A Chance to Save Lives
Mosaic proteins hold the promise of becoming the first viable vaccine to protect people from the virus that causes AIDS.
Bette Korber of the Los Alamos Theoretical Biology and Biophysics group types a command on her computer and brings up a graph replete with more than a dozen sets of vertical lines.
"This is what caught everyone's attention," she says softly, pointing to a set of multicolored lines strikingly longer than the others. "It shows that in rhesus monkeys, at least, the artificial mosaic proteins elicit a significantly broader immune response than natural proteins. As a vaccine they should offer greater protection against rapidly evolving viruses like HIV."
The human immunodeficiency virus, HIV, causes a weakening of the body's immune system. It infects—and cripples—the helper T cells (CD4+ cells) that are critical to a healthy immune response, and though people can live with HIV infection for years without treatment, at some point the disease progresses to acquired immune deficiency syndrome, or AIDS, the 100 percent fatal end stage that is characterized by an immune system too compromised to stave off various opportunistic infections. HIV is also known to act synergistically with other diseases, such as tuberculosis, and may be linked to the 4-fold increase in tuberculosis in countries where HIV is rampant.
In the thirty-some years since HIV was identified, the virus has wreaked a global pandemic of fearsome proportions: 27 million dead, 33 million infected, and rates of 3 million infections and 2 million deaths per year. Put another way, if the virus had plagued only the United States, more people than currently live in Texas and New Mexico would have already died from AIDS; everyone in the coastal states of New Jersey, Maryland, Delaware, Virginia, North Carolina, and the District of Columbia would be living with HIV; all of Chicago would likely become infected this year; and the nation should prepare to bury everyone in Indianapolis, Denver, and Albuquerque by year's end. So far, the health-research community has been unable to create a widely applicable vaccine that could curtail the spread of the virus. There are many reasons for that, the main one being that HIV is one of nature's great quick-change artists—the virus mutates so fast that some version always emerges that goes unrecognized by a person's immune system.
But there's newfound hope for containing or reversing the spread of HIV and eradicating AIDS. Korber's eclectic team of immunologists, biologists, physicists, and computer programmers developed the mosaic proteins—so named because they are constructed from many small protein pieces—specifically to help the immune system fend off HIV. They just might have succeeded.
An immunologist and evolutionary biologist of international distinction, Korber is quick to mention that research on a mosaic-protein vaccine is still in its early stages, that it isn't known yet whether the vaccine will be effective in humans. Phase I trials, led by Barton Haynes of Duke University, have barely gotten underway; they will check the safety of the proposed vaccine's components and see the efficacy of the immune response in humans. If the results are promising, a vaccine comprised of three or four mosaic proteins will be tested in a large-scale human trial to assess its level of protection.
"This has been my life's work," says Korber. "While there's much more to do, many indicators suggest we're on the right track."
That a vaccine can help protect us from specific disease-causing bacteria, viruses, or other organisms (pathogens) follows from how the body's defenses work. Whenever a pathogen is detected, some immune cells mature and become activated to fight the microbe. But the large-scale maturation of cells takes time, so the initial immune response to an unknown pathogen tends to be relatively slow and weak. We're sick until the response becomes strong enough to clear the infection.
The silver lining, however, is that after that first encounter, a small cadre of activated cells remain as long-term "memory" cells. Subsequent encounters with the pathogen send a call to action to the memory cells, which then expedite the production of cells and proteins that specifically target the invader.
The vaccine's job is to initiate a strong immune response and create a strong immunologic memory against the pathogen without making a person ill. It does so by introducing into the body a nonpathogenic surrogate, typically a weakened or dead form of the pathogen itself, or a part of the pathogen, such as some of its DNA or proteins.
To date, every successful vaccine (against typhoid, polio, small pox, measles, and others) has stimulated an antibody-based immune response. Once a pathogen has been recognized and targeted, small proteins (antibodies) are created that stick selectively to any pathogens found circulating through the blood or lymph systems, inhibiting their ability to infect cells or marking them for elimination by "cellular hit men." It's no surprise, then, that many attempts at an HIV vaccine have been and still are antibody-based. So far, those vaccines have failed to demonstrate significant, long-term protection against HIV.
Korber's mosaic vaccine is different. It is designed to stimulate primarily a cellular immune response, which targets and eliminates infected cells instead of mobile pathogens. The strategy revolves around the short protein segments called T-cell epitopes that a cell uses to talk to the immune system.
Proteins are the miraculous do-everything molecules of the cell, and every cell in the body produces and destroys them as necessary. Unneeded or damaged proteins are chopped into small pieces only a few to a dozen amino acids long. (Amino acids are small molecules that link together to form a protein.) Pieces that are 8–12 amino acids long can be captured and held, like a hotdog in a bun, by a large protein known as the major histocompatibility complex (MHC). This protein will ferry the T-cell epitope to the cell's outer surface, where immune system cells can examine it. The principal examiners are white blood cells known as killer T cells (CD8+ cells), which continually roam the body looking for cancerous or infected cells.
When a pathogen, say a virus, invades a cell, it forces its host to make the virus's proteins. These also get chopped up into T-cell epitopes and displayed on the cell's surface. A mature killer T cell, set to eliminate a particular type of pathogen, inspects the segment using highly specialized surface proteins that only bind to the pathogen's T-cell epitopes. (See "Immune System Activation" on page 16.) If binding occurs—condemning evidence that the targeted pathogen has invaded—the killer T cell attacks and kills the infected cell.
The idea behind the mosaic HIV protein is that while it mimics the overall size and shape of a natural HIV protein, it's constructed from the potential T-cell epitopes that are commonly found among the different natural HIV variants in circulation about the world. Regardless of how it is chopped up, a mosaic protein will produce those recurrent T-cell epitopes and, ideally, broad immunologic memory against the most prevalent HIV variants.
From the initial awareness of HIV/AIDS in 1981, about 60 million people have been infected with HIV and more than 27 million have died. The AIDS Memorial Quilt is a poignant reminder of the toll taken by the pandemic. It consists today of more than 44,000 individual 3-by-6-foot panels sewn together in sections, the vast majority of which commemorate the life of someone who has died of AIDS. The photo shows the AIDS Memorial Quilt in Washington, DC, 1987.
Virus of a Thousand Faces
The exotic mosaic protein is needed because of HIV's extraordinary diversity. By way of explanation, Korber brings another graphic up on her computer monitor, this one of a phylogenetic tree of the main (M) group of HIV variants. These are the viruses responsible for the global pandemic.
With nine distinct subdivisions, or clades, each representing in some cases thousands of genetically distinguishable viruses, the M-group is visual evidence of HIV's diversity. The Los Alamos HIV sequence database, a global resource to assist in HIV research and analysis that is run by Korber and colleagues Carla Kuiken, Karina Yusim, and Thomas Leitner, contains more than 2600 complete HIV genomes, and well over 330,000 gene sequences. (A gene is a segment of DNA that is the blueprint for a particular protein; the protein itself is a sequence of amino acids, and different versions of a gene have different amino acid sequences. Genome refers to all of an organism's genetic material—genes plus any non-protein-producing DNA segments.)
The driving force behind HIV's diversity is the inaccurate copying of its genome; the instructions for producing newly made viruses are almost always a little different than the original. If one were to isolate the same HIV protein from two different viruses and compare them side by side, up to 35% of the amino acids would be different. This diversity lands smack on the immunes system's Achilles heel, namely that a mature killer T cell will only bind to identical, or nearly identical, T-cell epitopes (likewise for antibodies). The natural variation of HIV proteins and the resulting variation within the epitopes means that many HIV variants will go unrecognized by the immune response.
A Beautiful Idea
It was 1994 when Korber began thinking that the extraordinary diversity of the virus would likely prevent conventional vaccines from working. She envisioned creating an artificial strain of the virus, with a genome that would place it at the center of the M-group's family tree. That central virus would be genetically closer to any other virus in the tree than viruses from different clades are to each other. Activating the immune system to fend off the central virus might allow it to fend off any HIV variant within the M-group. That result might also be achieved simply by immunizing a person with a central gene, either a resurrected ancestral gene or a consensus gene, which is essentially the average gene from a large number of HIV variants.
Korber's colleague, Feng Gao at the University of Alabama, was the first to test the central-gene idea in the lab. Gao's promising initial results were seconded after Haynes at Duke and Norman Letvin at Harvard proved that a centralized vaccine could induce a monkey's immune system to recognize multiple viral variants. This was the result Korber had hoped for, that a centralized vaccine could prepare the immune system to cope with a rapidly mutating virus.
The favorable early results encouraged Korber to consider a vaccine that would transport several central genes into a cell—leading to the production of central proteins and a set of T-cell epitopes—and how one might optimize the set to maximize the immune coverage. She was funded by a Los Alamos directed research project to form a team that would develop computational methods and design a small number of artificial proteins. The team began referring to their work as the mosaic vaccine problem.
"It's relatively easy to assemble several 'generic' versions of the protein, using only the most common forms of the possible epitopes," says Will Fischer, a scientist on the mosaic team. "But because epitopes can overlap, assembling the proteins turns out to be a hard optimization problem. The choice of which 'version' to use isn't always obvious. Plus, joining the pieces has to be done carefully, to avoid creating spurious epitopes at the junctions that do not exist in nature."
The mosaic vaccine problem was tackled by a team that included Fischer, Simon Perkins (currently at Google), Tanmoy Bhattacharya, and James Theiler. Somewhat appropriately, their computer-based solution relied on genetic algorithms.
A Mosaic Set
Genetic algorithms are a class of methods often used to solve problems that are too hard, or too expensive, to solve analytically. The methods are modeled after evolution: candidate solutions (a population) are evaluated and selected based on a set of criteria (evolutionary pressures), then randomized in some way (sex), to produce new solutions (offspring). The steps are repeated until an optimal solution emerges.
Perkins designed a genetic algorithm that elegantly generates near-optimal results. Candidate proteins are generated in the computer by recombining two versions of the same natural protein (interchanging arbitrary but equal-length pieces), each plucked at random from among thousands stored in a database. The recombined proteins are evaluated to see, for example, if they contain unnatural sequences, and then scored according to how many common epitopes they contain. The top-scoring proteins are grouped into a set, which is similarly scored. The proteins are then returned to the database, and at speeds that only a computer could achieve, the process repeated until a set of proteins is generated that will produce the largest number of frequently encountered HIV T-cell epitopes.
Korber recalls that initially, the suggestion that an artificial, designed-on-a-computer protein might be useful for a vaccine was too radical for the HIV research community. "In 1995," she says, "people were unwilling to even test the concept, as they felt an artificial protein would never fold up correctly or be stable and elicit good immune responses. Feng Gao and Beatrice Hahn, his mentor at the time, changed that perception with those first experiments. Mosaic vaccines were another stretch, and a priori were deemed so far out of reach that several grants were turned down before we got funded to do the theory for this work."
Hope for a Vaccine
Aside from the Haynes studies, other Phase I studies with mosaics are going forward under the leadership of Dan Barouch (at Harvard) and Nelson Michael (in the U.S. Army). Those studies will test different strategies for delivering a mosaic-protein set.
Other intriguing HIV vaccine strategies are being explored by other groups, including vaccines designed to elicit improved antibody responses and vaccines that carry the most conserved regions of HIV. Ultimately, it may take a combination of the best approaches to create an effective HIV vaccine.
Mosaic vaccines may prove useful against other rapidly evolving viruses. Says Carla Kuiken, head of the Los Alamos Hepatitis C (HCV) database, "We adapted the algorithm developed by Bette's team to HCV. The main issue we faced was the uneven sampling—some HCV clades have only a few published sequences. But we were able to construct some HCV mosaics that look very promising. The designs are also being tested by Barouch at Harvard."
A collaboration between Korber's mosaic team and Los Alamos biophysicist Paul Fenimore resulted in a set of mosaic vaccines against the filoviruses (which include Ebola virus). The vaccines are being tested in collaboration with John Dye, an infectious disease expert in the U.S. Army. If these designs continue to do well, the mosaic method can be applied to a long list of other variable viruses against which current vaccines do not work well.
While optimistic about the results of these studies, Korber has little sense of satisfaction. "This virus destroys lives," she says. "There's no cure for it, and without constant medication, it will eventually kill you. Preventing HIV infections through a vaccine is likely to be the best way to curb its impact, if we can find a way to reach that goal."
And if a way is found to not just curb, but eradicate the virus altogether? Korber won't miss it.
— Jay Schecker
In this issue...
- Wandering Worlds
THE MYSTERIOUS PLANETARY SYSTEMS AROUND OTHER STARS
- Secure Communication Now and Forever
QUANTUM ENCRYPTION FOR THE CONSUMER
- A Chance to Save Lives
A NEW VACCINE STRATEGY TO PROTECT AGAINST HIV/AIDS
- Global Security
THE GROWING CHALLENGE
BOUNDING THE OIL SPILL
DO THE TIME WARP
WARMING OCEANS, SHRINKING ICE