Characterization of an HLA-C-restricted CTL response in chronic HIV infection



HIV-specific CTL play an important role in the host control of HIV infection. HIV-nef may facilitate escape of HIV-infected cells from CTL recognition by selectively downregulating the expression of HLA-A and HLA-B molecules, while surface expression of HLA-C is unaffected. The HLA-C-restricted CTL responses have previously been largely ignored and poorly characterized. We examined the frequency, function, and phenotype of HLA-C-restricted CTL in ten antiretroviral therapy-naïve Caucasian and African individuals with chronic HIV-1 infection (for at least 8 years; CD4 cell counts in the range of 50–350) who carried the HLA-Cw04 allele. HLA-Cw04-restricted CTL that recognize a conserved epitope within HIV-1 envelope (aa 375-383 SF9) were analyzed using IFN-γ ELISPOT assays and phenotypic analysis was carried out by flow cytometry. HLA-C-restricted CTL play an important role in the HIV-specific response, and can account for as much as 54% of the total response. HLA-C-restricted CTL are functionally and phenotypically identical to HLA-A- and HLA-B-restricted CTL. HLA-C-restricted CTL in chronic HIV infection are memory cells of an intermediate phenotype, characterized by high CD27 and low CD28 expression and lack of perforin production.


HIV-specific CTL are important determinants of the host control of viral replication. The emergence of virus-specific CTL is temporally linked with the reduction of plasma viremia 1 with subsequent disease progression characterized by the loss of these virus-specific CTL 2, 3. In addition, infected long-term non-progressors maintain a low viral burden in the presence of a broad HIV-specific CTL response 2–4. Despite a strong CTL response against viral antigens, untreated individuals with HIV infection eventually progress to AIDS. Disease progression is multi-factorial and may involve mechanisms that HIV has developed to evade the host immune response including selective downregulation of MHC class I molecules by HIV-nef. HIV-nef selectively downregulates HLA-A and HLA-B alleles while not affecting HLA-C expression, and allowing infected cells to escape from immune surveillance by HLA-A- and HLA-B-restricted CTL. Selective downregulation facilitates inhibition of NK cells the inhibitory receptors of which are activated by HLA-C 5, while recognition of viral epitopes presented by HLA-C is unaffected.

HLA-C-restricted CTL responses were previously underestimated due to poor serological techniques for detecting HLA-C alleles, and close linkage disequilibrium between certain HLA-C and HLA-B alleles, which led to some HLA-C-restricted responses being assigned to the linked B molecules. Molecular techniques have allowed for more specific typing of C alleles. Structurally, HLA-C molecules closely resemble HLA-A and HLA-B molecules; however, cell surface expression of HLA-C is less than 10% of the level of expression of HLA-A and HLA-B molecules 6, 7. Despite low levels of HLA-C cell surface expression, it is becoming increasingly apparent that HLA-C molecules play an important role in both the innate and the adaptive immune response to tumors, viral infections, and autoimmune disorders 8, 9. In addition, recent genome wide studies have shown that single nucleotide polymorphisms near the HLA-C gene locus are associated with a lower HIV viral set point 10.

A limited number of HLA-C-restricted HIV-1 epitopes have been described. Among these is the HLA-Cw0401-restricted epitope (aa 375–383, sequence SFNCGGEFF (SF9)), an epitope that lies within a conserved region of gp120 8, 11. The HLA-Cw4 allele is commonly expressed, with an average frequency of 0.16 and 0.12 in black and white ethnic groups, respectively 12, 13. In a cohort of Africans and Caucasians expressing the HLA-Cw0401 allele, the magnitude and phenotypic characteristics of HLA-C-restricted CTL targeting of this epitope was evaluated.

Results and discussion

High frequency of HLA-C-restricted CD8+ T cells in HIV-1 infection

HLA type, dates of enrollment, and CD4 count at the time of the study are detailed in Table 1. All participants carried the HLA-Cw04 haploptype, which was found in linkage disequilibrium with HLA-B53 in 50% of participants. Most study participants were documented to be HIV positive for more than 10 years in the absence of antiretroviral therapy. CD4 counts at study entry were generally low with the majority of respondents having CD4 counts below 200.

Table 1. Baseline CD4 count and MHC class I alleles in study donors
Patient IDHLA class IDateCD4 cell count (cells/μL)
  • a)

    a) Date of first reported HIV-positive sample.

  • b)

    b) Date indicates date of enrollment; these individuals were HIV seropositive at the time of enrollment into the cohort.


HIV-specific responses evaluated using IFN-γ ELISPOT assays with overlapping clade A 15-mer peptides showed the presence of strong responses to epitopes within HIV-nef, p15, p17, p24, env, Tat, and rev. Although strong responses were directed at epitopes within gag restricted by HLA-A and HLA-B alleles, in four of the ten patients a significant proportion of the response was directed at the HLA-Cw0401-restricted SF9 epitope. The response to the SF9 nonamer peptide represented 46.8, 53.6, 26.7, and 41.1% of total HIV-1 peptide-specific CD8+ T-cell responses in ML612, ML1307, ML1837, and NP043 respectively (Fig. 1A). The responses were consistently strong and ranged from 160 to 3346 spot forming units/million PBMC (Fig. 1B).

Figure 1.

IFN-γ ELISPOT and intracellular cytokine staining responses to overlapping peptides in gag, nef, rev, tat, and SFNCFFEFF (SF9) peptide. (A) CD8+ T-cell responses to the HLA-Cw4-restricted CTL epitope (SF9) served as the immunodominant response in four donors, ML612, NP043, ML1837, and ML1307. The response to the SF9 peptide accounted for 46.8, 53.6, 26.7, and 41.1% of total CTL responses in ML612, ML1307, ML1837, and NP043, respectively. (B) Frequency of IFN-γ ELISPOT responses to the gp120 peptide SF9 following stimulation of fresh PBMC with 10 μM peptide. (SFU: spot forming units) (C) Intracellular cytokine staining for IFN-γ in patient ML1307; 5.15% of CD8+ T cells produced IFN-γ upon stimulation with 10 μM SF9 peptide in a 6-h assay (left panel). Staining of whole blood (NP043) with HLA-Cw4 tetramer folded around the gp120 epitope SF9; greater than 4% of CD8+ T cells were HLA-C-restricted HIV-specific cells (right panel).

Tetramer analysis of PBMC available from three of the four donors with strong responses to the SF9 peptide showed that 4.15, 1.59, and 1.84% of CD8+ T cells NP043, ML 612, and ML1008, respectively, were HLA-Cw4-SF9-restricted HIV-specific cells (NP043 is shown in Fig. 1C).

The phenotype of HLA-C-restricted CTL is similar to that of HLA-A- and HLA-B-restricted T cells

Tetramer staining in combination with antibodies to cell surface differentiation markers (CD45RA/RO, CD27, and CD28) was used to determine the differentiation phenotype in HLA-Cw4-restricted CTL in three donors (NP043, ML612, and ML 1008) in whom the response to the SF9 epitope was very strong. The HLA-Cw0401-restricted response was noted to be of intermediate differentiation with CD27+CD28CD45RACD45RO+ expression (representative data are shown in Fig. 2A–C). Expression of homing markers CCR7 and CD62L was absent (data not shown). Although these HLA-C-restricted CTL showed strong cytotoxic activity in standard chromium release cytotoxicity assays, they expressed low levels of perforin, coupled with high levels of granzyme (Fig. 2D). They were highly activated with upregulated expression of CD38 and HLA-DR (Fig. 2E). The cell cycle status of these highly activated cells was examined by looking at intracellular expression of Ki67, which is a surrogate marker for T-cell turnover. HLA-C-restricted CTL had low levels of Ki67 expression, exhibiting a pattern of expression that is similar to that seen in antigen-specific memory T cells (Fig. 2E) 14.

Figure 2.

Characterization of HLA-Cw4-restricted CTL in chronic HIV infection by flow cytometry using antigen-specific tetramers and surface antibodies, with representative staining from three donors NP043, ML612, and ML 1008. (A) CD27 and CD28 expression on whole CD8+ T-cell populations. (B) CD27 and CD28 expression on tetramer-positive populations. (C) CD27 and CD45RA expression on tetramer-positive populations. (D) Expression of cytotoxic factors with gating on the HLA-Cw4 SF9 tetramer-positive population. Granzyme A and perforin expression in NP043 is shown. The data are representative of data obtained with other donors. (E) Expression of activation markers CD38 and HLA-DR by HLA-Cw4 gp120 tetramer-positive CD8+ T cells. Expression of the cell proliferation markers Ki67 and of the anti-apoptosis factor Bcl-2 is shown.

CTL play an important role in host control of HIV infection 1, 15, 16. HIV-1-nef-mediated downregulation of HLA-A and HLA-B alleles has long been recognized as an important virulence mechanism for HIV-1 17–19. HIV-infected cells expressing viral epitopes presented by HLA-C continue to be susceptible to lysis by HLA-C-restricted CTL. Prior studies using T-cell clones have shown that HLA-C-restricted CTL have equivalent cytotoxic activity relative of HLA-A- and HLA-B-restricted CTL 19, 20. There have been fewer data reported on the quantitative and qualitative characteristics of HLA-C-restricted CTL ex vivo in chronic HIV infection.

In this study, we show that HLA-C-restricted CTL are present in relatively large numbers in patients with chronic HIV infection, accounting for up to 54% of the HIV peptide-specific CTL response. These HLA-C-restricted responses are likely to play a key role in the cellular response to HIV infection in chronically infected individuals. Therefore, the limited analysis of HLA-C-restricted responses in studies that have characterized the cellular response to HIV infection may have underestimated the breath and magnitude of the response. Using tetramer analysis, we were able to show that like HLA-A- and HLA-B-restricted CTL, HLA-C-restricted CTL in HIV infection are of an intermediate memory T-cell phenotype expressing high levels of CD27 and low levels of CD45RA and CD28 21. The absence of CD28 expression on CTL implies that the cells are incapable of full activation in the absence of the costimulatory signaling. Previous studies have shown that HIV infection is characterized by the expansion of the CD8+CD28 T-cell pool that is impaired in IL-2 production 22. CD8+ T-cell anergy has been observed in HIV and other chronic infections in both humans and mice resulting in failure to induce full activation of effector function, including perforin production, upon antigenic stimulation 23. In this study, we show that HLA-C-restricted HIV-specific CTL lack expression of CD28 and express low levels of perforin. In addition, the HLA-C-restricted CTL, although they are highly activated, show low levels of Ki67 expression. These HLA-C-restricted HIV-specific cells belong to a pool of memory T cells that share the same phenotypic characteristics as HLA-A- and HLA-B-restricted HIV-specific CD8+ T cells.

Concluding remarks

Despite low cell surface expression of HLA-C, HLA-C-restricted CTL that are generated during the course of HIV infection share similar properties to HLA-A- and HLA-B-restricted CTL. In addition, in several donors the HLA-C-restricted CTL accounted for a significant proportion of the HIV-specific immune response, eliciting very strong IFN-γ responses, which were present at a high frequency as detected by tetramer staining. In view of the potential for HIV-infected cells to evade recognition by CTL restricted by class I HLA-A and HLA-B molecules 18, a full analysis of the HIV-specific CTL response should also always include an analysis of CTL specific to HLA-C-restricted epitopes. These data also show that in the study of epitopes that can serve as potential vaccine candidates or help us to develop a better understanding of immune control of HIV infection, it may be important to include epitopes that generate strong HLA-C-restricted responses which will not be affected by selective downregulation of MHC class I by HIV-nef.

Materials and methods

Study subjects

PBMC were isolated from ten antiretroviral naïve individuals with the HLA-Cw4 genotype who were identified and recruited from well-established cohorts in London, UK, and Nairobi, Kenya. HLA typing was carried out by amplification refractory mutation system PCR using sequence-specific primers as previously described 24. Clinical and baseline laboratory data including CD4 count were obtained from a database of patients within the clinical cohorts. Ethical approval for this study was obtained from the relevant institutional review boards in London and Nairobi.

HIV peptides and peptide synthesis

The 15-mer or 20-mer overlapping HIV-1 clade B env, gag p15, gag p17, nef, rev and tat, and clade A gag p24 peptides were obtained from the National Institute of Biological Standards and Control AIDS Reagent Project. All other peptides were synthesized by F-moc chemistry using a Zinnser Analytical synthesizer (Advanced Chemtech, Louisville, KY) and purity was established by HPLC.

HLA-Cw4 tetramer synthesis

cDNA was made from mRNA extracted from C1R cells which lack the expression of HLA-A and HLA-B alleles but express normal levels of HLA-Cw4 25 PCR amplification of cDNA, and subsequent expression in Escherichia coli was carried out as previously described 26. Cw4 monomers were refolded and purified as previously described 26. Purified product was biotinylated in the presence of BirA enzyme (Avidin). Extravidin-PE (Sigma) was added in a 1–4 molar ratio, and tetramers were subsequently stored at 4°C.

Interferon-γ ELISPOT assays

Peptide-specific CTL were determined using the IFN-γ ELISPOT assay as previously described elsewhere 27. Assays were carried out in duplicate; 5×104–105 PBMC were stimulated in an overnight IFN-γ ELISPOT assay with 10 μM peptide – the gp 120 peptide SF9 and pools of overlapping 15-mer peptides from HIV-1 clade B nef, p17, Tat, env, p15, and rev, and clade A p24. Responses were numerated using an automated ELISPOT plate reader (Autoimmun Diagnostika, Strasburg, Germany) and data are represented as the number of spot forming units/million PBMC. Responses greater than 100 spots/million above background were considered to be positive.

Flow cytometry

Intracellular staining assays for IFN-γ, perforin, granzyme, and Ki67 were carried out. Cell surface staining for phenotypic markers (CD45RA, CD45RO, CD27, and CD28), activation markers (CD38 and HLA-DR), and tetramer staining were performed as previously described elsewhere 21. After staining cells were washed and fixed in PBS with 4% formaldehyde at 4°C overnight prior to acquisition with a FACSCalibur flow cytometer (Becton Dickson). Data analysis was carried out using CellQuest (Becton Dickson) software.


This work was funded by The Medical Research Council, Human Immunology Unit Grant, The Elizabeth Glaser Pediatric AIDS Foundation Grant and The Rhodes Trust.

Conflict of interest: The authors declare no financial or commercial conflict of interest.