Crowdfunded free HIV vaccine: as unlikely as it sounds
Science 17 February 14 by Liat Clark
An HIV vaccine is possible. Cynics of a decade ago have largely been beaten by clinical progress, and a decrease in the horrifying numbers -- 35.3 million people living with HIV -- now seems a prospect, not just a hope. "I think in the next five years we're going to see the first vaccines coming into clinical testing that might induce neutralising antibodies, and that's going to be incredibly exciting," Executive Director at the International AIDS Vaccine Initiative Jill Gilmour tells Wired.co.uk. This is about as optimistic a sentence as you will hear an HIV vaccine researcher utter. Yet it would be years of safety and efficacy trials later that a vaccine could eventually be widely manufactured and distributed.
It's a little surprising, then, when a video is posted online calling for the public to crowdfund a free HIV vaccine with these words: "In a few short years we can be distributing our vaccine to the millions who need it most. So pledge now to make a difference, and together let's end HIV/AIDS and save millions of lives."
The non-profit Immunity Project launched in January with backing from US seed accelerator Y Combinator. It sought $460,000 (£265,000) to carry out a trial of its vaccine in humanised mice so it could then apply to the FDA for a Phase 1 clinical trial. The campaign video talks about using machine learning to hack the DNA of "real superheroes" with "real superpowers" (so-called "elite controllers" with a natural immunity to HIV) with the aim of turning everyone into a superhero with their free vaccine. That vaccine would deliver a T-cell response (usually hijacked by the HIV virus) by prompting the immune system to target the same epitopes (protein markers on the virus surface) that are recognised and attacked by elite controllers' immune systems.
The hashtag #hackhiv launched and before you knew it Ashton Kutcher was telling his 15.7 million Twitter followers to support the cause by purchasing digital gift cards sold by Gyft, a startup backed by his VC that temporarily pledged profits to Immunity Project. The crowdfunding campaign was a huge success and hit its target with ten days to spare.
What's the problem then? Well, there are quite a few.
The vaccine creator Reid Rubsamen is an anaesthesiologist and president of a drug delivery device company. He has boundless enthusiasm for his work and a genuine desire to advance the field and make a difference. Speaking about the decision to go the startup route and launch a crowdfunding campaign, he told Wired.co.uk: "I decided I don't want to be on my death bed and say to myself, 'gee, I had the solution to this thing in my hand and we never got to test it'. I decided we have to think out of the box to move forward." The problem? He has no experience in HIV or AIDS research, nor does anyone on the Immunity Project team.
There's also been a great deal of miscommunication around the project. Take, for instance, this excerpt from a Fast Company article: "The vaccine was originally developed in a partnership between Dr Bruce Walker from Harvard, Dr. David Heckerman, inventor of the spam filter and machine learning/artificial intelligence scientist at Microsoft e-Science Research, and Dr Rubsamen." Both Walker, a renowned HIV and elite controller researcher at the Ragon Institute, and Heckerman have since had their names removed from the Immunity Project's white paper and maintain they have no connection to the project.
"It's possible that approach might be feasible, but they're a long ways from even showing they can generate immune responses"
Bruce Walker, Ragon Institute
Their imagined collaboration with Rubsamen has proven a common misconception. The two have been using machine learning to identify the epitopes targeted by elite controllers, and this work was repeatedly referred to in the original white paper as the basis for Immunity Project's vaccine.
"I found out I was listed as an advisor when I was contacted by a colleague, and immediately got in touch to say I was not and had not given consent to be one," Walker told Wired.co.uk. "I talk to a lot of people about approaches they're taking. That doesn't mean I sanction those approaches." Walker believes Rubsamen's plan is to use data from Heckerman's work (once it's published) to replicate the T-cell response. "It's possible that approach might be feasible, but they're a long way from even showing they can generate immune responses."
Heckerman did approach Rubsamen three years ago to see if the microsphere technology he was working on -- a biodegradable drug delivery system -- could be used to develop a vaccine to get an immune response from epitopes, without using a live virus vector. "We succeeded in doing this in mice using mouse epitopes and our results have been submitted to Vaccine which is reviewing our manuscript," Rubsamen tells us. The problem is, no one can see that data until it's published.
"I think they're using the fact they're seeking to publish as an excuse for not talking about their experiment, because Vaccine is not a journal that publishes major breakthroughs," UCL professor Greg Towers, a Wellcome Trust Fellow who last year published a paper identifying the "cloaking" mechanisms of the HIV virus, told Wired.co.uk.
There's another fairly big problem with taking on T-cell immune responses. This approach has already repeatedly failed.
"It's complete rubbish," Louis Picker of the Vaccine and Gene Therapy Institute at Oregon Health and Science University, told Wired.co.uk. "Many vaccines have been tried including some not too dissimilar from this, and none worked... They failed for very predictable reasons we understand now. You cannot do something in mice, get it published and say that allows you to go into clinical trial. We have two papers in Nature Medicine and are still not in clinical trials, because we're doing the due diligence." Picker says there's no AIDS model in mice, and any model in mice would be artificial and have no bearing on how the host parasite would interact in humans.
"It's sounds like a nice story to a layperson, but it's all smoke and mirrors. You can't get to an HIV vaccine with mouse studies, not even 4,000," Picker says.
Rubsamen's lab might be doing great research -- important research that at the small scale deserves to be done, either to rule certain things out or further identify markers. But the sales pitch has been totally wrong. Opening with the promise of turning the public into "superheroes" that block HIV infection is misleading. As Picker points out, elite controllers have particular HLA types that induce specific immune responses -- "but you can't make those responses happen in people that don't have those HLA types". HLAs are proteins that help the immune system tell the difference between the body's proteins and foreign ones by presenting T-cells in the thymus with tiny fragments of human protein -- it's like a training programme before the real event, so the immune cells don't wreak carnage on the body's own cells on release. T-cells respond by binding weakly with the proteins -- any that react too strongly commit cell suicide. It was Walker's team that in 2010 discovered HLA B57 is common to elite controllers. It presents a smaller diversity of protein fragments to T-cells, making them more likely to bond strongly to foreign virus and friendly proteins to kill them. Usually when HIV attacks, it hijacks the T-cells. Instead of attacking the virus and alerting B-cells to produce antibodies, the T-cells are destroyed or help the virus replicate. But T-cells in those with HLA B57 seem more likely to bind to and kill the virus.
We are only just beginning to understand the intricacies of this response, let alone start replicating and inducing it, though.
"Are these beneficial targets, [the epitopes], all you need to turn someone into a controller or not?" asks Rubsamen. "We think the answer is yes and for certain HLA types we can do that sooner."
"What they're saying doesn't make sense from an immunological point of view," Towers says. "Thirty percent of [elite controllers] have particular HLA. You might want to recapitulate that kind of protection in a vaccine, but we don't understand it well enough. Normally when you're writing a grant application you provide a detailed account of your plan so other scientists can evaluate whether it's likely to work or whether there's new information.Ó
The trouble is, there isn't really any new information. Rubsamen eventually explained all the founding data comes from other studies -- from "people in labs right now looking at targets" -- and that they do not have the key targets identified by Heckerman and repeatedly cited in the original white paper, because "he's not published yetÓ.
The literature feels littered with buzzwords -- the approach uses "computer modelling and machine learning to literally hack the HIV life cycle", Heckerman's algorithm is "based off of similar principles used in the most advanced spam filtering software" -- which distract from the fact that neither the algorithm nor the targets are anything to do with the Immunity Project's team. There are also plenty of promises that sound like foregone conclusions: "Our nasal dosage delivery mechanism will significantly decrease distribution challenges in Africa."
"I think their site overpromises from one approach and it's important to make it clear what the prospects for success and failure are," says Walker. "It was making claims about very promising results in animals and made it sound as if a vaccine had been made and was ready to move forward, and that's not the case [...] The approach is in very early stages."
Just like any crowdfunding project, the campaign was directed at laypeople, which might explain the miscommunication. It's this seemingly fact-light approach that's led some to react with suspicion, regardless of what good intentions are there.
Rubsamen put Walker's and Columbia University epidemiologist Salim Abdool Karim's requests to have their names removed from the Project down to an institutional discomfort.
"Generally this approach is foreign to some individuals and institutions that are used to being affiliated with works sponsored by government funding," Rubsamen told us. "It's a different approach. It's hard to explain what this is when it hasn't been done."
Picker would contest it has nothing to do with the institutionalisation of academics -- "if their idea would work I'd be happy to support it, it's not that the establishment wants to crush them because they have a novel idea. It's not novel, it's been thought about, tested and it's not ready for prime time. It's not even close."
However, Rubsamen's comments might be true of Abdool Karim. In December he wrote an opinion piece urging the public to a buy a Gyft card to help fight HIV, saying, "I'm pleased to be a part of a team, known as the Immunity Project."
It's unlikely Heckerman would be put off by the Immunity Project's crowdfund, however, considering he's so embedded in the Silicon Valley tech scene the Immunity Project wants to align itself with -- "It's the Silicon Valley mentality," says Rubsamen, "don't sit and complain you can't get your work done. Figure out how to do it".
When we asked the Immunity Project's director of communications Shawn Jain if we could speak with David Heckerman, still under the assumption he was involved, he replied: "Because of his affiliation with Microsoft, David is under a press embargo until his research papers publish." Later that day, we asked Jain, Y Combinator partner Sam Altman and David Heckerman, what exactly Heckerman's involvement was. Jain responded: "David has no official role with Immunity. We are only using publicly disclosed targets until his data publishes. The algorithm has been discussed in the public domain." Heckerman passed on the message that he has no involvement with the project. Altman, who told us the day before he would ask the Immunity Project team that same question, said: "I believe Heckerman has direct research involvement and still does." When we told him what we understood from Heckerman and Jain, he adjusted his answer.
Y Combinator invests a relatively small amount of money into a large number of startups each year. So it's not a disaster if one fails -- the seed accelerator is more interested in trying out new things while spreading its risk, so can afford to take a punt on what is, in reality, a remarkable new model for research fundraising.
"The idea you could do vaccine development as a startup, incredibly inexpensively relative to what big pharma does and then give it away -- that's the sort of hack we like," Altman told Wired.co.uk.
The team spent more time deliberating the Immunity Project than it usually would a tech startup -- "30 minutes instead of the usual ten". Compare that to the months it would take to pen a research paper, take it through the peer review process, then apply for grants.
"I was able to explain it to Sam [Altman] in four minutes," says Rubsamen. "They got it. This is a relatively straightforward story to tell a mathematician or computer scientist. The novel part is the IT." Y Combinator sent data from Rubsamen's unpublished paper to scientists to check it, and requested they release data from the funded trial (now ongoing) as it happens. Rubsamen wants to tell the story of his research in a transparent way, and plans to blog about it. He says he did not want to wait a year or two for traditional funding since, "we have a good idea, there's a way to test it and we don't need an insane amount of money". He believes it could be a repeatable paradigm for anyone doing small science, and wants to continue the route of the peer-review process but in a way that "rapidly disseminates data".
Picker says, "they'd get nowhere" if they'd tried to go the traditional funding route.
"I spend three quarters of my time dealing with the bureaucracy to get money and move forward in a very complicated environment. There's no question if somebody just gave me $100m I could probably do it a lot quicker -- but there's a safety net in that evaluation." It's this skipping of steps and overpromising that bothers Picker most. He's not against crowdfunding in principle, but asks "when does it become a new way to raise money versus almost a sham".
"I'm not saying this is a sham; it may well be very well intentioned. But there's a continuum there where people are going to be convinced to donate and hopes are going to be raised and who knows if people are going to give money they can afford or not -- and it's not going anywhere."
HIV and AIDS funding has suffered hugely since the recession though, with year-on-year donations down $41.1m (£24.5m) in 2011. So it makes sense to try new avenues.
But more than a decade working in the field, Picker is rightly sensitive to the community at large. "I get, from the papers we publish, heartbreaking letters all the time... from people that are desperate. They've just got HIV and they're going to kill themselves. Really heartbreaking letters from people looking for a cure; looking for a way out. In a sense, because of that experience, I see this as almost a way to exploit those people."
"The pact we have with the public to do good science and be honest, to me is very important. The thing that bothers me most about this is these people are breaking that pact. Maybe they're doing that because they believe in it. Nevertheless they're not abiding by the standards of scientific investigation most of us adhere to," Picker says, adding that the Immunity Project will get people's hopes up.
"There is absolutely no way in the world this is going to work the way they're envisioning, they're very na•ve. It's going to be another thing of 'scientist promises this and doesn't deliver'. So those of us working for years, bit by bit carefully building our case, will suffer because of bozos like this. I really have no respect for that."
Picker is adamant the approach won't work, and the most important thing is to conserve funding for approaches that have a chance, because "there's only so many phase 3 trials that can be done before the public loses faith; before the people in high incident areas have no more incentive to volunteer for studies".
You have to wonder what data the Y Combinator scientists saw to come out with such a contradictory opinion to Walker, Towers and Picker?
The probability of the vaccine succeeding matters most to these academics. But to understand how the Project got to this stage, you have to understand Altman and Rubsamen's motives. Altman is a tech entrepreneur and investor -- he believes in trying something new and doesn't mind failing lots to get there.
"If it doesn't work, it's not going to scare me from doing something similar again," he tells Wired.co.uk. "I think the idea of developing new tech in the way Y Combinator does is a good thing, so even if it doesn't work this time for medicine we should try again. Even if it has a one-in-a-thousand chance of working it's worth trying... if a company wants to vaccinate all of Africa for free [...] we want to try that." It's important to note Immunity Project's decision to give it away for free is not novel. The Bill and Melinda Gates Foundation has given more than $2.5 billion (£1.5bn) in HIV grants to worldwide organisations, including Bruce Walker's programme, and $1.4 (£834m) billion to the Global Fund to Fight AIDS, tuberculosis and malaria. Each grant comes with the proviso the organisation gives away any vaccine for free.
It's Rubsamen's drive, however, that clashes most with Towers' and Picker's.
"We need to set ourselves up to be pushed by our supporters to say, 'things are looking pretty good, why can't we move faster? What are the barriers -- financial, scientific, regulatory, public policy?' Maybe you can or can't accelerate those for a project like this. But I need to take the aggressive position about what could be. If the general consensus is it's going to take x-long to do this, that's probably how long it's going to take [...] I think it's better to say what could this be, how fast could we move [...] If you're not aggressive up front you are guaranteed to slip into a longer programme that might have been avoidable."
So whether you agree with the science, the process, the aspirations, know this: 1,463 people have pledged $462,570 on what could be. It's good to be a dreamer when you're an artist, but is it such a good idea to hinge vaccine declarations and fundraising campaigns on "could", when you're a scientist?