Abstract
To study whether an expansion of HIV-1-specific CTL is contributing to the skewed TCR repertoire in HIV-1-infection, we characterized the TCR usage of CTL clones specific for a conserved epitope in HIV-1 reverse transcriptase (RT/476-484). CTL clones from three HIV-1-infected patients displayed highly similar TCR usage and used the identical Vβ6.1 and Vα2.5 gene segments. CTL clones from two patients showed a very high degree of similarity within the TCR complementarity-determining region-3 (CDR-3). In accordance with the similar molecular structure, all three CTL clones also exhibited a similar functional activity with regard to recognition of variant peptides and cytokine secretion pattern. In one subject clonal expansion of a single CTL specificity could be shown over a 10-mo period. TCR spectratyping of PBMC from two patients revealed a marked expansion of CDR-3 segments of a certain length within the Vβ6-family. Sequence analysis of these CDR-3 yielded sequences identical to the RT/476-484-specific CTL previously isolated from the same patients. This analysis demonstrates that clonal expansion of HIV-1-specific CTL is contributing to the skewed TCR repertoire in HIV-1-infected patients.
The HLA-class I restricted CD8+ CTL play an important role in controlling HIV-1 replication in HIV-1-infected individuals. A strong CTL response has been associated with the suppression of HIV-1 viremia at primary HIV-1 infection (1, 2, 3, 4) and with a more benign course of HIV-1 infection (5, 6, 7, 8, 9, 10).
Specific CTL recognize, with their TCR, HIV-1 derived peptides bound to HLA-class I molecules at the surface of infected cells (11). During intrathymic TCR rearrangement, random N and P nucleotides are added and/or deleted to the D-J and V-D junctions, increasing further the combinatorial diversity created by the recombination of V, D, and J gene segments. The Ag specificity primarily seems to be focused to the complementarity-determining region-3 (CDR-3)4 comprising the V-(Dβ)-J junctional sequence.
Several reports have suggested that HIV-1 infection leads to a skewed T cell repertoire affecting both the CD4+ and CD8+ T cell pools and that these alteration may contribute to the pathogenesis of AIDS (12, 13, 14). In some patients with primary HIV-1 infection, a strong oligoclonal expansion of CD8+ cells has been reported that was associated with a worse prognosis than in patients with less pronounced but polyclonal T cell expansions (3). With progressive HIV-1 infection, a growing perturbation of the TCR repertoire is observed (12, 15). It has been suggested that HIV-1 encoded superantigens might be responsible both for the expansion and for the deletion of T cells with the usage of certain Vβ family-specific gene segments (16). Due to the unprecise recombination and random insertion of N segments, the length of the CDR-3 region can differ by up to 22 aa between various TCR. The analysis of the CDR-3 lengths by TCR spectratyping of peripheral T cells displays a Gaussian-like distribution for each of the 22 human TCR Vβ families in healthy subjects (17). In contrast, studies in HIV-1-infected patients revealed grossly disrupted CDR-3 distribution patterns (12, 15). CD4+ cell perturbations have been reported predominantly in advanced HIV-1 infection, whereas CD8+ cell perturbations were found in all stages of HIV-1 infection and were unique for the individual patients, arguing against a Vβ-directed bias of the CD8+ T cell repertoire (15). It has been suggested that the skewed TCR repertoire is due to the expansion of HIV-1-specific CD8+ T cells (18), and a recent study demonstrated in several HIV-1-infected patients the expansion of certain TCR-BV families due to clonal expansion of HIV-1-specific CTL clones using double staining with MHC tetramers and TCR-BV family-specific Abs (19).
In our study, we characterized the TCR repertoire against a conserved HIV-1-reverse transcriptase (RT) epitope. We found a striking homology of the CDR-3 regions of CTL clones from three HIV-1-infected patients, which correlated to a high degree to similar functional characteristics such as recognition of viral variants and cytokine production pattern. In addition, using TCR-CDR3 spectratyping we could provide further evidence that clonally expanded HIV-1-specific CTL contribute to the skewed TCR repertoire in HIV-1-infected patients.
Materials and Methods
Subjects
The subjects were three asymptomatic HIV-1-infected patients. Subject HA was treated with a combination of zidovodine and zalcitabine. His CD4 counts ranged from 286/μl to 585/μl during the study period, and his plasma HIV-1-p24-Ag levels in serum were negative. Patient RI still was without treatment and had experienced a recent drop of his CD4 counts from 593/μl 3 mo before the study to 344/μl at the time when the clone was isolated. He exhibited a plasma viremia of 11,000 RNA copies/ml. Patient MA was without antiretroviral treatment and had normal CD4 counts ranging from 505/μl to 592/μl; HIV-1-p24 Ag levels in serum were negative. The subjects gave written informed consent, and the study was approved by the medical faculty’s Human Studies Committee.
Cell lines and culture media
EBV-transformed B lymphoblastoid cell lines (B-LCL) were generated and maintained in R20 medium, consisting of RPMI 1640 medium containing 20% (v/v) heat-inactivated FCS and supplemented with l-glutamine, penicillin (50 U/ml), streptomycin (50 μg/ml), and HEPES (10 mM), as described previously (20). All media and supplements were obtained from Life Technologies (Paisley, U.K.). CTL clones were cultured in R10 (same composition as R20 except with only 10% FCS) supplemented with various concentrations of IL-2 (Eurocetus, Ratingen, Germany).
HLA typing
Synthetic peptides
Synthetic HIV-1 peptides were synthesized by Quality Controlled Biochemicals (Hopkington, MA) as C-terminal carboxamides. The sequences in single amino acid letter codes of the peptides used in the study were: RT/476-484: ILKEPVHGV, RT/476-484-E/D: ILKDPVHGV, RT/476-484-G/E: ILKEPVHEV. Lyophilized peptides were reconstituted at 2 mg/ml in sterile distilled water with 10% DMSO with 1 mM DTT.
Isolation of HIV-1-RT-specific CTL clones
A total of 5 × 106 PBMC obtained by Ficoll-Hypaque density gradient centrifugation (Pharmacia, Uppsala, Sweden) were cultured in 1 ml R10 medium supplemented with 10 U/ml IL-2 in a 24-well plate. Peptide RT/476-484 was added directly to the culture at a final concentration of 2 μg/ml. Every 3–4 days, the medium was partially exchanged and the IL-2 concentration was increased to 100 U/ml. After 2 wk, outgrowing cells were tested for specific recognition of RT/476-484 pulsed B-LCL in a standard chromium release assay. RT/476-484-specific CTL lines were restimulated every 2–3 wk with peptide-pulsed irradiated B-LCL with or without irradiated allogeneic feeder cells.
Flow cytometric analysis
Phenotypic analysis of CTL clones was performed with a FACS (EPICS-XL-MCL, Coulter, Hialleah, FL) using fluoresceine-conjugated anti-CD8 mAb and PE-conjugated anti-CD4 mAb or similarly labeled control mAb (Immunotech, Luminy, France) according to the manufacturer’s instructions.
Cytotoxicity assay
B-LCL were sensitized with synthetic peptides as described and tested in a 4-h chromium release assay (6). Spontaneous release was <30% of maximum release unless otherwise noted. For peptide titrations, chromium-labeled target cells were incubated with serial log dilutions of peptides in a 96-well plate for 1 h before adding effector cells.
Analysis of Ag-induced cytokine secretion
A total of 2.5 × 105 CTL, cultured in 1 ml R10 in a 24-well plate, were stimulated either by the addition of 1 × 106 peptide-pulsed autologous or HLA-matched B-LCL or by the addition of soluble peptide (final concentration, 2 μg/ml). B-LCL were incubated with peptides at a concentration of 2 μg/ml for 1 h, washed two times with R10, and then irradiated with 100 Gy before addition to the culture. After 48 h, the supernatants were harvested and the contents of cytokines were determined by ELISA. Following ELISA, IL-4 (PharMingen, San Diego, CA), RANTES, and IFN-γ (R&D Systems, Wiesbaden, Germany) were used.
RNA extraction
Total RNA was purified by the Trizol technique according to the manufacturer’s instruction (Life Technologies, Paisley, U.K.). Briefly, 5–10 × 106 PBMC or CTL were pelleted and lysed in 1 ml Trizol and 0.2 ml chloroform (HCCl3). The aqueous phase was then precipitated with isopropanol and resuspended in RNase-free water. The cDNA synthesis was performed using Moloney murine leukemia virus RT and random hexamer oligonucleotides (Stratagene, Heidelberg, Germany).
Amplification of TCR-α and TCR-β sequences by PCR
PCR was performed in a 10-μl volume with 120 μM dNTPs each, 0.5 μM primer each, 3 mM MgCl2, and 0.8 U Taq using a hot air rapid cycler (Idaho Technologies, Birfeld, Switzerland). PCR conditions were 35 cycles, denaturation at 94°C for 1 s, annealing at 60°C for 1 s, and elongation at 72°C for 20 s.
PCR amplification was performed as published by Choi and coworkers (21) using a panel of 22 oligonucleotides specific for Vβ families as the upstream primer. A 5′-Hex (6-carboxy-2′,4′,7′,4,7-hexachlorofluorescein)-labeled oligonucleotide (5′-HEX-CTC TGC TTC TGA TGG CTC AAA CAC-3′) served as the downstream primer.
PCR amplification, cloning, and sequencing of rearranged TCR-α and TCR-β transcripts
The cDNA was amplified in two stages using 28 TCR Vα region consensus primers (22) and nested C region primers. For stage I PCR (15 cycles), 0.5 μl cDNA was amplified using 500 nM of 5′Vα primer each and 500 nM Cα external primer. The stage II amplification (25 cycles) was performed with 0.5 μl of stage I reaction product using 500 nM of internal Cα primer (23).
TCR-β and TCR-α PCR products were size fractionated and excised from 4% Metaphor Agarose gel (FMC BioProducts, Oldendorf, Germany) and purified using a Qiaex gel extraction kit (Qiagen, Hilden, Germany). PCR products were sequenced directly or cloned with a pGEM vector System (Promega, Mannheim, Germany). Sequencing in both directions was performed using a Prism DyeDeoxy Terminator cycle sequencing kit and a 373A DNA sequencer (Applied Biosystems, Weiterstadt, Germany). Only the functional TCR sequences are shown in the results section. The identities of TCR gene-family sequences obtained here were established by comparison with published sequences.
Analysis of CDR-3 length polymorphism by TCR spectratyping
For this analysis, 1 μl of fluorescent PCR product was mixed with 3 μl of formamide, heat denatured at 95°C for 2 min, and chilled on ice. Samples were loaded on a prewarmed 5% sequencing gel and run on an automated DNA sequencer Type 373 at 30 W constant power for 6 h using Genescan Software (Applied Biosystems).
Results
Isolation of RT/476-484-specific CTL clones
To analyze the TCR repertoire against the highly conserved HLA-A2-restricted RT epitope ILKEPVHGV, we generated RT/476-484-specific CD8+ CTL lines from three different HIV-1-infected HLA-A2-positive patients by stimulating PBMC with peptide in the presence of IL-2. We obtained vigorously growing RT/476-484-specific HLA-A2-restricted CTL lines, which could be purified to clonality by repeated restimulation with peptide-pulsed autologous B-LCL. In subject HA, we could obtain RT/476-484-specific CTL clones at four different time points within a period of 10 mo.
RT/476-484-specific CTL clones from all three patients use TCR-Vα2.5 and Vβ6.1 gene segments
To study the TCR repertoire of RT/476-484-specific CTL, RNA of the CTL clones was extracted and reverse transcribed using an oligo(dT) primer. TCR gene utilization was analyzed by PCR of cDNA using primers specific for TCR V gene families. In all CTL clones from the three patients, a single PCR product could be demonstrated for α- as well as for β-chains on agarose gels indicating the clonality of the CTL. The RT/476-484-specific CTL lines from all three patients used the same TCR Vα2.5 and TCR Vβ6.1 gene segment. A representative gel is shown in Fig. 1.
Analysis of TCR usage of RT/476-484-specific CTL clone from patient HA. A single dominant PCR product could be detected by agarose gel electrophoresis both for the Vα-chain (Vα2.5; A) as well as for the β-chain (Vβ6; B), demonstrating clonality of the CTL clone.
Analysis of TCR usage of RT/476-484-specific CTL clone from patient HA. A single dominant PCR product could be detected by agarose gel electrophoresis both for the Vα-chain (Vα2.5; A) as well as for the β-chain (Vβ6; B), demonstrating clonality of the CTL clone.
Sequencing of the TCR-α- and TCR-β-chains of the RT/476-484-specific CTL clones revealed a high sequence similarity also within CDR-3 (Fig. 2). This was especially true for CTL clones from patient RI and MA, which demonstrated a high similarity with only a difference of 2 aa within the TCR-α-chain and of three amino acids in the TCR-β-chain. All differences resulted from inserted N nucleotides in the V-J junction (α-chain) or V-D junction (β-chain). Interestingly, the clones from both patients displayed the identical D-J junction of the β-chain without N-region diversification.
Sequence analysis of the CDR-3 of the TCR-α- and TCR-β-chains of RT/476-484-specific CTL. Sequences were derived after amplification of CTL clones with a panel of TCR Vα and Vβ primers and direct sequencing of PCR products as described in Material and Methods. Shown are the nucleotide sequences (A) and the amino acid sequences (B) from the TCR-α- and TCR-β-chains from the three patients HA, MA, and RI. B, The amino acids differing between MA and RI are displayed in bold face letters. These sequence data are available from EMBL/GenBank/DDBJ under the following accession numbers: TCR α-chains HA (AJ238414), MA (AJ238415), RI (AJ238416); TCR β-chains HA (AJ238417), MA (AJ238418), RI (AJ238419).
Sequence analysis of the CDR-3 of the TCR-α- and TCR-β-chains of RT/476-484-specific CTL. Sequences were derived after amplification of CTL clones with a panel of TCR Vα and Vβ primers and direct sequencing of PCR products as described in Material and Methods. Shown are the nucleotide sequences (A) and the amino acid sequences (B) from the TCR-α- and TCR-β-chains from the three patients HA, MA, and RI. B, The amino acids differing between MA and RI are displayed in bold face letters. These sequence data are available from EMBL/GenBank/DDBJ under the following accession numbers: TCR α-chains HA (AJ238414), MA (AJ238415), RI (AJ238416); TCR β-chains HA (AJ238417), MA (AJ238418), RI (AJ238419).
Biased TCR repertoire due to the expansion of RT/476-484-specific CTL clones
The TCR repertoire was studied by measuring the length distribution of the TCR CDR-3 from peripheral T cells by spectratyping. For this purpose, RNA from PBMC was isolated and transcribed to cDNA using an oligo(dT) primer. Then, the CDR-3 of the TCR-β-chains were amplified by RT-PCR using a panel of 22 Vβ family-specific primers and a fluorescence-labeled Cβ primer and separated on a high-resolution gel according to the CDR-3 length. In patient RI, the Vβ6 family showed a clear expansion of TCR of a certain CDR-3 length, whereas other Vβ families such as Vβ7, Vβ5.1, or Vβ5.2 showed a polyclonal distribution or only minor expansions (Fig. 3). In contrast, patient HA showed a grossly distorted CDR-3 length distribution, suggesting a nearly monoclonal T cell expansion within several TCR Vβ families such as Vβ1, Vβ12, Vβ13, Vβ15, Vβ16, and Vβ20 (Fig. 3). To test whether the CDR-3 of the Vβ6.1-positive RT/476-484-specific CTL clone from patients RI and HA had the same size as the clonally expanded CDR-3 within the Vβ6 family in peripheral T cells, cDNA from the CTL clones were mixed to the cDNA obtained from autologous peripheral T cells and amplified using Vβ6-specific primers. This resulted in additional increases just of the expanded CDR-3 peaks within the Vβ6 family (shown in Fig. 4), demonstrating that the lengths of the CDR-3 of the CTL clones were identical to the lengths of the CDR-3 of the expanded Vβ6-positive peripheral T cells. To confirm the identity of the TCR, the expanded peak within the Vβ6 family derived from PBMC was cut out from the gel, cloned, and sequenced. In both patients, three of five clones showed an identical sequence as the CDR-3 of the corresponding RT/476-484-specific CTL clones, demonstrating that the expansion of T cells with Vβ6-containing TCR mainly was due to a clonal expansion of the RT/476-484-specific CTL clone. In the case of patient HA, direct sequencing of the TCR Vβ6 CDR-3 PCR product confirmed the monoclonal T cell expansion and the identity of the TCR of the in vivo expanded T cells with the RT/476-484-specific CTL clone obtained from this patient several months earlier.
Analysis of junctional diversity of TCR from PBMC by “spectratyping.” In the PBMC of patient RI, the Vβ6 family showed an oligoclonal TCR expansion, whereas other Vβ families showed a polyclonal distribution or only minor expansions. In the PBMC of patient HA, grossly disrupted CDR-3 length distributions could be observed in several Vβ families including Vβ6, suggesting a strong oligoclonal expansion of T cells.
Analysis of junctional diversity of TCR from PBMC by “spectratyping.” In the PBMC of patient RI, the Vβ6 family showed an oligoclonal TCR expansion, whereas other Vβ families showed a polyclonal distribution or only minor expansions. In the PBMC of patient HA, grossly disrupted CDR-3 length distributions could be observed in several Vβ families including Vβ6, suggesting a strong oligoclonal expansion of T cells.
Analysis of CDR-3 length of TCR from HIV-1-specific CTL. TCR CDR-3 length analysis of the RT/476-484-specific autologous CTL clones revealed a single peak within the Vβ6 family (clone). In addition, cDNA from RT/476-484-specific autologous CTL clones was mixed before PCR to cDNA obtained from autologous PBMC. Amplification using Vβ6-specific primers resulted in additional increases just of the expanded CDR-3 peaks both for patients RI and HA (clone + PBMC).
Analysis of CDR-3 length of TCR from HIV-1-specific CTL. TCR CDR-3 length analysis of the RT/476-484-specific autologous CTL clones revealed a single peak within the Vβ6 family (clone). In addition, cDNA from RT/476-484-specific autologous CTL clones was mixed before PCR to cDNA obtained from autologous PBMC. Amplification using Vβ6-specific primers resulted in additional increases just of the expanded CDR-3 peaks both for patients RI and HA (clone + PBMC).
PBMC were not available from patient MA from the time points when the CTL clones were generated. One year later, the Vβ6.1-positive RT/476-484-specific CTL clone could not be detected any longer despite intensive efforts. TCR spectratyping, performed with PBMC from these later time points, showed a normal distribution within the Vβ6 family, paralleling the presumed loss of this CTL clone (data not shown).
RT/476-484 peptide-specific CTL clones from three different donors exhibit a similar recognition of variant peptides
As the TCR of the three RT/476-484-specific CTL clones showed striking sequence homologies, we wanted to test whether these structural similarities were reflected by similar functional characteristics. In peptide titration experiments, the CTL clones showed an equivalent recognition of the RT/476-484 peptide, which still sensitized target cells for lysis at concentrations down to 1 ng/ml (Fig. 5). The RT/476-484-epitope is a highly conserved epitope with only a few mutations in the Los Alamos data base (24). In the HIV-1 clade A, an aspartate is found at amino acid position 4 instead of the glutamate. Of the clade B viruses listed in the Los Alamos data base, only HIV-1 SF2 shows a mutation within the epitope, with a glycine to glutamate substitution at amino acid position 8. All three CTL clones showed an identical pattern of recognition of the variant peptides corresponding to these viral variants. They all recognized the nonconservative glycine to glutamate substitution at position 8 as well as the wild-type peptide. In contrast, the conservative glutamate to aspartate substitution at position 4 fully abrogated recognition by all three CTL clones (Fig. 5).
Peptide titrations of the RT/476-484 peptide and the RT/476-484-G/E variant. RT/476-484-specific CTL clones from the three patients were tested in a 4-h chromium release assay at an E:T ratio of 10:1 against autologous or HLA-A2-matched B cell lines sensitized with serial dilutions of peptides. The RT/476-484-E/D variant was used only at the highest concentration of 100 μg/ml.
Peptide titrations of the RT/476-484 peptide and the RT/476-484-G/E variant. RT/476-484-specific CTL clones from the three patients were tested in a 4-h chromium release assay at an E:T ratio of 10:1 against autologous or HLA-A2-matched B cell lines sensitized with serial dilutions of peptides. The RT/476-484-E/D variant was used only at the highest concentration of 100 μg/ml.
Analysis of cytokine production shows a similar TC1 cytokine profile in CTL clones from three different donors
After in vitro stimulation for 48 h with peptide RT/476-484 or with peptide-pulsed HLA-A2-positive B-LCL, all clones showed good Ag-specific production of IFN-γ and RANTES (Fig. 6), and all failed to secrete IL-4 into the culture supernatant (with a lower limit of detection of 100 pg/ml; data not shown). In all clones, cytokine secretion could be stimulated both by the wild-type peptide and the variant peptide with the nonconservative glycine to glutamate substitution at position 8, but not by the conservative glutamate to aspartate substitution at position 4 or by a control peptide (Fig. 6).
A, IFN-γ secretion in vitro of RT/476-484-specific CTL clones from the three patients stimulated for 48 h with irradiated autologous or HLA-A2-matched B cell lines sensitized with RT/476-484 wild-type peptide, variant peptides, or the control peptide (VIYQYMDDL). B, RANTES secretion in vitro of RT/476-484-specific CTL clones from the three patients stimulated for 48 h with RT/476-484 wild-type peptide, variant peptides, or the control peptide (VIYQYMDDL). The peptides were added directly to the culture without presenting B cell lines.
A, IFN-γ secretion in vitro of RT/476-484-specific CTL clones from the three patients stimulated for 48 h with irradiated autologous or HLA-A2-matched B cell lines sensitized with RT/476-484 wild-type peptide, variant peptides, or the control peptide (VIYQYMDDL). B, RANTES secretion in vitro of RT/476-484-specific CTL clones from the three patients stimulated for 48 h with RT/476-484 wild-type peptide, variant peptides, or the control peptide (VIYQYMDDL). The peptides were added directly to the culture without presenting B cell lines.
Longitudinal clonal expansion of RT/476-484-specific CTL
In patient HA, several RT/476-484-specific CTL clones could be obtained by peptide stimulation at four different time points within a 10-mo period. CTL clone HA1 was isolated in 4/95, HA2 was isolated in 8/95, HA3 was isolated in 10/95, and HA4 was isolated in 1/96. All four clones used TCR Vα2.5 and the Vβ6.1 gene segments and revealed identical sequences of the CDR-3 of the TCR-α- and TCR-β-chains. Parallel to the shared sequences of the TCR, all four clones showed identical functional characteristics with regard to recognition of viral variants. All clones efficiently lysed target cells sensitized with the index peptide RT/476-484 and the variant peptide with a glycine to glutamate mutation at the P8 position, but not the variant peptide with a glutamate to aspartate substitution at the P4 position. In addition, all clones showed the same pattern of Ag-induced cytokine secretion with production of IFN-γ and RANTES and lack of production of IL-4 (data not shown). These data suggest that CTL derived from a single CTL precursor can persist for prolonged periods of time and that the cytokine pattern of clonally expanded CTL seems to be preserved over time.
Discussion
Ag-specific primary and secondary cytotoxic T cell responses are central to the establishment of an antiviral immune response in HIV-1 infection. Tracking the fate of these Ag-specific CTL in vivo has been a significant technical problem, mainly because of their low frequency in infected individuals. To overcome this problem, we used CDR-3 length polymorphism analysis of peripheral blood in combination with molecular cloning of CTL responding to a defined HIV-1 peptide.
In this study, we found a restricted recognition of an highly conserved HLA-A2-restricted HIV-1-RT epitope by CTL clones derived from three different HIV-1-infected patients. All three CTL clones shared the use of Vβ6.1 and Vα2.5 gene segments, and two of them also showed a high degree of homology in the CDR-3. In accordance to the similar molecular structure of their TCR, all three CTL clones demonstrated very similar functional characteristics with regard to recognition of the index peptide and of altered peptides corresponding to viral variants published in the Los Alamos data base (24). The restricted use of TCR for recognition of the RT/476-484 epitope in our study suggests that the TCR have been selected in these chronically infected patients because of a favorable TCR-peptide avidity. All CTL clones analyzed in this study failed to recognize the conservative glutamate to aspartate substitution, which represents the most common sequence variation in this epitope (24). This is another example that, at least in highly conserved CTL epitopes, published mutations often correspond to CTL escape variants (6). Despite the conservative nature of the glutamate to aspartate exchange at position 4, this mutation fully abrogated TCR signaling both for target cell lysis and for cytokine secretion. This demonstrates that even minor sequence changes in CTL epitopes could confer CTL escape and that they may have profound consequences for the control of HIV-1 and other viruses.
A CTL clone against the RT/476-484 epitope has first been described by Walker et al. (20), but the TCR usage of that CTL clone has not yet been reported. Vessey et al. identified an RT/476-484 peptide-specific CTL clone also using a Vα2-chain (AV2S1A2), but a different Vβ-chain (BV1S1A2) (25). In a recent report by Wilson et al., HLA tetramer staining revealed the usage of Vβ3 (BV3) of RT/476-484 peptide-specific CTL (19). This demonstrates that other V-D-J combinations are able to form high-avidity TCR specific for this epitope. However, the RT/476-484 peptide-specific CTL reported by Vessey recognized only very poorly the glycine to glutamate exchange at the P8 position in the epitope (25), whereas the CTL clones in our study recognized this variant as well as the index peptide RT/476-484. This indicates that despite a similar recognition of the index peptide RT/476-484, usage of alternative TCR V-D-J combination can induce significant and important changes in the fine specificity of TCR. It remains speculative whether the differences in the TCR usage are due to selection by different viral variants or whether the TCR usage is shaped by host factors as allelic polymorphism in TCR gene segments (25) or recognition of autologous peptides by thymocytes in a different HLA context during negative selection in the thymus (26, 27).
In mice, repeated injection of defined i-Ek-restricted peptides confirmed selection for homogeneity in the CDR-3 length of responding T cells occurring before selection for some of the characteristic amino acids (28). There is only a limited number of studies in humans analyzing the TCR usage of CTL clones targeting the same viral epitope in different individuals. Studies of the TCR usage directed against an HLA-B14-restricted CTL epitope in HIV-1-gp41 showed a diverse TCR repertoire of epitope-specific CTL in different HIV-1-infected patients (22). Despite the use of different TCR gene segments, these clones had exhibited a similar ability to recognize sequence variants. Other studies on the TCR usage against viral epitopes revealed a more limited TCR repertoire. The majority of CTL clones specific for an HLA-A2-restricted CTL epitope in the influenza A matrix protein used Vβ17 (29). A conservation in the usage of Vα and Jα together with Vβ and NDNβ also was seen in CTL clones derived from three different donors with specificity for an HLA-B27-restricted CTL epitope in the influenza A matrix protein (30). A very restricted TCR usage also was found for CTL from different donors specific for an HLA-B8-restricted EBV epitope (31). However, consequent studies on the recognition of this epitope in individuals with primary infection showed the presence of the conserved TCR in the early infection but also detected the development of diverse TCR Vβ clonotypes during the course of infection (32). A report from the same group about the recognition of another HLA-B8-restricted CTL epitope located in the EBV immediate-early protein BZLF1 described a CDR-3 length restriction in all responding TCR-β-chains, but a diversity of TCR Vβ clonotypes in individuals with primary infection. Another study about TCR against HLA-A11-restricted CTL epitopes in the EBV nuclear Ag-4 revealed a highly restricted response against the subdominant epitope with conserved Vβ usage together with identical length and amino acid motifs in the CDR-3, while a broad repertoire of TCR was selected by the immunodominant epitope (33). It has been speculated by the authors of that study that these differences in the TCR repertoire could result from different levels of Ag load on APCs. The lower density of ligands could shape a homogenous TCR response against the subdominant epitope, whereas a higher peptide density of the dominant epitope would favor a broader response (33). Studies of tumor-specific CTL revealed a similar picture. Tyrosinase-specific CTL generated by peptide stimulation from healthy donors displayed a restricted usage of TCR-α and -β gene segments, but showed diversity within the CDR-3 resulting in a distinct fine specificity (34).
A recent study of Naumov et al. found after in vitro stimulation of PBMC with the influenza A matrix peptide M1 (58–66) a high polyclonality of CDR-3 within Vβ17-specific CTL in an HLA-A2-positive subject (35). As we generated the CTL clones by peptide stimulation, we cannot exclude that the patients could mount additional TCR against the RT/476-484 epitope that we failed to detect. It could be argued that the high homology of the TCR of the CTL clones in our study is the consequence of the in vitro outgrowth of RT/476-484-specific CTL that were especially well stimulated by a favorable TCR-peptide MHC interaction. However, sequencing of TCR from the expanded peaks within the Vβ6-family in PBMC of two patients revealed CDR-3 sequences identical to the CTL clones generated by peptide stimulation in vitro. These results clearly demonstrate that the peptide-stimulated CTL were not the products of methodological bias due to special culture conditions or of a primary stimulation of rare naive CTL precursors, but that they corresponded to an in vivo expanded CTL population.
So far, only very few studies have examined the course of Ag-specific TCR clonotypes over time in patients (19, 36, 37). None of these studies has examined the cytokine secretion pattern of these longitudinally expanded clonotypes. We could observe in subject HA a clonal expansion of a RT/476-484-specific CTL clonotype over a 10-mo period. In addition to identical TCR, all CTL clones generated at four different time points from this subject exhibited an identical TC1 cytokine secretion profile. Usually, the cytokine pattern of ex vivo-cultured T cell clones is very stable and could not be modified in vitro so far. The preserved cytokine profile of the RT/476-484-specific CTL within this 10-mo period suggests that it is already determined at the priming of the naive CTL at the first Ag encounter and that the memory cells preserve a stable phenotype with regard to cytokine secretion.
Expansions of CD8+ T cells with distinct TCR Vα- and Vβ-chains are a typical feature in HIV-1 infection (12), and they are already seen in primary infection (14) and in vertically infected children (38). In the study of Gorochov et al. (12) the perturbations within the TCR Vβ-repertoire of PBMC were unique for individual patients arguing against a selective TCR bias. So far, it can only be speculated whether the differences in the perturbations of the T cell repertoires in HIV-1-infected individuals result from differences of the HLA type, recognition of different T cell epitopes, usage of different TCR against immunodominant epitopes, or from differences in the magnitude of the expansion of certain TCR.
A recent study using staining with HLA tetrameric complexes in conjunction with anti-TCR Vβ-chain-specific Abs could demonstrate in three HIV-1-infected patients that chronically clonally expanded CD8+ T cells were HIV-1-specific (19) and that these clonotypes persisted for >2 years. Our study provides further evidence that clonally expanded HIV-1-specific CTL contribute to the skewed TCR repertoire in HIV-1-infected patients. The in vivo expansions of RT/478-484-specific CTL clones in two of the patients were sufficient to distort the particular Vβ gene family and were therefore accessible to molecular analysis despite marked differences in the peripheral individual T cell repertoires. Patient RI displayed a nearly polyclonal CDR-3 distribution in contrast to the oligoclonal appearance within the T cell repertoire of patient HA. These observations show that by analysis of the CDR-3 length distribution the representation of HIV-1-specific T cells with a defined TCR can be determined semiquantitatively, even from frozen samples.
This method is a very powerful tool for the longitudinal follow-up of HIV-1-specific CTL clones with a defined TCR as it is independent from in vitro culture conditions and it also can assess functional anergic CTL populations. Recently, HLA tetramer staining of peptide-specific T cells has provided important insights into the magnitude and course of Ag-specific T cell responses (19, 39, 40, 41). However, so far, this HLA tetramer technology is still limited to a few HLA alleles and, therefore, does not cover the full breadth of the TCR repertoire. The analysis of the CDR-3 length distribution can detect even minor expansions of TCR within specific TCR Vβ families and seems to be more sensitive in the detection of clonally expanded CTL clones than cytofluorometric analysis of PBMC using HLA tetramers in combination with TCR Vβ- or Vα-specific Abs. The use of molecular fingerprinting of HIV-1-specific CTL clones facilitates the monitoring of CTL and it will help to delineate the role and the fate of HIV-1-specific CTL during the course of HIV-1 infection.
Footnotes
This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 466) and the Bayerische Staatsministerium für Kultus, Erziehung, und Wissenschaft (T. Harrer) and a grant from the Bundesministerium für Bildung, Wissenschaft, Forschung, and Technologie (01GB9706 to M.H. and J.R.K.).
Abbreviations used in this paper: CDR, complementarity-determining region; RT, reverse transcriptase; B-LCL, B lymphoblastoid cell lines.