The Man Who Says Science Blew Its Best Shot at an AIDS Vaccine
And then the AIDS epidemic hit. The first hint that a new disease was killing people had come in 1981, in a publication called the Morbidity and Mortality Weekly Report. What followed were years, decades even, of tragedy and homophobia-tainted ignorance. Still, by 1987 the first vaccine trial was underway, the World Health Organization had launched a global fight against the contagion, the playwright Larry Kramer had started the activist group ACT UP, and the first antiretroviral drug, AZT, was available.
Even so, science was still woefully short of understanding the plague or coming up with a vaccine that could prevent it. By the end of the decade, more than 100,000 people were infected in the United States. Absent treatment, the mortality rate for these patients, then as now, was effectively 100 percent. Dorman knew vaccines; he started talking to other people in the vaccinology world about being part of the fight. Don Francis, a longtime disease hunter then with the Centers for Disease Control and one of the main characters in the book And the Band Played On, got in touch—Dorman had beaten feline leukemia, and it’s caused by the same type of virus that triggers AIDS. Why not try to tackle HIV?
Dorman was leery. The emerging effort looked chaotic. Dorman got his old vaccine team together for a meeting at his office; his son Sam remembers one at their house in Berkeley. “My image as a kid was of my dad feeling duty-bound, that there were people suffering and in danger, and they could do something,” Sam says.
Dorman decided to try. Today his name peeks through some of the stories of the early days of the epidemic and the hunt for a vaccine against the virus. He photobombs, metaphorically, books about the early years of the effort like Jon Cohen’s Shots in the Dark and Patricia Thomas’ Big Shot (both published in 2001). Since then, new therapies like antiretroviral drugs have made HIV infection into something it’s possible to live with rather than die from—at least, in the developed world. The search for a vaccine continues, a decadal, tidal ebb and flow of optimism followed by failure.
Dorman, too, is still at it. But he’s pushing an approach to developing a vaccine that he argues the entire scientific edifice has largely abandoned. Dorman advocates a path that you might broadly term “classical.” It’s almost trial-and-error, a methodology that goes back to smallpox and rabies. Like early vaccinologists—Jenner, Pasteur, Salk—Dorman is a tinkerer who figures out how to grow, kill, and administer viruses in a way that sparks an immune response. It can work—and indeed often has—without a researcher knowing much, if anything, about the underlying immunology.
Because of how HIV works—how the virus infects a cell, what kinds of cells it infects, how it mutates and reproduces—and because of how vaccines get tested and developed, most scientists working on HIV immunology don’t think that such a classical approach can work. Instead, they aim to break apart and rearrange the specific pieces of the virus, like the sugars and proteins embedded in its shell, and deliver those alongside enhancing agents. These approaches, arising from recombinant DNA and protein technologies, are by their very nature more hypothesis-driven. More rational. This is the approach that garners almost all research funding from government agencies and pharmaceutical companies.
Perfectly reasonable. And yet, in the 35 years since scientists isolated the virus that causes AIDS, 35 million people have died of the disease worldwide while waiting for a vaccine. “It’s a nice thing to argue that we will one day understand the biology well enough to do rational design of a vaccine,” Dorman says. “But an equally diligent effort should be made to extract what we can from methods that have already been invented.” That is what he has been saying for three decades. It hasn’t happened.
To the mainstream scientific community, Dorman’s quest is quixotic at best, tilting at windmills made of glycoproteins and RNA. But Dorman, now a spry 80, hasn’t given up. He’s convinced that if the rest of the scientific community had joined him decades ago, millions of lives would have been saved. They still could be. This isn’t a matter of science—at least, it’s not a matter of only science. It’s a matter of scientific culture—of a framework for making decisions about a research agenda.
That doesn’t mean Dorman is right and they’re wrong. He’ll be the first to say that he doesn’t know. But he’s also the first to point out: No one else knows, either. Not for sure.
That recovery hints that a vaccine is possible for those diseases—and many others. It’s a matter of creating a trigger that will speed up some processes that already happen in nature.
In the late 1700s, people knew that getting cowpox made you less likely to get smallpox, which in those days in Europe killed up to 400,000 people a year. Cowpox was a virus that their immune systems could fight off, and the cellular recording of that fight—the creation of a defense protocol in the immune system—would often ward off smallpox, too. This led the scientist Edward Jenner to try intentionally infecting a boy with cowpox as a prospective treatment. When the boy was subsequently found to be immune from smallpox, Jenner’s method took off, and the death toll from smallpox plummeted. The Latin name for cowpox is vaccinia, after the Latin for “cow,” so Jenner named his process “vaccination.”
Jump ahead to the mid-1950s, when 16,000 people, mostly children, got paralytic polio every year in the US. Working from the idea that a vaccine didn’t actually have to infect someone with a disease to jump-start immunity, Jonas Salk learned to kill poliovirus with formaldehyde and administer it. This “whole killed virus” vaccine worked, though people like Albert Sabin thought the formaldehyde would lead to a shorter period of immunity.
This is classical, empirical vaccinology, 20th century style. The road wasn’t always straight. As Cohen’s book Shots in the Dark lays out, Salk tested his vaccine on children—with few if any of the permissions and safeguards a researcher would need today, like FDA approvals or signoffs from Institutional Review Boards. He just kind of … did it. It worked. Polio has been almost eliminated on Earth, and smallpox no longer exists in the wild.
The past half-century has been miraculous—something like 50 vaccines exist for humans, and hundreds for nonhuman animals we live with. The process for making almost all those drugs, essentially, involved subjecting a pathogen to every tool a laboratory can bring to bear—how to grow it in culture, how to kill or attenuate it, what chemicals to administer alongside it, how many doses to give and in what interval. Ideally in the end you hit on a combination that confers immunity. It’s a process that a tech entrepreneur like Sam Dorman would identify pretty closely with product development. Lots of iteration. An engineering problem.
Why? First of all, HIV is a retrovirus, a type of virus that rarely infects humans. This crafty bug—protein-and-sugar molecules embedded in a fatty coat around a bundle of genetic material—invades a cell and copies its genetic material, RNA, turning it into DNA and then inserting it into the nuclei of the host’s own cells. The viral DNA becomes part of the person. Today antiretroviral drugs interrupt that process—they prevent the viral RNA from becoming DNA, or keep it from integrating into the cellular genome, or stop the cell from making new virus. But take the antiretrovirals away, and the virus starts churning out again.
Maybe even more importantly, HIV attacks cells that would otherwise mediate a response to a pathogen—among them, CD4 T-cells. CD4 is a protein on the outside of certain cells critical to the human immune response; it’s also the protein that HIV latches onto and uses to break into those cells. A hybrid sugar-protein on the outside of the virus called gp120 hooks onto CD4 like a key, opening the door for other HIV proteins and allowing the virus to fuse with the cell and inject its genes.
Not only does the virus make a lot of copies of itself very, very quickly, its many different strains also mutate. The human immune system will attack any invader, but it also learns to tailor a response to specific pathogens based on proteins on that pathogen’s outer shell, waving like the livery of an enemy combatant. Studying HIV, researchers learned that it was protean. The virus’ glycoproteins change slightly from generation to generation, allowing it to evade detection. It doffs its old livery, in other words, and puts on a slightly different outfit, one that somehow evades the immune system’s surveillance tactics.
These nuances have only revealed themselves over 35 years of research. Unlike other pathogens, nothing about HIV even hinted a vaccine was possible. “We have to be better than human,” says Larry Corey, principal investigator at the HIV Vaccine Trials Network. “We’re zero out of 65 million in self-cure.”
But one issue became clear early on. As people began to look at vaccine strategies, they lost confidence in their ability to kill or fully inactivate HIV—a necessary predicate not only for a vaccine but simply to test candidates.
When people sign up for trials—much more regulated than in Salk’s day, and much more subject to the vicissitudes of liability law—those research subjects don’t know if they’re going to get a placebo or the medicine. And because AIDS is near-universally fatal, researchers and ethicists want to be able to assure them that they won’t accidentally get the disease. “We can tell them, hand on heart, you won’t get HIV from the product,” says Mitchell Warren,1 executive director of AVAC. “That’s a really important message to give people enrolling in clinical trials, and we couldn’t do that working with a whole killed vaccine.”
Even into the 1990s, AIDS vaccine researchers were still ruminating on the possibilities of classical techniques. But the complexities of the virus, an evolving understanding of the human immune system’s response to it, and the fear of causing new infections all combined to lead them to reject that approach. That happened just as the tantalizing new field of genetic engineering, of recombinant DNA, offered the potential of not only an AIDS vaccine but perhaps an entirely new avenue of vaccine discovery.
Instead of the classical spadework Dorman advocates, scientists would use genetic and protein engineering techniques to build a vaccine from scratch, taking pieces of the HIV virus, bits of other viruses that the immune system kind of knows what to do with, “adjuvant” agents that boost an immune response…bits and pieces chained together into an exquisite corpse of immune-goosing biotechnology.
The approach has seen some success; it led to vaccines against hepatitis B and human papilloma virus. It takes longer, but it also allows scientists to learn more about the pathogen and immunology. And it has come to seem like the best route to an HIV vaccine.
He didn’t plan to go into the vaccine business; Dorman thought he’d become a researcher and a professor. “But before I finished my Cancer Society post-doc, I had a wife and four kids,” he says. Academia wasn’t going to pay enough to support a young family. So instead Dorman started what we’d now call a biotech company. Not that it was a surefire moneymaker. “There’s no such thing as a home-run product in animal health,” he says. And doing science outside the academy didn’t have the respect it does in today’s venture-capitalized hothouse. “In those days, if you left academia it felt like walking the plank,” Dorman says. But the business actually succeeded. That makes him something of an outlier in vaccinology: Burt Dorman has actually made vaccines. Most of the people working on HIV have not.
The business wheel turned. Dorman sold the vaccine company and switched to making diagnostic technology. Then, in 1988, he wrote a proposal to work on HIV—make virus, purify it, kill it, learn to formulate it into a vaccine, figure out dosages, a hundred different variables, levers to pull and tweak—and sent it to Anthony Fauci. Give me two years and $5 million, Dorman said, and he would have a vaccine ready for human trials. “The NIH clinical study section laughed at that,” he says. “My rebuttal said, OK, we’ll take four years and $10 million, but that just pissed them off worse.” (Fauci doesn’t recall the proposal or Dorman.)
Into the 1990s, as success continued to elude the vaccine research community and governments and NGOS began to make increasingly large financial commitments to research, Dorman was still pitching. The mid-1990s brought the International AIDS Vaccine Consortium (known as IAVI), funded by foundation money from Rockefeller and Bill Gates, among others. “By then I had written so many failed proposals to NIH that the staff was pretty self-conscious,” he says. “Over the years I pitched IAVI, I pitched Gates. I pitched the Grateful Dead.” All those funders would say killed viruses wouldn’t work against HIV.
Dorman argued back. Just let us try it, he’d say. Simply testing the idea might provide new knowledge about HIV and its so-called correlates of protection in the human immune system—so other vaccine makers would know what to look for in their own research. He kept pitching. He got letters from researchers who said he might have a point. He tried to get editorials into journals, unsuccessfully.
That went on until 2000. “And then I gave up,” Dorman says. “In the process I pretty much ruined my diagnostics company.”
From that perspective, vaccines are not great. For example, Francis wrote, it cost VaxGen $300 million to try and fail to develop an HIV vaccine in the 1990s. It cost Avirion $340 million to bring Flumist, the nasal flu vaccine, to market. (Another writer speculates it cost Sanofi Pasteur $1.4 billion to develop a recombinant, live-attenuated dengue vaccine—and it took 24 years.)
Maybe they didn’t expect to incur those costs. Maybe they believed in doing the right thing anyway. Sometimes vaccine makers also have NGO or philanthropic funding. Regardless, the incentives are all upside-down. Even if a manufacturer gets one made and approved, it’s hard to sell on an open market. Vaccines prevent rather than treat, and require just one or a few doses (flu vaccines being a notable exception).
That means the company makes only one sale, as opposed to keeping patients on the hook for a lifetime prescription. On the demand side, many people have a hard time internalizing the true risk of a disease they might someday get, maybe. And in the developing world, people are less likely to be able to pay for the product. And so on. So back-burnering vaccine development in favor of therapeutics and drugs to treat chronic conditions is just good business. Put it this way: In 2001 the entire market for all known vaccines was the same size as the market for the anticholesterol drug Lipitor.
A few billionaires have tried to fight this garbage fire by making it rain money. Francis’ article acknowledges that donations from Warren Buffett and the Bill and Melinda Gates Foundation introduced billions of dollars into vaccine development, but foundation money and grassroots efforts tend to tilt research in the direction of single vaccines for single diseases, potentially siloing off broadly useful knowledge. “The private sector works on things that are proprietary, patentable, and profitable,” Dorman says. “Everybody is doing exactly what you would have expected and wanted them to do. I don’t have criticisms of anybody’s role at all.”
Except, well. He kind of does. Burt Dorman has been at this for three decades. He’s an old, Berkeley-looking dude now. Glasses in a jacket pocket, pad and pen in a shirt pocket, close-cropped beard. Tends to give the same examples for why he’s right more than once, in long emails—as you might expect from someone who has been advocating the same thing for decades.
But what he is outlining is a tragedy made worse by a tantalizing possibility: What if he was right? Not now, today, talking about vaccines in a Berkeley coffee shop, but back then? Thirty years and 35 million deaths ago?
What if he was right?
He realized he had to try again. A couple years later, in 2010, after years in the world of technology and product development, Dorman’s son Sam decided to join him. Sam had noodled around with video and thought that might make a difference. “I think I just thought, maybe I can help my dad a little. He writes beautiful letters and beautiful papers, and he has a lot of faith in their ability to sway people,” Sam says. “But I thought if I could bring more modern, visual storytelling to this that it might be helpful.”
So they started a website, KillHIVNow.org. After decades of carrying around testimonial letters from disease hunters and trying to get articles published in journals, Dorman’s arguments and those of his supporters are online videos now, laying out the case.
His timing couldn’t be worse.
This year, for the first time in, well, ever, researchers are testing a vaccine in human beings that actually seems to kind of sort of work a little.
Like all the vaccines under investigation, it goes by a lot of different names. The Thai vaccine, because of where it first showed promise, also known as RV144 And, in a way, its history is a mix of classical and hypothesis-driven vaccinology. Researchers noticed that long-term nonprogressers have what are called cellular responses that control the disease. So it’d be good to try to induce those, went the thinking.
Researchers also knew that infusions of a particular kind of antibody prevented infection from an engineered, laboratory-built simian-human immunodeficiency virus … and that antibodies and T-cell responses could protect monkeys against infection from both S-HIV and simian immunodeficiency virus. Oh, and they had a Phase III trial of a vaccine made from the HIV glycoprotein gp120 that didn’t work, but it did hint at all sorts of immunological characteristics that might.
So RV144 mixed all that in a pot—combining four injections of a recombinant canarypox vector vaccine with two booster injections of the gp120 vaccine. In 2009, public health workers from the Thai Ministry of Public Health, Thai universities, NIAID, the US military, and lots of other places reported that RV144 showed an efficacy of 31.2 percent. That is to say, about a third fewer people in the test group got infected than the control group.
That seems like not a lot. It’s also better than any other HIV vaccine has ever done. And a new formulation of it is being tested in more than 5,000 people in South Africa.
Another study getting underway uses an adenovirus as a vector, and genes from several variants of HIV—a so-called mosaic—and a booster with a different combination of ingredients, or another envelope protein from HIV called gp140. Or both. It showed some protection in monkeys and people.
And yet a third study is taking an approach that would not have been possible in the days of classical vaccinology. Those researchers are working from the knowledge that a class of immune cell called a broadly neutralizing antibody can prevent infection with a major strain of HIV. Specifically the researchers are using an antibody developed at the National Institutes of Health’s Vaccine Research Center, called VRC01, infusing it directly into men and transgender people who have sex with men to see if it provides protection. And as the name implies, it’s only the first of many potential antibodies—a “proof of concept,” says Warren of AVAC.
All those years of immunology research have given scientists the ability to more quickly understand all sorts of outcomes. Today’s would-be vaccine-makers have a whole new set of tools that let them take results from small test groups and re-engineer formulations almost on the fly. “It’s called de-risking,” Corey says. “You can put in all this engineering. It’s complicated, but it’s not a biological concept.”
Indeed, 30 years of work has transformed immunology as a whole. It’s starting to look like product development again. “We have incredible, high-resolution tools, down to the molecular level in terms of understanding antibodies and the surface of the virus to know whether we’re achieving the kind of immune response that we’re targeting,” says Mark Feinberg, president and CEO of IAVI. He says that as early as the Phase I trials designed only to test the general safety of a new drug, vaccine researchers can now get a sense of whether they’re on the right track.
Fauci, meanwhile, still says a vaccine is coming—though perhaps in combination with antiretrovirals, circumcision, prophylactic drugs, and antibody infusions, it doesn’t have to be 100 percent effective to stop the epidemic, even in the developing world. “We would settle for a 55, 60 percent effective vaccine,” he says.
Not everyone buys all this. “They’re doing various gradations of RV144—an extra booster, a different adjuvant,” Levy says. “And we don’t even know if RV144 can be reproduced.” His frustration is palpable. “You’re talking to someone who has been complaining about this for a long time. You get locked into one program and put all your resources into that, so anything innovative has to tie into that one direction.”
Like energy from nuclear fusion, an HIV vaccine is always 10 years away and always has been. But all these new directions and new studies have stirred excitement among people who have been working on vaccines for years. Again. “I would argue 2018 is the most optimistic we’ve ever been,” Warren says.
It wasn’t easy. First of all, Kang says, HIV isn’t easy to grow in culture, and regulators don’t like the idea of someone having a big tank of HIV. Then, once he had the virus, Kang had another obstacle: the Food and Drug Administration. “FDA says, if there is a single live virus in the vaccine, it’s one too many, right?” Kang says. As a condition of the trial, the FDA told Kang he had to show complete, total, utter killing.
So Kang killed the hell out of it. His group first genetically engineered the virus so that it couldn’t infect cells anymore but could still replicate. Then they poured on a chemical called Aldrithiol-2, a standard virus killer. And then they exposed the poisoned, mutant virus to gamma radiation to break all its genes.
When Kang went back to the FDA with his irradiated, poisoned, mutant virus, “the FDA suggested we should use HIV-positive individuals, because the main objective of a human Phase I clinical trial is safety,” he says. “So we did that.”
The study, published in 2016, looked at just 33 volunteers; of the group who actually got the vaccine, all seemed to tolerate it well. “As a side result, we could also look at immune responses,” Kang says. “If the vaccine worked properly, it should also stimulate antibody production, and that’s what we saw.” People who got the vaccine had boosted immune responses, and an increased level of the broadly neutralizing antibodies HIV researchers are so optimistic about.
Kang says he’s hoping to run Phase II trials this year—one to maximize immune responses by varying the amount of antigen and the frequency of immunization, and then, he hopes, another using HIV-negative people. (Kang’s funding comes from various federal agencies in the US and Canada and from a biotech called Sumagen.)
Dorman remains skeptical, even of this approach. “I’ve tried to explain to Dr. Kang what I thought were issues that perhaps he didn’t appreciate yet,” Dorman says. “He picked a strain of the virus because it was convenient, and a cell system that was convenient. He did a cloning procedure because he had reasons to think it might be useful. The odds that what he has gone to the clinic with will be protective are not great. That’s not a criticism of him. It’s simply an acknowledgement that we don’t know how to make these choices, and there are a lot of them to make.”
But, I say, isn’t Kang’s work at least a sign of receptiveness to your ideas? You said nobody would fund a classical approach, but here it is.
“My concern is, if and when it is shown to fail, it will discredit the concept even further. And I said that to him. Of course, that doesn’t stop him,” Dorman answers.
He got published, I say. He got funding from government agencies in two countries, and a pharma company.
Dorman insists he’s rooting for Kang, but a lot of people get funding and get published with early-stage HIV vaccines. “You can’t think your way to a vaccine. You have to experiment your way,” he says. “And that is an idea that the Tony Faucis of the world have never digested. Is that a criticism? No. They’re in a different business, and their business does a tremendous amount of good. But it doesn’t get a vaccine in short order.”
But that doesn’t mean science and scientists are always right. Those methods are iterative; the idea is to get more right, or understand where you were wrong.
Science—the institution, not the set of methods—is made of people doing the hard work of apprehending the world and trying to change it. But those people are just as subject to social forces as any other group of humans. They are vulnerable to bias, blind spots, and groupthink. That’s not an excuse to stop believing in scientific conclusions. (Vaccines prevent disease. Life on Earth develops and continues to change through a process of evolution. Human industrial emission of carbon into the atmosphere is altering the planetary ecosystem.) It just means scientists aren’t always right.
They certainly haven’t been right about an AIDS vaccine. “We’re 30 years in and we don’t have anything. RV144, that’s it,” Levy says. “Burt has tried to get independent funding, and I think he still could, but it has to be from some pretty forward-thinking philanthropists, because you’re not going to get it from a foundation and you’re not going to get it from the government.”
Dorman thinks that the vaccine community’s resistance to classical vaccinology is an example of bias, of cleaving to a set of ideas in the absence of evidence for them. “Until fairly recently, expert opinion held that the earth was flat! Erroneous ideas sometimes are difficult to dislodge, as has been noted by many observers including Leo Tolstoy and Upton Sinclair,” he writes me in an email timestamped at 3:05 am the morning after I asked him about Kang’s research. To the email Dorman attached a PDF with quotes from Tolstoy and Sinclair. The Tolstoy was in the original Russian, with a translation.
I’d suggest that the problem here isn’t bias so much as what the philosopher Thomas Kuhn called a paradigm. Scientists establish paradigms, Kuhn wrote in a book called The Structure of Scientific Revolutions, and those sets of ideas inform and guide research until another paradigm overturns them. Those paradigm shifts are hard to predict, nearly impossible to engineer, and, when they happen, tectonic. (In some cases, literally—it took decades for geologists to accept the idea of plate tectonics.)
The dominant paradigm moved away from Burt Dorman 35 years ago. That dominant paradigm has not produced a vaccine. The fact that Dorman has never gotten to try his approach is certainly a missed opportunity. And, perhaps, a tragedy.
1 UPDATE 6/1/18 10:45 AM Corrected first name