Other immunologists, however, warn that the mystery of exactly how HIV causes immunodeficiency may be as impenetrable as ever. It is extremely difficult, they point out, to differentiate between these “naive” cells and the “memory” cells that respond to pathogens the body has encountered before. This means that the findings may be much less clear-cut than the Stanford researchers believe.

Roederer and his colleagues have been studying the immune system’s T cells. Naive T cells patrol the blood in search of the particular foreign molecule, or antigen, they are primed to recognise. The immune system generates an almost endless variety of such cells that recognise different foreign molecules – although most of the T cells never encounter their antigen. Those that do, however, trigger an immune response and turn into memory cells – so named because they remember the invader that carried the antigen. This memory can last for decades, as the cells divide through successive generations, enabling the immune system to react rapidly to the pathogen, should it invade again.

The Stanford researchers counted the naive and memory T cells in the blood of 285 people who had been infected with HIV for varying periods of time. Until now, most AIDS researchers have concentrated on the steady decline in numbers of one of the main sub-types of T cells, called CD4 cells, that occurs in HIV-positive people. But in the May issue of the Journal of Clinical Investigation (vol 95, p 2054 and 2061), Roederer and his colleagues claim that behind this general change lurks a more subtle shift in the relative proportions of naive and memory cells.

Healthy adults have equal numbers of naive and memory cells in both the CD4 sub-type and another class called CD8 cells, which kill other cells that are infected with an invading pathogen. Roederer found that naive T cells seem to decline steadily in both cell classes from the early stages of HIV infection. In patients with AIDS, the proportion of naive cells dropped to less than 30 per cent for CD4 cells, and to around 12 per cent for CD8 cells.

The loss of naive CD8 cells was especially surprising, because the total number of CD8 cells initially increases after infection with HIV, as the body battles against the virus. The new finding changes the whole picture of T cell loss in HIV infection, claims Leonore Herzenberg, who with her husband Leonard heads the lab in which Roederer works. “Measurements lumping together all types of CD8 cells failed to resolve the gross imbalance within that population,” she says.

Since the body relies primarily on naive T cells to ward off new infections, the Stanford researchers argue that the disappearance of these cells may be what eventually leaves HIV-infected people defenceless against new pathogens. “Immunodeficiency may result because the naive cells are no longer there,” says Roederer.

Other researchers are sceptical. “This study will help explain the mechanism by which immuno deficiency develops, but whether it has truly found the heart of the problem is not clear,” says Jonathan Kagan, an immunologist at the US National Institute of Allergy and Infectious Diseases in Bethesda, Maryland.

The new results are difficult to interpret, say immunologists, because naive and memory cells are difficult to tell apart. This is done by examining “marker” molecules carried on the surface of T cells which change as naive cells turn into memory cells. The Stanford researchers studied three markers, none of which individually is 100 per cent reliable: naive cells can be made to look like memory cells just by adding proteins that stimulate the cells to divide rapidly – a phenomenon called immune cell activation. Using cell surface markers to separate naive and memory cells, says Kagan, “is like defining me as a liberal because I have long hair”.

Nevertheless, if the Stanford researchers are correct, trials of new treatments for HIV and AIDS are ignoring important information. Researchers testing drugs or therapeutic vaccines in people with HIV try to predict the treatments’ effectiveness by counting their patients’ CD4 cells; if these cells increase, researchers take this as a sign that a treatment may be working.

Roederer says people running clinical trials should routinely monitor their patients’ naive cells. This could be vital for interpreting the results of trials of therapeutic vaccines. To be effective the vaccines might need a certain number of naive cells. If too few naive cells are present, the immune response provoked by a therapeutic vaccine may be too inflexible to cope with HIV’s tendency to mutate. This means that even if the overall results of a trial suggest that a therapeutic vaccine is ineffective, this could be masking the possibility that the vaccine works well in a fraction of patients with higher numbers of naive cells.

MS is just one of many debilitating diseases caused by auto-immune reactions, where the immune system attacks the body’s own tissues. It afflicts about a million people worldwide, two-thirds of them women. People usually first notice the disease’s symptoms in their twenties when immune cells begin to attack the protective sheath wrapped around nerve fibres in the brain and spinal cord. People with MS may eventually lose control over their muscles after repeated cycles of remission and relapse.

Researchers have long speculated that viruses and other infectious agents could trigger MS. Many pathogens are thought to evade the immune system using a cloaking strategy called molecular mimicry. The immune system does not normally attack “self” proteins, so by producing protein fragments, or peptides, that imitate the structure of those in the body’s own tissues, viruses and bacteria may avoid destruction. But about ten years ago, biologists pointed out that this strategy could go awry with disastrous consequences: if the immune system is no longer fooled by a pathogen’s disguise, it could begin to attack both the mimicking pathogen and whichever of the body’s tissues it is imitating.

Proving this hypothesis has been very difficult, as singling out the cloaking peptides among the many thousands of peptides produced by a typical virus or bacterium is a time-consuming and thankless task. The Harvard researchers began by analysing the three-dimensional structure of the small region of the nerve cell sheath protein that is targeted by immune cells in MS patients. Because proteins need not have exactly the same sequence of constituent amino acids to fold up into roughly the same shape, the researchers then had to work out the range of sequences that would mimic the structure of this human peptide.

Once they had done this, Wucherpfennig and Strominger searched through two protein sequence databases and found 600 viral and bacterial sequences that would imitate the key nerve cell sheath peptide. After discarding those from microorganisms that do not infect people, or which only occur in the tropics, where MS is uncommon, the researchers were left with a panel of 129 microbial peptides. They then added these, one by one, to cultures of self-reactive immune system T cells taken from MS patients. These are the cells which attack nerve cell sheaths.

Most of the suspect peptides had no effect on the cultured cells, but seven viral peptides and one peptide from a bacterium made the T cells start dividing. This jump-started cell division is called immune cell activation and is an important component of both normal immune responses and autoimmune reactions.

Wucherpfennig and Strominger suggest that MS begins when a viral or bacterial peptide activates some of the potentially self-reactive T cells that are constantly circulating in the body, but which are usually kept in a suppressed, inactive state. Once activated, the researchers argue, these cells can breach the blood-brain barrier that normally keeps immune cells away from the nervous system. They will then attract the nerve sheath peptide resembling the microbial peptide that originally activated them.

Although researchers have long suspected that viruses or bacteria may cause MS, the identity of the infectious agents implicated by the new study is surprising. MS is a relatively rare disease, but the peptides identified by Wucherpfennig and Strominger come from viruses and a bacterium that are very common. The viruses include the influenza virus, the cold sore virus herpes simplex, and two types of common cold virus called adenoviruses and reoviruses. The Harvard researchers have also implicated two cancer-causing viruses: human papilloma virus, which can cause cervical cancer, and Epstein- Barr virus, which infects about 95 per cent of people and has been linked to several types of tumour. The bacterial peptide found to activate MS patients’ T cells came from Pseudomonas aeruginosa, a common inhabitant of human skin which can infect wounds.

Clearly, simply coming down with flu or a herpes infection does not automatically make a person develop MS, so other factors must also be involved. One such factor may be a genetic predisposition to the disease. Scientists already know that the possession of certain immune system genes makes people more likely to develop MS, and it could be that these genes make people prone to autoimmune reactions when faced with microbial molecular mimicry.

Strominger stresses that more work needs to be done to prove that viruses and bacteria trigger MS, and to understand how these infectious agents interact with genetic factors. Nevertheless, he believes that the new research may eventually lead to strategies for preventing MS. “In cases where multiple sclerosis runs in families, one could identify children who have inherited high-risk genes, and give them preventive vaccines,” he says.