To Build a Killing Machine David Kirn can't turn his back on a century-old quest to pit oncolytic viruses against tumors. By Andrew Holtz

As he entered the cramped academic office at St. Mary's Hospital in London seven years ago, David Kirn wasn't sure anything would come out of meeting with one of the world's top experts on pox viruses. Kirn hoped that Imperial College professor Geoffrey Smith would help him determine if a member of the viral genus that includes smallpox - and the vaccinia virus used to eradicate that killer from the wild - could become the vector for a radical approach to treating cancer. Kirn knew he was asking Smith to turn his intimate knowledge of pox viruses in a new direction: away from perfecting tame viral vaccines that expose themselves to the immune system.

"We went in there and said, 'Geoff, we want you to work with us to come up with new viruses. We want to do the opposite. We want a killing machine,'" Kirn recalls. Kirn had recently moved to Imperial College after leaving Onyx Pharmaceuticals in Emeryville, Calif. He was trying to make a fresh start after Onyx abandoned plans for what could have been the first Phase III clinical trials of a cancerkilling or oncolytic virus. Although Onyx's O15 adenovirus had shown that it could preferentially replicate in, and eventually burst or lyse cancer cells in patients, leaving normal cells unharmed, the results of early trials hadn't been strong enough to convince a large pharmaceutical company that O15 could be approved and marketed successfully. And even Kirn had doubts about adenovirus.

Kirn's search for the ideal oncolytic virus led him to Smith, and so with piles of journals and other papers surrounding them in the professor's office, Kirn began asking about the pox viruses and genetic modifications that might produce a prolific cancer killer. "When David came along and we spoke, it was certainly a new direction from the sort of things we'd been doing hitherto. ... It was a new adventure," Smith says.

Smith knew which of the viral landmarks would be useful for plotting the new course. He had been studying ways in which pox viruses evade the immune system and looking at ways to disable them. "It's amazing what goodies it has in there," Smith says. Kirn wanted something that replicated fast and went largely undetected, and he wanted to tap Smith's vast library of viral strains. "He had pox viruses from China and from Russia and India that were used as vaccines. He had elephant pox, camel pox, you name it; so we could immediately tap into that treasure trove and try to identify the best anticancer strains," Kirn says.

Of particular interest to Kirn was a form of vaccinia known as extracellular enveloped virus (EEV), which appears tailored for long-distance travel through the blood stream and seems resistant to complement and neutralizing antibodies. He saw EEV as a key to achieving systemic spread of a therapeutic virus, in order to hunt down distant tumor metastases. "[Smith] was able to identify mutant forms that produce more of this EEV form that we could immediately get from him and test," Kirn says.

Although Kirn was trying to break new ground, the therapeutic concept that he was promoting was older than anyone in the room that day. In 1904, the American Journal of the Medical Sciences published a case report describing how the white blood cell counts of a patient with chronic myelogenous leukemia decreased dramatically during what was described as a flu-like illness.¹

Reports like this prompted many investigators in the middle of the last century to apply viral infections to tumors, but the available viruses, either taken from the wild or adapted from contemporary vaccines, were blunt instruments. Advances in bioengineering have provided the means to sharpen viral activity, fine tuning its targeting and adding transgenes that might aid in the destruction of cancer. Kirn is eager to get these infectious agents into humans.

On a sunny March morning, the San Francisco Bay and the Oakland Bay Bridge offer a shining view out the window, but the attention of the six staff members gathered around a conference table at Jennerex Biotherapeutics is focused on the young company's founder. Kirn, his receding hair and moustache goatee closely-cropped has his sleeves rolled up and his collar open. He's standing at a white board, jotting key points about the status of research in their field. He and his colleagues have just returned from the Fourth International Conference on Oncolytic Viruses as Cancer Therapeutics near Phoenix, which brought together 180 attendees and presenters representing 16 companies and more than 50 academic centers worldwide. Organizers say this meeting was double the size of the last conference two years ago.

Kirn says the meeting has evolved from an emphasis on lab and animal work to a focus on human clinical trials. "Clinical used to be the last session. Everyone was burned out. Scientists didn't care and it was just painful. Now it's the first session," he says. Now, "People are saying, ?Mouse tumors? Who cares?'" Kirn says.

He ticks off a list of presentations that included reports on a variety of viruses being tested in humans: measles, reovirus, herpes, Newcastle disease virus, Seneca Valley virus, and adenovirus. "It used to be us and that was it. Now you look and there are eight in the clinic and more that didn't present." The field is dominated by a cadre of university-based enterprises and small companies, including BioVex, Oncolytics Biotech, Neotropix, and Wellstat.

The Jennerex team clearly feels good about how its report on the vaccinia strain compared to the other clinical presentations at the Arizona meeting. Kirn says he offered Phase I safety data on Jennerex's JX-594 vaccinia virus in 11 patients with a variety of advanced cancers, including melanoma, liver, lung, renal, and stomach cancers.

JX-594, like all of Jennerex's current vectors, has a deletion in the thymidine kinase (tk) gene that makes it dependent on high cellular levels of thymidine, as would be found in rapidly dividing cancer cells. Two other versions of the virus, JX-963 and JX-929, are based on a different viral backbone that appears to be more potent. Neither has been tested in humans yet. In addition to the tk deletion, the other versions have a second restriction, a VGF gene deletion that prevents the virus from activating cellular epidermal growth factor receptor (EGFR). Both deletions provide safeguards for normal cells, but cancer cells undergoing rapid division and with an active EGFR pathway are vulnerable.

All the Jennerex viruses are engineered with what Kirn calls a "payload," a foreign gene that could produce antitumor effects. In JX-594 and JX-929, granulocyte monocyte-colony stimulating factor (GM-CSF), is intended to stimulate white blood cell production. Kirn says they have seen cell counts increase two to five times among patients in their Phase I/II trials receiving higher doses of the JX-594 virus. JX-963 codes cytosine deaminase, an enzyme that can deaminate the prodrug 5-fluorocytosine, converting it to 5-fluorouracil, which is cytotoxic and radiosensitizing.

A cancer therapy using transgenic, replication-competent viruses raises concerns. As difficult as any biologic is to get approval, the treatment closely resembles both cancer vaccine approaches and gene therapy, the latter of which has long suffered a safety image problem.

The virus seems safe in humans according to Kirn. Aside from flu-like symptoms, there appeared little negative reaction and no dose-limiting toxicity in the safety trial. Kirn says he told the Arizona meeting goers that in eight of the first 10 liver cancer patients in a current trial, JX-594 produced what he termed high levels of intratumoral replication - enough to sustain at least 105 viral genomes per milliliter of blood - despite the efforts of the patients' immune systems to clear the virus.

"We see distant spread of the virus as well. We've had a couple of patients who had superficial skin nodules [that] we could biopsy after we treat in the liver, to see if the virus got there. And we are two for two," Kirn says. "We treated hundreds of patients with adenovirus, and we never saw that. And we looked." Although participants in the Jennerex trial have had both local and distant tumor responses, with no controls little can be said about efficacy or survival.

Only three miles away from the Jennerex staff meeting, a University of California, San Francisco, graduate student briefs his lab chief on the oncolytic virus conference. "By end of conference, I think it really became clear that if you suppress the immune system, you get better results," Mike Nehil reports. "Yes!" exclaims Frank McCormick, pumping his fists, but revealing the tricky nature of eliciting an immune response: Some response, it would seem, helps the virus kill the tumor, but immunity can snuff out the treatment as well.

Nehil tells McCormick that, although viral infections could trigger an immune system attack on tumors, presenters at the oncolytic virus meeting were able to correlate measurements of viral levels with tumor responses, thus providing evidence of a direct effect. McCormick, who is now the director of UCSF's Comprehensive Cancer Center, founded Onyx Pharmaceuticals in 1992. David Kirn was also on the team that nearly a decade ago came close to bringing an oncolytic virus into Phase III trials. Onyx's leading virus, O15, contained an E1B-55K deletion that, it seemed, rendered the virus harmless in cells with active p53.

A series of Phase I and II trials demonstrated that O15 produced tumor shrinkage. Based on the scientific and clinical successes, preparations were underway for Phase III trials, and ultimately the commercialization of the oncolytic virus, when business realities intervened. Onyx was in a partnership with Warner-Lambert, but then in 2000 Pfizer purchased Warner-Lambert. Under Pfizer's direction, Warner-Lambert backed away from its promise to commit $40 million to Phase III trials of the O15 oncolytic virus.

Meanwhile, Onyx had a partnership with Bayer to develop a more conventional small-molecule cancer therapy called sorafenib. Onyx had to prioritize the products in its pipeline, and so in early 2003 Onyx officially cancelled clinical work on O15. "If the company had to live or die by O15, we probably would have pushed that study forward. It was only a $10 million step," says McCormick, who left the company to direct the UCSF cancer center in 1997 but continued to advise Onyx. "It's possible we could have gotten O15 approved for some indications if we just stuck to it. But we hadn't shown systemic efficacy, so it might have been a minor niche product. Who knows? We didn't fail, but we didn't succeed, either."

Systemic efficacy, the ability of a virus to find and attack distant metastases, is critical to showing an advantage over conventional therapies. McCormick says the decision to drop further studies of adenovirus O15 made sense at the time, and he notes that the approval of sorafenib (marketed as Nexavar) to treat renal cancer indicates they made the right business decision.

David Kirn left Onyx in 2000, taking an academic post at Imperial College London, where he was able to step back from the demands of a business. "I think having been forced to go back to the drawing board and start over, while painful, in the long run has taken us far beyond where we ever would have gotten by making small incremental steps with adenovirus," Kirn says.

In London, Kirn undertook a fundamental review of the concept of oncolytic viruses. "I think all of us in the field were na￯ve, in the sense that based on the way viruses behave in a dish, you can eradicate almost any cancer cell line with almost any virus," Kirn says. Thinking about how infections work in vivo versus in vitro set Kirn looking for infections that would circulate and spread in the complex tumor environment. "What we didn't appreciate then, but we certainly appreciate now, is the fact that a three-dimensional solid tumor in the body is a much more complex situation."

Even with more than a century of work on oncolytic viruses that hasn't spurred a major breakthrough, Kirn says that things are different now. The goals of research are more sharply defined, and the questions being asked in trials are more specific. Instead of merely documenting the effects of available or slightly-modified viruses, new candidate viruses can meet specific criteria.

Kirn outlines three criteria that he says will set apart vectors such as vaccinia: 1) intravenous spread - vaccinia is highly motile, unlike adenovirus, which spreads by diffusion through mucus membranes; 2) high rate of replication - to avoid clearance by the immune system and to stay ahead of tumor growth and metastases; and 3) relatively large genome - to carry additional therapeutic payloads, such as proteins, to stimulate the immune system or anticancer agents. Vaccinia has an encoding capacity of 25-50 kb in its genome, compared to 2 kb in adenovirus.

Kirn says he also learned that many viral proteins are multifunctional. For example, the E1B-55K deletion that gave Onyx O15 its selectivity also attenuated it significantly; thus, even though it grew only in tumor cells, it didn't grow or spread fast enough. Moreover, recent studies have shown that the deletion neither prevents infection in cells with p53, nor ensures replication in cells without p53. Evidence now supports an mRNA transport role as the mechanism for selectivity.² "So I think that's just a model for the field, to say, ?Be careful. These things are not black and white,'" Kirn says.

After two years in academia, considering the best way forward, and establishing ties with Smith and other experts in virology and cancer, Kirn left adenovirus behind. He decided that vaccinia would be the best vehicle and plunged back into the business of developing a practical and effective oncolytic virus. He created Jennerex in 2003, though the company's first significant funding arrived only in 2006. The company is named in honor of Edward Jenner, who two centuries ago championed vaccination against smallpox with vaccinia.

Helena Chaye, a virologist and Jennerex vice president for business development, says she admires Kirn's courage: "He put his life on the line to do this."

The large pharmaceutical companies, however, which have the resources to commercialize an oncolytic virus, appear to be waiting for evidence that the technique offers a substantial advantage over conventional treatments. Survival data will be essential, Kirn and others say, as will proving a reliable systematic effect - that the virus can resist immune system clearance, and travel to and kill distant metastases.

Nevertheless, some signs look promising: A large pharmaceutical company has recently licensed a vaccinia-based cancer immunotherapy. Sanofi Aventis signed a $690 million dollar deal with Oxford BioMedica for TroVax, which is being billed as the first gene therapy that Big Pharma has embraced. Though different from an oncolytic approach, it signals increased acceptance of biologics.

Earlier this year, an adenovirus therapy not too different from Onyx 015 - Sunway Biotech's H101 - was approved for use in humans in China. Kirn and others say that Chinese regulators relied on the type of tumor-response data that Onyx had produced years earlier. They add that they aren't sure the clinical data collection system in China will be able to conclusively answer questions about patient survival. Efforts to contact Sunway Biotech for comment were unsuccessful.

Jennerex announced its own Asian foray last fall. It is collaborating with investigators in South Korea on a Phase I/II trial of its JX-594 vaccinia virus in patients with liver cancer. Kirn says liver cancer is an appealing target, because the tumors are honeycombed with newly-formed blood vessels that have porous walls, allowing easy passage of viral particles.

The liver is somewhat shielded from the immune system, thus giving the virus a better chance to survive. Moreover, the EGF and RAS pathways are turned on, which provide favorable conditions for vaccinia. "So a local application in the liver might have a high likelihood of actually improving survival," Kirn says.

He says they plan to present results of the Phase I/II South Korean trial at the American Society of Gene Therapy meeting this month. Work is already underway designing Phase II trials that could begin in the United States and Canada later this year. But as the presentations at the recent Arizona conference demonstrated, other investigators have picked other viruses, and much work remains for all of them.

"We're looking for the next big leap to make the whole thing work," says McCormick, who led the way with Onyx's O15. Now, from his lab at UCSF, he is watching all the newcomers with great interest. "They are all dark horses, except for O15, which is sort of a dead horse," McCormick chuckles. "It's definitely at the cottage industry level of things right now. I think it'll take a pretty big home-run of clinical efficacy to get people's attention again."

Kirn agrees that it's too early to say which of the viruses might be the first to clear the scientific and practical hurdles, and of course he's rooting for vaccinia. "This platform has the perfect features. [The major pharmaceutical companies] just need the clinical proof to say, ?OK, now it's safe to jump in the water.' And once they do, then I think the field will just explode from the industry standpoint."