CD4⁺ T cells specific for glycoprotein B from cytomegalovirus exhibit extreme conservation of T-cell receptor usage between different individuals

The cytotoxic T-cell response has evolved primarily to control viral infection through the recognition of peptides derived from intracellular proteins. CD8⁺ cytotoxic T cells recognize peptides presented by human leukocyte antigen (HLA) class I molecules, and the repertoire of peptides recognized by CD8⁺ T cells is generally much more restricted than those that elicit a CD4⁺ T-cell response.¹  In addition, comparison of T-cell receptor usage from CD8⁺ cytotoxic T-cell clones taken from different individuals often shows marked conservation of T-cell receptor (TCR) structure, indicating a high degree of clonal selection in vivo. The precise factors that determine such immunodominance remain uncertain but include selection of high-avidity T-cell clones through competition for antigen at the surface of the antigen-presenting cell. These findings have important implications for attempts to enhance CD8⁺ T-cell cytotoxicity through vaccination or adoptive T-cell transfer

Cytotoxic CD4⁺ T cells have many phenotypic and functional features in common with CD8⁺ cells, including expression of perforin and granzyme, and are likely to play a role in the elimination of viral infection from antigen-presenting cells that express HLA class II.^2,3  At the current time, little is known as to the degree of clonal restriction within cytotoxic CD4⁺ T cells.

Cytomegalovirus (CMV) is a ubiquitous pathogen that infects the majority of the human population.⁴  The virus is never cleared from the host, and viral latency results from sustained immune control of lytic viral replication. The CD8⁺ T-cell response to CMV is among the largest that has been documented to any antigen and can exceed 40% of the CD8⁺ T-cell repertoire in elderly individuals.⁵  Less is known about the CD4⁺ cytotoxic T-cell response to CMV, although its role in both murine and human viral infection is now being investigated.^6,–8 

CD4⁺ T cells usually recognize an antigen that has been taken up by the antigen-presenting cell rather than presented directly from cytosol. Thus, virally infected cells might be unable themselves to present antigens through HLA class II molecules, which could limit the sterilizing potential of a cytotoxic CD4⁺ T-cell response. However, recent evidence has indicated that CMV glycoprotein B (gB) can be presented directly from infected cells, and gB-specific CD4⁺ T cells are able to lyse infected HLA class II⁺ targets without the need for cross-presentation.^9,10  Characterization of these T-cell responses identified the immunodominant peptide epitope DYSNTHSTRYV from gB, hereafter referred to as DYS, which is restricted through HLA-DRB*0701.¹⁰ 

Here, we have studied the T-cell response to DYS peptide in healthy individuals and show extreme conservation of T-cell receptor usage in both the TCRα and TCRβ chains. This pattern of TCR selection was directly comparable with that seen in previous studies of CD8⁺ cytotoxic T-cell immune responses and is the first example of TCR clonal conservation within CD4⁺ T-cell immune responses. Moreover, this TCR sequence is extremely homologous to that observed in cases of large granular lymphocytosis (LGL) in patients with the HLA-DRB*0701 allele¹¹  and suggests that CMV can act as an antigenic stimulus for the initiation or maintenance of the LGL expansion.

Institutional Review Board approval was granted by the South Birmingham Regional Ethics committee, and informed consent was obtained in accordance with the Declaration of Helsinki.

Human fetal fibroblasts (HFFs) were obtained from the European Collection of Animal Cell cultures and were maintained in the media described above. HFFs were used for virus propagation in vitro.

Primary human fibroblasts derived from skin biopsies were grown in Dulbecco modified Eagle medium (Invitrogen, Paisley, United Kingdom) supplemented with 2 mM glutamine, 100 IU/mL penicillin, 100 μg/mL streptomycin, and 10% fetal calf serum. Fibroblasts were activated to upregulate HLA expression by addition of 300 U/mL interferon-γ (IFN-γ; Sigma-Aldrich, Poole, United Kingdom) to culture medium 3 days prior to use as antigen-presenting cells.

Peripheral blood mononuclear cells (PBMCs) were separated from healthy, CMV-seropositive individuals by Ficoll-Hypaque centrifugation into RPMI 1640 medium (Invitrogen) supplemented with 2 mM glutamine, 100 IU/mL penicillin, 100 μg/mL streptomycin, and 10% fetal calf serum. Where specified, PBMCs were depleted of CD8⁺ T cells using anti-CD8 antibodies (BD Biosciences, Oxford, United Kingdom) and anti-phycoerythrin (anti-PE) microbeads (Miltenyi Biotech, Surrey, United Kingdom) and magnetic selection in accordance with the manufacturer's recommended protocol.

Dendritic cells (DCs) were prepared by 6-day culture of adherent PBMCs in the above medium supplemented with granulocyte-macrophage colony-stimulating factor (Peprotech, London, United Kingdom) and interleukin-4 (R&D Systems, Minneapolis, MN), and then maturation for 24 hours in 50 ng/mL tumor necrosis factor-α (Peprotech). Epstein-Barr virus (EBV)–transformed lymphoblastoid cell lines (LCLs) were prepared from PBMCs with prototype 1 strain B95.8 and were cultured in the medium described above with cyclosporin A.

CMV strains AD169 and RV798 were kind gifts from Gavin Wilkinson (Cardiff, United Kingdom) and Tom Jones (Wyett Research Institute), respectively. CMV lysate was produced from AD169-infected HFFs after 2 weeks of infection in vitro. Recombinant modified vaccinia virus Ankara (MVA) expressing gB was created as described previously. DYS peptide was synthesized by 9-fluorenylmethoxycarbonyl chemistry (Alta Bioscience, University of Birmingham, Birmingham, United Kingdom) dissolved in dimethyl sulfoxide (DMSO) in 10-mg/mL stock solutions.

CD4 T-cell clones were generated by stimulating freshly isolated PBMCs with DYS peptide (5 μg/mL) for 3 hours. Antigen-specific cells were then selected using the IFN-γ Cytokine Capture Assay (Miltenyi Biotech) in accordance with the manufacturer's protocol. Magnetically selected cells were cloned by limiting dilution at 0.3 and 3 cells per well with autologous or HLA-matched, gamma-irradiated LCL loaded with DYS peptide (10⁴/well) and allogeneic gamma-irradiated, phytohemagglutinin-treated PBMCs (10⁵/well) in RPMI supplemented with 2 mM glutamine, 100 IU/mL penicillin, 100 μg/mL streptomycin, 100 U/mL interleukin-2 (Chiron, Emeryville, CA), and 10% human AB serum (HD Supplies, Aylesbury, United Kingdom).

CD4⁺ T-cell clones and PBMCs were tested for cytotoxicity in 6-hour chromium release assays. Effector-target ratios were 2.5:1 unless otherwise stated. Targets were autologous, partially HLA matched or HLA mismatched LCLs, DCs, or IFN-γ–activated fibroblasts. Targets were pulsed with 2 μg/mL peptide for 2 hours prior to the assay (DMSO solvent was used for control targets) or pulsed overnight with CMV lysate before the assay. Virus infections with MVA-gB or MVA-pSC11 (at multiplicity of infection [MOI] of 2:1) were carried out 24 hours before assays, and infections with CMV strains AD169 or RV798 (at MOI 5:1) were for 48 hours prior to the assay. Supernatants were sampled and γ-emission quantified using a Packard Cobra gamma counter (Packard, Berks, United Kingdom). Results were expressed as percentage of specific lysis of the target cell line. For inhibition of perforin, DYS-specific clones were incubated with 100 nM concanamycin A (ConA; Sigma) for 30 minutes before the addition of target cells. The assay was then carried out as described.

CD4⁺ T-cell clones were incubated with autologous, HLA-matched or HLA-mismatched LCLs that had been preexposed for 2 hours to peptide (2 μg/mL, unless otherwise stated) or preexposed overnight to CMV lysate or infected with MVA-gB or MVA-pSC11 at an MOI of 2:1 at 24 hours prior to use. The supernatant medium was harvested after 18 hours and assayed for IFN-γ by ELISA (Endogen; Rockford, IL) in accordance with the manufacturer's recommended protocol.

RNA was extracted from DYS-specific clones (1–5 × 10⁶ cells) using the RNeasy kit (Qiagen, Crawley, United Kingdom) according to manufacturer's instructions. Complementary DNA synthesis and amplification were performed using the SMART RACE cDNA amplification kit (Clontech, Basingstoke, United Kingdom) with TCRα and TCRβ chain–specific primers. Polymerase chain reaction products were cloned into a TOPO vector (Invitrogen) and DNA sequencing was performed using a Bigdye Terminator cycle-sequencing kit (Applied Biosystems, Warrington, United Kingdom).

Antigen-specific, IFN-γ–producing cells were detected by stimulating 1 × 10⁶ PBMCs with DYS peptide (2 μg/mL) or CMV lysate for 6 hours, with the addition of 10 μg/mL Brefeldin A (Sigma-Aldrich) after 1 hour. Cells were washed and incubated with anti-CD4PC5 and anti-CD3PE (Beckman Coulter, High Wycombe, United Kingdom) or, for phenotypic analysis, anti-CD4PC5 and one of a panel of fluorescein isothiocyanate (FITC)–conjugated anti-CD45RA, anti-CD45RO, anti-CD28, anti-CD27 (all BD Biosciences, Oxford, United Kingdom), or anti-CD57 (Beckman Coulter). Cells were fixed and permeabilized using reagents from the Intraprep Kit (Beckman Coulter) according to the manufacturer's instructions before being incubated with PE or FITC-conjugated anti–IFN-γ (BD Biosciences).

For detection of intracellular perforin and granzyme B, clones were stained for surface antigens as before and then fixed by addition of 2% paraformaldehyde/phosphate-buffered saline. Anti–perforin-PE and anti–granzyme B FITC antibodies (BD Biosciences) were added to the cells in the presence of saponin buffer (0.5% saponin [Sigma], 5% fetal calf serum, phosphate-buffered saline).

For TCRβ-chain variable region gene analysis, indirect staining was performed using an anti–TRBV6–5 IgG antibody (Beckman Coulter) and a goat anti–mouse IgG FITC (DAKO Cytomation, Ely, United Kingdom) before staining with anti-CD4PC5. Cells were fixed, permeabilized, and stained with anti–IFN-γ FITC (BD Biosciences) as described. All samples were analyzed on a Coulter XL flow cytometer and later using WinMDI (version 2.8) software (Scripps Institute; http//:facs.scripps.edu/software.html).

PBMCs from 10 HLA-DRB*0701–positive donors were stimulated with DYS peptide from CMV gB or whole CMV viral lysate prepared from infected fibroblast cultures. Antigen-specific responses were detected using cytokine flow cytometry (CFC) for detection of IFN-γ production within the CD4⁺ T-cell population (Figure 1A). No IFN-γ production was detected following stimulation of HLA-DRB*0701–positive, CMV-seronegative donors or with use of mock CMV lysate generated from uninfected fibroblasts (data not shown). Furthermore, stimulation of PBMCs from HLA-DRB*0701–negative, CMV-seropositive donors with DYS peptide showed no response. All HLA-DRB*0701–positive, CMV-seropositive donors mounted strong CD4⁺ T-cell responses to CMV lysate that averaged 5.72% of total CD4⁺ T cells (range, 0.54%-13.6%; Figure 1B). However, stimulation with the DYS peptide alone also induced a strong CD4⁺ T-cell response. As much as 16% of the total CD4⁺ T-cell population was specific for this peptide, and 7 of 10 donors showed epitope-specific responses of more than 1%, indicating marked focusing of the immune response (overall mean, 3.64%; range, 0.09%-16.8%). Interestingly, 3 of the 10 donors demonstrated a greater response to DYS peptide than to whole viral lysate. It is noteworthy that these individuals were among the 4 donors with the largest CD4⁺ T-cell response to DYS peptide, and it is likely that this finding reflects limiting amounts of DYS peptide within the lysate mixture.

The magnitude of the DYS-specific immune response is the largest antigen-specific CD4⁺ T-cell response that has been detected in humans; therefore, we sought to characterize DYS-specific T cells in more detail by determining their cell surface phenotype. PBMCs from HLA-DRB*0701–positive, CMV-seropositive donors were stimulated with DYS peptide or CMV lysate using CFC and costained with antibodies specific for cell-surface markers CD45RA, CD45RO, CD28, CD27, and CD57. DYS-specific cells expressed high levels of CD45RO (mean, 86.9%; range, 66%-99%), with low expression of CD45RA (mean, 9.4%; range, 1.0%-27%) and CD27 (mean, 14.4%; range, 0.3%-73%; Figure 2A,B). Levels of CD28 expression were low for the majority of donors, but in some cases more than 50% of the responding population was CD28 positive (mean, 32%; range, 0.17%-97%). Expression of CD57, a characteristic marker of highly differentiated T cells, showed marked variation between donors (mean, 38%; range, 3%-91%). This phenotype is indicative of an effector memory T-cell subset¹²  and was similar to the phenotype of CMV lysate–specific CD4⁺ T-cell clones from the same donors (Figure 2C). Interestingly, a small percentage of the antigen-specific T cells demonstrated CD45RA expression, indicating reversion from CD45RO expression similar to that described for CD8 T cells.^13,–15  In order to assess the direct cytotoxic potential of DYS-specific T-cell clones without prestimulation, a cytotoxicity assay was performed using freshly isolated peripheral blood as the effector cell population. PBMCs were depleted of CD8⁺ T cells and used as effectors in a chromium release assay on LCL targets (Figure 2D). Significant lysis was observed against targets either pulsed with peptide or infected with MVA-gB, indicating that circulating DYS-specific cells are directly cytotoxic against virally infected target cells.

To examine the functional properties of DYS-specific CD4⁺ T cells in more detail, we established DYS-specific T-cell clones from 5 donors using the IFN-γ–cytokine secretion assay. All clones showed specific recognition of peptide-pulsed LCL but were also able to recognize processed antigen when target cells were loaded with CMV lysate or infected with MVA-gB (Figure 3A). In order to determine whether endogenous gene transcription was required for antigen presentation, target cells were pulsed with ultraviolet-inactivated MVA-gB. These targets were not recognized by DYS-specific T cells, confirming that recognition was not due to exogenous uptake of gB protein. The HLA restriction of the clones was confirmed using HLA-DRB*0701–matched and mismatched target cells (Figure 3A).

The functional avidity of DYS-specific T-cell clones was assessed using peptide titration analysis, with serial dilutions of peptide ranging from 10⁻⁴ to 10⁻¹¹ M. The amount of peptide required to produce 50% maximal recognition ranged from 10 nM to 100 nM for each of the 8 clones, comparable to other HLA class II–restricted responses^16,17  (Figure 3B).

The chromium release assay was used to measure cytotoxic activity of the CD4⁺ clones, which all demonstrated significant lysis of peptide-loaded LCL (Figure 4A). The mechanism of cytotoxicity was determined with use of the perforin inhibitor ConA, which reduced specific lysis to background levels (Figure 4A). This indicates that cytotoxicity is mediated through the perforin–granzyme system and is supported by detection of intracellular perforin and granzyme B (Figure 4B). In order to investigate the potential biologic spectrum of antigen presentation to DYS-specific CD4⁺ T cells, we infected fibroblasts and DCs with CMV strain AD169 prior to cytotoxicity analysis. CMV-encoded immune evasion genes carry the potential to downregulate HLA class II–restricted peptide presentation, and so the recombinant RV798 viral strain, which carries a deletion in the US2-US11 region, was also studied. Fibroblasts were pretreated with IFN-γ to induce expression of MHC class II prior to infection with AD169 or RV798, or were pulsed with DYS peptide. Fibroblast targets were effectively killed in all cases and demonstrated a high degree of lysis (Figure 4C). DCs express constitutively high levels of MHC class II and were efficiently lysed by DYS-specific cytotoxic T lymphocytes (CTLs) following infection with AD169 or RV798 (Figure 4D). Indeed, cytotoxicity following direct infection with virus was substantially greater than that seen following pulsing of DCs with exogenous DYS peptide and contradicts the idea that human CMV infection of DCs blocks T-cell recognition. Thus, DYS-specific CTLs are able to lyse antigen-presenting cells with either constitutive or inducible expression of HLA-DRB*0701. In addition, the immune evasion genes US2-11 have limited effects on the presentation of this peptide, since recognition of targets infected with RV798 was no greater than recognition of targets infected with AD169.

Cytotoxic CD8⁺ T cells frequently demonstrate highly conserved T-cell receptor gene usage between different individuals. In order to determine whether a similar phenomenon was present in DYS-specific CD4⁺ cells, T-cell receptor sequences were cloned and sequenced from T-cell clones derived from 4 donors. TCR usage of DYS-specific clones was extremely conserved within both the TCRα and TCRβ chains. All of the clones used the TRBV6-5 gene segment, and a highly conserved Q-G amino acid motif was present in the non–germline-encoded hypervariable CDR3 region (Figure 5). An identical TCRβ chain sequence was shared by clones from 2 different donors, and a highly conserved length of the CDR3 loop was present in 11 of 12 sequences.

TCRα chain sequences could be cloned from 7 of the 12 T-cell clones and demonstrated restriction in variable region usage with only 2 different variable region segments being used (Figure 6). Clones from donors 01 and 05 used only the TRAV25 gene segment, whereas clones from donor 02 used TRAV25 or TRAV9-2. Two different transcripts were expressed by clones from donor 02, whereas clones from donor 01 and donor 05 used a single transcript. Junctional segment selection was also conserved with usage of only 3 different segments, and clones showed use of the same junctional region, despite having different variable gene segments.

In order to investigate whether the TCR sequence conservation of DYS-specific clones was subject to potential bias during T-cell cloning, we undertook flow cytometric analysis of TCRβ chain expression on DYS-specific cells isolated directly from peripheral blood. PBMCs from donor 01 were stimulated with DYS peptide and stained for surface expression of TRBV6-5. A total of 57% of the antigen-specific CD4⁺ cells expressed TRBV6-5, confirming conservation of TCR usage by this population (Figure 7A). TRBV6-5 was the predominant TCR gene segment expressed within PBMCs from a further 2 donors, accounting for up to 88% of DYS-specific cells (Figure 7B,C).

Clonal focusing of cytotoxic cells generally occurs during maturation of chronic immune responses. The clonal expansion and immunodominance of the DYS-peptide specific response may have potentially developed over a number of years since primary CMV infection. Primary infection with CMV is almost always silent in immunocompetent individuals, and such subjects are thus difficult to identify. Despite this, we were able to study the DYS-specific response in an immunocompetent patient who developed symptomatic CMV primary infection (donor 11). The DYS-specific T-cell response represented 0.8% of the CD4⁺ T-cell pool within the first month after infection, and more than 40% of this population used the TRBV6-5 gene segment, demonstrating that significant expansion and clonal restriction of this response can occur early after primary infection (Figure 7D).

Immunodominance refers to the restricted antigenic specificity of T cells in an immune response to a complex antigen.¹  In addition, clonal selection within a peptide-specific immune response further limits the breadth of the T-cell repertoire, and these phenomena are commonly seen with CD8⁺ cytotoxic T-cell responses to murine or human viruses.^18,,–21  CD8⁺ T-cell immunity to the influenza matrix peptide is highly dependent on expression of the TCRBV17 gene segment,²²  and structural analysis has indicated that this reflects optimal side chain interactions with the peptide-HLA complex.²³  CD8⁺ T cells specific for a peptide from EBV restricted by HLA-B8 show complete TCR conservation between individuals,^20,24  and less complete clonal conservation is seen in CD8⁺ T-cell recognition of epitopes from CMV.^15,25,26 

In contrast, CD4⁺ T-cell responses are characterized by much broader recognition of peptide epitopes^27,28  and, rather than undergoing clonal competition, CD4⁺ T cells of different specificities may act cooperatively.²⁹  It is not clear why the CD4⁺ T-cell response is less focused than that of the CD8 population, although HLA class II molecules have less stringent anchor positions than HLA class I and are able to bind and present a greater variety of peptide epitopes.

Cytotoxic CD4⁺ T cells have been identified recently as being of considerable importance in the clearance of viral infection³⁰  and bear many phenotypic similarities to their cytotoxic CD8⁺ counterparts. Antigen-presenting cell populations of DCs and B cells have constitutive expression of HLA class II proteins, whereas other cell types, such as endothelium, are induced to express HLA class II molecules by inflammatory mediators. These cell types are tropic for CMV infection, and thus become targets for CD4-mediated immune control. The majority of HLA class II–restricted peptide presentation comes from uptake of exogenous proteins, and this pathway is unlikely to play an important role in the direct recognition of virally infected cells. However, recent work has shown that proteins from CMV and EBV can be presented directly by HLA class II molecules after viral infection.^9,31,32  This observation greatly increases the potential importance of cytotoxic CD4⁺ T-cell populations in the control of lytic viral replication.

CD4⁺ T cells with cytotoxic phenotype are commonly observed in cases of TCRαβ⁺/CD4⁺ T-LGL associated with cytopenia. Garrido et al recently studied TCR usage in 36 cases of LGL and found highly conserved TCR usage in 15 cases, all of which were associated with the HLA-DRB*0701 allele.¹¹  The authors speculated that this might result from antigenic simulation through a common peptide but did not identify the etiologic agent. We have focused on the CD4⁺ T-cell response to DYS peptide from cytomegalovirus glycoprotein B restricted through HLA-DRB*0701.¹⁰  The magnitude of the DYS-specific CD4⁺ T-cell response averages 3.6% of the total CD4⁺ T-cell repertoire, and comparison with the response to whole viral lysate suggests that the DYS-specific immune response constitutes a very large proportion of the total CMV-specific CD4⁺ T-cell immune response in HLA-DRB*0701–positive donors. These findings are noteworthy for a number of reasons. First, this magnitude of CD4⁺ immunity to a single peptide is, to our knowledge, higher than any other peptide-specific immune response that has been reported to date. Moreover, the CMV-specific CD4⁺ cytotoxic response shows extreme immunodominance in HLA-DRB*0701–positive individuals and reveals that immunodominance within CD4⁺ T-cell responses can be at least as marked as that seen within the CD8⁺ population.

In order to investigate the possible factors that might determine immunodominance, we cloned DYS-specific T cells, and peptide titration analysis revealed avidity that was comparable with CD8⁺ cytotoxic clones. T-cell clones lysed target cells directly infected with vaccinia virus encoding the gB gene, demonstrating that gB epitopes can be presented from the cytosol of infected cells. Finally, we were able to infect target cells with CMV and demonstrate cytotoxicity mediated through DYS-restricted T-cell clones.

T-cell receptors from a number of DYS-specific T-cell clones were cloned and sequenced. These showed extreme conservation of T-cell–receptor sequence between different individuals, a finding not previously reported for peptide-specific CD4⁺ T cells.³³  The pattern of T-cell–receptor conservation within TRA and TRB genes is entirely analogous to that seen in the study of TCR usage of CD8⁺ cytotoxic T cells. In particular, the combination of unique TRBV gene-segment usage and CDR3 sequence and length together with strong, but less absolute, selection of TRAV and TRAJ segments is a highly characteristic pattern.³⁴  Structural analysis of T-cell receptor binding to HLA-peptide complexes has been performed for several CD8⁺ virus–specific T-cell clones and demonstrates a consistent pattern of TCR orientation over the MHC complex.³⁵  Crystallographic studies of TCRs isolated from cytotoxic CD4⁺ T cells are not yet available, but the T-cell receptor sequence conservation seen in the CD4⁺ cytotoxic T-cell response to the DYS peptide strongly suggests that TCR orientation will be conserved in this HLA class II–restricted T-cell response.

The biologic basis for this extreme immunodominance and clonal selection is unclear, but a number of possibilities require investigation. The first is that the T-cell receptor interaction with the DYS peptide is of unusually high affinity,³⁶  although peptide titration analyses do not provide convincing evidence for this. Equally, different CMV proteins may have differential access to direct presentation by HLA class II molecules following viral infection, and as gB accumulates in endosomes during viral assembly, peptides from this protein may gain selective access to an HLA class II pathway. It is tempting to speculate that the cytotoxic potential of this CD4⁺ T-cell response contributes directly to a wide degree of clonal selection observed within DYS-specific T cells. Experimental limitation of the number of antigen-presenting cells leads directly to selection of high-avidity T-cell clones due to competition of antigen at the cell surface. Cytotoxic T cells have the ability to lyse antigen-presenting cells,³⁷  and this may be one factor that helps to control ongoing immune responses. Indeed, it is possible that the HLA class II restriction of CD4⁺ cytotoxic T cells actually enhances their potential for clonal selection in comparison to CD8⁺ cytotoxic cells as expression of HLA class II is focused on antigen-presenting cell populations. As such, more pronounced and rapid clonal competition may occur within CD4⁺ cytotoxic T-cell responses than is widely believed.^8,38  Our observation that clonal focusing can occur very early after acute infection is in line with evidence from murine models in which immunodominance was established within the first few hours of an immune response.³⁹ 

Of similar importance is the observation that the TCR usage is highly homologous to that reported in patients with TCRαβ⁺/CD4⁺ T-LGL associated with the HLA-DRB*0701 allele.¹¹  The TCRβ sequence uses the same TCRVB gene segment (albeit with different nomenclature) and retains the Q-G amino acid motif in the highly variable CDR3 region. It would be virtually impossible for such conservation to be observed by chance and, together with the absolute requirement for HLA-DRB*0701 expression in the LGL cases, identifies the CMV DYS peptide as the antigen that has driven the CD4⁺ T-cell clonal expansion. This is the first identification of an infectious agent involved in the initiation or maintenance of LGL and suggests that CMV may also be involved in the pathogenesis of TCRαβ⁺/CD8⁺ or γδ LGL. Treatment with antiviral agents may thus be a reasonable therapeutic option in these cases.

In conclusion, we demonstrate that immunodominance and clonal restriction can occur to a very marked extent in a CD4⁺ cytotoxic immune response. The clinical implications of such clonal restriction may be considerable and should be considered in the development of immunotherapeutic strategies designed to amplify CD4⁺ cytotoxic immune responses. Such a potent and focused immune response may not be of benefit to the host. Indeed, HLA-DRB*0701–positive patients demonstrate an increased risk of CMV reactivation following organ transplantation,⁴⁰  suggesting that the DYS-specific response may contribute directly to immunopathology by an unknown mechanism. Excessive destruction of host antigen-presenting cells is one mechanism that deserves investigation.

This work was supported by the Medical Research Council.

Contribution: L.C. and N.K. performed the laboratory research, N.K. and P.M. designed the research, R.K. and L.N. contributed vital new reagents, L.C. analyzed the data, and P.M. and L.C. wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Prof Paul A.H. Moss, CRUK Institute for Cancer Studies, University of Birmingham, Vincent Drive, Edgbaston B15 2TT, United Kingdom; e-mail: p.moss@bham.ac.uk.