<PAGE>
EXHIBIT 99.2
Abstract
P53-selectively replicating adenoviruses expressing the cytosine deaminase
prodrug converting enzyme
S. Laquerre, M.D. Young, L. Johnson, J. Nye, L. Hawkins, S. Weber, P. Holman, M.
Lemmon, S. McCoy, L. Cohen, W. Collard and T. Hermiston
Onyx Pharmaceuticals, Inc, Richmond, CA 94806
Cancer is a major cause of death in the United States and new approaches are
clearly needed to prevent and/or treat the disease. Gene transfer of
chemosensitization genes is a promising cancer therapy strategy, with the
potential for higher active drug concentrations within tumors while
minimizing systemic toxicity. To take advantage of the prodrug converting
enzyme strategy, we modified a selectively replicating adenovirus CI-1042
(ONYX-015) to encode the E. COLI cytosine deaminase (CD) gene. The CD gene
was placed in multiple, distinct loci within the virus and expressed using
endogenous, viral promoters. Analysis of these viruses demonstrated that i)
the CD prodrug converting enzyme was expressed in infected cells and was
capable of converting 5-fluorocytosine (5-FC) to 5-fluorouracil (5-FU), ii)
the IN VITRO oncolytic effect of the virus expressing CD is increased in the
presence of 5-FC, and iii) the levels of 5-FU achieved after conversion of
5-FC within the tumor were higher than those reported to have a
chemotherapeutic effect. In contrast the plasma levels of 5-FU were much
lower than the tumor levels, implying reduced systemic toxicity.
Introduction
CANCER:
Although an increasing number of genes are found to be altered in human cancers,
mutations in components of the Ras signal transduction pathway, the RB and p53
tumor suppressor pathway, and the activation of telomerase are common features
of nearly all human cancers.
CI-1042 (ONYX-015) MECHANISM OF ACTION:
At Onyx Pharmaceuticals, we have pioneered the use of selectively replicating
adenoviruses as new therapeutic agents for the treatment of human cancer. The
first of these viruses, CI-1042 (ONYX-015), is a human adenovirus that has
been genetically engineered to lack the E1B55K gene, the product of which
binds to cellular p53 protein. In concert with another adenovirus encoded
protein, E4ORF6, E1B55K functions to target p53 for degradation in infected
cells. CI-1042 (ONYX-015) is therefore incapable of inactivating p53
function. Infection of normal human cells with CI-1042 (ONYX-015) results in
an upregulation of p53 protein and an attenuation of the virus' ability to
produce new viral progeny. Infection of human cancer cells with CI-1042
(ONYX-015) results in the unchecked continuation of the virus replication
cycle, the cancer cells are killed, new virus progeny are produced, and
neighboring cancer cells are infected to start the virus life cycle anew. To
date,
<PAGE>
CI-1042 (ONYX-015) has been studied in Phase II clinical trials in head and
neck cancer patients; in Phase I/II trials in patients with colorectal cancer
and with metastases to the liver and pancreatic cancer patients; and in Phase
I studies in patients with ovarian and lung cancers.
PRODRUG CONVERTING ENZYMES:
Although the results of these clinical trials are very encouraging, we have
developed a system permitting insertion of therapeutic transgenes into the
virus genome in order to expand the range of cancer indications served by our
selectively replicating viruses including gp19K, ADP and E3b. This approach
takes advantage of the endogenous viral promoters and the virus replication
cycle to express the transgenes, thereby coupling transgene expression to
virus replication. The significant features of this technology are selective
expression and amplification of the transgenes within the local tumor
environment. A number of anti-cancer chemotherapeutic agents are prodrugs.
Prodrugs are the inactive form of a drug that requires specific enzymatic
modifications to be converted to their active form. The E. COLI cytosine
deaminase gene, CD, is one such enzyme and it is capable of converting
5-fluorocytosine (5-FC), an antifungal agent into 5-flurouracil (5-FU), an
anti-cancer chemotherapeutic agent commonly used for the treatment of head
and neck and colorectal cancers.
Strategy
Our strategy has been to insert and express the E. coli CD gene from the
p53-selective replicating virus' genome. We envision a development strategy
whereby the CD-armed selectively replicating virus is administered to patients
and in the presence of the prodrug 5-FC, the CD-armed virus would not only
reproduce itself, but also result in the expression of high levels of CD
transcripts. Because the virus replicates selectively in the cancer tissue, in
the presence of administered 5-FC, it should also steadily increase the
concentration of 5-FU in the cancer mass. This should expose the cancer cells to
higher levels of 5-FU than is possible with systemic administration of 5-FU.
Replication and production of new virus progeny should also result in higher
levels of transduction and transgene expression than is possible via any other
mode of gene delivery. Selective expression of CD and conversion of 5-FC to 5-FU
at the tumor site should also minimize systemic exposure and toxicity.
Conclusion
1) 5-FC was detected in tumor and plasma of CI-1042 (ONYX-015) injected
animals, and as expected no 5-FU was detected.
2) In ONYX-700 injected tumor as well as in their plasma, the level of
5-FC was similar to that detected in CI-1042 (ONYX-015) treated
animal, however a small conversion was observed at 2 h post 5-FC
injection.
3) The level of 5-FC detected in tumor of animal treated with ONYX-740 was
lower than that detected in tumor of animal treated with CI-1042
(ONYX-015) or ONYX-700, presumably due to the high level of 5-FC to
5-FU conversion in these tumors. The level of 5-FC in plasma of animal
treated with ONYX-740
<PAGE>
was similar than that of animal treated with CI-1042 (ONYX-015) and
ONYX-700 and little 5-FU was detected in the plasma of ONYX-740
treated animal.
Thus, by incorporating a gene encoding a pro-drug converting enzyme into a
selectively replicating adenovirus, we have increased the potency of this
replicating virus, as determined by both in vitro and in vivo assays. This
"Armed Therapeutic Virus-TM-" is expected to have a higher therapeutic index
when used to treat human cancer.
FORWARD-LOOKING STATEMENT
This abstract contains certain forward-looking statements regarding the
development of potential human therapeutic products that involve a number of
risks and uncertainties. Actual events may differ from Onyx's expectations.
