A Virus and a Vector, Evolving

A decade and a half ago a team of researchers came together in Portland to pursue the use of cytomegalovirus as a potential vector for vaccines. Its future will be determined by upcoming human clinical trials.

By Michael Dumiak

Oregon Health & Science University (OHSU) researchers in the state’s pine-forested primate research center are preparing to do new things with something very old. Teams there are bolstering techniques and manufacturing practices as they ready an HIV vaccine candidate for safety studies in humans. The candidate employs cytomegalovirus (CMV) as the vaccine vector, a virus with a long history.

Probing further, OHSU researchers are manipulating CMV vectors in the hope of uncovering a tool for tailoring or “programming” different kinds of immune responses. Working with the large non-human primate (NHP) research program there, researchers hope to use CMV to create a “platform” approach to vaccine development, where a single vector is used for candidates against a variety of maladies. This work has gained the notice of high-flying venture capitalists. A US$150 million venture called Vir Biotechnologies launched earlier this year to support and grow the effort. Now it remains to be seen if the CMV candidate generating so much promising data in monkeys will deliver similar results in humans.

Something old, something new

The cytomegalovirus family tree stretches to the Triassic: the period of early amphibians and ferns. The virus has since undergone 200 million to 240 million years of Darwinian selection, making it very good at what it does. CMV infects 50 to 80 percent of Americans, most of whom show no symptoms. It is widespread in the developing world as well. Like other herpes viruses, CMV establishes lifelong latency in its host after infection.

Its persistent and widespread nature is part of what makes CMV an interesting vaccine vector. OHSU researchers are modifying the viral genome and hijacking it to carry HIV antigens, hoping CMV’s persistence might also mean lifelong protection against HIV. The vector’s other advantages include its low pathogenicity and efficient ability to reinfect, which should mean a CMV-based vaccine would be effective even if someone had already developed immunity to the virus through natural infection. Researchers also favor its large genome, which offers great potential for manipulation—it can express more than 200 proteins.

Louis Picker, associate director of OHSU’s Vaccine and Gene Therapy Institute, is a familiar name in the HIV vaccine research field. Jay Nelson, founder and director of the Institute, and Klaus Frueh, an immunologist-turned-virologist who is a senior scientist there, are also integral parts of the operation. Each have their own labs in Portland. All came to the Pacific Northwest by following CMV as a beacon. Nelson was already established as a CMV expert by 1992, coming to OHSU as a molecular virologist interested in the pathogenesis of the virus. The operation then really began to develop in the late 90’s and early aughts.

Frueh had been working first at The Scripps Research Institute in La Jolla and then for Johnson & Johnson (J&J), where he was director of an antiviral pharmaceutical research program. He was, at first, interested in studying how viruses evade detection from CD8+ T cells. Peter Ghazal, a molecular geneticist and former trainee of Jay Nelson’s, was working then with a viral protein from herpes simplex. Ghazal was the one who brought Frueh’s attention to a growing body of work about CMV. Frueh then began focusing on how CMV counteracts CD8+ T cells. Nelson recruited him to Portland in 2000.

By then Picker was already there, coming from Dallas where in 1999 he’d been working as a pathologist monitoring immune responses in specimens from AIDS patients at the University of Texas Southwestern Medical Center. Picker first directly encountered HIV as a resident at Boston’s Beth Israel in early 1983, where he was performing postmortems on patients dying of a mysterious virus prompting unusual, awful symptoms and devastating immune system collapse. What would come to be called AIDS continued to affect Picker personally. As the southern California native continued in his medical studies at the University of California, San Francisco, he lost classmates and acquaintances. He published his first clinical pathologic paper on HIV in 1985 and studied human T cell biology for the next decade. He started working on HIV from that point on.

While in Dallas, Picker became drawn to CMV and its potential in vaccine development. It stood out to him because the virus is largely benign and because it induces very strong lifelong immune responses. In long-term, healthy carriers of CMV, upwards of 10 percent of T cells that distinguish self from non-self are devoted specifically to CMV proteins, says Peter Barry, director of the Center for Comparative Medicine at the University of California, Davis. Barry’s lab also works with CMV and collaborates with the OHSU researchers.

Picker wanted to act against HIV, and Jay Nelson, CMV expert, already headed the Vaccine Institute. It was a perfect match. Picker, Frueh, and Nelson began collaborating on antigen design. Frueh and Nelson grew experimental batches of CMV vector candidates, while Picker designed and conducted trials with the center’s rhesus macaque population. The continuing experiments built up a collection of data illustrating how CMV induces immune responses in monkeys and how these immune responses affect SIV.

In 2009 Picker and colleagues showed that a rhesus cytomegalovirus (rhCMV) vector expressing the simian immunodeficiency virus (SIV) proteins Gag, Rev/Nef/Tat, and Env primed and maintained effective SIV-specific immune responses in 12 vaccinated rhesus macaques challenged with the highly pathenogenic SIVmac239 strain (Nat. Med. 15, 293, 2009). While all of the control group of 16 animals became progressively infected with SIV after repeated low-dose challenges, four of the vaccinated macaques never showed sustained SIV infection. It turned out they completely controlled it.

This suggested that the CMV-based vaccine immediately and completely controlled viral infection, with experiments later showing that this control is followed by clearance of the infection. Results from a study published the next year were the first to show that downregulation of major histocompatibility complex (MHC)-1, a set of cell surface proteins that tell an immune cell about foreign molecules, enabled CMV superinfection. This suggested that widespread pre-existing immunity to CMV would not hamper its use as a vector (Science 328, 102, 2010 and see CMV Superinfection No Longer Shrouded in Mystery, IAVI Report, Vol. 14, Issue 2, p. 17).

In 2011, researchers showed that the rhCMV vector could produce a ‘functional cure’ of SIV-infected macaques, likely due to a memory T-cell response (Nature 473, 523, 2011). A follow-up study from Picker’s group showed the experimental CMV vaccine also led to undetectable SIV levels in about 50 percent of macaques challenged vaginally and intravenously, as well as intrarectally (Nature 502, 100, 2013). This indicated that the vaccine-induced immune responses can control viral load in the blood as well as the lymphoid tissues where SIV and HIV establish infection, Picker says. Most importantly the study showed clearance of the SIV infection in protected monkeys over time. This was a startling and highly publicized finding, one still generating new experiments. At the most recent Conference on Retroviruses and Opportunistic Infections (CROI) in Seattle, Picker outlined how his group’s current efforts to determine whether the SIV reservoir is progressively eliminated by vaccine-induced SIV-specific immune responses, or whether these immune responses initially limit the formation of the viral reservoir to such an extent that it dwindles to nothing over time (see Rallying CROI, IAVI Report, Vol. 21, Issue 1, p.13). Experiments so far indicate that the rhCMV immunization is limiting the formation of the viral reservoir, a finding that could have a bearing on research into an eventual HIV cure.

Even now though Picker cautions that a cure strategy is on a much slower track than prophylactic CMV-based vaccines. “An HIV cure is sort of a separate issue. We’re asking whether it will work for that. It’s a very different kind of application than a prophylactic vaccine, even though the virus is the same. It’s just like Hepatitis B. We have a prophylactic vaccine for Hepatitis B. We all take it. It works very well. But we don’t have a vaccine that gets rid of Hepatitis B in people that are chronically infected. Therapeutic and prophylactic are two very different things,” he says. “We’re way more advanced in the prophylactic vaccine.”

The birth of a vector  

Cytomegalovirus (CMV) may have a pedigree stretching back beyond the pterodactyl, but the roots for using CMV as a vector lay in a technique developed in Germany in the 1990s by Ulrich Koszinowski and Martin Messerle.

Messerle and Koszinowski developed a strategy for the cloning and mutagenesis (or recoding of the genetic information of an organism) of an infectious herpesvirus genome. After modification, it is maintained as a bacterial artificial chromosome, which is a DNA construct that can replicate when inserted into bacteria. Messerle and Koszinowski then showed that they could take a mouse CMV genome and maintain it as a 230 base-pair bacterial artificial chromosome when inserted into E. coli (Proc. Natl. Acad. Sci. 94, 14759, 1997). Doing so allowed scientists to modify the genome—mutate, insert, or delete pieces of its genetic code—more easily than before. —MD


Programming an immune response

All the key data on rhCMV as a vaccine vector so far come from Picker’s OHSU team, as they are the only ones to publish on it so far, Barry says. OHSU was uniquely positioned to do this work. Researchers there can draw on the resources of the primate center, as well as the scientists with immunology and virology expertise in CMV.

This expertise was assembled step by step, but the run of successful experiments started with a stroke of luck, Picker says. The rhCMV vector that the Oregon researchers began experimenting with has a genetic configuration unlike any other CMV vector. It contains specific gene deletions in discontiguous places that turn out to be required for inducing the specific immune responses required for protection. “If we’d started out with a different CMV variant, we might not have seen this,” Picker says. He has five papers lined up on his desk detailing this work, but the ongoing effort to translate the monkey work into human trials is taking precedence these days.

“When we started vaccinating monkeys and ultimately challenging them, we saw a CD8+ T-cell response to SIV,” Picker explains. “We had no inkling that there was going to be anything unusual. We saw this efficacy, and to try and figure out correlates for this efficacy we started doing detailcmv model ohsued analysis of the immune response.” The CD8+ T-cell response to SIV is well characterized following vaccination and natural infection. There are canonical epitopes in this response: a monkey that is said to be Mamu-A01+, which is an analog for the human coding complex for MHC proteins, will typically respond to a certain peptide (Immunogenetics 38, 141, 1993; Virology 77, 9029, 2003).

But when the group vaccinated with a CMV vector, though, they didn’t see this kind of response to the canonical epitopes (Science 340, 6135, 2013). “That was an important clue telling us that we should look at the epitopes and what was going on there, because maybe that was important,” Picker recalls. The Oregon researchers suspected the control they observed in the vaccinated macaques came from a more promiscuous immune response involving multiple epitopes.

So Frueh and Nelson continued to construct vectors, and Picker tested them. The results created an understanding of the gene coding in CMV that elicits both unconventional and conventional responses. Picker has said publicly in meetings that the unconventional responses are what seem to be required for efficacy. Both Frueh and Picker were recently in the Netherlands at a CMV workshop discussing ongoing research. It appears they are on the cusp of characterizing which viral genes are required to get specific vaccine-induced immune responses. This boosts CMV’s prospects not just for HIV but as a potential vaccine platform.

The group is now working to show that genetically altering CMV can actually “program” highly diverse CD8+ T-cell responses that differ in their epitope targeting. “The issue is understanding how that works and understanding not only how those responses are generated by CMV, but also what they are good for,” Picker explains. “At this point, we know that they’re good for a prophylactic SIV vaccine and, presumably, an HIV vaccine. And in data that we’re going to publish hopefully soon, it seems to work for tuberculosis as well.” The unconventional responses elicited by their CMV vectors are dependent on the deletion of 157.5/.4 genes (Science 351, 714, 2016), as well as other genes about which the group has not yet published.

Ubiquitous but unique

While the momentum in HIV vaccine research over the last decade has been much on the side of designing an antigen that would prompt or “coach” the body to start producing antibodies against the disease, specifically antibodies that neutralize a broad spectrum of HIV strains, a CMV-based vaccine would work a little differently. It is not intended to induce antibodies, and in its current configuration, it doesn’t. It would instead rely on the body’s ability to mount a strong T-cell response to CMV. The persistence of the response allows the T cells that it elicits to avoid going into a resting state, so they remain ready to manifest antiviral activity. Picker describes it as the ‘early intercept’ hypothesis. “If you meet it at the beachhead, so to speak, it never has the chance to use its programmed host evasion mechanisms, and therefore it is vulnerable.”

But some of CMV’s advantages could also prove to be concerns: a persistent vector that causes harm would be problematic. While CMV is widespread and largely benign, it can be dangerous in pregnant women and associated with accelerated senescence of the immune system (Virus Res. 157, 175-9, 2011). CMV is also linked by some researchers to poor outcomes in elderly people. These effects seem linked to the persistent immune stimulus caused by the virus’s enduring potency.

The Oregon researchers are at pains to say safety is the priority. “We’re all very aware of that literature. The way to prevent that is basically to limit the ability of CMV to reactivate and disseminate, and that’s why our clinical vectors will be attenuated. All of this literature really depends on the virus being able to reactivate in old people and disseminate, and our vectors won’t be able to do that,” Frueh says. “The literature on involvement of CMV in immunosenescence is highly controversial. We also need to recognize that basically CMV is part of our immune system. It has been for a long time.”

Peter Hunt, who researches the inflammatory consequences of HIV infection from his post as associate professor of medicine at the University of California, San Francisco, also thinks the immunosenescence problem is overstated. “While the dogma that ‘CMV accelerates aging’ has been in the aging literature for some time, this has been widely misinterpreted and paints too broad a brush,” he says. “I have very few concerns about Louis’s fibroblast-adapted vector,” Hunt adds, referring to Picker and the OHSU vaccine candidate. “His data are some of the most exciting preventative vaccine data to emerge from non-human primate models of HIV infection.”

Which may be why the CMV program has drawn the attention of venture capitalists. At the start of this year, a company calling itself Vir Biotechnology made its debut in San Francisco with the backing of a top-flight venture capital firm run by biotechnology impresario Robert Nelsen, as well as funding from the Bill & Melinda Gates Foundation. Set to run the firm is George Scangos, the former chief executive of Biogen, a high-flying Boston biotech founded by Nobel Prize winners. Vir has grand plans. At the heart of the company is CMV.

By 2010, Frueh says, about the time the impressive data was emerging from their CMV vector experiments, the group—Nelson, Picker, Ulrich Koszinowski, and Frueh—began to talk about starting a company to spin off their CMV vector. The research effort was growing larger and it became clear that successful results might lead to a marketable product. In the fall of that year they launched a spinoff called TomegaVax. It fell to Frueh to guide it. He’s the one who’d previously worked in industry. “I did have that, though I have no experience in building a startup company. Neither did any of the other founders. We were like the blind leading the blind in some ways,” Frueh says. What he had done as part of his job at J&J, though, was to evaluate the stream of biotech interests that approached the company in search of an industry partner: a deal. The experience was quite valuable down the road as Frueh began striding the circuit with TomegaVax, looking for backers.

There are compelling reasons to start a spinoff. One is that as an academic the university owns what you do, for better or worse. In the US companies are also eligible for small business grants. In 2014 TomegaVax landed a $225,000 federal small business research and development grant to pursue a CMV-based vaccine against human papillomavirus, which is a prerequisite infection for the development of cervical cancer.

About this time the Gates Foundation and the US National Institutes of Health were also paying close attention to what was happening at the OHSU primate center. By 2014 Picker’s lab received $25 million from the Gates Foundation; last year the National Institutes of Health provided $14 million.

Once OHSU’s results began suggesting that by modifying cytomegaloviral determinants that control unconventional T-cell priming it is possible to uniquely tailor the CD8+ T-cell response for each individual disease target, it was clear they were outgrowing TomegaVax’s capabilities. “It’s not something that we anticipated when we started, that we would be able to do this. We’re rewriting textbook immunology,” Frueh says. “We were actually asked by the Gates Foundation to find a corporate partner.” This is a measure of success, but also a challenge. TomegaVax had set up shop in a Portland biotech hub in a warehouse district along the Willamette River. “The problem there was that everyone loved the technology. It was clear that this was something new. But we didn’t have experienced management,” Frueh says.

Last September, though, Bob Nelsen came calling. Nelsen, a co-founder and managing director at the venture capital firm ARCH Venture Partners, is based in Seattle and is known at Forbes as ‘Biotech’s Top Venture Capitalist.’ Nelsen had a vision for CMV. “It’s a much bigger vision than we had for the company,” Frueh says. “He was really saying, ‘Let’s build a big startup right out of the gate that has the capability of trying out different platforms and has enough money to do multiple clinical trials.’”

ARCH Venture brought $150 million to the launch of Vir Biotechnologies, with the Bill & Melinda Gates Foundation also contributing as a lead investor. Other funding, according to Vir’s January launch announcement, will come from sovereign wealth funds, public mutual funds, philanthropists, and family offices. “Vir seeks to take a new approach, using breakthroughs in immune programming to manipulate pathogen-host interactions. The company will take a multi-program, multi-platform approach to applying these breakthroughs, guided by rigorous science and driven by medical need,” the company says on its website. It is adopting a broad technological portfolio, including everything about CMV that was obtained with TomegaVax. Though it’s still early days, the roles at Vir Biotechnology are pretty clear: Scangos is chief executive, Picker is scientific advisor, and Frueh is a director.

Inflection points

With the launch of Vir Biotechnologies the Oregon team has reached an inflection point: having secured resources, it is now all about translating the promising data they’ve collected in monkeys to humans.

In a cramped office off a hospital hallway at OHSU’s sprawling teaching hospital in Portland, Marcel Curlin, a young, bushy-haired researcher and a new member of the team with a background in HIV clinical trial research in Bangkok, is leading the effort there to screen potential volunteers for the vector’s Phase I safety trials. Curlin expects to screen 400 Portland-area residents to make sure they already carry CMV (and that their partners do as well) and that they are not pregnant or carry other risk factors. The goal is to enroll 75 volunteers for a Phase I trial. The start date, though, is a little fuzzy. There have been challenges along the way involving the CMV vector itself.

One is achieving the right level of attenuation to limit the ability of CMV to reactivate and disseminate, in other words to ensure its safety. The team developed a “safety valve” for the vector to accomplish this by deleting a gene that is essential for virus dissemination. Deleting this gene eliminates safety concerns, but the key is to balance safety with efficacy—an over-attenuated vector may render it less able to induce the types of immune responses that were protective in animal studies.

Another delay is because of the findings on immune response programming. Programming could wind up as a stellar advantage for building a vaccine platform, the very thing that interests Vir in CMV. The researchers don’t want to use vectors in humans that have different modifications compared to the original monkey work, so clinical vectors had to be re-designed according to the most recent results. Researchers found another essential change last year that had to be added to the clinical vector.

All told, these efforts have caused the timeline for human trials to slip from an expected start this summer. Picker now says the earliest that trials will get underway is the end of 2018 and that will probably move into early 2019.

“We’re rushing to finish up this characterization and to get the correct human vectors. We only figured out in the last year how to make them and they have to get to manufacturing standard,” Picker says. “We’re running hard to try and do that. Obviously, the monkey model is great, but,” he trails off. If CMV’s promise doesn’t translate to humans, it would be a huge setback. If it works similarly, ARCH estimates it may deliver $500 million to the overall CMV project.

Frueh has emerged as the point person in solving some of the manufacturing issues involved in making a vaccine a real possibility. “I’m actually not a trained virologist,” he says. “I’m now the virologist in this program.” Yet Frueh is relaxed in his corner office, looking over the primate center’s central services complex, a midcentury mod-style building. The primate houses are just beyond, where hundreds of monkeys are either leaping around in cages or in a big open oval, socializing.

Although Picker’s primary interest is the basic science, he is finding the translation of the CMV work to humans an interesting (and necessary) challenge. “The reason I haven’t burnt out yet is that this isn’t just development of a vaccine where you completely understand how it works and the energy is in making it manufacturable,” he says. “To do this translation you have to understand the basic science. It’s interesting and different and weird immunology. It creates problems—you stretch your colleagues’ credulity. But when you come down with solid findings, it’s fun.”

Michael Dumiak reports on global science, public health and technology and is based in Berlin.