[Editor's note: We've updated this story, originally published in Jan 2017, to reflect a newly published study.]
There’s a small collection of test tubes stored in a laboratory freezer at Fred Hutchinson Cancer Research Center. Hundreds of these freezers host thousands of tiny tubes, but this collection is different.
This one houses a multitude of possibilities.
In each tube, suspended in a few drops of water, is a collection of millions of mutant HIV viruses. Some of these mutants are very good at infecting human cells, no different from their “natural” viral ancestor. Some of them are worse.
Together, they can teach researchers something important about HIV in humans, say the collection’s creators, Fred Hutch evolutionary biologists Hugh Haddox, Adam Dingens and Dr. Jesse Bloom.
In two recent studies, the researchers describe their library of mutant HIVs, its creation, and how they used it to show that a certain type of HIV vaccine, designed to elicit the immune molecules known as broadly neutralizing antibodies, could be on the right track.
Although other researchers in the HIV field have conducted studies of viral mutants before, these collections are what's known as “deep mutational scanning,” Bloom said. Rather than study the mutations that naturally arise in HIV or other viruses, Bloom and his research team methodically create a catalog of every possible mutation and then ask how each mutation affects the virus.
In their studies, the researchers created genetic mutations in HIV to change each single amino acid — the 20 different building blocks of proteins — to every other possible amino acid in the viral protein known as envelope, or Env. In one study, published in December 2016 in the journal PLOS Pathogens, the researchers infected human T cells in the lab with each member of the library — millions of viruses representing more than 13,000 different Env mutants — and asked which mutations affected the virus’ ability to infect and replicate inside the cells.
More recently, in a study published Thursday in the journal Cell Host & Microbe, the researchers asked how a similar mutant library fared in the presence of a single broadly neutralizing antibody, known as PGT151. Broadly neutralizing antibodies are a type of immune molecule that is often more potent than a typical antibody. As the name implies, they can potentially eliminate a broad swath of HIV strains by binding to parts of the virus that are unchanging between variants. Such antibodies do crop up naturally in some HIV-positive people, but too late to prevent their own infection.
As the protein that covers HIV’s surface, Env is the primary part of the virus that our immune system can see and act on. HIV is a particularly notorious foe for both the natural immune system and for immune reactions spurred by vaccines in part because Env mutates so rapidly, slipping away from immune recognition before the body can effectively rid itself of the infection.
“The protein probably evolves faster than any other protein that we know about,” Bloom said. “In a single infected person in one year, this protein becomes more different than the typical protein shared between a human and a chimp. The amount of evolution that separates a human and a chimp — which took at least 5 million years — in the envelope protein of HIV, that takes one year.”
In other words, HIV is a shape-shifter.
But the virus’ propensity for change is not infinite. Interspersed throughout the rapidly evolving Env protein are pockets of consistency, regions of the protein that have never or rarely been seen to mutate.
“Vaccine efforts have tried to target these essential regions that can’t change. The problem is that our understanding of what these regions are is incomplete,” said Haddox, a doctoral student working in Bloom’s laboratory and first author on the 2016 study. “We lack a thorough understanding of which sites are unable to change.”
To date, there is no HIV vaccine effective enough to bring to market. But some recent studies at the Hutch and elsewhere have focused on candidate vaccines that could elicit the elusive broadly neutralizing antibodies. In the lab, these rare antibodies can protect animal models from infection, implying that, if produced early enough, they might do the same for people.
Although vaccine researchers believed — and hoped — that the regions of Env targeted by broadly neutralizing antibodies are unchangeable because HIV can’t change them without a cost to its own survival, scientists had to consider the possibility that these regions don’t mutate simply because the virus has never needed them to. It’s possible that mutations in the antibodies’ binding regions haven’t been seen because the antibodies themselves are relatively rare.
“Before our work, it was impossible to completely rule out this alternative hypothesis,” Haddox said.
The researchers found that viruses carrying mutations in the areas of Env targeted by broadly neutralizing antibodies indeed are hobbled in the laboratory setting, less able to grow inside infected cells than other strains.
This is good news for the HIV-vaccine field, which is already in pursuit of vaccines to elicit such antibodies or, in one new study co-led by Fred Hutch President and Director Emeritus Dr. Larry Corey, directly testing whether giving intravenous infusions of one broadly neutralizing antibody can protect against HIV infection.
Their findings mean that if an HIV vaccine succeeds in eliciting broadly neutralizing antibodies in people before they are infected, there’s a good chance that HIV won’t be able to mutate away from that immune response, Bloom said.
Their data show “that the virus actually has a hard time changing [these regions],” he said. “There are no absolutes in evolution, but certainly it won’t be able to change them easily.”
Fred Hutch vaccine researcher Dr. Leo Stamatatos is working on the first Hutch-developed HIV vaccine candidate, a vaccine designed to elicit broadly neutralizing antibodies. Haddox and Bloom’s study lends credence to the approach his team is pursuing, he said, and points to potentially important avenues of vaccine design.
“The pathways of escape from broadly neutralizing antibodies that the virus can follow are not infinite … and in fact some pathways are ‘dead-end,’” Stamatatos said. “Two goals of an effective vaccine against HIV are to elicit broadly neutralizing antibodies that will prevent infection and force any virus that got through to go down a dead-end pathway of escape. This study provides evidence that the pathways of escape from broadly neutralizing antibodies are limited.”
To get at the question of how broadly neutralizing antibodies affect HIV evolution, the scientists teamed up with Fred Hutch HIV researcher Dr. Julie Overbaugh in their most recent study. Using a strain of HIV directly isolated from an infected child, the researchers made another library with every possible Env mutation and infected T cells in the presence of the broadly neutralizing antibody PGT151.
The researchers then sequenced the mutant viral strains that were able to infect cells in a petri dish in the presence of this antibody to see which mutations allowed HIV to escape PGT151.
This is important, Bloom said, for researchers to better understand the evolutionary paths HIV could take to escape a broadly neutralizing antibody — and, ultimately, to understand how to head off those paths through smarter vaccine design.
Studies like this also yield, indirectly, an inferred map of where a given antibody binds to a given virus — and the method is simpler than 3-D crystallography, the gold standard in the field of understanding how two proteins interact. The mutations that allow HIV to slip through the antibody binding and infect cells in the lab are by their nature going to be in sites important for how the viral protein connects with the antibody.
All of this falls under the rubric of better understanding HIV’s biology to continue building a foundation for better vaccine design, Bloom said. We have many working vaccines against other viruses that researchers don’t fully understand, he said, but for really tricky viruses like HIV, “we’re going to need to be more rational about how we make vaccines.”
“If we can understand where the antibodies bind, we can engineer vaccines that elicit more of those types of antibodies,” Bloom said. “And when we find really good antibodies, we can understand what parts of the virus are sites of vulnerability.”
Bloom is quick to point out that although their research and this new method could teach researchers something about how HIV behaves in nature, the study is not designed to perfectly recapitulate how HIV naturally evolves in its human host.
“There’s always this trade-off in science,” Bloom said. Either you observe what’s happening in nature, where you can see what’s happening but you don’t know why, or you perform manipulations in the lab, as his team as done, where you can control the "why" but you can’t perfectly recreate the natural setting.
“The goal of our experiment is not to exactly mimic what’s happening in nature, it’s to start to deconstruct the different forces that are responsible for evolution in nature,” Bloom said.
Rachel Tompa is a former staff writer at Fred Hutchinson Cancer Research Center. She has a Ph.D. in molecular biology from the University of California, San Francisco and a certificate in science writing from the University of California, Santa Cruz. Follow her on Twitter @Rachel_Tompa.