BioWorld International Correspondent

LONDON It may one day be possible to immunize people who suffer from autoimmune diseases, such as multiple sclerosis and diabetes, by giving them small fragments of protein specifically recognized by the cells of the immune system that are thought to be attacking and damaging healthy cells.

In animal models, there already has been some success with this approach. In a mouse model of multiple sclerosis, for example, giving the animals an appropriate peptide can have opposing effects, depending on how it is given. Administering the peptide with an adjuvant can induce the disease. If, however, the mice first receive the peptide on its own in solution, this appears to switch off the T cells that cause the damage, while giving the peptide with an adjuvant later has no effect.

Stephen Anderton and colleagues at the Institute of Cell, Animal and Population Biology in Edinburgh, Scotland, together with collaborators at the University of Bristol in the UK, were therefore intrigued to find that one of the peptides they were studying did not have this effect.

Anderton, who is a Medical Research Council research fellow at the institute, told BioWorld International: “We found a peptide that could cause disease in a mouse model of multiple sclerosis, but when we tried to switch off the disease by first giving the peptide in solution, it did not work. Obviously, when trying to develop good, efficacious regimens for treating human disease, it is just as important to study the times when something does not work as it is to study the times when it does, so we decided to investigate further.”

The peptide they were studying comprises the amino acids at positions 89 to 101 on a protein of the central nervous system named myelin basic protein (MBP). They embarked on a thorough examination of which T cells could recognize this part of the protein.

They report their results in a paper in Nature Immunology (advance online publication at and February 2002 print issue) titled, “Influence of a dominant cryptic epitope on autoimmune T cell tolerance.”

“We discovered that there is not just one T-cell population that can recognize this peptide, but three,” Anderton told BioWorld International.

In vivo, before a T cell can recognize and respond to an antigen, the antigen has to be taken up by an antigen-presenting cell, such as a dendritic cell. So, for example, a dendritic cell will engulf a protein and digest it, before displaying fragments of the protein on its surface, held in the clasp of molecules known as class II MHC molecules. In experimental situations, however, a peptide given straight into the bloodstream of an animal such as a mouse can be bound directly by the MHC molecules, bypassing the normal route of passage through and processing by an antigen-presenting cell.

Anderton and his colleagues found that two out of the three T-cell populations they identified that could recognize MBP(89-101) could “see” the processed MBP antigen, while the third could “see” MBP(89-101) only when this was given as a peptide, but not after processing the whole MBP.

“We found that the T cells that only respond to the peptide form are not the cells that induce disease, because they don’t see’ the peptide that results when the whole antigen is processed by antigen-presenting cells,” Anderton said. “Instead, it is the two other populations of T cells, which do see the processed form of intact MBP, that are relevant to disease. Importantly, treatment with the MBP(89-101) peptide can switch off the T cells that are not relevant to disease, but leaves the disease-causing T cells unaffected.”

The reason, Anderton said, is that a naturally occurring enzyme is present in antigen-presenting cells, which breaks MBP between amino acids 94 and 95. A second paper in the same issue of Nature Immunology, by Colin Watts and colleagues at the University of Dundee in Scotland, reports that this enzyme is asparaginyl endopeptidase. This explains, Anderton added, why T cells recognizing the MBP(92-98) sequence do not play a role in triggering disease in animals.

What are the implications of these results for those trying to develop therapies for autoimmune diseases, including multiple sclerosis? “What this means is that the immune system is not geared to respond to synthetic peptides; it is geared to recognize peptides that are generated within antigen-presenting cells by breaking down whole proteins,” Anderton said. “It is crucial to consider this process, particularly which enzymes are involved in breaking down self antigens to generate peptides that are available for T cells to see.”

Researchers, he added, will need to use peptides that mimic naturally generated peptides “and not ones that, although they contain the right sequences, also contain other sequences that preferentially stimulate T cells that would not normally be relevant to disease.”