Andrew McMichael and Paul Klenerman*
HLA Leaves Its Footprints on HIV
The reasons for the poor control of HIV infection by the mammalian immune system are gradually being unraveled. The ability of certain HIV proteins to mutate and thus to elude immune detection is increasingly seen as crucial. On page 1439 of this issue, Moore et al. (1) provide new evidence for critical involvement of HLA proteins of the human histocompatibility complex in shaping variations in HIV proteins and possibly evolution of the virus itself.
During both the acute and chronic phases of HIV infection, production of cytotoxic T lymphocytes (CTLs) by the host immune system exerts a strong inhibitory effect on HIV growth and replication. Therefore, it is not surprising that there is strong selective pressure for survival of HIV mutants that escape the CTL response (2-7). Although escape from the host antibody response is well accepted (8), escape from CTL responses has until recently been more controversial. Objections to the idea have centered around whether a CTL response against several HIV epitopes could be undermined by escape of only one epitope. With the host immune system trying to control a swarm of rapidly replicating viruses, a viral variant that is slightly less well controlled because it carries a mutation in one of several crucial epitopes would still have a competitive edge. Often the host immune response has an unfortunate tendency to focus on a small number of "immunodominant" HIV epitopes, sometimes only one, making viral escape even easier.
Mutation in an epitope can have a number of consequences (2). If mutation affects the ability of HLA molecules expressed by CTLs to bind to the viral protein, then the epitope may simply become invisible to CTLs. In other cases, mutation may alter the interaction of the T cell receptor with the viral protein, resulting in altered recognition. This could be dealt with by the immune system selecting new, more specific T cell clones--but in practice that does not always happen (3). This phenomenon, "original antigenic sin," could reflect the fact that a lower stringency of antigenic stimulation is required for maintaining established CTL populations than for generating new ones. A failure to make primary CTL responses to new epitopes is accentuated by HIV-mediated damage to T helper cells, dendritic cells, and the architecture of the lymphoid system. Also, in some cases, HIV variants might antagonize the original T cell response (4, 5). If viral escape impairs immune control of HIV, then an increase in the number of viruses would be expected and indeed has been observed (6, 7).
Striking evidence that the ability of HIV to dodge the CTL response is important in determining the outcome of infection has come from studies in macaques infected with cloned simian immunodeficiency virus (SIV) (9, 10). Viral escape occurred rapidly at several different epitopes, all selected in response to pressure exerted by CTLs. Similar data from HIV-infected humans are beginning to accumulate (7), although it is not clear how widespread this process is. Moore et al. (1) make a timely contribution to this discussion by suggesting that HIV evasion of the CTL response is both common and important. They assessed amino acid sequence heterogeneity of the HIV pol protein in viral isolates from a large cohort of patients infected with HIV-1. The authors then analyzed the sequences according to the HLA type of the donor. They identified areas of variability in pol that were HLA class I dependent. Some of these areas map to known epitopes that engage the HLA molecules expressed by T cells; others do not, and may represent previously unidentified epitopes. Although the authors do not test whether these mutations impair CTL responses, this does seem likely.
Intriguingly, some areas of HIV pol were less variable in patients with particular HLA types. The reason for this is not clear. One possibility is that these areas are epitope regions where common HLA types have already selected a series of "optimized" mutations by passage through many infected individuals of the same HLA type. Overall, this study supports the idea that immune escape is common in HIV-infected persons and may be an integral part of the pathogenic process, and also that it is patchy, depending largely on the epitope (see the figure).
HLA to the rescue? HLA molecules on the surface of CTLs direct the CTL response to specific epitopes within the HIV pol protein. Some of these CTL responses (center) exert strong selective pressure on the virus. Viral variant epitopes that escape these CTLs by failing to bind to HLA or to interact with T cell receptors have an advantage in vivo. (Left) Other responses are less effective and are not associated with obvious escape. (Right) If an effective CTL response is directed against an epitope region that is constrained functionally or structurally, there may be no selection or slow selection because a mosaic of mutations is required for escape to occur.
If this view is correct, there should be regions of HIV proteins that for structural or functional reasons are harder to mutate without compromising virus survival. One might predict that persons with HLA types that select such epitopes might fare better when it comes to fighting infection. This seems to be the case for HLA B27 (11), which recognizes an immunodominant epitope in a conserved part of the p24 viral capsid protein. Immune escape requires at least two mutations in this region of p24 (12). HLA B57 also is associated with longer survival of HIV patients (13) and selects an epitope in this highly conserved region.
If an HLA type is common in the community, there is a good chance that an escape mutant virus could be transferred to persons of the same HLA type (14). This implies that people carrying common HLA types might do worse when infected with HIV, and that those with rare types might do better (frequency-dependent selection). Given that different populations have markedly different frequencies of HLA types, such selection could contribute to the generation of locally distinct viruses, although probably not the major clades, which diverged quite early in the history of HIV in central equatorial Africa (15).
The fact that HIV is an astounding escape artist has now been confirmed at a population level. This is worrying news for long-term immune control of HIV infection and for vaccine development. There have been clear warning signs that although CTL-inducing vaccines can offer some control of SIV infection in macaque models, such control can be undermined by these processes (16). HIV vaccines will have to match the relevant circulating virus, but must also elicit very broad responses to multiple epitopes in order to stay one step ahead of HIV variation.
- C. B. Moore et al., Science 296, 1439 (2002).
- P. Goulder et al., Immunol. Rev. 159, 17 (1997) [Medline].
- S. McAdam et al., J. Immunol. 155, 2729 (1995) [Medline].
- P. Klenerman et al., Nature 369, 403 (1994) [Medline].
- U. C. Meier et al., Science 270, 1360 (1995).
- P. J. Goulder et al., Nature Med. 3, 212 (1997) [Medline].
- P. Borrow et al., Nature Med. 3, 205 (1997) [Medline].
- P. W. Parren et al., AIDS 13 (Suppl. A), S137 (1999) [Medline].
- D. T. Evans et al., Nature Med. 5, 1270 (1999) [Medline].
- T. M. Allen et al., Nature 407, 386 (2000) [Medline].
- X. Gao et al., N. Engl. J. Med. 344, 1668 (2001) [Medline].
- A. D. Kelleher et al., J. Exp. Med. 193, 375 (2001) [Medline].
- S. A. Migueles et al., Proc. Natl. Acad. Sci. U.S.A. 97, 2709 (2000) [Medline].
- P. J. Goulder et al., Nature 412, 334 (2001) [Medline].
- B. Korber et al., Science 288, 1789 (2000).
- D. H. Barouch et al., Nature 415, 335 (2002) [Medline].
A. McMichael is at the Weatherall Institute of Molecular Medicine, Oxford OX3 9DS, UK. E-mail: firstname.lastname@example.org P. Klenerman is with the Medawar Centre for Pathogen Research, University of Oxford, Oxford OX1 IS5, UK.
Related articles in Science:
- Evidence of HIV-1 Adaptation to HLA-Restricted Immune Responses at a Population Level
Corey B. Moore, Mina John, Ian R. James, Frank T. Christiansen, Campbell S. Witt, and Simon A. Mallal
Science 2002 296: 1439-1443.
(in Research Articles)
Issue of 24 May 2002,
Copyright © 2002 by The American Association for the Advancement of Science. All rights reserved.