UCSF finding suggests mechanism of experimental cancer drug, hinting at broader application for the

By Jennifer O'Brien

University of California San Francisco researchers may have discovered why the
experimental anti-cancer drug Onyx-015 works more broadly than had been
expected, a finding that could solidify and expand use of the drug and which
reveals a potential new target for therapy that could prompt the development of
other cancer drugs.

Onyx-015 is a genetically modified version of a cold virus that was designed to
home in on cancer cells, specifically those in which the p53 tumor-suppressor
gene - a key sentinel of cell health—does not function. P53 is mutated and
does not function in 60 percent of cancers.  The theory has been that, without
the p53 defense, Onyx-015 could replicate in cancer cells.

However, researchers have observed in clinical trials and cell culture studies
that Onyx-015 also works against tumors in which the p53 tumor-suppressor gene
appears to remain intact. This has created heated debate in the scientific
community about the drug’s efficacy and mechanism of action.

Now, the UCSF researchers believe they have solved the conundrum: In the
October issue of Nature Medicine, they report that while some tumors responding
to Onyx-015 have mutations in the p53 gene itself, others have a defect in
another gene, known as p14 ARF, which is located “upstream” in the p53 pathway
and which indirectly regulates p53 function. They showed that loss of p14 ARF
leads to the deregulation of another protein, Mdm2, which in turn inhibits p53.

As a result, p53 cannot shut down the engineered virus’s attempt to take over
the cellular machinery, replicate incessantly and eventually kill the cell—
action that, counterintuitive as it may be, is just what researchers want to
happen when it comes to cancer cells.

The finding “solves a big puzzle about the drug’s mechanism of action,” says a
co-author of the study and the drug’s creator, Frank McCormick, PhD, director
of the UCSF Comprehensive Cancer Center, and co-founder of Onyx Pharmaceutical
Inc., which develops the drug. “Inactivating p14ARF is just another way of
turning off p53.”

The finding also suggests the possibility of new applications for Onyx-015,
says the senior author of the study, Michael Korn, MD, UCSF assistant adjunct
professor of medicine. “Our work supports the idea of exploring the use of
Onyx-015 in tumors that we presumed would be immune to the drug because they
were thought to frequently have p53 genes, such as such as melanoma and
glioblastoma [a form of brain cancer]. Clinical trials have to show if the p14
ARF and p53 status could serve as a reliable predictor of response to treatment
with Onyx-015.”

The finding could also lead researchers to consider how to design other
genetically engineered viruses that take the p14 ARF mechanism into
consideration, says Korn.

Moreover, he says, the discovery suggests that p14 ARF could potentially become
a therapeutic target in itself. A drug that could reactivate p14 ARF would
reactivate p53, thereby reinvigorating the cell’s defenses.

To date, researchers have focused clinical trials of Onyx-015 on cancer types
that previous studies have shown often have inactivated p53, though the drug
has been administered randomly. The new findings, says Korn, will encourage the
researchers to clarify the status of 53 and p14ARF in the individual tumors

In a recent multi-institutional phase II trial of Onyx-015 led by investigators
at University of Texas M.D. Anderson Cancer Center, 30 patients with advanced
head and neck cancer, most of whom had failed surgery or radiation therapy,
were treated. The trial was designed to determine the appropriate dosage for
the drug. In 63 percent of the patients, the tumors shrank by 50 percent or
more, and in eight of these 19 responding patients the tumors disappeared
altogether. (Nature Medicine, August 2000). The patients were treated with a
combination of Onyx-015 and chemotherapy.

Because researchers later determined that the response to the combination
therapy occurred both in tumors in which p53 was mutated and in which it was in
tact, the finding left open the possibility that chemotherapy was the critical
factor in the shrinking of tumors. However, says Korn, the new finding raises
the possibility that the tumors in which p53 remained intact might have been
defective in p14 ARF, and were, therefore, responding to Onyx-015. Based on
this hypothesis, he says, “It will be important to assess the effect of
Onyx-015 without chemotherapy, to see if the drug has the same potent effect.”

According to Onyx Pharmaceuticals, researchers participating in ongoing phase
II trials of Onyx-015 in different cancer types, including colorectal and
pancreatic cancer, have observed a response to therapy in some patients. As is
the case with all experimental cancer drugs, Onyx-015 still must be evaluated
in a phase III trial, a much larger trial designed to test the efficacy of a
drug in a large population. A phase III, multi-institutional clinical trial of
Onyx-015 for head and neck cancer has begun.

In another study, Onyx Pharmaceuticals will evaluate the p53 status of tumors
in patients who are participating in a phase II clinical trial in which
Onyx-015 will be administered intravenously. To date, the drug has been
administered in different ways: by direct injection into the tumor by
installation into the peritoneal cavity in patients with ovarian cancer, and by
injection into the artery that supports the liver in patients with
gastrointestinal cancer metastatic to the liver.

Onyx-015 is an example of the new wave of cancer therapies emerging based on
researchers’ growing understanding of the molecular mechanisms of cancer and
viruses. Researchers know that p53 plays the crucial role of halting cell
division when it detects that a cell’s DNA has been damaged, as occurs in
cancer or when a virus enters a cell. In many cases, loss of p53 function is
one of the first of several steps leading to the disruption of a cell’s
regulatory processes.

McCormick conceived the drug based on this knowledge of p53 - and the fact that
when the adenovirus cold virus, a relatively harmless irritant, enters a cell,
its first trick is to dismantle the p53 tumor suppressor gene, preventing it
from signaling the cell to self-destruct in the face of an invader. Once the
normal cold virus has taken over the cell’s molecular machinery, it begins
replicating, ultimately killing the cell. The adenovirus’s victory leads to
nothing more than sneezing and congestion.

But McCormick realized that if he disabled the protein within the cold virus
that sabotages p53, the virus would not be able to replicate in normal cells,
because p53 would remain intact and thwart the virus’s efforts. The virus
would, however, be able to grow in cancer cells in which the p53 gene was
already disabled. In this way, the crippled adenovirus would discriminate
between cancer cells and normal cells - a Holy Grail of cancer therapy.

In their effort to elucidate how Onyx-015 works in cells that appear to have
normal p53 genes, the researchers examined, in culture, colorectal cancer cells
that respond to Onyx-015 even though their p53 gene appears to remain in tact.
They observed that when p14 ARF was mutated, p53 did not function, even though
the p53 gene itself was not damaged. But when14 ARF was reintroduced in cells,
p53 began functioning again and Onyx-015 was no longer able to replicate in the
tumor cells. The explanation, they demonstrated, is that the loss of p14 ARF
leads to the deregulation of another protein, Mdm2, which in turn inhibits p53.
In this way, p53 is prevented from inducing the protective effects that would
otherwise shut down the engineered virus.
To confirm their findings, the researchers will examine tumor samples from
cancer patients who have intact p53 genes but respond to Onyx-015, to see
whether the p14 ARF pathway is, in fact, defective in these cells, says the
lead author of the study, Stefan Ries, PhD, a postdoctoral fellow in McCormick’
s lab.

In addition, researchers in McCormick’s lab are attempting to create a mouse
version of the Onyx-015 virus. If they succeed, they will then “knock out”
either the p14ARF gene or the p53 gene in mouse tumors to see how the drug

“We’re definitely making progress in understanding Onyx-015,” says Ries. “This
was a very nice discovery.”

Co-authors of the study are Christian H. Brandts, MD, a postdoctoral fellow,
Alicia Chung, a staff research associate, Carola H. Biederer, PhD, a
postdoctoral fellow, and Ettie M. Lipner, staff research associate, all in the
McCormick laboratory, and Bryon C. Hann, MD, PhD, a specialist in the
laboratory of Alan Balmain, PhD, UCSF professor of biochemistry.

The ongoing research is funded by Onyx Pharmaceuticals, Inc. (NASDAQ: Onxx).
McCormick remains a paid scientific advisor to the company, and is a
stockholder in the company. Korn is a paid clinical advisor to the company.