful”.  
Page: 211  
[593] Dr Griffon confirmed in his July/August 2003 progress report that the 2'-anhydro  
nucleoside strategy using Bu4NH2F derived from Dr Coe's report was unsuccessful.  
[594] Dr Griffon's progress reports for September 2003 – January 2004 do not record him doing  
any work on the 2’-C-Me/F project. The project was discontinued at the Idenix Montpellier  
Chemistry Meeting held on 5 November 2003.  
[595] In February 2004, Dr Griffon did further work on the 2’-C-Me/F project, attempting a 2'-  
ethenyl nucleoside strategy involving the fluorination of nucleosides with protected bases.  
However, Dr Griffon reported that this strategy had failed in February 2004.  
[596] In March 2004, Dr Griffon sent Dr Gosselin a report regarding the various strategies  
which he had attempted and set out proposed sugar strategies. Dr Gosselin then sent the report to  
other members of the Montpellier team asking for their personal contributions to the project.  
[597] In the summary of the April 21, 2004 Chemistry Meeting between Dr Storer, Dr Gosselin  
and Dr Griffon, it was noted that three different sugar strategies were in progress in order to  
synthesize a 2’-C-Me/F sugar. The proposed sugar strategies were:  
a. Opening of a spiro α-chloroepoxide in a sugar.  
b. Addition of an electrophilic fluorinating agent on a sugar 1,2-double bond.  
c. Total synthesis of the sugar synthon.  
Page: 212  
[598] In his April 2004 progress report, Dr Griffon recorded work on these strategies. He  
discussed the opening of a spiro α-chloroepoxide on a sugar, and described the strategy as having  
been elaborated in collaboration with David Dukhan, a PhD chemist at Idenix. Dr Griffon  
reported that at that date they had not been able to synthesize the desired precursor for the  
fluorination step.  
[599] The electrophilic fluorination of a sugar 1,2 double bond strategy using Selectfluor was  
also in progress in April 2004, and Dr Griffon reported that up to that point the strategy had not  
been successful.  
[600] Finally Dr Griffon's April 2004 progress report also discussed the total synthesis of the  
sugar synthon strategy. This was the sugar strategy based on the method for making gemcitabine.  
That strategy was in progress as of April 2004.  
[601] By May 2004, the opening of a spiro α-chloroepoxide sugar strategy was on “stand by”  
which meant that Dr Griffon did not perform any other attempts on that strategy.  
[602] Dr Griffon also noted in his May 2004 progress report that the addition of an electrophilic  
fluorinating reagent on a 1,2-double bond sugar had failed.  
[603] Work on the total synthesis of the sugar synthon (gemcitabine) strategy continued in June  
2004.  
Page: 213  
[604] Dr Griffon made a further attempt at the electrophilic fluorination of a 1,2-double bond  
sugar in July 2004, but this did not produce a compound incorporating fluorine.  
[605] Further work was also done on the total synthesis of the sugar synthon (gemcitabine)  
strategy in July 2004, but the aldol condensation failed.  
[606] Dr Griffon's later progress reports suggest that no work was done on the 2’-C-Me/F  
project after July 2004.  
[607] At the end of the project, Dr Griffon prepared a summary report setting out all the work  
that had been attempted in relation to the synthesis of 2’-C-Me/F nucleosides. After discussing  
literature searches, the attempted nucleoside routes, and the sugar routes, the conclusion of the  
report stated that “all the strategies that were attempted to introduce a methyl group at the 2'-  
'(up)' position and a fluorine atom at the 2'-(down) position failed”. Dr Griffon sent his report to  
the Idenix chemistry team in Cambridge, MA.  
(2)  
Did Dr Griffon Meet the Standards of a Skilled Person?  
[608] The major hurdle to Idenix’s sufficiency of its enabling disclosure is the inconvenient fact  
that Dr Griffon failed to make the 2’-C-Me/F compound, which Idenix attempts to prove was not  
the case, and that he made it but failed to recognize his success. As pointed out, if the Court  
cannot accept Idenix’s arguments challenging the competence of its own employee’s work, then  
any argument based on common general knowledge and normal trial and error experimentation  
must also fail, given his efforts to synthesize the compound without success.  
Page: 214  
[609] Idenix’s strategy therefore, is to demonstrate that Dr Griffon made the compound  
supporting the contention that both he and Jeremy Clark succeeded demonstrating it was not  
difficult to do, at the same time to show he was incompetent, apparently to demonstrate that even  
incompetent chemists could make this compound with enough experimentation. This argument is  
then supplemented by the further submissions that he prevented his superiors and fellow chemists  
of the Idenix discovery chemical team from learning of his incompetence by misinforming them  
of his results.  
[610] Idenix has made a number of allegations against Dr Griffon with the view to  
demonstrating that he was not a skilled chemist, which I set out below from its written  
submissions:  
[447]While Griffon’s credentials may meet the level of the  
skilled person (in the Court’s view, they exceeded them), his work  
or his conduct does not meet the level of what would have been  
expected of the skilled person. In particular, Griffon:  
a)  
Performed a flawed literature search and subsequently was  
led down a suboptimal path;  
b)  
Did not follow the two most pertinent results from his  
literature search;  
c)  
Routinely did not bother to analyze the products that he  
obtained from a reaction;  
d)  
Failed to follow the experimental conditions as described in  
prior art publications that he attempted to rely upon;  
e)  
Did not follow the explicit instructions from Drs Coe and  
Fleet and from the fluorination course;  
f)  
Drew incorrect conclusions and misrepresented the  
outcomes of his experiments in his monthly progress reports;  
Page: 215  
g)  
Did not give his supervisors the opportunity to provide  
meaningful advice; his supervisors were only shown his incorrect  
progress reports, rather than his lab notebooks to know what  
Griffon or his assistant or student was actually doing;  
h)  
Exhibited a poor understanding of basic chemistry  
principles; and  
i)  
Exhibited poor judgment by allowing undergraduate  
students to carry out critical reactions.  
[611] I do not find all of the allegations to be totally germane, or at least supportive of Idenix’s  
argument. I therefore, take the liberty of restating what I understand its argument is in relation to  
these allegations, providing some preliminary comments where I think useful.  
(a)  
Conducted a flawed search, including not following the two most pertinent  
results from his flawed literature search, and as a result, failed to  
immediately adopt a synthesis strategy using DAST (Points a) and b));  
[612] Dr Griffon conducted his search of the literature applying the regularly used search tool  
SciFinder. Dr Barrett indicates that he found no references to articles or other materials that  
referred to the 2’-C-Me/F compound. He understood that he was being requested to conduct a  
novel synthesis without any literature advising how to go about making the target compound.  
[613] Dr Damha and Barrett claim that Dr Griffon incorrectly refined his search to exclude  
nucleosides or sugars by using the “CH 2” groups rather than “C” or “CH”. They submit that he  
should have been aware that the target compound does not have such structural features and that  
this supposedly excluded many useful references. However, they did not point to any particularly  
important reference that was overlooked as a result of this refinement of his search. I do not see  
Page: 216  
this criticism as having any effect on Dr Griffon’s decision to not initiate his synthesis using the  
DAST reagent.  
[614] I am also troubled by the fact that it does not appear that Dr Griffon was given an  
opportunity to comment on this alleged failure on his part. All that Idenix did was confirm the  
nature of his searches. He should have been presented with the opinions of Drs Damha and  
Barrett and asked for an explanation of whether he agreed with their criticisms and whether it  
made any difference in the outcome of his work.  
[615] Additionally, Dr Griffon could not recall whether this was a subject that he discussed  
with his superiors, given that more than a decade has passed since these events occurred. His  
searches were described in his reports and at least one presentation. This is just one of the topics  
that his superiors should have been called to speak to, if unsatisfied with Dr Griffon’s work in  
2002.  
[616] Dr Griffon was also criticized for not following the two most pertinent results from his  
literature search. Dr Griffon admitted that he did not see the reference in the Olah article relating  
to the fluorination of a tertiary alcohol. However, this is not relevant to the use of the DAST  
reagent, as the article is a discussion concerning the reacting of HF pyridine with a tertiary  
alcohol. None of the experts commented on it.  
[617] The report of his search results extends over 12 pages of various articles and materials.  
This might explain somewhat his overlooking an apparently irrelevant article.  
Page: 217  
[618] Dr Griffon was also criticized for not following the Van Robays article that turned up in  
his search, referred to above, in which the authors describe a tertiary alcohol being successfully  
converted into a tertiary fluoride with inversion of stereochemistry using DAST. Dr Wnuk noted  
that among the number of experiments described in the article, often the major product was a  
rearrangement, and not the desired fluorinated product. This is supported by Dr Coe’s indication  
that fluorine reactions were “messy”.  
[619] In addition, I have already pointed out the very different structure of the starting material  
used in the article. Dr Wnuk considered it an important distinction that the article did not involve  
a nucleoside or ribose sugar. I accept his evidence that the Van Robays article described a  
successful substitution of the fluorine for a tertiary alcohol in a structure bearing none of the  
stereochemical challenges of the nucleoside or the ribo sugar. I also agree that it could not  
represent the common general knowledge or bear sufficient weight to direct Dr Griffon to use  
DAST as the reagent to synthesize the target compound.  
[620] Moreover, Dr Griffon testified that his own work indicated that DAST gave very low  
yields, which is why he used Deoxo-Fluor® instead. This appears an accurate conclusion given  
the yields from Mr Clark’s experiments. Idenix also hinted that Dr Griffon ought to have used  
DAST instead of Deoxo-Fluor® in his experiments. I accept the evidence that that Deoxo-  
Fluor® is a surrogate of DAST, and Dr Griffon’s opinion that it was known at that time to be  
more stable and safer than DAST, and in some cases, more powerful. In any event, Idenix now  
claims that Dr Griffon successfully synthesized the 2’-C-Me/F compound using Deoxo-Fluor®,  
which puts that argument to rest.  
Page: 218  
[621] In this regard, I also reject any suggestion that the literature pointed to using the sugar  
approach rather than the nucleoside route, as Dr Damha tries to argue in his reference to the  
Wachtmeister article. Again this appears to be another irrelevant criticism, given that Idenix  
alleges that Dr Griffon was successful in fluorinating the ribose nucleoside, while Jeremy Clark  
apparently made the 2’-C-Me/F compound via both pathways.  
[622] In addition, Dr Griffon described the technical problems faced at Montpellier in simply  
obtaining literature found in his searches. He testified that after the set of searches was done, the  
recovery of all the publications was a long process because at that time, the Idenix chemists had  
limited online access to materials. For the others, it was a tedious process either to go to the  
University library to physically make a Xerox copy of the publications, or alternatively with  
greater delays, to run the bureaucratic gauntlet by ordering the publication through the library  
department of the University.  
[623] I have already found that the common general knowledge does not necessarily point to  
employing DAST or Deoxo-Fluor®. Moreover, Dr Griffon’s supervisors and other experts  
consulted raised no objections with recommendations of fluorination by both routes. I do not find  
the evidence of Idenix’s experts sufficient to demonstrate a failure on his part to conduct  
appropriate searches, or otherwise direct him to begin his synthesis using a nucleophilic  
fluorinating reagent.  
[624] I also take stock in Dr Griffon’s extensive experience in using different fluorinating  
agents, including using DAST to fluorinate a 2' or 3' secondary alcohol, with the other hydroxyls  
Page: 219  
protected with a benzoyl group. The topic of his thesis was to introduce fluorines at the 2' or 3'  
position of nucleosides. He was considerably more experienced and knowledgeable than the  
skilled chemist in the fluorination of nucleosides. He could be expected to rely on his own  
research and experience, particularly where, as Dr Wnuk stated, there was generally little  
expertise in the field in 2002-3.  
[625] There is evidence that Dr Griffon had previously discussed his work with Drs Storer and  
Gosselin, and likely also his fellow chemists at Idenix. Dr Fleet had also been consulted and had  
not recommended a route using DAST. This evidence only confirms that there was no common  
knowledge that would necessarily have led Dr Griffon to start off on the DAST approach. It also  
confirms that he was attempting to synthesize a novel compound, with little direction. This in  
turn justifies the efforts of several months before he attempted the Deoxo-Fluor® experiments in  
February 2003. By that time, Dr Griffon had been working on the 2’-C-Me/F nucleoside project  
for about 10 months. He had tried a number of strategies and different fluorinating agents. Dr  
Coe’s recommendations only arrived in April 2003 and they did not recommend fluorination of a  
tertiary alcohol on a sugar with DAST in any of the routes he proposed. Even if Dr Griffon had  
succeeded in his Deoxo-Fluor® experiment in February 2003, this would have only been after an  
excessive period of experimentation.  
[626] I therefore reject the submissions of Idenix that Dr Griffon did not adhere to the  
approaches of the skilled chemist in not commencing his synthesis project by using DAST to  
fluorinate the 2’-C-OH/Me nucleoside, or that his literature search played any significant role in  
his choice of synthetic pathways.  
Page: 220  
[627] I also adopt Dr Wnuk’s opinion that the routes selected by Dr Griffon were reasonable in  
the circumstances and required excessive experimentation if he was to succeed in inventing the  
2’-C-Me/F compound  
(b)  
Routinely did not bother to analyze the products that he obtained from a  
reaction (Point c); and drew incorrect conclusions (Point f);  
[628] Idenix alleges that Dr Griffon was successful in synthesizing the 2’-C-Me/F compound  
when he employed fluorination approach using Deoxo-Fluor®, but due to inadequate analysis of  
the products failed to realize this.  
[629] Two attempts were made by Dr Griffon to make the compound using Deoxo-Fluor® in  
February 2003. The first was on February 13, 2003 and the second on February 19, 2003. Only  
the first attempt appears to be the subject of comment by the parties. The experiment involves a  
number of steps, including the preparation or use of the intermediate 2’-C-OH/Me compound.  
None of his work in these tests appears to be the subject of criticism.  
[630] Idenix argues that Dr Griffon’s failure was in not analyzing the product of his synthesis,  
as the skilled drug discovery chemists would have done to determine all the compounds in the  
reaction mixture. Had he done so it argues, he would have realized that he had synthesized the  
2’-C-Me/F compound. It supports these claims by the successful synthesis in the alleged  
simulation of Dr Griffon’s Deoxo-Fluor® experiments that AMRI carried out in 2014.  
Page: 221  
[631] I do not accept these arguments for three reasons. First, I conclude that Dr Griffon  
exercised appropriate skill in conducting his experiments and in his decision not to analyze all of  
the reaction mixtures of his Deoxo-Fluor® experiment.  
[632] Second, I am not satisfied that the Deoxo-Fluor® experiments conducted by Dr Griffon  
succeeded in synthesizing the 2’-C-Me/F compound, as claimed by Idenix.  
[633] Third, Dr Griffon was constrained by his working environment at the University of  
Montpellier, which I find was a contributing factor to his not having employed mass spectra or  
NMR to analyze the reaction mixture claimed by Idenix to contain the target compound.  
(i)  
Dr Griffon’s Decision Not to Characterize a Reaction Product  
Because of its Lack of “Charring”  
[634] After completing the Deoxo-Fluor® experiment, Dr Griffon carried out an HPLC  
analysis. This enabled him to determine that there were six or seven different compounds in the  
reaction mixture. He could not identify the compounds using this technology. The detection of  
compounds required an analysis using a Mass Spectra or NMR testing apparatus. It is a  
significant fact that Montpellier owned these testing apparatuses. It imposed restrictions on  
Idenix’s capability to analyze compounds synthesized from its chemists’ experiments.  
[635] Dr Griffon purified and separated the crude reaction mixture using silica gel column  
chromatography. He eventually combined the separated product into three fractions intended to  
concentrate the compounds from the reaction. He identified the vials from the initial separation,  
Page: 222  
which he combined to make fractions. Fraction 1 of 27 mg was composed of vials 5-7, Fraction 2  
of 20 mg was the combination of vials 8-9, and Fraction 3 of 72 mg was vials 10-14 (the starting  
material).  
[636] Dr Griffon placed a drop from each fraction on the bottom of the TLC slide and drew a  
line above the three spots as a starting line. He placed the slide into a solvent which moves the  
products up the slide based on differences in solubility, polarity and absorption of compounds.  
The location of the compounds on the TLC plate can be determined by holding it under UV light.  
[637] Additionally, he stained the plate with sulphuric acid. This causes the “spots” of the  
compound to become visible. This type of stain preferentially charcoals carbohydrates such as  
sugars.  
[638] Spots containing sugars would be expected to char darkly when using a sulphuric acid  
stain. Compounds that do not contain carbohydrates (such as a base that was not successfully  
coupled to a sugar) will not char as darkly, whereas the 2’-C-Me/F nucleoside is a sugar that  
would char. TLC does not identify which compounds have been made, but gives an idea of  
whether a reaction occurred and can give some information about the nature of any products by  
the effect of the charring of stain carbohydrate spots.  
[639] The final TLC plate from the experiment using the Deoxo-Fluor® as the reagent is set out  
below. It is important to note that it is depicted as taken from Idenix’s written submissions along  
Page: 223  
with its explanatory labels, about which more will be said below. The compounds contained in  
three fractions are in the vertical lanes with the visible compounds being circled in pencil.  
[640] The evidence indicates that the compounds in the right-hand (starting material) and  
middle lane charred being indicative of sugar carbohydrates. There is no evidence of charring of  
the compound on the left lane.  
[641] Dr Griffon acknowledged that he was not able to identify the compounds in the fractions  
from his experiment using the final TLC plate. To do so required analysis of the reaction  
fractions using the mass spectra or NMR testing apparatus.  
[642] Idenix’s comments identify the uncharred left-hand column as containing the product  
which it argues that the skilled chemist would have analyzed using mass spectra or NMR. Dr  
Griffon identified the product of Fraction 2 in the middle lane as a carbohydrate that could be a  
ribose nucleoside. It was analyzed using the University’s mass spectra and NMR equipment and  
found not to contain the target compound.  
[643] The right hand lane, although apparently a carbohydrate mixture by its charring, was not  
analyzed. Dr Griffon assumed it to be starting material. This does not appear to be in dispute.  
Page: 224  
[644] Dr Griffon was asked why he did not analyze all the compounds using mass spectra and  
NMR. His explanation was as follows:  
Q.  
Now why like, if you had three fractions separated, why  
didn't you run mass spec or NMR on all three fractions?  
A.  
Because, okay, because on this one I assumed that it was  
the starting material based on the TLC analysis.  
MR. JUSTICE ANNIS: Now just help me there, did you light  
something up on the screen?  
THE WITNESS: This one.  
MR. JUSTICE ANNIS: The bottom one, yes, okay.  
THE WITNESS: This one was identified  
MR. JUSTICE ANNIS: That is what you thought was the starting  
material, the one on the right hand lane?  
THE WITNESS: Yes, yes.  
MR. JUSTICE ANNIS: All right.  
Page: 225  
THE WITNESS: And this one  
MR. JUSTICE ANNIS: You are looking at the one in the left lane  
now?  
THE WITNESS: Yes, was assumed as the base alone without the  
sugar, this one. And the reason why it is because it is not charring,  
so definitely there is no sugar attached on the base here.  
[Emphasis added]  
[645] I find Dr Griffon’s testimony to be clear on this issue. He concluded that the target 2’-C-  
Me/F nucleoside was not the product in the left lane of his experiment. His statement “definitely  
there is no sugar attached on the base here” was an expression of a strongly held opinion: there  
was no charring, therefore, it was not a carbohydrate, thereby eliminating the possibility of it  
being a 2’-C-Me/F nucleoside.  
[646] Dr Wnuk agreed this was a reasonable assumption for Dr Griffon to make. The relevant  
portions of his opinion are found at para 15 of his Sur-Reply as follows:  
[…]In my opinion, these were reasonable assumptions for Dr  
Griffon to make. In particular, it would have been reasonable to  
assume that fractions 5 to 7 contain the uracil base as the  
corresponding TLC spot does not appeared to have charred darkly.  
[…]Spots containing sugars would be expected to char darkly  
when using a sulfuric acid stain. As the spots corresponding to  
fractions 5 to7 does not appear to be very dark, it suggest that  
sugars are not present in large quantities in the spots.  
[647] Idenix’s counsel did not attempt to pursue any discussion of Dr Griffon’s assumption that  
the third lane product did not contain a nucleoside based on his comparative charring analysis.  
Page: 226  
Admittedly, he was called as Idenix’s witness, which might have drawn objections had his  
evidence been pursued. However, the issue of the charring of carbohydrates was not considered  
in any of Idenix’s experts’ reports or testimony of its witnesses, nor was Dr Wnuk cross-  
examined on the subject. It should be noted that the issue also arose in the United Kingdom  
Sovaldi decision, so it is not as though Idenix would have been surprised by Dr Griffon’s  
testimony.  
[648] Instead, Idenix appears to have adopted a novel strategy in its final submissions. It  
distinguishes between the two-circled compounds in the left-hand lane. For the first time that the  
Court is aware of, it claims that Dr Griffon ought to have analyzed the upper spot or spots from  
the Fraction 1 mixture in the left lane.  
[649] As far as the Court is able to determine, previous references in this case, and that in the  
United Kingdom, contained no distinction between the spots in the left-hand lane. Rather, the  
compounds in the left-hand lane were considered together for assessing the reasonableness of Dr  
Griffon’s assumption that there was no sign of charring in the reaction mixture.  
[650] However, from the labeling of the Idenix slide described above, it appears that Idenix  
indicates in its final written submissions provided after oral argument that the upper smaller  
circled compound ought to have been subject to further analysis by mass spectra or NMR. Idenix  
describes the upper spot as “unanalyzed” and references Dr Wnuk’s evidence as calling it  
“another product.”  
Page: 227  
[651] I disagree that Dr Wnuk, when referring to another product in the left-hand lane, was  
distinguishing between the larger lower and smaller upper spots.  
[652] This testimony arose as a result of the Court’s questioning in an attempt to better  
understand Dr Wnuk’s evidence, particularly given my inexperience with TLC analysis.  
MR. JUSTICE ANNIS: So he analyzed 8-9, which is the one in the  
middle?  
MR. MARKWELL: Yes.  
MR. JUSTICE ANNIS: All right.  
BY MR. MARKWELL:  
Q. Do you believe that a skilled chemist would believe that that  
was a reasonable decision?  
A. Yes.  
Q. Are you able to determine what Dr. Griffon obtained?  
A. I think in his report --  
MR. JUSTICE ANNIS: Can you explain why it is reasonable?  
MR. MARKWELL: Sure.  
BY MR. MARKWELL:  
Q. You have come to the view that Dr. Griffon -- a skilled chemist  
would have believed that what Dr. Griffon chose to analyze is  
reasonable; why do you say that?  
A. Because analyzing [the] polarity of the compound, hydroxyl  
group was exchanged -- if the reaction was successful, [it] would  
have been exchanged with fluorine. So the product would be less  
polar and will move higher on TLC than the starting material.  
MR. JUSTICE ANNIS: What do you mean by “higher”? On the  
slide?  
THE WITNESS: It would be middle spot.  
Page: 228  
[653] Dr Wnuk was referring to the two large charred stains in the middle and right hand lanes  
when discussing Dr Griffon’s decision to analyze the compounds in the middle lane, not those in  
the left lane. They both evinced evidence of being nucleoside carbohydrates by their charring,  
while the middle lane, because it had experienced the greater displacement of the two, would be  
indicative of the fluorine replacing the hydroxyl and thereby being situated above the right-hand  
compound on the TLC plate.  
[654] Dr Wnuk was not referring to the left-hand lane, when addressing this issue. He had  
already agreed with Dr Griffon’s assumption that it was unlikely to contain the fluorinated  
carbohydrate because it had not charred.  
[655] In its final submissions, premised on the labeling of the upper compound in the left-hand  
lane on the TLC plate depicted above, Idenix now argues that the higher position in the left-hand  
lane meant that the spot could not have been a base which would not contain fluorine as it would  
be positioned lower on the TLC slide.  
[656] Dr Barrett, who testified after Dr Griffon, did not address Dr Griffon’s assumption that  
the left lane Fraction exhibited no charring. Indeed, even when addressing the issue of polarity  
and its impact on the movement of the spot containing the uracil base while commenting on the  
TLC depicted above, he did not point to any of the spots in the left-hand lane indicating that the  
highest one would likely be the spot containing the 2’-C-Me/F compound. His evidence was only  
that “it doesn’t move very far, it doesn't go right to the top of the plate. It sticks more or less at  
the bottom.”  
Page: 229  
[657] Given the evidence before the Court at the time he testified, I conclude that Dr Barrett  
was not prepared to opine that the upper spot likely contained the 2’-C-Me/F compound. Given  
that he also did not challenge the charring theory, I find that his evidence cannot be said to  
contradict Dr Griffon’s assumption that the 2’-C-Me/F nucleoside with the sugar base was not  
present in Fraction 1.  
[658] This submission appears to have arisen from a further question intended to assist the  
Court, as to the product in the third lane, as follows:  
MR. MARKWELL: And maybe I am obtuse, but what is the spot  
in the left column?  
A.  
Q.  
In the left column is another product.  
No, no, no. There are three columns on the TLC slide.  
MR. SMITH: You just gave the answer to him.  
I don't know what the answer is; I couldn't answer it.  
MR. JUSTICE ANNIS: He says it's another product. Another  
compound, I assume.  
[659] As mentioned, all previous discussion focused on the large spots in the lanes which were  
being compared. There was no concept of the smaller upper spot or spots in the left-hand lane  
being the source of comparison. All the compounds in the left-hand lane had already been  
eliminated from consideration for lack of charring.  
[660] In my view, it was up to Idenix to clarify that it was distinguishing between the circled  
compound spots on the left-hand lane in its evidence led from Dr Griffon or its experts. It was  
Page: 230  
aware of the controversy around the charring on the TLC plate from the previous litigation in  
other jurisdictions.  
[661] It also was required to address the issue of the polarity of the base and its position on the  
TLC slide with Dr Griffon so that he could respond. I therefore attach no weight to this evidence.  
I accept Dr Griffon’s assumption that no target compound was synthesized in experiments  
conducted on February 13, 2003.  
[662] Any lack of success in synthesizing the target 2’-C-Me/F compound cannot be laid at the  
feet of Dr Griffon. It occurred for whatever reason related to all the variables such as the  
solutions and temperatures that affect the fluorination of a tertiary carbon on a ribose nucleoside.  
Perhaps it was due to the fact that fluorination processes are unpredictable, or “messy”, as stated  
by Dr Coe in his letter. Maybe a faulty product prevented the synthesis, such as occurred with the  
DAST reagent that apparently affected experiments conducted by Pharmasset, described by Dr  
Patterson. Or else it was due to the fact that on a certain day, inexplicably a reaction did not work  
as was recognized by Dr Damha.  
[663] In any event, the evidence does not establish that the 2’-C-Me/F nucleoside was  
synthesized on February 13, 2003, or that any lack of skill on the part of Dr Griffon contributed  
to the unsuccessful outcome.  
[664] I conclude therefore, based on all the foregoing evidence that Dr Griffon carried out the  
fluorination experiment as a skilled chemist using the Deoxo-Fluor®o reagent and arrived at a  
Page: 231  
reasonable conclusion that the 2’-C-Me/F nucleoside had not been synthesized in the resultant  
product.  
(ii)  
The Requirement to Characterize All Products of a Reaction in 2003  
[665] I also disagree with the evidence of Dr Damha who, when discussing the TLC plates  
obtained by AMRI, stated “seeing those TLC plates, a skilled person would have proceeded to  
purify and characterize the products formed in the reaction.” I accept Dr Wnuk’s opinion which  
is more realistic that the skilled person would not have purified and characterized all products in  
2003. It was pointed out that the TLC plates of both AMRI and Dr Griffon show that multiple  
products were observed. In some instances, at least 10 spots are observed in the TLC plate.  
Further, a number of the spots are more polar than the starting material. This suggests that  
protecting groups were lost from the starting material and/or reaction products or by-products.  
[666] I accept that the skilled person would not have purified and characterized every product  
formed by such a reaction, if at all. Particularly, in observing the degree of manual effort to  
purify using silica gel chromatography, it would not have been efficient for Dr Griffon to purify  
and characterize every product formed in reactions when there were 7 to 10 compounds. I also  
accept that the more advanced purification technology of HLPC was not routinely used as a first  
purification method for every bench top reaction, as to do so would have been expensive and  
unnecessarily complicated. The skilled person would likely have used a silica gel column to  
purify the crude reaction mixture.  
Page: 232  
[667] In any event, the Court has always understood that it was the larger spot in the left-hand  
lane that Idenix argued should have been characterized. Dr Griffon came to the conclusion that it  
was not a sugar as it had not been charred.  
(iii)  
Constraints on Analysis by Montpellier  
[668] I also comment on Dr Griffon’s evidence that he was constrained by Montpellier’s  
protocols that required reaction samples to exhibit a degree of purity and yield for  
characterization by its mass spectra and NMR equipment.  
[669] When this issue was first raised by Dr Griffon, he indicated that the standard procedure  
imposed by the University was to perform NMR and mass spectra on pure compounds, by which  
he meant after being purified by silica gel chromatography.  
[670] Later, when Idenix’s counsel followed up on this answer, Dr Griffon reiterated that he  
was required to provide “pure compounds”. He added that to perform a NMR analysis on  
fluorine “more than 12 milligrams pure” was needed. For mass spectra, he indicated that only 1  
mg was required. He did not state, nor did anyone question him on, the degree of purity required  
to meet the University’s standards.  
[671] The Court has little information on why these limitations were imposed, or the details of  
how these constraints were applied to prevent analysis of the mixture depicted by the spots in the  
left-hand lane of the TLC slide. However, Dr Griffon referred twice during his testimony to these  
limitations as a factor that prevented a follow-up analysis being conducted using mass spectra or  
Page: 233  
NMR. This is in addition to his being satisfied that the untested product was not a sugar, because  
it showed no sign of charring.  
[672] Idenix argues that it is unclear why this limitation was imposed by the University. It also  
argues, with obvious logic, that if unable to run these MS and NMR experiments, Dr Griffon  
would be left in a position where he did not have access to the basic analytical instruments that a  
skilled person would, should have had at his or her disposal.  
[673] This may be so, nevertheless, this evidence was entirely within the control and knowledge  
of Idenix and its witnesses who worked in collaboration with the University. The evidence was  
out of the mouth of Dr Griffon, an employee of Idenix called to testify on its behalf. Other  
witnesses should have been aware of these limitations. Dr Gosselin taught at Montpellier, while  
other Idenix staff would have been required to adhere to the same protocol. In the circumstances,  
without contradictory evidence, and finding Dr Griffon to be a credible witness, I accept that he  
was functioning in conditions that limited his access to the Montpellier’s compound  
characterizing equipment. These problems would be those of Idenix, and not the responsibility of  
Dr Griffon.  
(c)  
Dr Griffon Misrepresented His Work Effort Denying His Supervisors the  
Opportunity to Provide Meaningful Advice (Point f)  
[674] Before I deal with the specifics of a criticism of a misrepresentation, or the remaining  
claims of incompetence made against Dr Griffon, I highlight that Dr Griffon’s supervisors are in  
the best position to prove these allegations. Drs Gosselin and Storer supervised Dr Griffon during  
Page: 234  
this period. They are themselves highly regarded and competent chemists. There is no need to  
retain Drs Barrett and Damha to criticize Dr Griffon. The Court would give precedence to those  
involved in the events at hand who are in a position to describe the facts and provide their  
opinions of Dr Griffon’s failures. Without their testifying, it is not clear how the Court could find  
against Dr Griffon regarding the following allegations:  
1. Unknown to them Dr Griffon had conducted a flawed literature search;  
2. He was acting without their knowledge or instruction on how he carried on his work;  
They were misled by Dr Griffon’s reports and findings;  
3. His failure to analyze all reaction compounds, and particularly those of the Deoxy-Fluor  
experiments was not standard practice at Montpellier and was a significant factor in  
Idenix’s failure to synthesize the Compound;  
4. Dr Griffon placed his research assistant working with him on DAST experiments, in what  
Dr Barrett thought were dangerous circumstances; and  
5. Generally, that Dr Griffon had demonstrated a failure to understand basic chemistry  
principles.  
[675] I also admit to having concerns about Idenix’s treatment of Dr Griffon. He remains an  
employee of Idenix. Yet it is clear that he was called to testify to his own incompetence. I find  
this to be an intolerable position in which to place an employee. By my assessment, Dr Griffon  
came to court with the intention of helping out his employer, so long as he was not required to  
state any untruths, which I find to be the case.  
Page: 235  
[676] I have difficulty believing however, that he would have agreed to the subject of an all-out  
attack on his professional competence by his employer, its lawyers and highly regarded expert  
witnesses, making damaging allegations of the nature as set out above. I do not think he would  
have accepted to have someone of Dr Barrett’s stature testify that his demonstrated incompetence  
was such that he would not have lasted two weeks in his employ. I also find it unlikely that he  
would have cooperated with Idenix’s lawyers, if he had known that he would be addressed as  
“Griffon” throughout its submissions, when every other PhD involved in the case was addressed  
by their hard-earned title of Doctor.  
[677] But regardless of the perceived lack of fairness in the treatment of Dr Griffon, I  
emphasize my substantive conclusion that without his supervisors testifying, there is no  
evidentiary basis for most of the claims against Dr Griffon. Only Dr Griffon’s supervisors can  
provide the foundational facts upon which most of the experts’ opinions rest. I find it a  
reasonable assumption that Dr Griffon’s supervisors were not prepared to testify against a long-  
standing, apparently competent employee whom they do not blame for Idenix’s failure to  
synthesize the 2’-C-Me/F compound.  
[678] Despite the evidentiary shortcomings of the evidence against Dr Griffon, I will  
nonetheless respond to the key criticisms being made against him, particularly as I conclude that  
they are groundless or irrelevant to the issues at hand.  
[679] The evidence on Dr Griffon’s alleged misrepresentation is as follows:  
Page: 236  
Q.  
And do you see the fourth line from the bottom of the  
paragraphs at the bottom?  
A.  
Q.  
A.  
Q.  
Yes.  
One new compound was formed during the reaction?  
Yes.  
If we looked at either the HPLC or the fractions that you  
separated?  
A.  
Q.  
A.  
Q.  
Yes.  
Your report, it doesn't seem to align?  
Yes. No. No, it doesn't.  
Why did you -- did you ever tell Dr Gosselin or Dr Storer  
that you actually obtained more than one compound during this?  
A.  
Q.  
A.  
No.  
Why not?  
Because this was a summary, and I wanted [went] to the  
end point of my conclusion, and my conclusion was the main, the  
only compound as a main product formed was the -- was not the  
target compound.  
Q.  
And that is not consistent with the results of your  
experiment; is it?  
A.  
No, it's not.  
[680] This is what I would describe as an example of Dr Griffon going along with his employer  
by not attempting to provide any real explanation of what occurred.  
[681] First, from the foregoing evidence it is apparent that Dr Griffon concluded that the only  
compound that could have been the target compound, i.e. the middle lane, was not the target  
Page: 237  
compound. He did not need to state that the other compounds were eliminated because they did  
not char.  
[682] Second, from the comments of Drs Damha and Barrett about the 7 to 10 compounds  
described on the HLPC printouts and those of the AMRI experiments, it seems obvious his  
supervisors would have expected more than one compound from a reaction mixture involving the  
fluorination of a nucleoside. It is difficult to believe that his supervisors and the other members  
of the drug discovery team, with whom the evidence shows Dr Griffon discussed his work,  
would not know that many minor products would result from his experiments. A factor in this  
conclusion is that Idenix’s complaint is now addressed not at the large uncharred spot on the left  
lane, but the smaller ones up above it.  
[683] Third, Dr Griffon has testified that only larger yield purified reaction mixtures could be  
submitted to Montpellier’s Chemical Department, where apparently Idenix stood in line with  
students and everyone else in being required to fill out forms and wait a number of days to  
receive results of their analysis. It is difficult to believe that his supervisors and his co-workers  
would not to know about these operational restrictions limiting the number of compounds that  
would be analyzed in a reaction.  
[684] Fourth, Dr Griffon testified, as is expected, that he has no memory of the events in 2003.  
He could recall whether or not he had discussions with his supervisors and other chemists about  
the Deoxo-Fluor® experiment. It is unfair to Dr Griffon to raise an issue like this a decade later  
when his ability to respond is limited.  
Page: 238  
[685] Finally, there is no motive or reason for Dr Griffon to mislead his supervisors. His  
notebooks demonstrate that he was a meticulous chemist. It is also apparent that he was  
dedicated to achieving the synthesis of the target compound. There is no basis to conclude that he  
would leave out information that was not known to be required to be shared with workers with  
whom he had worked closely over a number of years.  
[686] I do not find that Idenix has established that Dr Griffon either misreported his results by  
leaving out information that his supervisors expected to be presented to them, or that they were  
misled by his reports, particularly as they did not testify.  
(d)  
Failed to follow the experimental conditions as described in prior art  
publications that he attempted to rely upon (Point d); Did not follow the  
explicit instructions from Drs. Coe and Fleet and from the fluorination  
course (Point e); Exhibited a poor understanding of basic chemistry  
principles (Point h); Exhibited poor judgment by allowing undergraduate  
students to carry out critical reactions (Point i)  
[687] I do not propose to respond to any further attempts by Idenix to destroy its own  
employee’s professional reputation, apart from some very brief remarks that Dr Griffon did not  
follow Dr Fleet or Dr Coe’s advice.  
[688] I am not aware of any specific advice Dr Fleet may have provided Dr Griffon. I have  
already pointed out that Dr Coe did not in fact recommend DAST in the schemes proposed, but  
only mentioned it in an introductory general statement in passing as an example of his advice to  
adopt the sugar ring approach as opposed to the nucleoside starting material. It turns out to be an  
Page: 239  
irrelevant criticism, but in any event Dr Griffon gave valid reasons to prefer Deoxo-Fluor® over  
DAST, including being based upon his own experience in working with both products.  
[689] In conclusion, with respect to the allegations against Dr Griffon, I find that it has not been  
established that he failed to act as a skilled chemist in the following respects:  
1. in his searches and conclusions drawn from his searches;  
2. in not immediately adopting a synthesis route using DAST or Deoxo-Fluor®;  
3. in purifying and characterizing the results of his experiments;  
4. in the information provided to his supervisors related to the results of his searches and  
experiments; and  
5. in his attempts to synthesize the 2’-C-Me/F compound, which I find it most likely were  
unsuccessful for reasons unrelated to any lack of competence on his part.  
[690] I also conclude that if Dr Griffon had synthesized the 2’-C-Me/F compound in February  
2003 it required undue research and experimentation to do so.  
F.  
The AMRI Experiments  
[691] The second leg of Idenix’s strategy to demonstrate that Dr Griffon successfully  
synthesized the 2’-C-Me/F compound relies upon alleged simulations of his experiment on  
February 13, 2003 by AMRI in 2014. AMRI successfully synthesized 2’-C-Me/F compounds in  
Page: 240  
both sets of experiments. This apparently is intended to support its argument that the synthesis of  
the compound by a DAST or Deoxo-Fluor® catalyst was a matter of common general  
knowledge.  
[692] However, I find that the experiments did not faithfully simulate those of Dr Griffon. I  
conclude that AMRI had access to technology that allowed it to both synthesize and characterize  
the 2’-C-Me/F compound more readily than could Dr Griffon in 2003. I also give less weight to  
the tests due to concerns about their confidentiality as directed by Idenix.  
(1)  
First Tests Conducted in Secrecy  
[693] AMRI carried out two simulated experiments of Dr Griffon’s February 2003 Deoxo  
Fluor® synthesis procedure. Gilead had no knowledge of the first test conducted in June 2014.  
Idenix later advised Gilead of the first experiment and was invited to attend the second one.  
Gilead sent a representative to observe the August 2014 experiment.  
[694] Gilead justifiably argued that the Court should disregard the June experiments because  
they were conducted in contravention of the Federal Court’s Notice to the Parties and the  
Profession regarding Experimental Testing, even though issued on November 27, 2014. Given  
that I find that neither test establishes that Dr Griffon successfully synthesized the 2’-C-Me/F  
compound in February 2003, I find no reason to reject the first test. I do, however, rely upon the  
first test being carried out in secrecy as a factor undermining the weight given to the second test.  
Page: 241  
(2)  
Failure to Consult Dr Griffon  
[695] I have reduced the weight accorded to these experiments due to Idenix’s failure to  
provide Dr Griffon with an opportunity to comment on the tests. In my view, Idenix should have  
obtained Dr Griffon’s input generally, and specifically on whether the techniques used by AMRI  
would have provided an enhanced means to synthesize or characterize the products of the tests he  
had conducted in 2003. Who better to advise the Court on the fidelity of the simulations of  
AMRI’s experiments with those in 2003 in terms of the successful synthesis and characterization  
of the product of the tests, than the person who conducted them in the first place?  
(3)  
The Failure of AMRI to Simulate Dr Griffon’s Test  
[696] Dr Wnuk provided evidence, which the Court accepts, that the AMRI experiments  
differed significantly from the protocol described in Dr Griffon’s notebook and Witness  
Statements in two major ways and other cumulative minor ways. His evidence is as follows:  
(a)  
Different Reaction Monitoring  
(i)  
AMRI Used More Powerful Monitoring Techniques  
[697] AMRI used HPLC attached to a mass spectrometer, the combination of which is referred  
to as LC/MS. I reject Idenix’s explanation that the use of LC/MS would not impact on whether  
the target product was identified in the reaction product, in comparison with the situation in  
2003. I accept Dr Wnuk’s evidence that this technology allowed AMRI to locate “to fish out” a  
Page: 242  
specific reaction product with a specific mass from a complex reaction mixture. The AMRI  
scientist could then focus on a particular peak that eluted from the HPLC, purify this fraction  
many times, and then characterize it. I acknowledge however, that it is not apparent from the  
evidence provided on the experiments that this occurred.  
[698] However, Dr Wnuk noted that it appeared during the AMRI 1 experiment that many  
analytical runs were tested on the AMRI LC/MS machinery in order to develop parameters that  
would provide adequate separation of the compounds present in the reaction mixture.  
[699] Dr Griffon did not have this technology available to him until 2006. He relied upon thin  
TLC which does not allow the same level of analysis and fraction identification. Idenix did not  
take any serious issue with this evidence. I reject Idenix’s explanation that the use of LC/MS  
would not impact on whether the target product was obtained.  
(ii)  
Different Purification Techniques  
[700] Idenix used reverse phase [RP]HPLC. This is a method of separating and purifying the  
compounds in the crude reaction mixture. Dr Griffon used a regular silica gel column. While  
both are separation of purification techniques, they are not functionally equivalent as to the  
detection limits and separation precision they provide. The compound AMRI made was fairly  
pure. Dr Damha believes it to be more than 90% pure.  
[701] Dr Wnuk opines that the more sophisticated HPLC technique used by AMRI may  
generate different results than those obtained by purification on regular silica gel column. HPLC  
Page: 243  
was not a purification method that would have been used by the skilled chemist in 2003. Jeremy  
Clark used silica gel column chromatography to purify the fluorinated nucleoside compound  
made from a Deoxo-Fluor fluorination in 2003. The simulation should have used the same  
technology, to truly simulate the practice that constituted the common general knowledge in  
2003. This would appear to be even more significant, given Idenix’s late identification of the  
minor product, rather than the major product requiring purification in the left lane as that  
synthesized by Dr Griffon.  
[702] Dr Wnuk concludes that the HPLC and LC/MS monitor may have allowed for isolation  
and identification of a desired compound from a complex reaction mixture that could not have  
been identified and isolated from the same reaction mixture using a regular silica gel column and  
TLC. In short, AMRI used more sophisticated techniques than Dr Griffon had access to in 2003,  
which may have generated results that could not have been obtained by the skilled person with  
the technology available, or regularly used in 2003.  
(iii)  
Cumulative Minor Differences  
[703] Dr Wnuk summarized additional differences between the protocols in Exhibit “A”  
attached to his report, which he would not classify as major differences. Nonetheless, he  
indicated that it is not possible to know the cumulative effect of the various minor changes in  
protocol on the experimental outcomes. These include issues on details such as: drying and  
desiccating glassware; whether all reagents and solvents were from open bottles; whether the  
syringes were rinsed that were used to dispense pyridine and Deoxo-Fluor; the absence of a  
Page: 244  
thermocouple used to monitor the internal reaction temperature; and no indication that the  
resulting solution was refluxed under argon.  
(b)  
Differences in the Use of the TLC Plates  
(i)  
AMRI did not Stain the Plates with Sulphuric Acid  
[704] Dr Griffon used a sulphuric acid stain the TLC plates, whereas AMRI used an iodine  
stain. Sulphuric acid preferentially chars (stains) carbohydrates such as sugars. In comparison,  
iodine staining preferentially binds to double bonds, but without the charring effect.  
[705] Idenix argues that different visualization techniques are not part of the reaction and did  
not affect what compound was being made in the reaction mixture. This is not the sole issue.  
Whether the reaction product contained the 2’-C’-Me/F nucleoside is less relevant in this matter  
than whether the technology and procedures followed by Dr Griffon in 2003 were consistent with  
the common general knowledge and sufficient to permit an initial screening out of a non-sugar  
compound. That was his basis for not characterizing the major product in the third lane, which up  
until the end of this trial had been the focus of the parties.  
[706] Dr Griffon used the TLC plate as an initial simplified screening process based on his  
years of working with the nucleosides to eliminate a compound that visually did not contain  
sugar. It appears to have been his regular practice. Idenix did not lead evidence to suggest that it  
was not a practice generally followed by its chemists, or by synthesis chemists generally in 2003  
when dealing with products isolated by silica gel column chromatography. He concluded that the  
Page: 245  
major product was not a sugar in comparison with the charring of the reaction products in the  
other two lanes. Had AMRI used sulphuric acid, a direct comparison could have been made to  
see whether Dr Griffon’s conclusion on sugar charring was correct, as well as in comparing the  
outcomes on the two plates. Given that the product was 90 percent pure 2’-C’-Me/F nucleoside,  
it would have been relevant to the Court to know whether it would have charred had it been  
treated with sulfuric acid. Idenix was aware of the significance of the charring issue in Dr  
Griffon’s decision not to characterize the product in the third lane, particularly as this relates to  
the allegations of incompetence it directs at Dr Griffon.  
(ii)  
Different Yield of Compounds  
[707] In addition, the TLC plates prepared by AMRI were taken at different reaction time  
points which differed from those prepared by Dr Griffon. Besides the purity of the yield of the  
2’-C’-Me/F nucleoside product, they also do not appear to have the same number of spots. In the  
case of the second AMRI tests, the TLC plates contain smears that may represent one compound  
or multiple compounds. Dr Wnuk found that the fact that Dr Griffon’s experiment and those  
conducted by AMRI yielded multiple different reaction products prevents meaningful  
comparison of the TLC plates from each experiment.  
(c)  
AMRI Testing Protocol Made in Secrecy  
[708] No explanation is offered why AMRI did not precisely follow the protocol used by Dr  
Griffon. Dr Clemens, the only witness to give factual evidence about the conduct of the AMRI  
experiments, had no information beyond being given the protocol and told to follow it.  
Page: 246  
[709] Additionally, from the documents that Idenix disclosed during the trial, which had been  
withheld on grounds of solicitor-client privilege, it was revealed that Idenix’s lawyers were  
directing Dr Clemens. Idenix provided no explanation why lawyers were directing AMRI on a  
technical subject matter such as this. It appears to the Court that Idenix used lawyers in order to  
maintain the secrecy of how the protocol was developed under the cloak of litigation privilege.  
Given Idenix’s failure to provide an explanation for these differences in procedure, the Court is  
prepared to infer that Dr Griffon’s protocol was not followed in order to improve Idenix’s  
chances of obtaining the desired result.  
(d)  
Idenix’s Argument that Gilead Should Have Carried out Its Own Tests  
[710] In reply, Idenix argues that if Gilead was of the opinion that the protocols followed by  
AMRI did not properly replicate those followed by Dr Griffon, it could have undertaken the  
experiment to prove otherwise. This is a facetious argument at best. It is also a red herring to  
Gilead’s arguments demanding explanations as to why Idenix undertook its tests in secrecy and  
failed to adhere in important ways to the protocol followed by Idenix’s own employee, who  
Idenix is attempting to prove was incompetent.  
[711] These facts bear no similarity to those in Abbvie Corporation v Janssen Inc, 2014 FC 55  
cited by Idenix. In that matter, the testing was of a product that was the property of Janssen,  
whom Justice Roger Hughes found had the means to perform the necessary tests on its own  
product. In this matter, Idenix is trying to demonstrate that complex multistep novel experiments  
conducted by its employee over a decade ago, (and who remains its employee) resulted in the  
Page: 247  
synthesis of a product, that it synthesized using more sophisticated tools to synthesize after the  
means of synthesis has been disclosed to the world.  
[712] Moreover, I would think that if Gilead had set out to replicate the tests of Dr Griffon,  
Gilead would have needed access to Dr Griffon to review in detail how he conducted his tests  
and the assumptions that he made as the testing went forward. Idenix’s tests were as much to  
demonstrate that Dr Griffon did not meet the standards of the skilled chemist, as they were to  
demonstrate that he synthesized the target compound.  
[713] There is no reason why Gilead should incur the expense of undertaking its own testing,  
when Idenix could have provided a better solution, simply by being transparent in its own  
process. Idenix should have offered to retain an outside testing agency mutually agreed upon by  
Gilead. Similarly, it could have been transparent on working up a protocol that replicated all the  
steps followed by Dr Griffon and by permitting Gilead to participate fully in all the testing, all at  
Idenix’s costs. In these circumstances, it would have had a plausible argument, even if Gilead  
rejected the offers, that the results of the tests should serve considerable probative value.  
[714] Instead, Idenix carried out the design of the protocol for testing in secrecy behind the  
cloak of its litigation privilege. It conducted the first test in secrecy without Gilead or the Court  
knowing who put the protocol together, or what adjustments may have been made in order to  
obtain the desired result.  
Page: 248  
[715] This was followed by a second test, all while maintaining secrecy over how the testing  
protocol was developed. Even then Idenix failed to properly simulate Dr Griffon’s process in  
important ways, all without rational explanation, allowing Gilead to argue that it teased out the  
desired result using sophisticated techniques not available to Dr Griffon.  
[716] In these circumstances, there is no reason why Gilead would undertake its own testing  
given the obvious, multiple and significant shortcomings of Idenix’s experiments intended to  
simulate those of Dr Griffon. Gilead could gain nothing from undertaking expensive tests when  
satisfied that Idenix could not reasonably convince the Court that the AMRI tests could  
demonstrate that the synthesis of the 2’-C’-Me/F nucleoside was a matter of common general  
knowledge and routine experimentation in 2003.  
[717] For all the above reasons, I conclude that the AMRI tests have little probative value in  
this matter.  
G.  
Work of Other Idenix Discovery Chemists  
(1)  
The Admissibility of Transcripts of Dr Stewart from the UK Proceedings  
[718] At the end of trial, Gilead sought leave to file the Witness Statement of Dr Stewart dated  
July 11, 2014 [Witness Statement] and a transcript of a related cross examination under oath  
dated October 9, 2014 [Transcript]. These documents are part of the record of the prior civil  
action between these two parties in the UK High Court of Justice - Chancery Division (Patents  
Page: 249  
Court), Claim No. HP14D1069 [the UK Proceeding] and relate to the same issues as in this  
matter.  
[719] Dr Stewart was a chemist with Idenix from September 2003 to December 2014, when  
shortly before this trial his employment terminated for unstated reasons.  
[720] Dr Stewart worked on medicinal chemistry projects for discovery purposes at Idenix. In  
or around May 2004, Dr Stewart, working at Idenix’s Cambridge, MA facility, became involved  
in the project to synthesize a 2’-C-Me/F nucleoside. He was also involved in extensive  
interactions with a patent law firm in ongoing patent litigation and interference cases for Idenix  
providing historical perspective and scientific input, inter alia.  
[721] Around December 2004, Ms Wang, another chemist at the Cambridge, MA facility also  
commenced work assisting Dr Stewart on the project.  
[722] Neither testified. Idenix sought and obtained Gilead’s consent to file the witness  
statement of Ms Wang, dated July 11, 2014, and the related cross examination transcript dated  
October 7, 2014 from the UK Proceeding.  
[723] However, Idenix did not call Dr Stewart to lead evidence as part of its case in this matter.  
Furthermore, it refused to consent to file Dr Stewart’s prior sworn testimony from the UK  
Proceeding at this trial.  
Page: 250  
[724] There is no issue as to the authenticity or accuracy of the transcripts. The fact that they  
were tendered and made under oath in open court involving the same issue between the same  
parties relating to the synthesis of the 2’-C-Me/F nucleoside, only confirms their reliability and  
trustworthiness.  
[725] There is also no issue of the credibility of Dr Stewart, or loss of advantage to this Court  
from not being able to observe the demeanor of the witness that was afforded the judge in the UK  
Proceeding. Any role such an advantage provided to the initial judge would have no application  
where any implied bias by an employee called by Idenix would be attributed in favour of its  
interests. The Transcript on its face, should serve its full probative value with little diminution for  
the fact that it was not provided in the presence of the Court.  
[726] As mentioned, when Dr Stewart testified on behalf of Idenix in the UK Proceeding, he  
was an employee of Idenix.  
[727] During discussions at trial, the Court offered the view that the statement and transcript  
were admissible as an exception to the hearsay rule to the extent that they were documents  
containing admissions against the interest of a party made by an employee with authority to make  
the admissions: S. N. Lederman, A.W. Bryant & M.K. Fuerst, The Law of Evidence in Canada,  
4th ed. (Toronto: Lexis-Nexis Canada Inc, 2014), ss 6.453 [The Law of Evidence].  
[728] Although not pursued by the parties in their submissions, I continue to be of opinion that  
the Witness Statement and Transcript of Dr Stewart are admissible by Gilead as an exception to  
Page: 251  
the hearsay rule for the truth of their contents as documents containing admissions of Idenix by  
an authorized agent.  
[729] Perhaps only admissible to the extent of the admissions against the interests of Idenix,  
based upon the requirement for context, I admit the totality of the Witness Statement and  
Transcript.  
[730] In its submissions, Gilead made reference to American and New Brunswick statues to  
allow prior testimony in a subsequent proceeding as an exception to the hearsay rule [The Law of  
Evidence, at s.6.371-6.379]. The exception allows a party to tender the testimony in a previous  
matter of a witness (as opposed to the party or its authorized agent) called by the opposing party  
on the same issue.  
[731] The parties disputed whether the testimony is admissible under the “principled approach”  
to hearsay exceptions, if found to be both necessary and reliable: R v Starr, 2000 SCC 40. The  
dispute between the parties focused only on the issue of necessity.  
[732] Gilead contends that Dr Stewart’ evidence is necessary because, even though no longer  
an employee, he remains under the control of Idenix. It tendered affidavit evidence that Dr  
Stewart advised Gilead’s counsel that he had continuing obligations to Idenix under a severance  
agreement. Upon being so advised, Gilead’s counsel terminated the discussions with Dr Stewart.  
Page: 252  
[733] Gilead submits that it is not in a position to elicit testimony from Dr Stewart where Idenix  
still exerts some type of control over him, but refuses to lead his evidence in chief.  
[734] I am not aware of any prohibition against a party calling employees or agents of other  
parties. Obviously this may pose some challenges, not of concern in cross-examination of the  
same witness. Nevertheless, if Gilead had sought permission of the Court, it would have been  
permitted procedures to allow Dr Stewart to testify in this matter. It also did not request Idenix to  
release Dr Stewart from any constraints on testifying that may have applied.  
[735] Gilead offered no jurisprudence supporting its submission that a party may not call an  
employee of an adverse party. I think that is because the principle of necessity does not apply to  
those circumstances.  
[736] I conclude therefore, that it was not necessary to admit into evidence the Witness  
Statement and Transcript of Dr Stewart as an exception to the hearsay rule based upon the  
principles of reliability in necessity.  
[737] They are admissible however, as an exception to the common law rule against hearsay as  
containing admissions of a party.  
[738] Were I to rule the transcript of Dr Stewart inadmissible at Idenix’s insistence, I would  
draw an adverse inference against Idenix for failing to call a witness who would have knowledge  
Page: 253  
of the facts and would be assumed to be willing to assist that party: (The Law of Evidence, at s.  
6.450, R v Jolivet, 2000 SCC 20 para 25).  
[739] Because, Ms Wang worked under Dr Stewart’s direction, this adverse evidence would  
also lessen the weight I would attribute to her evidence for lack of expected corroboration from  
Dr Stewart.  
[740] I add, as well, that the evidence of either witness would have no causal impact on my  
conclusion drawn from all the evidence that the fluorination step of the synthesis of the 2’-C-  
Me/F nucleoside in 2003 was not sufficiently disclosed by the common general knowledge and  
routine experimentation.  
(2)  
The Evidence of Dr Stewart and Ms Wang  
[741] I summarize the evidence from the UK Proceeding of the two witnesses as follows:  
(a)  
Dr Stewart received a Master’s degree in chemistry from  
the University of Durham in 1999. He obtained a Doctorate in  
organic chemistry from the University of Oxford in 2003, having  
studied under Professor George Fleet;  
(b)  
Both Dr Stewart and Ms Wang were provided with a report  
summarising the work of Dr Griffon;  
(c)  
Both obtained starting materials from the Idenix  
Montpellier site;  
(d)  
On May 10, 2004, Dr Stewart met with and received input  
from Dr Fleet regarding the 2’-C-Me/F synthesis project.  
(e)  
Dr Stewart also had discussions with his colleagues in the  
group on a weekly basis regarding the 2’-C-Me/F project.  
Page: 254  
(f)  
Dr Stewart conducted an extensive literature search  
amounting to an average of two hours a day for six months in  
respect of the 2’-C-Me/F project.  
(g)  
Dr Stewart’s first planned route, developed with Dr Fleet in  
May 2004, involved fluorination using TBAF or TASF, not DAST.  
However the proposed starting material to be obtained from Dr.  
Fleet was too unstable to send.  
(h)  
The initial reaction Dr Stewart undertook in November  
2004 on the 2’-C-Me/F project was to fluorinate a 2’-C’-Me/OH  
nucleoside sugar (as Dr Coe had recommended) resulting in a ‘2  
fluor (up), methyl (down) compound (i.e. the wrong stereoisomer  
of the compound). Although this was the inverse of the ultimate  
goal, Dr Stewart felt the reaction would provide useful information.  
Dr Stewart described this reaction as a “breakthrough” when he  
discovered that it had been successful.  
(i)  
Although he succeeded in making the “wrong” sugar, he  
viewed the results of the “correct” fluorination as uncertain saying  
“It is by no means certain that the [correct stereochemistry] will  
behave in the same way for the fluorination since the steric  
hindrance and reactivity of the alcohol functionalities is different.”  
(j)  
When Dr Stewart tried the fluorination on the “correct”  
starting material, the result was a complicated mixture revealed by  
TLC analysis. Dr Stewart considered this reaction a failure.  
(k)  
(l)  
Dr Stewart encountered a number of failed DAST reactions.  
It was not until after Idenix received some information from  
a Pharmasset employee being interviewed for a role at Idenix that it  
was decided to shift the focus back onto the nucleoside route and  
lower the temperature of the DAST reaction on the nucleoside.  
(m)  
On January 12, 2005 Dr Storer informed Dr Stewart by  
email that Dr Gosselin had told Dr Storer that "someone [Dr.  
Gosselin] interviewed just after Christmas who worked at  
Pharmasset told him they made the compound from a nucleoside  
which may be good news for the other approach which Alistair  
[Stewart] discussed.”  
(n)  
Dr Stewart responded saying this information could be  
“very handy and might narrow things down a bit”. He indicated  
that based on this information, he would switch the focus back to  
the nucleoside route and vary a number of conditions including the  
temperature.  
Page: 255  
(m)  
Meanwhile Ms Wang had recently started with Idenix.  
During the month of December 2004, Ms Wang attempted to  
synthesize a fluorinated 2’-C-Me/F nucleoside with DAST (in a  
DCM solvent) using a nucleoside compound containing uracil as a  
starting material. It failed due to reorganization.  
(o)  
I conclude that the information Dr Stewart received from  
the Pharmasset employee was likely communicated to Ms Wang  
who started a reaction with DAST and a lower temperature the  
very next day using another nucleoside compound bearing cytosine  
instead of uracil as the starting material. Idenix claims that Ms.  
Wang successfully made the 2’-C-Me/F nucleoside in January  
2005. It is acknowledged however, that she placed a question mark  
beside the compound in her monthly report because it was unclear  
whether the desired compound had in fact been made.  
(p)  
The evidence shows that thereafter, a nucleoside 2’-C-Me/F  
compound was synthesized using DAST in March, but following  
all the conditions taught by the WO Clark publication of January  
13, 2005.  
(3)  
Analysis  
[742] Idenix argues that Dr Stewart and Ms Wang, having received Dr Griffon’s notes, would  
have been put off the trail of proceeding directly to the DAST fluorination. I agree with Gilead  
that this is speculation and is contradicted by the available evidence.  
[743] Dr Stewart followed the advice of Dr Fleet, and moreover, undertook an extensive review  
of the literature prior to commencing his experiments. My view is that after working more than  
18 months and having experimented with a vast array of fluorinating reagents, Dr Griffon would  
not have been in a position to provide any advice or direction on synthetic routes to follow.  
Page: 256  
[744] Given the extensive amount of time and analysis that Dr Stewart put into his work and his  
unsuccessful experiments, Idenix obviously focuses on the synthesis efforts of Ms Wang. Idenix  
attempts to argue that after less than one month of being involved with the project, with no prior  
experience in fluorination reactions or working with DAST, Ms Wang proceeded through all the  
steps required to synthesize the 2’-C-Me/F molecule all on her own, without Idenix even  
acknowledging that she was reporting to Dr Stewart.  
[745] Idenix also argues that her success was not due to learning information from the  
Pharmasset employee or its patent. I find that her recollection of these events is too vague to be  
reliable, to the point of not being able to recall how she became involved in the project.  
[746] Dr Stewart moved forward only after he received information that had been gleaned from  
the former employee of Pharmasset. He switched strategies back to fluorinating a nucleoside and  
working with DAST at a lower temperature. Ms Wang who recently arrived with no experience  
in fluorination or working with DAST, began experimenting using the same fluorination project  
as she describes “in the sense that we discussed what each of us was doing” while being under  
his direction.  
[747] I am satisfied that the convenient coincidence of garnering information from a  
Pharmasset employee and thereafter its patent application played a role in Idenix’s successful  
synthesis of the 2’-C-Me/F nucleoside. I rely on Dr Wnuk’s evidence with respect to these  
employees’ reliance upon Pharmasset information to assist in synthesizing the target compound. I  
Page: 257  
attribute very little weight to any success that Idenix may have had in synthesizing the 2’-C-Me/F  
compound in 2005.  
[748] In any event, the work of Dr Stewart and that of Ms Wang, whose work I find was largely  
based on Dr Stewart’s work, confirm that extensive research was required prior to initiating  
experiments, including consulting with others. This evidence confirms the obvious that it was not  
possible to synthesize the 2’-C-Me/F compound in reliance on the common general knowledge  
and without extensive research and trial and error experimentation.  
H.  
Jeremy Clark’s Work  
[749] Idenix relies upon the evidence that Mr Clark was able to make a 2’-C-Me/F nucleoside  
analogue relatively speedily and easily using DAST in reliance upon the Matsuda article to assist  
in synthesizing the intermediate 2’-C-OH/Me compound.  
[750] By way of background, Mr Clark was hired at Pharmasset after obtaining a Master’s of  
Science in Chemistry at the University of Tennessee in 2001. His employment at Pharmasset  
started on July 23, 2001 without any work experience on fluoro nucleosides. On December 6,  
2002, Mr Clark’s notebook first mentions a 2’-C-Me/F nucleoside on page 25.  
[751] Idenix points out that despite the fact that Mr Clark had no prior experience with fluoro  
nucleosides and did not have the education level of the skilled chemist, he nevertheless  
succeeded in making a 2’-C-Me/F compound on his first DAST fluorination on January 27,  
2003. On February 17, 2003, Mr Clark repeated a DAST fluorination on the same protected  
Page: 258  
sugar. There is some confusion on these dates, inasmuch as the 2005 article published by Clark et  
al where he indicated that he first invented a 2’-C-Me/F nucleoside. This is contrary however, to  
Mr Clark’s notebooks, which he very poorly maintained.  
[752] Idenix attempts, as well, to rely upon a statement by Mr Clark that the entire DAST  
synthesis was a 15-minute reaction. Without being too hard on Mr Clark, who was not called to  
testify, the Court nevertheless accepts Dr Patterson’s description of him from various somewhat  
disobliging comments as someone who had a tendency to be very outspoken about his own  
talents and opinions. An example of Mr Clark’s idiosyncratic behaviour is perhaps demonstrated  
by his report of February 28, 2003, recording his nucleoside DAST reaction scheme stating,  
“This reaction sequence will make a good chapter in the Anals of Shitpot Science. There’s  
nothing elegant in this chemistry”, but that “this is acceptable at this point however.”  
[753] Dr Patterson testified that Dr Kris Pankiewicz advised Mr Clark to use DAST on a 2’-C-  
OH/Me compound because DAST was commonly used to convert alcohols to fluorides in tertiary  
systems. Accordingly, Idenix argues that the chemists at Pharmasset followed the same approach  
to make the 2’-C-Me/F compound as described by the chemistry experts in this proceeding.  
[754] Mr Clark prepared the 2’-C-Me/F Compound using the nucleoside route. In so doing, he  
indicates in his lab notebook that he was following the “exact procedure as reported by  
Matsuda”. Idenix argues that this further supports the earlier discussion where it was shown that  
the skilled person would be aware of, or would find, Matsuda and would follow its teachings.  
Page: 259  
[755] Mr Clark also made further references to the same Matsuda article on pages 81, 83, and  
92 of his Notebook 2. In a summary report made by Mr Clark on July 21, 2003, he emphasized  
that the game plan for making a 2’-C-Me/F nucleosides “was inspired by the work of Matsuda”.  
Thus, Idenix argues that to make the claimed compound, all it took was Matsuda and fifteen  
minutes.  
[756] Pharmasset undertook efforts to increase yield, which are reflected in the minutes of the  
Pharmasset Chemistry Meeting on August 22, 2003. Idenix claims that the minutes record that a  
literature search on fluorinating tertiary carbons was done and that over 100 articles were found.  
Idenix indicates that this is in contrast to Griffon’s incorrect search, where he only found two  
references.  
[757] However, the Court is not aware of the source of Idenix’s information. The exhibit  
referred to by Idenix in support of these claims indicates only “[p] resented results from a  
literature search on fluorinating agents.” The minutes recorded a discussion involving a number  
of fluorinating agents including DAST and Deoxo-Fluor® as well as a HF and NFTh. There is  
nothing in the exhibit indicating that the search was limited to fluorinating tertiary carbons,  
although there was a comment that “several references were found but not presented on high-  
yielding tertiary carbon fluorination but without any mechanism being mentioned.” The  
emphasis of the meeting appeared to be on obtaining higher yields.  
[758] Idenix also states that this Chemistry Meeting minute recorded that Dr Wojciech J. Stec,  
the head of Chemistry at Pharmasset, said that everyone continues to use DAST, and that DAST  
Page: 260  
was the “conservative approach”. This suggests that DAST was already extensively used at  
Pharmasset prior to the literature search that found 100+ articles.  
[759] However, in consulting the same exhibit, Dr Stec emphasized that when he first arrived  
he distributed articles on fluorination, “but despite that everyone continues to use DAST”. He  
pointed out that DAST was not used in the industry. He advised the chemists, “[d]on’t be  
conservative (by using DAST)”. He suggested using different fluorinating agents to increase the  
yield. This evidence suggests that Pharmasset had a preordained tendency to use DAST among  
many other fluorinating agents available. This predilection for DAST may have contributed to it  
being recommended by senior chemists for use by Mr Clark as a matter of habit. This clearly was  
not the case at Idenix, or represents the views of other experienced chemists involved or  
consulted on its synthesis.  
[760] On September 19, 2003, Mr Clark used Deoxo-Fluor® to make a benzoyl-protected PSI-  
6130 on a large scale. Mr Clark reported that a Deoxo-Fluor® fluorination resulted in 23% yield  
to give a protected nucleoside. The underlying experiment was found on page 118 of Clark  
Notebook 2 (Exhibit 23-I), where column chromatography was used to purify the protected target  
product. Therefore, Mr Clark also succeeded in making the target compound fluorinate on his  
first try using Deoxo-Fluor®.  
[761] I find that Idenix paints too rosy a picture of the ease with which Mr Clark synthesized  
the 2’-C-Me/F compound. In the first place, Dr Patterson suggests that Dr Watanabe was the first  
person at Pharmasset to write down the structure of the 2’-C-Me/F compound. He presented it as  
Page: 261  
one of the analogues of the Idenix compound at a group of chemists, indicating that he wanted  
members of the chemistry team to consider attempting its synthesis.  
[762] Second, in terms of pathways, Mr Clark first tried the gemcitabine route and only moved  
on to the carbohydrate route after that failed. In addition, Dr Patterson testified that he received  
advice from a senior chemist with experience to make the compound using the nucleoside  
approach after he was having trouble making progress. There is no need for a skilled chemist to  
conduct a literature search, if you are following the advice of a senior chemist in the  
organization. Dr Patterson testified as follows:  
There were, you know, discussions. It took him a long time to get  
anywhere on it. People began giving him advice, particularly Chris  
Pankiewicz, asking him to make it from cytosine, you make it from  
the nucleoside rather than trying to do the coupling.  
And I believe that was a much faster, more viable route because  
you don't have to worry about, you know, preparation of all the  
stereocentres that would arise from synthesis of the carbohydrate,  
so nature has done much of that for you.  
[763] Dr Patterson repeated this point including that Mr Clark did not follow the advice as  
follows:  
Q. So, now, when Mr Clark is working on his first chemistry, I  
think you said the carbohydrate or the fluorination of that, can you  
just let us know if advice was being provided by someone to him  
on that?  
A. Oh, yes.  
Q. And what was the advice?  
A. The advice was make it from the nucleoside, start from C.  
Q. And how was Mr Clark about receiving instructions?  
Page: 262  
A. Not good.  
[764] In other words, Mr Clark started with the gemcitabine approach, was advised by senior  
members of the chemistry department to use the nucleoside approach and apparently synthesized  
the first compound using the sugar ring approach, and thereafter repeated his success using the  
nucleoside approach with higher yields.  
[765] Similarly, Mr Clark did not apparently come upon fluorinating with DAST by himself,  
but was advised again by senior chemists with whom he worked, according to Dr Patterson as  
follows:  
Dr Patterson, you mentioned that advice was provided concerning,  
I think you said the nucleoside and the cytosine?  
A. Correct.  
Q. Was advice also provided on the fluorinating agents?  
A. I think so.  
Q. And what was that advice?  
A. DAST or some similar, you know, aminosulphur trifluoride  
[766] It also seems clear that Mr Clark received advice from numerous members of the  
Pharmasset chemistry group. The 2005 article published with him as the head author included the  
following co-authors: Laurent Hollecker, J. Christian Mason, Lieven J. Stuyver, Phillip M.  
Tharnish, Stefania Lostia, Tamara R. McBrayer, Raymond F. Schinazi, Kyoichi A. Watanabe,  
Michael J. Otto, Phillip A. Furman, Wojciech J. Stec, Steven E. Patterson, and Krzysztof W.  
Pankiewicz.  
Page: 263  
[767] It should be noted as well, that Dr Patterson testified that everyone was very excited when  
Mr Clark found the “hole” in the Idenix patent. The evidence is clear that synthesizing the 2’-C-  
Me/F compound was a high priority for Idenix. I do not believe it would be overly speculative to  
conclude that Pharmasset would have recognized the opportunity left open by the omission of the  
fluoride analogue in the Idenix patent.  
[768] According to Dr Patterson, Mr Clark was assigned the task because he spotted the  
coverage omission in the patent. He was having a lot of trouble making any progress and  
according to Dr Patterson was receiving advice from various senior chemists. Given his  
inexperience, and indeed lack of formal training from the poor state of his notebooks, I conclude  
it was unlikely that he was left to himself to synthesize the compound. Dr Otto described the  
group as a close knit discovery chemist team housed in close quarters. There is evidence that Dr  
Stuyver may have assisted Mr Clark in selecting the target for synthesis as well.  
[769] While Mr Clark may not have had the training of the skilled chemist, if integrated into an  
effective, highly experienced discovery chemistry team, it could be expected that he would be  
operating at a level of experience and knowledge beyond that of the skilled chemist.  
[770] However, I do not wish to speak poorly of Mr Clark when in fact , it seems that he was  
the sort of person who marched to his own drum, possessing a large measure of self-assurance  
and perhaps possessing some of the natural traits of an inventor. Dr Patterson describes Mr Clark  
as someone with very strong ideas and self-assurance: “Jeremy Clark did have some very  
concrete ideas about how to make that molecule. He was very aggressive in pursuing that  
Page: 264  
synthesis and he would tell anybody that.” Similarly, Dr Patterson testified as follows: “Well, Mr  
Clark was very exuberant. And he was, when he was sure of something, he was sure of it, and  
everybody who knew him knew that.” That does not seem to describe your non-inventive skilled  
bench chemist.  
[771] In the sister case between these parties in the United Kingdom, Justice Arnold suggested  
that Mr Clark may have been lucky in the invention of the 2’-C-Me/F compound. On the other  
hand, it may well be that Mr Clark is one of those individuals who is particularly talented in  
quickly absorbing great amounts of information and intuitively working out inventive solutions  
from a variety of options that were presented to him. He appeared to pick out the omission in  
coverage of the 2’-C-Me/F compounds in the Idenix patent and followed advice only when it  
suited him. He also settled on toluene from a number of solvents and reaction conditions.  
[772] In addition to the help that I find he was provided by fellow employees, it is not possible  
to determine to what extent Mr Clark’s particular inventive talents or luck led him to the  
synthesis of the 2’-C-Me/F compound.  
[773] Whatever the explanation for Mr Clark’s relatively quick synthesis of the 2’-C-Me/F  
compound, after the false start on the Gemcitabine route, it does not convince the Court that his  
success was the result of, or reflects, the common general knowledge and routine  
experimentation alleged by Idenix to account for its synthesis. As said, it may well reflect a  
variety of factors, including a “conservative” pre-disposition at Pharmasset to use DAST, not  
necessarily followed by skilled chemists.  
Page: 265  
[774] Mr Clark’s performance most certainly does not overcome the results of the analysis of  
the many issues and evidence considered above that demonstrate no common general knowledge  
regarding the synthesis of the 2’-C-Me/F nucleosides, nor the use of DAST to do so. This is also  
confirmed by the evidence regarding the efforts and contributions of other experienced chemists  
in 2002 and 2003.  
[775] The Court concludes that Idenix’s claim that common general knowledge and  
permissible experimentation supplemented the written disclosure, thereby providing enabling  
disclosure of how to synthesize the 2’-C’-Me/F nucleoside is not supported by the evidence. The  
disclosure of the synthesis of the 2’-C-Me/F compound in the ‘191 Patent is therefore,  
insufficient.  
IX.  
A.  
Overbreadth  
Introduction  
[776] Gilead altered somewhat its argument on overbreadth from what was originally argued in  
order to advance two novel arguments to support such a claim. The new arguments relate to  
Idenix’s failure to make the 2’-C-Me/F compound at the time of filing and publication dates.  
[777] It was originally argued that Idenix had, within the scope of its claims, compounds which  
were not made or could not be made. This related to the numerous compounds in Claim 1 and the  
compound’s Base. With the Court’s acceptance of the abandonment of Claim 1 and the chemist’s  
Page: 266  
focused construction on the scope of compounds in the claimed base, I find that there remains no  
case on these grounds for an argument of overbreadth.  
B.  
General Principles  
[778] Overbreadth arises by the requirement under section 27(4) of the Act that the specification  
end “with a claim or claims defining distinctly and in explicit terms the subject matter if the  
invention for which an exclusive privilege or property is claimed.”  
[779] It is common ground that “No inventor is entitled to a monopoly on more, or even a little  
more, than he invents”, (Radio Corporation of America v Hazeltine Corporation (1981), 56 CPR  
(2d) 170 at 188). As stated by the Supreme Court of Canada in Burton at para 16.  
It is stressed in many cases that an inventor is free to make his  
claims as narrow as he sees fit in order to protect himself from the  
invalidity which will ensue if he makes them too broad. From a  
practical point of view, this freedom is really quite limited because  
if, in order to guard against possible invalidity, some area is left  
open between what is the invention as disclosed and what is  
covered by the claims, the patent may be just as worthless as if it  
was invalid.  
C.  
The Parties Submissions  
[780] Gilead advances its argument that the claims of the ‘191 Patent are overbroad by a  
rationale that I would summarize as follows:  
Page: 267  
(i)  
At the time the patent application was filed and published, Idenix had been  
unsuccessful in making a compound within the scope of the claims; as they did not  
have a way of making the claimed compounds, they cannot be said to have completed  
the act of invention; and  
(ii)  
Not having invented the compounds, any claim to any such compounds is by  
definition overbroad.  
[781] The argument is novel. For that reason, I will set it out in verbatim, as follows:  
a. This issue raises two fundamental questions: (i) when is an  
invention made; and (ii) when can someone be said to have  
invented a class of compounds?  
b. In the early case of Christiani v Rice, the Supreme Court  
articulated the following principles which have been repeatedly  
referenced by the jurisprudence:  
It is not enough for a man to say that an idea floated  
through his brain…  
[…]  
[T]he date of discovery of the invention is meant the  
date at which the inventor can provide he has first  
formulated, either in writing or verbally, a  
description which affords the means of making that  
which is invented.  
[Gilead’s Emphasis]  
c.  
There is a dearth of Canadian case law on the threshold  
issue of when a compound can be said to have been invented.  
However, there is considerable American jurisprudence on the  
subject owing in large part to the fact that until very recently, the  
US used a “first to invent” system which often required the Court  
to establish dates of inventorship.  
Page: 268  
d.  
Under US law, a compound is not said to have been  
invented until the inventor has the idea of a compound and an  
operable way of making it. (See for example Oka et al v  
Youssefyeh et al, 849 F.2d 581 (Fed Cir) at 3).  
e.  
During the course of the trial, Idenix provided the Court  
with a number of US authorities. Each of these cases is  
distinguishable from the current case. None of these cases stand for  
the proposition that a patent may be obtained for something the  
inventor does not know how to make. While US (and Canadian)  
law does not require that the inventor has actually made each of the  
compounds claimed, the law requires that the inventor has a way of  
making them, and discloses that method in the patent.  
f.  
This is consistent with the general principle espoused in  
Christiani v. Rice, [1930] SCR 443 at 454, 456. It is also consistent  
with the policy stated by the Supreme Court in AZT (Apotex Inc v  
Wellcome Foundation Ltd, 2002 SCC 77 at para 84):  
i.  
An applicant does not merit a patent on an  
almost-invention, where the public receives only a  
promise that a hypothesis might later prove useful;  
this would permit, and encourage, applicants to put  
placeholders on intriguing ideas to wait for the  
science to catch up and make it so. The patentee  
would enjoy the property right of excluding others  
from making, selling, using or improving that idea  
without the public’s having derived anything useful  
in return.  
[Gilead’s emphasis]  
g.  
As of the filing and publication dates of the ‘191 Patent, as  
far as the named inventors knew, they had only failed in their  
efforts to synthesize a 2’-fluoro(down)-2’methyl(up) nucleoside.  
As they were not aware of any manner in which any compound  
within the scope of the claim could be synthesized, they did not  
describe any such method in the patent.  
h.  
The named inventors cannot be said to have been in  
possession of an invention at the time the ‘191 Patent was filed or  
published. To borrow the language of the Supreme Court’s  
decision in AZT, the inventors had an intriguing idea for which the  
science had not yet caught up. The ‘191 Patent allowed Idenix to  
enjoy a monopoly right without giving the public anything useful  
in return.  
Page: 269  
i.  
Because Idenix had not actually invented 2’-fluoro(down)-  
2’methyl(up) nucleoside compounds, any claim to any such  
compounds is by definition overbroad.  
[My emphasis]  
D.  
Analysis  
[782] Gilead is in effect, adding a new condition that the inventor must have a way of making  
the invention to the requirement of being able to describe how to make it. As is apparent by the  
quote from Wellcome/AZT that patents are not granted when the public receives only a promise  
that a hypothesis might later prove useful, Gilead is arguing that the patentee is claiming a  
monopoly for a research project, something for which it has not paid the hard coinage.  
[783] Idenix’s response, considering the logic of paragraph “g” above, is that the definition of  
the invention, also found in Christiani v Rice, [1930] SCR 443, is that there is no need for the  
inventor to have a way of making the compound (“a description which affords the means of  
making”), if the inventor is able to sufficiently disclose how to make the compound, and  
assuming he or she can soundly predict its utility.  
[784] Idenix argues simply that if the claims are soundly predicted and there has been sufficient  
disclosure of how to make the invention, then there can be no overbreadth of claims. I agree with  
this submission based on present Canadian law.  
[785] Any difficulties that arise from the fact that the inventor did not make the compound are  
supposed to be reflected in the limits on patents for unmade compounds by the requirements that  
Page: 270  
the inventor be able to soundly predict the invention and to sufficiently disclose its synthesis. It  
turns out that these limitations work quite well in this matter, as I find the ‘191 Patent invalid on  
both grounds. This is not however, a peremptory rule such as Gilead is arguing here that will  
permit the dismissal of the action without requiring a trial of an action along with all the  
challenging evidentiary and legal issues that normally ensue.  
[786] Although sympathetic with Gilead’s underlying submission that Idenix should not be  
rewarded for its failure to make a compound, I do not think that the new principle Gilead seeks to  
advance can be proclaimed under the tenets of overbreadth. Overbreadth is a concept that  
acknowledges the patent may be valid, but the inventor has claimed too much territory. It is that  
space between validity and overreaching, depending on the wording of the claim, as described in  
the passage quoted above from Burton.  
[787] Gilead’s true argument is that this case is one where there is no basis for the Patent in the  
first place, because the invention was never made despite efforts to do so, and the patent should  
never have been granted. If the rule is that the inventor must have a way of making the claimed  
compounds, in addition to being able to disclose how to make the compound, my concern is that  
it opens the floodgates somewhat to new challenges requiring a means to make the inventions for  
all unmade inventions relating in timing to disclosure, but affecting soundly predicted inventions.  
I also think that this is a one-off situation that is unlikely to recur.  
[788] It might be arguable as a logical and non-disruptive corollary to the doctrine of sound  
prediction that you cannot patent an invention that the inventor has tried and failed to invent.  
Page: 271  
This would be accomplished by conjoining the inventor’s knowledge required for the sound  
prediction of utility on filing date, with the simultaneous knowledge that the inventor has tried  
and could not make the invention. But this approach was not argued by Gilead, and is not  
therefore before the Court. As mentioned it also does not appear necessary, as Idenix’s ‘191  
Patent is being found invalid on the basis of well-established principles in patent law.  
[789] In conclusion, I do not find that an overbreadth claim may be made against Idenix in this  
case on the grounds alleged by Gilead.  
X.  
A.  
Idenix’s Counterclaim of Anticipation  
Introduction  
[790] Idenix’s counterclaim includes a claim that that each of the Claims of the ‘657 Patent was  
anticipated by the disclosure and enablement of their subject matter by the ‘191 Patent as filed on  
June 27, 2003 and as published on January 8, 2004. It seeks a declaration under section 60(1) of  
the Patent Act that the ‘657 Patent is invalid under section 28.2(1) of the Act  
[791] Subsection 28.2(1) of the Patent Act provides, inter alia:  
28.2(1) The subject-matter  
defined by a claim in an  
application for a patent in  
Canada (the “pending  
application”) must not have  
been disclosed  
28.2(1) L’objet que définit la  
revendication d’une demande  
de brevet ne doit pas :  
[…]  
[…]  
Page: 272  
(b) before the claim date by a  
person not mentioned in  
paragraph (a), in such a  
b) avant la date de la  
revendication, avoir fait, de la  
part d’une autre personne,  
manner that the subject-matter l’objet d’une communication  
became available to the public qui l’a rendu accessible au  
in Canada or elsewhere;  
public au Canada ou ailleurs;  
(c) in an application for a  
patent that is filed in Canada  
by a person other than the  
c) avoir été divulgué dans une  
demande de brevet qui a été  
déposée au Canada par une  
applicant, and has a filing date personne autre que le  
that is before the claim date. demandeur et dont la date de  
dépôt est antérieure à la date  
de la revendication de la  
demande visée à l’alinéa (1)a);  
[792] For the sole purpose of simplifying this action in Canada, the parties agree that the claim  
date in respect of the ‘657 Patent is the filing date, April 21, 2004.  
[793] The enablement issue of disclosure for anticipation has been largely resolved by my  
analysis concluding that there was insufficient disclosure of how to make the 2’-C’-Me/F  
nucleoside in the ‘191 Patent. That analysis applies in a similar fashion to enabling disclosure for  
anticipation.  
[794] What follows therefore, is a brief construction analysis of the claims in the ‘657 Patent,  
followed by some discussion of the issues concerning the disclosure of the invention, which seem  
somewhat differently argued by Gilead for anticipation from its submissions regarding  
insufficiency.  
Page: 273  
(1)  
Claims Construction of the ‘657 Patent  
[795] The skilled person would have understood that each of the claims of the ‘657 Patent is  
directed to 2’-fluoro(down)-2’-methyl(up) cytidine and uridine nucleosides and nucleotides and  
that the claims narrow down to two discrete compounds, (2’R)-2’-deoxy-2’-fluoro-2’-C-  
methyluridine and (2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine.  
(a)  
Claim 1 of the ‘657 Patent  
[796] The person of ordinary skill in the art would have understood that the compounds of  
Claim 1 could generally be described as being 2’-C-Me/F nucleosides with either a cytosine or  
uracil base and could be depicted in the following way:  
[797] Further, the person of ordinary skill in the art would have understood that Claim 1 of the  
Gilead ‘657 Patent is directed to compounds (or their pharmaceutically acceptable salts) having  
the following components:  
Page: 274  
(i) a natural cytosine or uracil base;  
(ii) a sugar ring containing an oxygen (O);  
(iii) a methyl group (CH3 ) in the 2’ (up) position;  
(iv) a fluorine atom (F) in the 2’ (down) position  
(v) one of the options presented for substituents at the 3’ position, including hydroxyl,  
monophosphate, diphosphate, triphosphate, H-phosphonate and others; and  
(vi) one of the options presented for substituents at the 5’ position, including hydroxyl,  
monophosphate, diphosphate, triphosphate, H-phosphonate and others.  
(b)  
Claim 2 of the ‘657 Patent  
[798] Claim 2 narrows the options at the 3’ (down) position to hydroxyl. Consequently, the  
person of ordinary skill in the art would have understood that the compounds covered by Claim 2  
can be represented as follows:  
Page: 275  
[799] Further, R1 is limited to three possible options (monophosphate, diphosphate or  
triphosphate). Thus, the person of ordinary skill in the art would have understood that Claim 2  
includes six compounds (or pharmaceutically acceptable salts thereof):  
(i) (2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine 5’-monophosphate;  
(ii) (2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine 5’-monophosphate;  
(iii) (2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine 5’-diphosphate;  
(iv) (2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine 5’-diphosphate;  
(v) (2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine 5’-triphosphate; and  
(vi) (2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine 5’-triphosphate.  
(c)  
Claim 3 of the ‘657 Patent  
[800] Claim 3 limits the compounds of Claim 1 by specifying that R7 is H, R1 is a diphosphate  
or a triphosphate. The person of ordinary skill in the art would have understood that Claim 3  
encompasses a smaller number of compounds than Claims 1 and 2. In particular, the person of  
ordinary skill in the art would have understood that Claim 3 includes the following four  
compounds (or pharmaceutically acceptable salts thereof):  
(i)  
(ii)  
(2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine 5’-diphosphate;  
(2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine 5’-diphosphate;  
(2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine 5’-triphosphate; and  
(2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine 5’-triphosphate.  
(iii)  
(iv)  
Page: 276  
(d)  
Claim 4 of the ‘657 Patent  
[801] Claim 4 limits the compounds of Claim 1 by specifying that R7 is H and R1 is a  
triphosphate. The person of ordinary skill in the art would have understood that Claim 4 includes  
the following two compounds:  
(i) (2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine 5’-triphosphate; and  
(ii) (2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine 5’-triphosphate.  
(e)  
Claim 5 of the ‘657 Patent  
[802] Claim 5 limits the compounds of Claim 1 by specifying that both R1 and R7 are H. The  
person of ordinary skill in the art would have understood that Claim 5 includes the following two  
compounds:  
(i) (2’R)-2’-deoxy-2’-fluoro-2’-C-methylcytidine; and  
(ii) (2’R)-2’-deoxy-2’-fluoro-2’-C-methyluridine.  
(f)  
Claim 6 of the ‘657 Patent  
[803] Claim 6 limits the compounds of Claim 1 to a single compound, (2’R)-2’-deoxy-2’-  
fluoro-2’-C-methylcytidine (or pharmaceutically acceptable salt thereof):  
Page: 277  
(g)  
Claim 15 of the ‘657 Patent  
[804] Claim 15 limits the compounds of Claim 1 to a single compound, (2’R)-2’-deoxy-2’-  
fluoro-2’-C-methyluridine (or pharmaceutically acceptable salt thereof):  
(h)  
Other Claims of the ‘657 Patent  
[805] The remaining claims of the ‘657 Patent incorporate by reference the claimed  
compounds. In particular:  
(a) Claims 7 to 12 are directed to pharmaceutical compositions comprising the  
nucleosides of Claims 1 to 6 (Claims 7 to 12 respectively), or their pharmaceutically  
acceptable salt(s), and a pharmaceutically acceptable carrier;  
Page: 278  
(b) Claims 13 and 14 are directed to methods of synthesizing the 2’-fluoro(down)-  
2’methyl(up) cytidine and uridine nucleosides of Claim 1;  
(c) Claim 16 is directed to a pharmaceutical composition comprising the nucleoside of  
Claim 15, or a pharmaceutically acceptable salt, and a pharmaceutically acceptable  
carrier;  
(d) Claims 17 to 19 are directed to liposomal compositions comprising the compound of  
Claim 1, Claim 6, or Claim 15 (Claims 17, 18, and 19, respectively) and optionally a  
pharmaceutically acceptable carrier; and  
(e) Claims 20 to 26 are directed to uses of claimed nucleosides, or compositions  
containing the claimed nucleosides, as antiviral agents.  
(2)  
Anticipation  
(a)  
The Test for Anticipation  
[806] The test for anticipation was set out by the Supreme Court of Canada in [Plavix/Sanofi].  
It comprises two steps:  
(a) Disclosure – The prior patent must disclose information which, if performed,  
would necessarily result in infringement. The person of ordinary skill in the art  
reads the particular piece of prior art to understand whether it discloses the second  
invention. The skilled person is “taken to be trying to understand what the author  
Page: 279  
of the prior patent meant.” There is no room for trial and error or experimentation  
at this stage of the process. The person skilled in the art is simply reading the  
patent for the purposes of understanding it. The evidence to be considered is  
comprised solely of the prior art, as the person of ordinary skill in the art would  
understand it.  
(b) Enablement – If the disclosure requirement is satisfied, then the Court must  
determine whether the ordinary person skilled in the art would have been able to  
perform the invention without undue burden. Routine experimentation is  
permitted at the enablement stage; however, prolonged or arduous trial and error  
would not be considered routine. For purposes of enablement, the question is no  
longer what the person of ordinary skill in the art would think the disclosure of the  
prior patent meant, but whether he or she would be able to work the invention.  
(b)  
Analysis of Gilead’s Submissions  
[807] Gilead claims that the ‘191 Patent does not disclose the claimed subject-matter claimed in  
the ‘657 Patent:  
(i)  
The skilled reader would not be able to identify the active compound in the ‘191  
Patent.  
As stated above, based on the evidence of Drs Damha and Barrett, I am satisfied that  
the skilled reader “taken to be trying to understand what the author of the prior patent  
meant” would understand from examining the claims and the relevant materials in the  
Page: 280  
‘191 Patent would recognize that the active compounds in the ‘191 Patent were the  
2’-C-Me/F nucleotide and nucleoside compounds identified in Claims 2 and 3 and  
their derivatives, with natural bases similar to those in the ‘657 Patent. The core of  
both inventions is clearly at the ‘2 position of the sugar ring. It would be apparent to  
the skilled reader of the ‘191 Patent that other areas of the invention would be more  
widely stated around this structure which forms the heart of the disclosure.  
(ii)  
The ‘191 Patent as filed and published did not focus on Formula IX compounds. In  
fact, the claims of the ‘191 Patent as filed and as published covered all of the  
Formulae and were not limited to Formula IX compounds.  
For the same reason based on an analysis of the claims, I accept the views of Drs  
Damha and Barrett that the skilled reader would realize that the ‘191 Patent was based  
on Formula IX compounds, recognizing that the Patent contains much irrelevant  
material which should be disregarded.  
(iii)  
The ‘191 Patent does not contain a clear description of 2’-C-Me/F nucleosides with  
either a cytosine or uracil base and does not direct the person of ordinary skill in the  
art to select cytosine and uracil bases out of the myriad of bases discussed therein.  
There are no examples, schemes or data in respect of 2’-C-Me/F nucleosides with  
either a cytosine or uracil base.  
On the basis of the same careful reading the skilled reader would be aware from the  
common general knowledge that that the nucleoside analogues used as an antiviral  
drug would be linked to the one of the four natural bases which would include the  
Page: 281  
cytosine or uracil bases. This would be a sufficient direction to disclose the invention.  
Gilead has not argued that the natural purine bases were not active.  
[808] The disclosure of the ‘191 Patent indicates that the compounds claimed by the ‘657  
Patent or a pharmaceutical composition comprising the claimed compounds can be used as an  
antiviral agent or, in the case of claimed nucleosides, they can be used to make prodrugs. Both  
patents also indicate that the compounds can be used to treat Flaviviridae viruses including HCV.  
Additionally, both patents indicate that the compounds can be used in combination with other  
antiviral compounds. As such, the uses that are disclosed and claimed in the ‘657 Patent are the  
same uses that are disclosed in the ‘191 Patent. As discussed above, the same compounds are  
disclosed in both patents, particularly the narrower compounds claimed in both patents that form  
the genus class.  
[809] I am satisfied therefore, that the same invention of the ‘657 Patent is disclosed in the ‘191  
Patent. In the result, whether there is anticipation of the subject matter of the claims of the ‘657  
Patent would be determined by whether the ‘191 Patent provides an enabling disclosure thus  
allowing the skilled chemist to be able to make compounds within the scope of the claims of the  
‘657 Patent.  
[810] As I have already found that there is no enabling disclosure of how to synthesize the  
compounds in the ‘191 Patent, Idenix’s claim of anticipation is rejected.  
[811] The ‘191 patent does not enable the claimed subject-matter claimed in the ‘657 Patent:  
Page: 282  
There are no examples or schemes in respect of how to make 2’-C-Me/F nucleosides with  
either a cytosine or uracil base;  
Synthesis of 2’-C-Me/F nucleosides was not known in 2004;  
There was no common general knowledge as to how to prepare nucleosides with a  
tertiary fluorine at the 2’ position; and  
The person of ordinary skill in the art would not have been able to prepare 2’-C-Me/F  
nucleosides based upon the ‘191 Patent and the common general knowledge without  
undue experimentation, if at all.  
XI.  
A.  
Idenix’s Counterclaim of Infringement  
Introduction - The Bifurcation Order  
[812] Idenix counterclaims that Gilead infringed certain claims of its ‘191 Patent by its making,  
using and selling to others the subject matter claimed in the ‘191 Patent by the production of  
sofosbuvir and its sale in Canada as Sovaldi which commenced on January 6, 2014.  
[813] Despite my conclusions that the ‘191 Patent is invalid, for the purposes of a potential  
appeal, I am nevertheless required to consider Idenix’s infringement argument. It assumes that it  
soundly predicted the utility of the 2’-C-Me/F nucleosides and sufficiently disclosed how to  
synthesize them. On that premise, the Court’s function is to determine the agreed-upon question  
Page: 283  
as contained in Prothonotary Tabib’s Amended Bifurcation Order dated September 9, 2014 [the  
Bifurcation Order], which is as follows:  
1.  
The following issues of this action shall be heard and  
determined by a trial separate from, and after the trial scheduled to  
being on January 12, 2015 (the “First Trial”), if necessary,  
depending on the outcome of the First Trial:  
a. when, where and for how long infringement occurred, and  
who committed the acts of infringement;  
b. whether Idenix is entitled to an injunction, declaratory  
relief, delivery up and accounting of profit or damages of  
any act of infringement;  
c. the extent of any infringement, including the application of  
statutory or common law exceptions;  
d. the quantum of Idenix’s damages or reasonable  
compensation from any infringement, and the quantum of  
the profits of any infringement.  
2.  
For greater certainty, the parties are to proceed to the First  
Trial on the issue of whether sofosbuvir and/or Sovaldi and/or the  
compound identified in paragraph 64A of the Amended Statement  
of Defence and Counterclaim [purported manufacturing  
Intermediate] fall within the scope of the ‘191 Patent.  
[814] The issues for determination in this matter is whether sofosbuvir and/or Sovaldi and/or  
the compound identified in paragraph 64A [the “intermediate” compound] of the Amended  
Statement of Defence and Counterclaim fall within the scope of the ‘191 Patent.  
Page: 284  
B.  
The Parties’ Submissions on Infringement and Conclusions  
[815] Idenix makes three submissions by which it claims Gilead has infringed the ‘191 Patent.  
The first involves a claim of infringement of the “purported manufacturing intermediate”, which  
is as follows:  
(a) Gilead’s 2’-C-Me/F nucleoside with hydroxyls at the 3’ and 5’  
positions on the sugar ring, as an intermediate in the preparation of  
sofosbuvir, falls within the scope of claim 3 of the ‘191 Patent [the  
“Intermediate” compound issue].  
[816] Gilead challenges the Court’s jurisdiction to consider this issue in light of the Bifurcation  
Order. It submits that Idenix’s infringement claim concerns the use of the claimed compounds, as  
opposed to whether the intermediate falls within the scope of the ‘191 Patent. Otherwise, Gilead  
makes no submissions on the infringement by the use of the 2’-C-Me/F nucleoside in sofosbuvir,  
apart from there being no infringement claim because the ‘191 Patent is invalid.  
[817] Idenix’s other two claims of infringement turn around the construction of the description  
of the R1 substituent at the 5’ position on the sugar ringer which reads as follows:  
… phosphate; … or a pharmaceutically acceptable leaving group  
which when administered in vivo is capable of providing a  
compound wherein R1 … is … phosphate.  
[818] Idenix’s submissions on this point are as follows:  
(b)  
Gilead’s SOVALDI sofosbuvir is within the scope of Claim  
2 of the ‘191 Patent in that it has a phosphoramidate prodrug  
moiety at the 5ꞌ position, which is a “phosphate” as well as a  
Page: 285  
“pharmaceutically acceptable leaving group which when  
administered in vivo is capable of providing a compound wherein  
R 1 is phosphate [Phosphoramidate use].  
(c) After Gilead’s SOVALDI sofosbuvir is administered, the  
phosphoramidate prodrug component leaves resulting in a  
monophosphate compound within the scope of Claim 2 of the ‘191  
Patent, which subsequently is phosphorylated in vivo to form the  
diphosphate and triphosphate compounds, equally falling within  
the scope of R1 being a “phosphate” as defined in claim 2 of the  
‘191 Patent [in vivo metabolization].  
[819] Gilead’s submissions in reply to these arguments are as follows:  
(a)  
phosphate when used in the claims of the ‘191 Patent  
means PO4 (monophosphate);  
(i)  
phosphate when used in the claims of the ‘191  
Patent does not include diphosphate, triphosphate, stabilized  
phosphate and stabilized phosphate prodrug;  
(ii)  
phosphate when used in the claims of the ‘191  
Patent does not include phosphoramidate;  
(b)  
phosphoramidate (the group at the 5’ position of  
sofosbuvir) is not covered by R1 in the claims of the ‘191 Patent --  
it is not phosphate or a pharmaceutically acceptable leaving group  
which when administered in vivo is capable of providing a  
compound wherein R1 and /or R2 is independently H or phosphate;  
(c)  
SOVALDI tablets are not covered by claims of the ‘191  
Patent, as its medicinal ingredient, sofosbuvir, is not covered by the  
claims;  
(d)  
the monophosphate, diphosphate and triphosphate  
metabolites of sofosbuvir are not covered by the claims of the ‘191  
Patent as the ‘191 Patent is not directed to metabolites and the term  
“phosphate” in the claims of the ‘191 Patent does not include  
diphosphates or triphosphates; and  
(e)  
the use of PSI-6206 in the manufacture of sofosbuvir is not  
covered by the claims of the ‘191 Patent as the ‘191 Patent is not  
directed to manufacturing intermediates.  
Page: 286  
[820] In the analysis that follows, I first consider Idenix’s submission that Gilead’s intermediate  
compound falls within the scope of the ‘191 Patent. I find this to be the case based on Gilead’s  
implicit admission that the structures of the compounds are identical.  
[821] Thereafter, I consider the interpretive issue involving phosphate, which I conclude  
extends to the monophosphate, diphosphate and triphosphate metabolites of sofosbuvir. I also  
conclude that phosphoramidate is a pharmaceutically acceptable leaving group as that term is  
used to describe the R1 substituent at the 5’ position on the sugar ring.  
C.  
The Intermediate Compound Identified in Paragraph 64A of the Statement of Defense and  
Counterclaim  
[822] Idenix’s preferred submission is that the intermediate compound used in the manufacture  
of sofosbuvir and Sovaldi falls within the ‘191 Patent. Gilead argues that this issue is about  
whether the act of making, selling or using sofosbuvir or its metabolites, Sovaldi, or the  
manufacturing intermediate constitutes an act of infringement. It claims that those questions have  
all been bifurcated to a later date. The only issue before this Court is one of claim construction:  
whether sofosbuvir, Sovaldi or the manufacturing intermediate fall within the scope of claims of  
the ‘191 Patent.  
[823] I agree that Idenix has argued this issue in terms of infringement by use. For instance, it  
relies on jurisprudence citing the “Saccharin doctrine” from Saccharin Corporation Ltd v Anglo-  
Continental Chemical Works Ltd (1900), 17 RPC 307 (Eng Ch Div) and Wilderman v Berk  
(1925), 42 RPC 79(Eng Ch Div) and other cases to the same effect. These stand for the principle  
Page: 287  
that the importation and sale of a product which was made abroad using a patented process or  
intermediate constitutes “use” (and thus infringement). I do not believe that this line of argument  
is relevant to the issue as framed by the parties in the Bifurcation Order.  
[824] However, Idenix includes submissions with respect to the compound identified in  
paragraph 64A of the Amended Statement of Defence and Counterclaim [purported  
manufacturing intermediate] falling within the scope of the ‘191 Patent.  
[825] Gilead’s admission of the use of the Claimed Compound is consistent with a Sofia article  
published by Pharmasset. In the article, the phosphoramidate group is added to the uridine  
nucleoside analogue (Compound 5) such that the nucleoside is used as an intermediate in  
Gilead’s manufacture of sofosbuvir. The use of the intermediate within the scope of Claim 3 can  
be depicted as follows (at p. 7204):  
Page: 288  
[826] Idenix has pointed out Gilead’s admission that the structure of the compound described in  
paragraph 64A of the Statement of Defence and Counterclaim as depicted below on the right is  
the same as Compound 5 in the above scheme. I therefore conclude that Idenix is correct in its  
determination that the manufacturing intermediate falls within the scope of claims of the ‘191  
Patent.  
[827] The phosphoramidate group is added to the uridine nucleoside analogue such that the  
nucleoside is used as an intermediate in Gilead’s manufacture of sofosbuvir. The fact that the  
compound is framed as being used as an intermediate, does not alter the admission as to what the  
compound is or its structure. Gilead acknowledges that it is in the form below on the left, which  
is the same as that of Claim 3 in the ‘191 Patent in the middle, and similar to Claim 15 in the  
‘657 Patent on the right except for the base.  
[828] Gilead argues that the ‘191 Patent only addresses nucleosides, or a nucleotide (one with a  
monophosphate at the 5’ position on the sugar ring). The intermediate compound used by Gilead  
is a nucleoside and it clearly falls within the scope of the ‘191 Patent as the foundational  
nucleoside compound claimed at Claim 3 which is responsible for the antiviral activity of the  
claimed invention. I accept Dr Damha’s evidence that a person skilled in the art would have  
Page: 289  
known that nucleoside analogues in general could be used to prepare nucleoside/nucleotide  
prodrugs.  
[829] I also reject Gilead’s argument that the use of the nucleoside in the manufacture of  
sofosbuvir is not covered by the claims of the ‘191 Patent as the ‘191 Patent is not directed to  
manufacturing intermediates. If accepted, this would eliminate all consideration or coverage by  
the Patent for indirect use of the subject matter of the claims. The Patent does not have to be  
directed at every possible use of the claims. It would be common general knowledge that these  
nucleosides would be used as intermediates in the manufacture of drugs or they could not serve  
their intended purpose.  
[830] I conclude that Gilead’s 2’-C-Me/F nucleoside with hydroxyls at the 3’ and 5’ positions  
on the sugar ring, identified in paragraph 64A of the Statement of Defence and Counterclaim as  
an intermediate in the preparation of sofosbuvir, falls within the scope of Claim 3 of the ‘191  
Patent.  
D.  
SOVALDI sofosbuvir – Phosphate and Pharmaceutically Acceptable Leaving Group  
(1)  
The Preferred Skilled Reader  
[831] The construction of the terms “phosphate” and “a pharmaceutically acceptable leaving  
group which when administered in vivo is capable of providing a compound wherein R1 is …  
phosphate” is largely determinative of the two remaining issues agreed upon by the parties  
regarding Idenix’s infringement claim.  
Page: 290  
[832] It is common ground that because the terms in issue are technical in nature, the Court  
must rely on the opinions of experts to assist in deciding what interpretation should be applied in  
its claims analysis.  
[833] As is to be expected, the parties’ experts provide contradictory opinions. I find this is  
largely due to the dissimilarities in their expertise that is reflected in the attributes that they  
propose for the skilled reader, whose perspective the Court is urged to adopt.  
[834] Gilead sees the issue as one of prodrugs and argues that someone with expertise in the  
field of pharmaceutical sciences and prodrugs should be relied on, like Dr Krise, who is highly  
qualified to speak to these issues. Conversely, Drs Damha, Barrett and Lamarre describe the  
skilled chemist as someone on the discovery team working to invent new compounds. Dr Wnuk  
is part of that team, however, he opines that someone from the pharmaceutical sciences and  
prodrugs should be relied on, like Dr Krise.  
[835] The Idenix experts take the perspective of the skilled discovery chemists who, having  
accomplished their task, see prodrugs as an ancillary matter to be referred to in broad general  
terms, such as are found in both the ‘191 and ‘657 Patents. Recognizing that the choice of the  
prodrug to be matched to the antiviral compound is an aspect of the future development of the  
invention, it is something other experts will have to work on, long after the drug design of the  
antiviral component is complete.  
Page: 291  
[836] Dr Krise on the other hand, delves into the metabolism of the prodrug in the specifics of  
how leaving groups depart from the phosphoramidate prodrug. His opinions are fixated only on  
the wording of the R1 substituent of Claim 1. His contextual considerations are limited to those  
that reflect on the metabolism of prodrugs.  
[837] In my view, the skilled reader, or readers of choice should be the chemist to whom the  
‘191 Patent is most particularly addressed. This is Dr Barrett’s view, which I think reflects his  
experience in these matters and is the most reasonable as it makes little sense to rely upon  
chemists who would not be interested in the art that is the subject matter of the invention.  
[838] Based on this preference criterion, I have little difficulty concluding that the ‘191 Patent  
is addressed most particularly to persons skilled in the art or science of novel drug discovery as  
opposed to those who work on the after-development of the drug, once invented.  
[839] Accordingly, I find that Dr Krise is the only expert who I would situate outside of the  
drug discovery team. This is not to discredit Dr Krise’s expertise in the development of novel  
prodrugs. But prodrugs are an ancillary subject matter in the ‘191 Patent. Indeed, Gilead’s  
criticism of the “few pages” devoted to prodrugs, confirms that the Patent was not addressed to  
the skilled reader with an expertise in the subject matter.  
[840] I could not see Dr Krise gleaning much useful information of interest to his expertise  
from a patent that disclosed a raw nucleoside at its initial stage of invention. His skilled reader  
would recognize that the patent could not contain useful information on prodrugs at this early  
Page: 292  
stage of first discovery with all the development that lay ahead, including developing an  
appropriate prodrug strategy. Indeed, I would think that the pharmaceutical skilled reader would  
share the same view as those in drug discovery or other areas of drug development, namely to  
expect only the broadest most general references to prodrugs without any intention to limit or  
specify the candidates eventually chosen sometime in the future. These remarks are supported by  
Gilead’s ‘657 Patent, which also reflects much generalized references to prodrugs, citing various  
papers with little specificity. Like the ‘191 patent, it only refers to phosphoramidate in passing as  
found in the literature.  
[841] Dr Wnuk included a pharmacologist amongst his skilled team in describing the person  
skilled in the art. My sense is that this addition to his drug discovery team fits very nicely with  
Gilead’s decision to call a pharmacologist as its expert regarding infringement. I prefer Dr  
Barrett’s conclusion contradicting somewhat Dr Wnuk’s opinion that the pharmacologist would  
have expertise in the “drug delivery and metabolism aspects of drug discovery”, but that “the  
medicinal chemist and virologist would have the collective expertise in this area.” Moreover, I  
note that Dr Wnuk was careful to link the pharmacologist in some form to “drug discovery”. To  
the extent that there may be prodrug discovery related to this invention, none is apparent to me.  
[842] The perspective of the drug discovery chemist is consistent with the limited descriptions  
provided in both the ‘191 and ‘657 Patents on prodrugs. The essential aspects of the invention  
turn on the discovery and testing of the new compounds. They do not describe any intention to  
identify or limit what prodrugs might eventually be applied to the nucleoside as long as those  
eventually chosen result in their gaining entry to the cell with a monophosphate moiety at the R1  
Page: 293  
position ready to set off on its journey to destroy a virus after being metabolized into a  
triphosphate nucleoside.  
(2)  
Phosphate  
[843] It appears that the parties agree that “phosphate” does not include a pharmaceutically  
acceptable leaving group on which there was much evidence by Dr Krise. The wording of the  
section must exclude a definition of phosphate that would include a pharmaceutically acceptable  
leaving group, because in vivo, phosphate is what remains after the leaving group leaves the  
prodrug to provide a phosphate. The context of the description of the substituents separated by  
“or” equally supports the conclusion that “phosphate” and “pharmaceutically acceptable leaving  
group” are independent terms.  
[844] Accordingly, the debate on the interpretation of “phosphate” is whether it is a mono-  
phosphate as Gilead argues, or includes a di-, or tri-phosphate. I agree with Gilead that normally  
there would be an interpretive limitation on the use of the term phosphate. If used twice in the  
same definition, it should have the same meaning when in the drug and after when in vivo, which  
suggests limits on its definition. I also agree that the term is ambiguous inasmuch there is much  
confusion in the various references to phosphates in the specification and the different  
expressions as a mono, or all three forms, which are used interchangeably and indiscriminately.  
[845] That said, I am persuaded that the skilled chemist would have to understand that in vivo,  
the monophosphate in the nucleoside, must be metabolized by phosphorylation into di- and  
finally tri-phosphates, if the drug is to have any antiviral activity. I am in agreement with Drs  
Page: 294  
Barrett and Damha, therefore, that the interpretation of phosphate offered by Gilead would make  
no sense to the skilled chemist or virologist to limit phosphate to its monophosphate form. It  
would be a contradiction of the common general knowledge of how the nucleoside compound  
was intended to be effective. I therefore, conclude that “phosphate” includes di-and triphosphates  
in the description for the R1 substituent.  
[846] For that reason, I cannot accept Gilead’s submission that the monophosphate, diphosphate  
and triphosphate metabolites of sofosbuvir are not covered by the claims of the ‘191 Patent as it  
is not directed to metabolites. The skilled chemist would understand that what is left behind,  
described as a “phosphate” after the leaving group leaves, is a monophosphate that metabolizes  
into a diphosphate and triphosphate metabolites for incorporation into the virus replication  
process. It is also possible that in the future a triphosphate analogue may be manufactured that  
may be rendered into a prodrug form. There is no reason why the patent would not extend to  
protect this form of nucleotide.  
[847] Conversely however, I disagree with Idenix that the skilled person would understand that  
a phosphoramidate, although containing phosphate, is a “phosphate” as the defined term,  
particularly in the context of the description for the substituent R1. I accept the evidence of Dr  
Krise that a phosphoramidate refers to a class of prodrugs that serve a different purpose, basically  
to overcome identified barriers that limit the therapeutic usefulness of the drug, and particularly  
in this case to overcome barriers associated with nucleotide delivery so as to deliver a nucleotide  
in triphosphate form to the active site. The term phosphoramidate is not interchangeable with  
phosphate. Additionally, the skilled reader considering the distinction between phosphate and a  
Page: 295  
pharmaceutically acceptable leaving group would understand that the latter definition was  
intended to apply to prodrug forms of the molecule, which normally are formed at the 5’ position  
on the sugar ring, in counter distinction to it being considered a phosphate.  
(3)  
Leaving Group  
[848] “Leaving group” in its context is as follows: “A pharmaceutically acceptable leaving  
group which when administered in vivo is capable of providing a compound wherein R1 is …  
phosphate.”  
[849] Five experts provided opinions on the appropriate interpretation of what is meant by a  
pharmaceutically acceptable leaving group, being Dr Wnuk and Dr Krise on behalf of Gilead and  
Dr Damha, Dr Barrett and Dr Lamarre for Idenix. Dr Krise and Dr Damha were the primary  
witnesses who provided more fulsome explanations for their views.  
(a)  
Dr Wnuk  
[850] In his first report, Dr Wnuk acknowledges that much of the discussion of salts and  
prodrugs in the ‘191 patent “is outside the scope of my area of expertise.” Similarly, when  
questioned about how to make a nucleoside prodrug, he acknowledges that he really was “not an  
expert in this literature.” He nevertheless felt he had the expertise in the terms of chemistry to  
weigh into the discussion on the meaning of a “leaving group.” Dr Wnuk stated that in the  
context of the ‘191 Patent, which obviously is that of salts or prodrugs, he believed that “leaving  
group” refers to a group on a nucleoside which, when the nucleoside is administered in vivo, will  
Page: 296  
entirely cleave off and subsequently replaced by another atom or group of atoms. His report  
provided examples of the leaving groups being cleaved off. It included one where two separate  
moieties were cleaved off from a compound, apparently in one step, which appears somewhat  
inconsistent with his views.  
[851] He also opined that “pharmaceutically acceptable” has no standard meaning in chemistry  
and is ambiguous. The Court notes however that that the ‘657 patent is replete with this term, as  
well as being found throughout the ‘191 Patent. He supports his opinions by an expert in Fox &  
Whitesell, Organic Chemistry 2nd ed. 1995. It defines a leaving group as “a group displaced from  
a reactant in a substitution or elimination reaction.” This says nothing however, of whether the  
reaction must take place in one or three steps, or whether one describes a leaving group as being  
composed of a single compound or more than a single compound, such as to differentiate  
between his example of two moieties leaving at the same time as opposed to leaving one after the  
other.  
[852] Dr Wnuk completes his opinion in the first report, stating that a phosphoramidate is not a  
“pharmaceutically acceptable leaving group…, as it is not entirely removed or cleaved in its  
entirety when sofosbuvir is administered to patients in vivo.” To back up this opinion he offers  
that the skilled chemist would have understood that the isopropyl ester, the phenyl ring, and the  
alanine portions of the phosphoramidate are separate leaving groups, as they are each cleaved off  
in three separate and distinct reactions, in vivo.  
Page: 297  
(b)  
Dr Krise  
[853] Dr Krise provides the perspective of a pharmaceutical chemist who specializes in all  
aspects of prodrug development and application. The essence of his opinion is that because  
Sofosbuvir is metabolized to a 2’-C-Me/F nucleotide in 3 steps, instead of 1, it is not a true  
leaving group, which by his definition must leave the 2’-C-Me/F nucleotide in one step.  
[854] Dr Krise supports his opinion with a detailed analysis of the three-step metabolization of  
sofosbuvir wherein the isopropyl alcohol leaves catalyzed by human cathepsin A (CatA) and  
carboxylesterase 1 (CES1), followed by the second step of an intramolecular cyclization wherein  
the leaving group is phenol, and then being completed by the third step, wherein the leaving  
group is the amino acid alanine that is catalyzed by an enzyme (Hint1) leaving compound PSI-  
6206-MP inside the infected liver cells. This explanation is depicted in Figure 6 in his report  
entitled Metabolism of Sofosbuvir, as follows:  
Page: 298  
[855] Compound PSI-6206-MP at the lower left-hand position of Figure 6 is the same 2’-C-  
Me/F nucleotide with a monophosphate at the 5 position on the sugar ring that the ‘191 Patent  
describes after metabolization by the cleaving of the leaving group. As Dr Krise properly  
concludes this compound “is effectively trapped in the liver cells and is subsequently  
phosphorylated to generate PSI-6206-TP (the triphosphate metabolite), which is antivirally  
active.”  
[856] Dr Krise has no difficulty with the term “pharmaceutically acceptable”, but states that a  
leaving group is “a group that is entirely cleaved off in the body to generate the parent  
compound”, elaborating as follows:  
If the group installed at R1 is a “pharmaceutically acceptable  
leaving group”, then it must leave in its entirety and generate a  
nucleoside with an unconjugated hydroxyl group at the 5’ position  
(meaning there is nothing bonded to the hydroxyl group at the 5’  
position).  
[857] Accordingly, he concludes that the ‘191 Patent could not be referring to a  
phosphoramidate prodrug because it describes a leaving group that is cleaved in only one step.  
[858] He also points out that to his knowledge as of January 8, 2004, phosphoramidate prodrug  
strategies had been considered in the literature as an approach to deliver nucleotide analogues,  
but no compound utilizing this approach had been successfully brought to market. To his  
knowledge sofosbuvir was the first safe and effective phosphoramidate prodrug to be approved  
for sale in the US for the treatment of viral infections.  
Page: 299  
[859] Dr Krise offers no substantiation for his opinion from pharmaceutical texts or other  
literature on prodrugs or biological chemistry. The fact that Dr Krise’s opinion is unsupported is  
of concern to the Court. If the distinction he makes between one-step and more-than-one-step  
metabolization of prodrugs is common general knowledge, I would have expected the literature  
to make the type of distinction that Dr Krise urges the Court should apply to constrain Idenix’s  
possible monopoly. I would have expected some reference to a scientific nomenclature that  
distinguishes between prodrugs based on the number of in vivo steps involved in their  
metabolization, either as a function of some form of classification, or some practical  
experimental, or other scientific necessity to describe the sharp distinction Dr Krise attributes to  
the words of the ‘191 Patent.  
[860] Instead, he attempts to fortify his opinion by the words of the ‘191 Patent. I set out Dr  
Krise’s supporting argument at para 91 of his report as follows:  
[91] This view is reinforced when considering the other groups  
that are listed in Claim 1 as possible substituents at the R1 and R2  
position. In particular, many of the other groups listed as possible  
substituents, including phosphate, acyls, lipids, amino acids, etc.,  
appear to me to be groups that could be cleaved off of the parent  
nucleoside. For example, phosphatases, which are ubiquitous in the  
human body, are known to catalyze the removal of phosphate  
groups from molecules. Similarly, esterases are found in human  
bodies and would facilitate the removal of acyls, etc. Following  
this pattern, the skilled person would have understood that the  
phrase “pharmaceutically acceptable leaving group…” was  
intended to capture similarly cleavable moieties that were not  
otherwise specifically listed in the claim.  
[Emphasis added]  
Page: 300  
[861] I find that Dr Krise’s reasoning is not sufficiently purposive. Apparently, by pluralizing  
the term “group” to “groups”, such that the phrase would read “pharmaceutically acceptable  
leaving groups”, Idenix’s description would have encompassed a phosphoramidate prodrug that  
passes through a three-step cleaving metabolization. This type of interpretation is what Lord  
Diplock admonished should be eschewed as a “purely literal one derived from applying to it the  
kind of meticulous verbal analysis in which lawyers are too often tempted by their training to  
indulge”, Catnic. (I again note that the parties did not present any submissions regarding non-  
essential or inventive features of an invention based upon the “Catnic Principle” as endorsed by  
the Supreme Court in the Free World and Whirlpool decisions, and none are applied in this  
analysis.  
[862] I also do not understand the relevance in Dr Krise’s comment about the lack of specific  
reference to phosphoramidate prodrugs in the ‘191 Patent, or any indication that it was in the  
contemplation of the inventors. As he points out, phosphoramidate prodrugs had apparently not  
been successfully applied to nucleotide analogues in 2004. Moreover, Gilead’s ‘657 Patent made  
no specific reference to phosphoramidate prodrugs, the only mention I am aware of, being in a  
literature references, similar to in the ‘191 Patent.  
[863] I also disagree with the logic of Dr Krise when he concludes that his argument is  
bolstered by the fact that many of the other groups listed as possible substituents of the ‘191  
Patent are prodrugs with single-step leaving groups, thereby attributing that interpretation to the  
meaning of “leaving group” in the Claims. Any reference to the indeterminate descriptor “many”  
applied to one set of circumstances logically entails that that it does not apply to all  
Page: 301  
circumstances; other circumstances will be different. The description in the ‘191 Patent of the  
characteristic of how the leaving group metabolizes must therefore be interpreted to refer to both  
single and multistep processes in the Claims language.  
[864] Moreover, accepting Dr Krise’s reasoning that distinguishes between prodrugs on the  
basis of the number of cleaving reactions required to produce the monophosphate nucleotide,  
would mean that Idenix’s monopoly under the ‘191 Patent would apply to some of prodrugs, but  
not others. I think this is an unreasonable and unacceptable result that the Court could not accept  
as arising from a prodrug of the invention.  
[865] In conclusion, there is no suggestion in Dr Krise’s interpretations of a mind willing to  
understand the meaning of the terms – one that necessarily pays close attention to the purpose  
and intent of the author as best ensures the attainment of the patent’s objects as is normally  
manifested by context. To the extent that any context is applied by Dr Krise, I conclude that it is  
based upon his personal expertise as a pharmaceutical scientist and not that of a discovery or  
medical chemist.  
(c)  
Dr Lamarre  
[866] Dr Lamarre, on the half of Idenix, limited his opinion to the conclusion that a  
pharmaceutically acceptable leaving group would include a nucleotide prodrug which can be  
metabolized and phosphorylated to produce the active compound in the host. He stated that  
phosphoramidate nucleotide prodrugs were an example of such a pharmaceutically acceptable  
leaving group which was known prior to 2004.  
Page: 302  
(d)  
Dr Barrett  
[867] Dr Barrett concluded that the term “a pharmaceutically acceptable leaving group which  
when administered in vivo is capable of providing a compound wherein R1 is phosphate” allowed  
for pro-drugs at the R1 that result in the R1 positions being phosphate once the pro-drug moiety is  
cleaved in vivo. Contrary to the Wnuk Report, he disagreed that this phrase was “ambiguous”  
when read in the context of the ‘191 Patent. Rather, he stated it would be immediately recognized  
as describing a pro-drug and that prodrugs are molecules that are cleaved in vivo to produce the  
active form of the product. The cleaved portion of the molecule can be termed a “leaving group”.  
[868] He stated that prodrugs involve moieties that are added to a drug molecule to improve  
some sort of characteristic of the drug, normally one of the ADME (absorption, distribution,  
metabolism or excretion) characteristics. The moiety is cleaved in the body after administration.  
Beyond this general overview, he noted that pro-drugs were well within the domain of the  
medicinal chemist as of January 2004.  
(e)  
Dr Damha  
[869] In his September 19, 2014 report, Dr Damha acknowledged that he was “not a prodrug  
expert in terms of following metabolism and clearance of these compounds.” But he was  
acknowledged as an expert in prodrugs to the extent necessary for drug design. He possessed the  
expertise necessary to describe the skilled drug discovery chemist’s reading and the  
understanding of the information gained from the Patent. Similarly, the skilled reader in the  
Page: 303  
synthesis side of drug discovery would know the basics as to why the nucleotide analogue would  
need to be transformed into a prodrug nucleotide.  
[870] Dr Damha provided evidence describing the skilled drug discovery chemist’s  
understanding of the necessity and role of prodrugs. His evidence of that knowledge describes  
why the nucleoside must be phosphorylated into a nucleotide outside of the cell by the addition  
of a mono-phosphate at the 5’ position on the sugar ring, but that doing so reduces the antiviral  
effectiveness of the drug by limiting the cells uptake of the drug, thus the need for a prodrug.  
[871] Dr Damha discusses aspects of the ‘191 Patent that generally discuss prodrugs. He points  
out that at the page 107 (lines 12-18) pharmaceutically acceptable prodrugs are described as  
compounds that are metabolized in the host to form an active compound, listing the numerous  
examples. Specifically, he notes that the ‘191 Patent discusses administering a nucleotide or  
nucleoside prodrug to increase the activity, bioavailability, stability, or otherwise alter the  
properties of the nucleoside, including increasing the passage into cells (page 108, lines 16-20).  
Moreover, in respect of phosphoramidate prodrugs, many nucleotide prodrugs were described in  
the literature including in Minireview: nucleotide prodrugs by Jones and Bischoferger (1995),  
Antiviral Research, 27:1-17, which describes a series of nucleotide prodrugs that had been made  
for zidovudine (azidothymidine, AZT), including the phosphoramidate prodrug formed between  
AZT-phosphate and alanine methyl ester. This prodrug amongst others was contained in the  
article and is depicted below [The third structure drawing from the left is a phosphoramidate]:  
Page: 304  
[872] In his November 7, 2014 report responding to Dr Wnuk’s opinions, Dr Damha criticized  
Dr Wnuk for applying a literal chemical definition that was out of context and from the  
perspective of the pure synthetic chemist and which ignores the biological context in which these  
terms are used in the claims of the ‘191 Patent. I agree with this comment.  
[873] In this context the terms “pharmaceutically acceptable” and “administered in vivo” both  
denote the biochemical process of having the active compound reach its biological target. It was  
well known that a prodrug moiety would eventually leave the compound, at least in part, once the  
compound enters the environment of its biological target (McGuigan C. et al (1992), Antiviral  
Research 17: 311-321). The meaning of “leaving group” should be viewed from this biochemical  
perspective rather than the pure chemical definition as suggested by Dr Wnuk; in other words, a  
skilled person would equate “pharmaceutically acceptable” leaving group which when  
administered in vivo is capable of providing a compound wherein R1 and/or R2 is independently  
H or “phosphate” to mean a prodrug.  
[874] He also stated that using a purposive approach, the term “leaving group” would not  
require a prodrug to leave in one step, because that was inconsistent with how prodrugs were  
known to work. Dr Damha states rather, from a purposive perspective, in a biological system,  
Page: 305  
what is important is for a prodrug to serve its function and then to be removed from the  
nucleoside or nucleotide. Whether it leaves as one group or in parts makes no difference. He  
notes that this is discussed in the prodrug review articles and other references found at pages  
108-112 of the ‘191 Patent.  
[875] In conclusion, I accept the opinions of Drs Damha and Barrett, in addition to Dr  
Lamarre’s restraint in offering no opinion on the subject, that “leaving group” should have been  
interpreted to encompass all forms of prodrug metabolization, so long as they are  
pharmaceutically acceptable and produce an in vivo result of a monophosphate nucleoside or  
nucleotide in preparation for further metabolization: in other words, an appropriate prodrug of no  
specified or limited makeup that properly performs the task of a prodrug. Accordingly, leaving  
group includes a phosphoramidate prodrug, including in its manufactured state of sofosbuvir and  
in tablets of Sovaldi and are covered by the ‘191 Patent.  
[876] In conclusion, sofosbuvir and/or Sovaldi and/, including in their metabolized forms, and  
the compound identified in paragraph 64A of the Amended Statement of Defence and  
Counterclaim [the manufacturing intermediate] fall within the scope of the ‘191 Patent.  
XII. Counterclaim -- Section 53 -- Wilful Misrepresentation of the Inventor  
A.  
Introduction and Principles  
[877] In the further alternative Idenix pleads that the ‘657 Patent is invalid under section 53 of  
the Patent Act for knowingly omitting naming Dr Lieven Stuyver as inventor, which omission  
Page: 306  
Idenix claims was a misrepresentation willfully made for the purpose of misleading and not made  
in good faith.  
[878] Section 53 of the Patent Act, RSC 1985, c P-4, reads as follows:  
53(1) A patent is void if any  
material allegation in the  
petition of the applicant in  
respect of the patent is untrue,  
or if the specification and  
drawings contain more or less  
than is necessary for obtaining  
53(1) Le brevet est nul si la  
pétition du demandeur, relative  
à ce brevet, contient quelque  
allégation importante qui n’est  
pas conforme à la vérité, ou si  
le mémoire descriptif et les  
dessins contiennent plus ou  
the end for which they purport moins qu’il n’est nécessaire  
to be made, and the omission pour démontrer ce qu’ils sont  
or addition is wilfully made for censés démontrer, et si  
the purpose of misleading.  
l’omission ou l’addition est  
volontairement faite pour  
induire en erreur.  
(2) Where it appears to a court (2) S’il apparaît au tribunal que  
that the omission or addition  
pareille omission ou addition  
referred to in subsection (1)  
est le résultat d’une erreur  
was an involuntary error and it involontaire, et s’il est prouvé  
is proved that the patentee is que le breveté a droit au reste  
entitled to the remainder of his de son brevet, le tribunal rend  
patent, the court shall render a jugement selon les faits et  
judgment in accordance with  
the facts, and shall determine  
the costs, and the patent shall  
be held valid for that part of  
the invention described to  
which the patentee is so found  
to be entitled.  
statue sur les frais. Le brevet  
est réputé valide quant à la  
partie de l’invention décrite à  
laquelle le breveté est reconnu  
avoir droit.  
[879] Section 53 implies the notion of fraud (Eli Lilly Canada Inc v Apotex Inc, 2008 FC 142 at  
para 62). In order to succeed in its allegation, Idenix must prove that:  
Page: 307  
(a) an untrue allegation was made (i.e. that Dr Stuyver was an inventor and should  
have been listed as such on the ‘657 Patent in Canada);  
(b) that the failure to name Dr Stuyver as an inventor was a “material” untrue  
allegation; and  
(c) that this was “wilfully made for the purpose of misleading” (671905 Alberta Inc v  
Q'Max Solutions Inc (2003), 27 CPR (4th) 385 at 400 (FCA)).  
[880] There remains some issue as to what constitutes a material allegation, which is considered  
below.  
B.  
Is the Omission of Dr Stuyver as an Inventor an Untrue Allegation in the ‘657 Patent?  
(1)  
Co-inventorship  
[881] In order to determine who the inventor is, the Court must determine who is responsible  
for the inventive concept. As was previously pointed out, the parties appear to have studiously  
avoided discussing the issue of an inventive concept or inventive steps in either Patent. The  
inventive concept appears to be assumed in this case to be the conception of the 2’-C-Me/F  
compound as an antiviral. Neither party is interested in discussing whether the synthesis of the  
2’-C-Me/F compound may have involved an inventive step. The question framed by the parties  
therefore, is who is or are, the persons who conceived the 2’-C-Me/F nucleoside as an antiviral. It  
is understood that Dr Stuyver is not an inventor on account of his role in the testing of the  
Page: 308  
invention. This was stipulated by the Supreme Court in Wellcome/AZT at paras 96 and 97 as  
follows:  
Inventorship is not defined in the Act, and it must therefore be  
inferred from various sections. From the definition of “invention”  
in s. 2 for example, we infer that the inventor is the person or  
persons who conceived of the “new and useful” art, process,  
machine, manufacture or composition of matter, or any “new and  
useful” improvement thereto. The ultimate question must therefore  
be: who is responsible for the inventive concept?  
Section 34(1) requires that at least at the time the patent application  
is filed, the specification "correctly and fully describe the invention  
... to enable any person skilled in the art or science to which it  
appertains ... to ... use it". It is therefore not enough to have a good  
idea (or, as was said in Christiani, supra, at p. 454, "for a man to  
say that an idea floated through his brain"); the ingenious idea must  
be "reduced ... to a definite and practical shape" (ibid.). Of course,  
in the steps leading from conception to patentability, the  
inventor(s) may utilize the services of others, who may be highly  
skilled, but those others will not be co-inventors unless they  
participated in the conception as opposed to its verification. As  
Jenkins J. notes in May & Baker Ltd. v. Ciba Ltd. (1948), 65  
R.P.C. 255 (Ch. D.), at p. 281, the requisite "useful qualities" of an  
invention, "must be the inventor's own discovery as opposed to  
mere verification by him of previous predictions". 934  
[Emphasis added]  
[882] Idenix acknowledges that there can be multiple inventors, but each must make a  
contribution to the inventive concept Wellcome/AZT at para 99, citing Gerrard Wire Tying  
Machines Co v Cary Manufacturing Co, [1926] Ex CR 170):  
Nor is a patent to joint inventors invalidated by the fact that one of  
them only first perceived the crude form of the elements and the  
possibility of their adaptation to complete the result desired. In fact  
the conception of the entire device may be attributed to one, but if  
the other makes suggestions of practical value, which assist in  
working out the main idea and making it operative, or contributes  
Page: 309  
an independent part of the entire invention which helps to create  
the whole, he is a joint inventor even though his contribution be of  
minor importance.  
[883] Idenix contends that Dr Stuyver conceived the 2’-C-Me/F compound as an antiviral for  
HCV. It did not plead however, that Jeremy Clark is not an inventor. At one point in its written  
submissions it commented “just because Jeremy Clark was the first person at Pharamsset to  
synthesize a Claimed Compound, it does not necessarily follow that he is the inventor”. I do not  
accept this innuendo statement however, as a plea or submission that Jeremy Clark was not an  
inventor, which issue has never been put before the Court.  
[884] Accordingly, for the purposes of this issue, if I conclude that Dr Stuyver contributed to  
the idea of Jeremy Clark making the invention, he would be a co-inventor.  
(2)  
The Evidence of Dr Stuyver of his Contribution to the Invention in the ‘657 Patent  
[885] Dr Stuyver was vice-president of biological sciences for Pharmasset. He headed up the  
biology team which was responsible for the testing of compounds. Dr Stuyver testified that he  
came up with the idea of the 2’-C-Me/F formula from his previous work with fluorides in the 2’  
(down) position on the Ribose ring and his duties of keeping a library of scientific materials for  
Pharmasset where he researched compounds with antiviral activity. He also claimed to be  
inspired by the Roche Patent Application WO 02/18404. It confirmed activity for one of  
Pharmasset’s compounds containing a fluorine in the 2’ (down) position that it was in the throes  
of patenting. He claims that Dr Schinazi was upset with the fact that he was working on  
Page: 310  
compounds with fluorines at the 2' position, and was not pleased with the fact that he discovered  
something with activity at the 2' (down) position.  
[886] Moreover, Dr Stuyver alleges that he found the omission of the fluorine at the 2’ (down)  
position in the Idenix application referred to above. He testified that it was discussed a few times  
over the summer, but Dr Schinazi never gave the instructions to start working on the synthesis of  
this compound. Because he was a little bit disappointed in this type of approach, he went to  
Jeremy Clark because he is the type of person that would “start working on this under the radar”.  
He claims his discussions with Mr Clark is confirmed at page 25 of the Laboratory Notebook of  
Jeremy Clark (trial exhibit I-23) dated December 6, 2002 signed by both of them. The page in its  
entirety is set out below:  
Page: 311  
[887] This page displays a note in Dr Stuyver’s handwriting that reads “candidate compounds  
as anti-HCV.” He claims this refers to the three compound figures on the page that he discussed  
with Jeremy Clark, which includes as Compound 3, the 2’-C-Me/F nucleoside. There is also a  
Page: 312  
reference on the page of Jeremy Clark writing “Retrosynthetic plan for LS’s [Lieven Stuyver]  
idea.” Dr Stuyver claims this refers to Compound 3, being the 2’-C-Me/F cytidine nucleoside  
with the note beside it “for HCV.” Dr Stuyver alleges that this is evidence of his contribution of  
the idea that led Jeremy Clark to synthesize the 2’-C-Me/F nucleoside.  
[888] The page in Jeremy Clark’s notebook is difficult to interpret, particularly the reference  
“Retrosynthetic plan for LS’s idea.” There is also a comment in the box below that stating “L.  
Stuyver proposition per 12/6/02 where n=1.” None of this bears any relationship to the 2’-C-  
Me/F nucleoside. In referring to that compound, it shows only that Jeremy Clark was struggling  
to find a means to synthesize it, which was clearly not at hand from those drawings. Dr Stuyver  
acknowledges that he had no reliable recollection of his conversation with Mr Clark, such that  
his memory of events is not certain.  
[889] There were protocols at Pharmasset for the immediate recording of ideas that could prove  
inventive. These were not adhered to, if in fact Dr Stuyver came up with the idea to synthesize  
the 2’-C-Me/F nucleoside. Notebook pages were to be witnessed by persons who did not  
consider themselves to be co-discoverers. Furthermore, Dr Stuyver suggested that he came up  
with the idea for the Compound sometime in the summer of 2002, which is not reflected in the  
documentation.  
[890] While I admit that this evidence describes some discussions between Dr Stuyver and  
Jeremy Clark acknowledging that he had contributed an idea, I am not convinced that it is  
sufficiently persuasive to overcome the other evidence that points to Mr Clark having decided to  
Page: 313  
make the compound, probably on instructions of Dr Watanabe, because he found that Idenix had  
omitted the 2’-C-Me/F nucleoside analogue in its patent application.  
[891] In particular, I find it difficult to accept Dr Stuyver’s claim that he was the person who  
discovered the “hole” in the Idenix patent by which it omitted to claim the 2’-C-Me/F nucleoside  
as a possible analogue to the 2’-C-Me/OH compound. This evidence appears to be highly  
coincidental with his evidence as coming up with the idea independently from his own work and  
his apparent need to have approached Jeremy Clark under cover, contrary to Dr Schinazi’s  
permission. I also find this latter testimony makes little sense, except to attack Dr Schinazi,  
which was the flavour of much of Dr Stuyver’s evidence.  
[892] Mostly however, Dr Stuyver’s evidence that he found the omission in the Idenix  
compound is largely contradicted by the evidence of Dr Patterson’s account as to how the 2’-C-  
Me/F nucleoside was chosen as a candidate for synthesis.  
[893] He testified that Dr Schinazi came to the chemists with the news that the Idenix 2’-C-  
Me/OH nucleoside in its patent was active against hepatitis C and he wanted them to consider  
synthesizing some analogues of it. Dr Watanabe apparently presented some analogues of interest.  
Dr Patterson stated that nobody was particularly interested in making analogues because they  
were certain that Idenix would have covered all the reasonable analogues.  
[894] As seen above, this in fact was the case as evidenced by the notes on the Maui meeting in  
December 2001 whereat it was decided that Idenix would attempt to synthesize a number of  
Page: 314  
analogues of the 2’-C-Me/OH nucleoside with new substituents at the 2’ (down) position. Dr  
Patterson further testified as to his belief that Jeremy Clark discovered that the 2’-C-Me/F  
molecules had been omitted from the Idenix patent. This caused much excitement at Pharmasset  
as to why it had been omitted, so that there was an effort undertaken to make that molecule.  
[895] Obviously this evidence suggests that the concept of the 2’-C-Me/F nucleoside having  
possible antiviral activity was not inventive, nor was the idea to proceed to attempt its synthesis  
the idea of either Mr Clark or Dr Stuyver. But neither party is advancing this argument, which  
would turn the focus on whether the synthesis of the compound was the inventive concept or  
step.  
[896] Dr Patterson also indicated that substituting the fluoride for the hydroxyl was his first  
choice because of the similarities in structure. However, I am somewhat concerned that hindsight  
might play a role in this opinion. It is however, a different basis for the reasoning of Dr Stuyver  
who apparently came up with the idea from his past work with Pharmasset’s libraries of patents  
and assaying other chemists’ synthesized compounds.  
[897] He recalled that Jeremy Clark was the one who was the most vocal about finding the  
omission in the patent, so he took it upon himself to attempt to synthesize it. He recalls that  
Jeremy Clark wanted to prepare it by synthesis of the carbohydrate part of the molecule first and  
then coupling it to the nucleoside. He also stated that it took him a long time to get anywhere on  
it, so people began giving him advice.  
Page: 315  
[898] Dr Patterson demonstrated a superior memory and ability to recount events of the various  
witnesses who testified on this subject. He also was independent, despite being called by Idenix.  
I prefer his evidence over that of Dr Stuyver. It also fits with the common theme shared by both  
Dr Stuyver and Dr Otto that the 2’-C-Me/F nucleoside was chosen for synthesis because it was  
an analogue of the antiviral 2’-C-Me/OH compound, which had been omitted from coverage  
under Idenix’s patent.  
[899] It also fits with the narrative of a junior chemist at Pharmasset being assigned the role of  
synthesizing the compound, which others thought had likely been attempted by Idenix, and as a  
reward of sorts for having spotted the omission in the patent.  
[900] I also find it notable that Idenix did not ask Dr Patterson whether he was aware of Dr  
Stuyver being involved in the events leading to the synthesis of the 2’-C-Me/F nucleoside.  
Pharmasset was a small organization and it would appear likely that he would have been aware  
of contrary claims to the discovery of the omission. He also appears to indicate that Dr Watanabe  
may have suggested the compound as one of the possible analogues.  
[901] [REDACTED]  
[902] Similarly, after Dr Stuyver expressed concerns regarding inventorship, [REDACTED].  
Page: 316  
[903] Idenix argues that Jeremy Clark should have been called by Gilead. Its reluctance to do so  
was explained by Dr Otto. I have already indicated that I am not prepared to find any  
presumption from the failure of Jeremy Clark to testify.  
[904] Based on the evidence before the Court, I am not satisfied that Idenix has demonstrated  
that Dr Stuyver likely contributed to the invention of the 2’-C-Me/F nucleoside.  
(3)  
Dr Stuyver’s removal as a co-inventor on the ‘657 Application  
[905] Dr Stuyver was named as a co-inventor on Pharmasset’s provisional 368 patent  
application which is the priority filing for Gilead’s Canadian application. Before filing the PCT  
Application which resulted in the ‘657 Patent, Dr Stuyver decided to leave Pharmasset to return  
to Belgium. Idenix originally claimed that Dr Schinazi, the founder and CEO of Pharmasset, was  
angered by Dr Stuyver’s decision to leave and had his name removed as an inventor on the PCT  
Application. This line of argument was not pursued in final written submissions and I reject it as  
being entirely unsupported by the evidence.  
[906] I find it reasonable that Dr Stuyver was initially named on the US ‘368 Provisional  
Application because of his role in overseeing the testing PSI-6130, and not because of any  
perception that Dr Stuyver had come up with the idea for the compound. This is consistent with  
Dr Stuyver’s summary statement in the passage above which refers to his contribution as having  
conducted the assays of the 2’-C-Me/F nucleoside.  
Page: 317  
[907] There is also evidence that the decision as to who would be named as the inventor of the  
compound was done in an expedited fashion in order to file the application with the intention to  
review who the inventors were at a later date. [REDACTED]  
[908] However, Idenix maintains its submission that Dr Stuyver signed the declaration  
removing himself as co-inventor under duress. Dr Stuyver stated that the declaration was not  
accurate when it stated that he was not an inventor.  
[909] He testified that Dr Otto told him that if he did not sign the declaration, that “that was  
extremely bad news for the company and also very bad news for myself.” In respect of the  
negative impact on Dr Stuyver, he stated, “that there was no option for me, I had to sign. If not,  
they were about to, what I understood, they were about to take away my options.”  
[910] Dr Otto denies making any such threat. Although his testimony suffered certain memory  
lapses which are understandable in attempting to recount events approaching a decade ago, he  
was a credible witness. Drs Otto and Stuyver had a close personal relationship, vacationing  
together with their spouses, including around the time of the alleged threat. It does not seem  
plausible that Dr Otto would treat an old friend so deceitfully, when there is no apparent basis on  
the record anywhere that would demonstrate Pharmasset was not simply attempting to determine  
the inventors of the 2’-C-Me/F nucleoside, or that Dr Stuyver’s options could somehow have  
been in jeopardy.  
Page: 318  
[911] Neither witness could recall the specifics of any discussion between them. Given the onus  
on Dr Stuyver to prove duress, this does not assist his position. [REDACTED]  
[912] [REDACTED]  
[913] [REDACTED]  
[914] In addition, I find Dr Stuyver’s testimony tainted somewhat by his allegations against Dr  
Schinazi having removed his name because he left Pharmasset. This occurred after the compound  
was synthesized and Dr Stuyver had demonstrated its HCV antiviral potency. Dr Stuyver makes  
this allegation based on extremely weak hearsay evidence of an unnamed co-worker. It is  
contradicted by the documentary evidence, which indicates that he left Pharmasset on good terms  
for personal reasons and with his contributions to Pharmasset being greatly appreciated and his  
stock options fully intact.  
[915] Had he not cashed in his stock options early, he would have enjoyed a substantial return  
for his contributions. He sought further compensation at a later time, which was turned down. I  
do not think it pure speculation to suggest that given the phenomenally massive returns enjoyed  
by Pharmasset from the invention and its acknowledgment of his contributions to these results,  
he may be somewhat embittered towards everyone whom he believes denied him his just  
desserts, including Dr Otto, even if not justified.  
Page: 319  
[916] While it is unfortunate that he has not profited from his work to the same extent as other  
contributors at Pharmasset, I conclude that he was not a co-inventor of the compound, nor did he  
sign the declaration under duress acknowledging this fact in 2005.  
(4)  
Inventorship as “Material Allegation in the Petition” and an Alternate  
Interpretation of Section 53(1)  
(a)  
Materiality  
[917] Idenix argues that the omission to name Dr Stuyver was a “material” untrue allegation in  
the petition of the applicant in respect of the ‘657 patent within the meaning of section 53 of the  
Act. It submits that the omission is material for the public, such as being able to contact named  
inventors to discuss the invention and possible improvements, which in turn may result in further  
advancements in the art. It also argued that recognition of the inventors by the Canadian Patent  
Office promotes the integrity of the Commissioner’s office and the Canadian patent system, as  
well as adherence to Canada’s international obligations.  
[918] Idenix cites the Court of Appeal decision in Corlac Inc v Weatherford Canada Inc, 2010  
FC 602, varied 2011 FCA 228 at paras 123-124 [Corlac] to the effect that the benefits to the  
inventors, the public, and the Canadian patent system are not to be minimized.  
[919] In my view however, a fair reading of Corlac indicates that the Court very much  
minimized the impact of improperly omitting a co-inventor on the determination of materiality in  
section 53(1).  
Page: 320  
[920] Justice Michael Phelan in the Federal Court decision of Corlac at para 337, held that a co-  
inventor of the subject matter of the patent at issue had been improperly omitted. However, he  
concluded that the omission was not material because, at the relevant time, the patentee held all  
of the rights and interest in the patent and would have done so irrespective of whether the co-  
inventor had been named.  
[921] On appeal, the Court in upholding the trial Judge, explained at paras 123-125 the limited  
significance of failing to include a co-inventor as follows:  
[123] The third justification is founded on various public policy  
arguments and provisions of the Act that address the identification  
of inventors, the need to promote integrity of the Commissioner’s  
office and the Canadian patent system, Canada’s international  
obligations, and the personal benefits to which inventors are  
entitled in respect of their inventions. While the appellants’  
arguments are not to be minimized, it is highly doubtful, in my  
view, that they will be determinative for purposes of interpreting  
materiality in the context of subsection 53(1). I refer again to  
Q’Max where the absolute voiding of the patent on the basis of  
misstated inventorship was regarded as a “draconian remedy.” The  
appellants’ policy arguments must be balanced against this result.  
If the appellants’ position is correct, it would yield an anomalous  
result. That is, other inventors would effectively lose their interests  
in the patent monopoly rather than be able to access what they had  
previously been denied.  
[124] The most compelling of the appellants’ arguments in this  
regard is their point that accurate disclosure of the inventors’  
identity provides a number of benefits to the public. They maintain  
that the identification of inventors permits members of the public  
to contact those inventors to discuss the invention and possible  
improvements which, in turn, will assist in advancing the art.  
Further, they argue that parties to a patent infringement action are  
entitled to examine inventors as assignors of their patent rights  
under rule 237(4) of the Federal Courts Rules, SOR/98-106 (the  
Federal Courts Rules). Failure to disclose the inventors’ identity is  
said to inhibit a defendant’s right to make full answer and defence  
to allegations of infringement in such circumstances.  
Page: 321  
[125] These arguments, while interesting, are far from conclusive.  
The appellants do not suggest that members of the public have any  
particular right to communicate with inventors listed in the patent  
registry or that inventors are in any way obliged to respond to  
attempted communications. The examination of an inventor  
pursuant to rule 237(4) of the Federal Courts Rules may prove  
useful, but it is a pre-trial questioning of a potential witness, unlike  
an examination for discovery of a party: Teledyne Industries Inc. v.  
Lido Industrial Products Ltd. (1978), [1979] 1 F.C. 310, para 11  
(C.A.). Recourse to rule 238 remains available to litigants.  
[Emphasis added]  
[922] The Federal Court of appeal in Apotex Inc v Wellcome Foundation Ltd (2000), 195 DLR  
(4th) 641 expressed a similar sentiment that the failure to name a co-inventor is not a material  
misrepresentation at para 48 as follows:  
Examination of A&N's contention that a failure to name a co-  
inventor is a material misrepresentation leading to invalidity  
demonstrates that this is an illogical proposition. If such was a  
material misrepresentation, a true inventor who went unnamed in a  
patent would have no remedy to share in the monopoly of his or  
her invention.  
[Emphasis added]  
[923] Idenix has led no evidence to establish any manner by which the alleged failure to name  
Dr Stuyver affected the term, substance or ownership of the ‘657 Patent, or the public’s ability to  
use the invention. To the contrary, Idenix admits that the naming of Dr Stuyver on the ‘657  
Patent would have had no impact on the ownership of the patent pleading that:  
By virtue of their employment agreements with Pharmasset, the  
rights of Jeremy Clark and Lieven Stuyver in their purported  
invention were transferred to the same corporate entity.  
Page: 322  
[924] There was a suggestion during Dr Otto’s cross-examination on the last day of trial that Dr  
Stuyver may hypothetically have been under an obligation to assign his invention rights to the  
U.S. government on account of grant funding in relation to his research. It argued that this could  
have had a significant impact on the public’s use of the ‘657 Patent and would be “material” to  
the entitlement of the invention.  
[925] This contention was not supported with any evidence, nor was this scenario pleaded. It  
also runs counter to Idenix’s admission and the evidence that Pharmasset would have held all of  
the relevant rights even had Dr Stuyver been named. I disregard this evidence as having any  
relevance to the issue of materiality.  
[926] Idenix argues that this case is unprecedented because of the wilfulness of the omission.  
Idenix suggests that if it is able to prove that Dr Stuyver’s name was omitted “wilfully for the  
purpose of misleading,” then the materiality requirement would be automatically satisfied. It  
referred to comments of Justice James O’Reilly in Merck & Co v Canada (Health), 2010 FC  
1042, 88 C.P.R. (4th) 98, at para 56 [Merck/dorzolamide], where a similar section 53(1)  
argument “would have had considerable force” had the judge not found that there was no untrue  
allegation concerning inventorship.  
[927] Given my finding that there was no duress or willfulness in removing Dr Stuyver as an  
inventor in the Petition, I do not have to resolve this issue. I note however, that Gilead counters  
this submission on the basis that it is contrary to the ordinary rule of statutory construction that  
words in a statute are presumed to have a specific role to play in advancing the legislative  
Page: 323  
purpose (Ruth Sullivan, Sullivan on the Construction of Statutes 5th ed. (Markham: Lexis Nexis  
Canada Inc, 2008) at 210).  
[928] If Idenix’s theory were correct, it would be superfluous to include the term “material” if  
the legislature had intended that section 53 could be satisfied by evidencing only that a  
misstatement was willfully made for the purpose of misleading. I find this reasoning persuasive,  
and moreover, is supported by a similar analysis conducted by the Supreme Court in R v Kelly,  
[1992] 2 SCR 170 at 187-188.  
(5)  
Conclusion  
[929] For all the reasons above, I reject Idenix’s claim that the ‘657 Patent is invalid under  
section 53(1).  
XIII. CONCLUSIONSAND COSTS  
[930] In conclusion, I find that Gilead’s allegations as to the invalidity of the ‘191 Patent to be  
well-founded. I declare that the ‘191 Patent and each of its claims to be invalid, void and of no  
force or effect in rem on grounds of lack of utility and insufficiency.  
[931] I further dismiss Idenix’s counterclaim in its entirety.  
Page: 324  
[932] Gilead shall have its costs of the action and counterclaim, which shall include costs on  
motions conducted during and before trial on which no cost order was previously made.  
[933] The parties shall provide the Court their submissions on costs which shall be made in  
accordance with a timetable to be agreed upon and provided to the Court within four weeks from  
the date of release of this judgment.  
[934] The timetable will require Gilead to file its initial submissions, with Idenix to respond  
thereafter and Gilead to reply to any new matters arising from Idenix’s submissions.  
Page: 325  
JUDGMENT  
THIS COURT’S JUDGMENT is that  
1. Gilead has standing to bring this action pursuant to s 60(1) of the Patent Act.  
2. The ‘191 Patent and each of its claims are declared invalid, void and of no force  
or effect in rem.  
3. Idenix’s counterclaim is dismissed.  
4. Gilead is awarded its costs of the action and counterclaim payable by Idenix,  
which shall include costs on motions conducted during and before trial on which  
no cost order was previously made, all such costs to be determined upon the filing  
of submissions of the parties and the hearing of oral submissions, if requested and  
ordered by the Court.  
"Peter Annis"  
Judge  
FEDERAL COURT  
SOLICITORS OF RECORD  
DOCKET:  
T-1156-12  
STYLE OF CAUSE:  
GILEAD SCIENCES, INC. AND GILEAD SCIENCES  
CANADA, INC. v IDENIX PHARMACEUTICALS,  
INC., UNIVERSITA DEGLI STUDI DI CAGLIARI,  
L'UNIVERSITÉ MONTPELLIER II AND, CENTRE  
NATIONAL DE LA RECHERCHE SCIENTIFIQUE  
AND UNIVERSITA DEGLI STUDI DI CAGLIARI,  
L'UNIVERSITE MONTPELLIER II AND CENTRE  
NATIONAL DE LA RECHERCHE SCIENTIFIQUE  
PLACE OF HEARING:  
DATE OF HEARING:  
TORONTO, ONTARIO  
JANUARY 12, 2015  
ANNIS J.  
PUBLIC JUDGMENT AND  
REASONS:  
DATED:  
NOVEMBER 2, 2015  
APPEARANCES:  
Jason Markwell, Jordana Sanft,  
Amy Grenon, Adam Haller  
FOR THE PLAINTIFFS  
FOR THE DEFENDANTS  
Patrick Smith, Alex Gloor,Viktor  
Haramina, Ivy Tsui, Melissa Binns,  
ChrisVanBarr  
SOLICITORS OF RECORD:  
Norton Rose Fulbright Canada  
LLP  
Barristers and Solicitors  
FOR THE PLAINTIFFS  
FOR THE DEFENDANTS  
Gowling Lafleur Henderson LLP  
Barristers and Solicitors  


© 2022 IncJournal is not affiliated with or endorsed by the U.S. Securities and Exchange Commission