HTLV-1 infection: role of CTL efficiency

Phenotypic examination by Sabouri and colleagues of CD8⁺ T cells in HTLV-1–infected people supports genetic and immunologic evidence that an inefficient CTL response to HTLV-1 results in a high proviral load and the inflammatory disease HAM/TSP.

In addition to causing the aggressive CD4⁺ T-cell malignancy known as adult T-cell leukemia, human T-lymphotropic virus type-1 (HTLV-1) causes a chronic debilitating inflammatory disease of the central nervous system known as HTLV-1–associated myelopathy/tropical spastic paraparesis (HAM/TSP) in 0.25% to 2% of infected people. HTLV-1 isolates vary little in sequence, so host factors appear to be decisive in determining the risk of HAM/TSP. The active inflammatory lesions in the spinal cord contain mononuclear cell infiltrates, and much research has been carried out on the T-cell response to HTLV-1 and its possible role in HAM/TSP.

The cytotoxic T lymphocyte (CTL) response to HTLV-1 has been a special focus of this work. The early indications were that the CTL response to HTLV-1 was robust in patients with HAM/TSP.¹  However, a strong, chronically activated CTL response to HTLV-1 was then found in asymptomatic healthy carriers of HTLV-1 as well.²  Subsequent lines of evidence from the study of virus and host genetics, gene-expression microarrays, and ex vivo assays of CTL function have favored the hypothesis that the CTL response plays a critical part in limiting HTLV-1 replication in vivo, and that genetically determined differences in the efficiency of the CTL response account for the observed differences between infected individuals in the risk of development of HAM/TSP and in the proviral load.³  This conclusion does not exclude the possibility that activated HTLV-1–specific CTLs might also contribute to the tissue damage seen in HAM/TSP.⁴ 

In this issue of Blood, Sabouri and colleagues tested the hypothesis that variation among individuals in the efficiency of the anti–HTLV-1 CTL response is reflected in the expression of T-cell molecules involved in CTL lysis (perforin, granzyme B, and CD107) and costimulation (CD27, CD28, CD80, CD86, CD152). The results revealed a higher frequency of CD8⁺ T cells that were negative for these costimulatory molecules in patients with HAM/TSP than in age-matched uninfected controls, but there was no such difference between healthy HTLV-1 carriers and the uninfected controls. Sabouri et al also found a significantly lower frequency of perforin⁺ cells and granzyme B⁺ cells in the CD8⁺ population in HTLV-1–infected subjects than in uninfected controls, although there was no significant difference between patients with HAM/TSP and healthy carriers.

These 2 observations suggest that the CD8⁺ T cells were subjected to significantly greater antigenic stimulation in vivo in HTLV-1–infected people, especially in patients with HAM/TSP, leading to a discharge of perforin and granzyme B and to the characteristic loss of expression of costimulatory molecules that accompanies T-cell differentiation.

Next, making an important link between the phenotype and the function of HTLV-1–specific CTLs, the authors found a significant inverse correlation between the proviral load and the frequency of perforin⁺ CD8⁺ T cells in all HTLV-1–infected people. This inverse correlation was stronger and more statistically significant in people with HLA-A2, which was found in 1999 to be associated with a lower proviral load and with protection against HAM/TSP in southern Japan.⁵  Interestingly, Sabouri et al found this negative correlation to be statistically significant in HLA-A2⁺ healthy carriers alone, but not in HLA-A2⁺ HAM/TSP patients, suggesting that the class 1 HLA-restricted T-cell response to HTLV-1 is more effective in healthy carriers.

The emerging picture is that HTLV-1 infection, which was previously thought to be latent, is in fact a persistent active infection, and that the HTLV-1–specific CTL response plays a critical role in limiting the replication of HTLV-1, the proviral load, and the risk of the inflammatory disease HAM/TSP. But in this dynamic equilibrium between host and virus, the frequency and the differentiation state of virus-specific T cells are both the cause and the effect of the level of the proviral load. This dynamic complexity makes it difficult to distinguish cause and effect: for example, in patients with HAM/TSP, is the perforin content of the CTLs low because they are subject to frequent stimulation (and consequent degranulation) by the high antigen load, or conversely does the high viral load in such patients result from the low perforin content in inefficient CTLs? Or both?

The role of host genetics is perhaps the key to avoid this potential circularity. The association of a single class 1 HLA allele such as HLA-A2 with protection against HAM/TSP and with a lower proviral load⁵  strongly suggests that the CTL response to HTLV-1 is a dominant determinant of the outcome of HTLV-1 infection, not a passive follower of the proviral load: HTLV-1 infection cannot, of course, determine the host genotype.

Sabouri and colleagues then compared the lytic capacity of HTLV-1–specific CTLs between patients with HAM/TSP and healthy carriers by staining T cells for CD107. CD107 is expressed on the CTL surface when the lytic granules are discharged during target-cell lysis; CD107 staining can be used to quantify the recent killing history of CTLs. Sabouri et al found that stimulation of CD8⁺ T cells from HLA-A2⁺ subjects with the immunodominant Tax_11–19 peptide elicited lower CD107 staining in HTLV-1 antigen-specific CD8⁺ T cells from patients with HAM/TSP than in those from healthy carriers (see figure). The lower CD107 staining in CD8⁺ T cells from HAM/TSP patients observed by Sabouri et al appears at first sight to conflict with a report in this issue of Blood from Enose-Akahata et al⁶  in which CD8⁺ T cells from HAM/TSP patients gave greater staining for CD107 than those from healthy carriers. However, there was a crucial difference between the 2 studies: Enose-Akahata et al quantified CD107 expression as the fraction of CD107⁺ PBMCs, whereas Sabouri et al quantified CD107 expression as the fraction of CD107⁺ antigen-specific CD8⁺ T cells (those that bound the HLA-A2/Tax_11-19 tetramer). Putting these 2 results together, the message of Sabouri and coauthors seems clear: patients with HAM/TSP have a high frequency of HTLV-1–specific CD8⁺ T cells with poor lytic capacity, whereas healthy carriers have a lower frequency of cells with high lytic capacity.