By Alla Katsnelson
A computer model proposes a solution to a long-standing mystery in HIV research -- why a small percentage of people infected with the virus never develop full-blown AIDS. The answer lies in how the immune cells that recognize invaders are educated, and suggests new strategies for designing an HIV vaccine.
The human immune system detects foreign cells with the help of cell-surface proteins called human leukocyte antigens (HLAs). Each person's cells carry a particular set of HLA molecules -- the person's HLA type -- which bind fragments of virus or bacterial protein and 'present' them to T cells, the immune cells that recognize and attack infected cells. But before T cells are ready to perform their killer function, they are in effect trained on fragments of the body's own proteins -- self-peptides -- in an organ called the thymus. To 'graduate' from the thymus, a T cell must be able to recognize at least one combination of HLA molecule and self-peptide, which provides the template for its subsequent immune response against a foreign peptide bound to that HLA molecule. T cells that bind to self-peptides very strongly, however, are rejected, as they would attack the body's own cells.
Researchers in Massachusetts and California began with two observations. First, HIV-infected people who manage to keep the virus in check -- so-called 'elite controllers' -- often carry a particular HLA gene variant, HLA B57. Second, people with this gene also have a higher risk of developing autoimmune diseases, in which the immune system does produce a harmful response against the body's own proteins.
Arup Chakraborty, an immunologist at the Massachusetts Institute of Technology in Cambridge, and one of the lead authors of the new study, published online in Nature today, thought the two observations might be related. He had not previously studied HIV, but he had studied how T cells are selected in the thymus by their ability to recognize specific HLA molecules and the peptides bound to them. He surmised that the HLA molecules of elite controllers might be binding a relatively small number of self-peptides.
Indeed, a look through a database of the binding properties of HLA molecules revealed that HLA B57, along with HLA B27 -- which also protects against HIV -- binds a much smaller proportion of self-peptides than HLAs that are not protective. The researchers then used a computer algorithm to predict how this would affect T-cell maturation.
T cells that develop in people with the HLA B57 gene would be presented with a smaller variety of peptides in the thymus. Their model showed these cells have broader activity and would be likely to recognize HIV even if the virus mutates, allowing the immune system of elite controllers to keep the infection under control. But that same property would also make them more likely to turn on the body's own cells, explaining why HLA B57 leads to a higher risk of developing autoimmune diseases. "If you have a smaller diversity of self-peptides in the thymus," explains Chakraborty, "there's a higher probability that T cells with a stronger reactivity and cross-reactivity" might be released.
The mechanism of protection identified by the study was a complete surprise, says Bruce Walker, director of the Ragon Institute for HIV research at Massachusetts General Hospital in Boston, who is also a lead author of the study. "I actually had to pull out text books," he says, to recall the process of T-cell selection in the thymus, which has not previously been associated with the immune response to HIV.
Testing their model on data from 1,900 HIV-infected individuals with known HLA types, 1,100 of which were elite controllers, the researchers found that the progression of the disease was strongly correlated with the number of self-peptides an HLA molecule was able to bind.
"I think it's a remarkably interesting hypothesis," says Sarah Schlesinger, an HIV vaccine researcher at the Rockefeller University in New York. It "explains clinical observations made for over a decade", she adds.
Extra-reactive T cells are more numerous in carriers of HLA B57, says Walker, but everyone has them in low numbers so it might be possible to design a vaccine that actively selects them. "What we need to do with a vaccine is train bigger T-cell armies that will be there when a person first encounters the HIV virus."
How researchers might design such a vaccine is unclear, but the study's findings are likely to be helpful, says Schlesinger. In the mid-1990s, she notes, scientists identified another mechanism that provides a degree of natural immunity to HIV -- a mutation resulting in the lack of a receptor that HIV commonly hijacks to enter a cell. "It was not entirely clear how that was going to be clinically useful," she says, but it resulted in a drug called maraviroc, made by the pharmaceutical company Pfizer, which prevents the virus from binding to this receptor and entering the cell.
"I think it's becoming more and more apparent that conventional ways of designing vaccines are not going to work for HIV," says Helen Horton, an HIV vaccine researcher at the Seattle Biomedical Research Institute in Washington state. Studying the body's own tricks for controlling infection could provide just the shot in the arm that HIV vaccine efforts need, she says. "I think this is definitely where we need to be headed."