A Subset of Human Autoreactive CD1c-Restricted T Cells Preferentially Expresses TRBV4-1⁺ TCRs

Whereas CD1d is the only CD1 protein found in mice, the genomes of humans and many other mammals encode multiple members of this protein family (1). In humans, the CD1 family consists of CD1a, CD1b, CD1c, CD1d, and CD1e, of which CD1a, CD1b, CD1c, and CD1d present lipid Ags at the cell surface (2–4). CD1e is an intracellular chaperone involved in the processing and presentation of lipids by other CD1 proteins (5, 6). Lipid-presenting CD1 molecules are further divided into group I (CD1a, CD1b, and CD1c) and group II (CD1d), on the basis of their homology. The two groups also differ in their tissue expression pattern: group I CD1 proteins are restricted to professional APCs and thymocytes, whereas CD1d is also expressed on certain epithelial tissues (7, 8). CD1d and CD1d-restricted NKT cells have been extensively studied in mice and humans. A subset of human NKT cells is molecularly defined by the expression of the invariant TRAV10-TRAJ18 TCRα-chain paired with semivariant TRBV25 TCRβ-chains. The recognition of self-lipids is important for the thymic selection, peripheral maintenance, and activation of invariant NKT (iNKT) cells (9–11).

CD1c-restricted T cells have been understudied relative to iNKT cells. Nevertheless, several lines of evidence in noninfectious diseases suggest the potential importance of self-recognition by CD1c-restricted T cells. CD1c-restricted autoreactive T cells isolated from systemic lupus erythematosus patients have been found to enhance the production of IgG by B cells (12). Moreover, CD1c⁺ APCs and CD1a- and CD1c-restricted T cells have been found to infiltrate the thyroid in patients with Graves’ or Hashimoto’s disease (13). Group I CD1 proteins have also been detected in atherosclerotic arteries by immunohistochemistry and have been found to colocalize with CD68 (14). Finally, malignant cells of hematologic origin express CD1c, and a tumor-associated self-lipid isolated from leukemic cells has been found to activate CD1c-restricted T cells (15).

CD1c tetramers were recently developed to identify mycobacterial lipid-specific populations ex vivo (16). Using this technology, Roy et al. (17) isolated TRDV1⁺ γδ T cells stained with the CD1c-phosphomycoketide tetramer, and demonstrated that some of the clones also recognized CD1c presenting self-lipids such as sulfatides and lysophospholipids. However, the molecular identity of autoreactive CD1c-restricted αβ T cells remains largely unknown. Based on single-cell cloning, the frequency of autoreactive CD1c-restricted αβ T cells was estimated to range from 0 to 7% of CD4⁺ T cells (18), thus representing a significant population in certain individuals. Elucidating the molecular basis of self-antigen recognition by CD1c-restricted T cells will strengthen understanding of the fundamental biology of these cells and may facilitate the development of therapeutic receptors targeting CD1c-lipid complexes as an HLA-unrestricted form of immunotherapy (19, 20).

We have previously developed an artificial APC (aAPC) system based on the K562 human cell line, which lacks endogenous expression of MHC class I, MHC class II, and CD1 molecules. K562 has been engineered to be immunogenic through expression of the costimulatory molecules CD80 and CD83. Various Ag-presenting molecules have been individually introduced into CD80⁺CD83⁺ K562 cells to design aAPCs that can activate a cognate Ag-reactive T cell population of interest (21–25). Recently, we have demonstrated that CD1d⁺ aAPCs presenting endogenous lipids are able to expand a polyclonal T cell population in a CD1d-dependent manner (26, 27). From the expanded population, we isolated a large panel of clonotypic TRBV25 iNKT cell receptors exhibiting a wide of range of autoreactivity. In this study, we used the aAPC system to study CD1c-restricted autoreactive T cells and identified TRBV4-1 as a conserved molecular feature of the responding TCR repertoire. Without knowing the exact identity of individual self-ligands, our strategy enables the identification of T cells that react with dominant autoantigens presented by CD1c. The TRBV4-1⁺ T cells expressed heterogeneous TCRα genes and depended on critical germline-encoded TRBV4-1 CDR1 and CDR2 residues to recognize CD1c. Together, our data contribute to the fundamental understanding of the molecular biology of CD1c-restricted T cells.

Peripheral blood samples were obtained following institutional review board approval and written informed consent from all donors. Primary T cells, Jurkat 76 cells (28), K562, T2, and their derivatives were cultured in RPMI 1640 supplemented with 10% FCS and gentamicin (Life Technologies, Carlsbad, CA).

T cells were purified from PBMCs by untouched selection using magnetic beads (Miltenyi Biotec, Auburn, CA). aAPCs were irradiated (200 Gy) and added to sorted T cells at the indicated T cell/aAPC ratios. Cultures were supplemented with 20 IU/ml recombinant human IL-2 (Chiron, Emeryville, CA) every 3–4 d. T cells were analyzed or restimulated every 7 d. Anti-CD1c blocking (clone L161) and isotype control mAbs from BioLegend (San Diego, CA) were used at 10 μg/ml.

The following mAbs recognizing the indicated Ags were used: TRBV panel (IOTest β Mark), TRBV4 (clone ZOE), and TRAV10 (clone C15) from Beckman Coulter (Mississauga, ON, Canada); CD54 (clone D2) from Ancell (Bayport, MN); CD80 (clone L307) and CD58 (clone 1C3) from BD Biosciences (Mississauga, ON, Canada); CD83 (clone HB15e) and CD45RA (clone MB1) from Caltag Medsystems (Buckingham, U.K.); and CD1a (clone HI149), CD1b (clone SN13), CD1c (clone L161), CD1d (clone 51.1), CD3 (clone UCHT1), CD69 (clone FN50), CD45RO (clone UCHL1), CD154 (clone 24-31), pan-TCRαβ (clone IP26), IL-2 (clone MQ1-17H12), TNF-α (clone MAb11), and isotype controls from BioLegend. Viability was determined using 7-aminoactinomycin D (BioLegend) in surface staining. For intracellular detection of cytokines, T cells were stimulated with T2 target cells at 1:1 E:T ratio for 5 h. Brefeldin A (BioLegend) was added 1 h poststimulation. Fixable Aqua Dead Cell Stain Kit (Thermo Fisher Scientific, Waltham, MA) was used to determine viability. Cells were permeabilized and fixed using the BD Cytofix/Cytoperm kit (BD Biosciences). All data shown were gated on singlets and live cells, and as further specified in the respective figure captions.

Full-length cDNAs encoding the TRBV4 TCRβ, CD1c, CD4, and CD8 genes were molecularly cloned via RT-PCR using gene-specific primers into the pMX vector. TCRα genes were cloned by 5′ RACE as previously described (29–32). Alanine mutants were generated by PCR-based site-directed mutagenesis (33). Nucleotide sequencing was performed at the Centre for Applied Genomics, The Hospital for Sick Children (Toronto, ON, Canada). TCR gene segments and CDR sequences were defined according to the International Immunogenetics Information System (http://www.imgt.org).

Using the 293GPG-based retrovirus system (34), we generated CD1c⁺ aAPCs and T2/CD1c by transducing CD80⁺CD83⁺ K562 and T2 cells with CD1c, respectively. TCRα⁻β⁻ Jurkat 76 cells were similarly transduced with a β₂-microglobulin short hairpin RNA (OriGene, Rockville, MD) and the CD8αβ and CD4 genes. The Jurkat 76.clone 10 (Cl10) expressing coreceptors and low levels of CD1 was established by the limiting dilution method. Cl10 cells were further transduced with a pair of clonotypic TCRαβ genes to reconstitute TCRs. Control TCR, clone A6, recognizes a peptide derived from the human T cell leukemia virus type 1 Tax protein in an HLA-A2–restricted manner (35). CD3⁺ cells were collected by magnetic or flow-assisted cell sorting to purify TCR transfectants.

CD1c or CD1d monomer (5 μg/ml; NIH Tetramer Core Facility), or anti-CD3 mAb (clone OKT3, 2 μg/ml; in-house) was coated onto 96-well ELISA plates (Thermo Fisher Scientific) with anti-human CD28 mAb (clone CD28.2, 1 μg/ml; BioLegend) for 16–18 h at 37°C. Anti-CD28 mAb alone was coated as no CD1 control. After washing twice with PBS and once with complete culture media, 5 × 10⁴ Cl10 transfectants were seeded into each well and cultured for 5–6 h. The upregulation of CD69 surface expression was analyzed by flow cytometry.

Statistical analyses were performed using GraphPad Prism version 6.0. All tests were two-tailed, and p < 0.05 was considered significant. All statistical analyses of primary T cell experiments were performed in a pairwise fashion comparing within donors. Detailed descriptions of the statistical tests are found in the respective figure captions.

Our previous experience in using CD1d⁺ aAPCs to successfully enrich for iNKT cell receptor-expressing T cells led us to use a similar system to study CD1c-restricted T cells. We cloned the full-length native CD1c gene from the Jurkat cell line and stably expressed it in the CD80⁺CD83⁺ K562-based parental aAPCs. K562 cells lack endogenous surface expression of MHC and CD1 molecules, but endogenously express the surface adhesion molecules CD54 and CD58 (24). This novel CD1c⁺ aAPC expressed surface CD54, CD58, CD80, and CD83 molecules at levels comparable with those in the parental CD1-null aAPCs (Fig. 1A).

When we stimulated peripheral blood T cells with the irradiated aAPCs at a relatively high T cell/APC ratio of 5:1, we consistently detected a population of activated CD4⁺ T cells, as measured by upregulated CD154 expression at day 7 (Fig. 1B). Notably, the aAPCs were not loaded with exogenous lipids but instead presented a mixture of endogenous lipids. The T cells activated by the CD1c⁺ aAPCs expressed αβ TCRs, which were not biased toward TRAV10, as shown in Fig. 1C. Furthermore, the activated T cells were largely positive for CD45RO at day 7, thus suggesting a recall response (Fig. 1C).

To characterize the TCR repertoire of the CD1c-reactive T cells, we stained the cells with a panel of commercially available TRBV-specific mAbs. Interestingly, we observed significant differences in the frequency of TRBV4, TRBV5-1, TRBV11-2, and TRBV20-1 usage between activated and nonactivated CD4⁺ T cells. Whereas the frequency of TRBV4 and TRBV11-2 usage increased among the activated T cells, TRBV5-1 and TRBV20-1 usage decreased (Fig. 1D). These data suggested that TRBV4 and TRBV11-2 are preferentially expressed by autoreactive CD1c-restricted T cells.

The TRBV4⁺ CD1c-reactive population was of interest because TRBV4 has been observed to be encoded by previously isolated autoreactive CD1c-restricted T cell clones (18, 36–38). To confirm the CD1c dependency of TRBV4 enrichment in activated T cells, we performed the stimulation experiment with anti-CD1c blocking or isotype control mAbs. In the presence of the blocking mAb, the frequency of CD154⁺ cells was significantly lower than the isotype control condition and was reduced to baseline (Fig. 2A, top, 2B). TRBV4 frequency was also reduced among CD154⁺ T cells when anti-CD1c blocking mAb was added with CD1c⁺ aAPCs instead of isotype, although it was still significantly increased compared with CD1-null aAPC stimulation (Fig. 2A, bottom, 2C). This may be because of incomplete blocking of CD1c with this high number of aAPCs. The percent of TRBV4⁺ cells did not change in unactivated CD154⁻CD4⁺ T cells (Fig. 2C), indicating that the addition of Ab in the culture had no effect on the overall frequency of this population.

Given that we observed a robust response, albeit with a potential bystander effect at the 5:1 T cell/aAPC condition, we then performed stimulations at a lower T cell/aAPC ratio of 20:1 with two successive stimulations over 14 d. This protocol allows for more specific enrichment for the population of interest and has been the standard condition in our previous studies using K562-based aAPCs (21, 24–26, 30, 31). After one or two stimulations at this lower ratio, we found that the frequency of TRBV4 usage was also significantly higher in the CD154⁺ T cells stimulated by CD1c⁺ aAPCs with the isotype mAb, compared with that with anti-CD1c Ab and control aAPCs. Using this stimulation condition of lower number of APCs, the enrichment of TRBV4 was completely blocked by the anti-CD1c mAb (Fig. 2D).

In addition to CD154, we analyzed CD1c-restricted T cell responses using intracellular cytokine production as a readout. Two times stimulated T cells from the 20:1 T/aAPC condition were restimulated on day 14 with T2, or T2 transduced with CD1c (T2/CD1c). Significant increases in IL-2 and TNF-α production were detected only when intact CD1c was available in both the initial stimulations and rechallenge (Fig. 3A). In eight of the nine donors with detectable CD1c-restricted cytokine responses, we analyzed the TRBV4 frequency in the cytokine-positive and -negative fractions of CD4⁺ cells. Consistent with the data obtained using CD154 as an activation marker, TRBV4⁺ cells were significantly enriched in the IL-2⁺ and TNF-α⁺ populations compared with the nonresponding fraction (Fig. 3B). CD1c expression levels in parental T2 and T2/CD1c cells are shown in Fig. 3C. Together, these data confirmed an association between TRBV4 and CD1c-restricted autoreactivity.

The TRBV4⁺ population after two stimulations was sorted from three donors (0927A, 0430A, and 0610A) by flow cytometry–guided sorting, with a purity >95%, and subjected to more detailed molecular analyses (Fig. 4A). Using the purified TRBV4⁺ T cells from the three donors, we cloned ∼50 independent TCRα or TCRβ genes from each sample. TCRβ genes were isolated with a TRBV4-specific primer, which allowed for cloning of all three TRBV4 genes, TRBV4-1, TRBV4-2, and TRBV4-3. The TCRα genes were cloned using the 5′ RACE method, because we had no prior knowledge of TRAV usage. The observed V/J segment usage, CDR3 sequences, and copy numbers of the isolated unique genes are shown in Supplemental Tables I and II. None of the TCRα or TCRβ genes were shared across donors. The cloned TCRα genes were highly diverse with regard to TRAV and TRAJ usage and CDR3α amino acid length (Fig. 4B, left). Although the isolated TCRβ genes also encoded heterogeneous TRBJ segments and CDR3β length, all encoded TRBV4-1 (Fig. 4B, right). These results suggested that TRBV4-1, but not TRBV4-2 or TRBV4-3, is the molecular motif associated with CD1c-restricted autoreactivity.

Next, we reconstituted TCRs consisting of the cloned TRBV4-1 TCRβ and TCRα genes and determined their CD1c-restricted autoreactivity. We generated a CD1c^lowCD4⁺CD8⁺ Jurkat 76.Cl10 (hereafter termed Cl10) T cell line as host cells (Supplemental Fig. 1). Surface CD1 was knocked down to prevent fraternally induced activation after transduction of autoreactive CD1c-restricted TCRs. Using these cells as a basis, we reconstituted unique TCRα and TCRβ genes for which we had isolated more than one copy within each donor in all possible pairings. In total, we generated 84 clonotypic TCR Cl10 transfectants, including 12 from donor 0927A, 32 from donor 0430A, and 40 from donor 0610A. We also reconstituted the HLA-A2/TAX–restricted TCR, clone A6, as a negative control (35).

We stimulated each transfectant with the CD1c⁺ or control aAPC and measured their activation by assessing CD69 upregulation. Forty-one of the 84 clonotypic TCRs demonstrated detectable CD1c reactivity, as defined by at least a 1.5-fold increase in CD69 positivity after stimulation with CD1c⁺ aAPCs compared with control aAPCs (Supplemental Fig. 2). As shown in Fig. 5A, responses from all of the CD1c-reactive transfectants were significantly blocked by the anti-CD1c mAb (p < 0.0001 for all clones comparing anti-CD1c and isotype control mAb treatment, by two-way ANOVA followed by Bonferroni post hoc analysis). The magnitude of the response was not associated with any particular TRAV usage (Fig. 5A, Supplemental Table I). CD1c-restricted and control TCRs were expressed at compared levels on Cl10 cells (Fig. 5B) and elicited similar maximal responses (Fig. 5C), demonstrating no intrinsic differences in these transfectants to upregulate CD69.

Collectively, these data suggested that the extent of CD1c-restricted autoreactivity was regulated by the unique CDR3 sequences of the TCR. For example, the clonotypic TCRs C13.7A-18/C13.7B-5 and C13.7A-18/C13.7B-61, which share the same TCRα gene and differ only in their CDR3β sequences, differed greatly in their CD1c reactivity (Fig. 5A, donor 0927A). Moreover, the clonotypic TCRs C15.7A-21/C15.7B-112 and C15.7A-130/C15.7B-112, which share the same TCRβ-chain and TRAV usage, also differed noticeably in their response (Fig. 5A, donor 0430A). To determine the contribution of the unique CDR3β residues in CD1c recognition, we performed alanine scanning of this region for two TRBV4-1 TCRβ genes: C15.7B-32 and C17.7B-25. These two TCRβ-chains were chosen because they encoded distinct CDR3 residues, length, and TRBJ (Supplemental Table II). The wild-type and mutant versions of C15.7B-32 were reconstituted with C15.7A-21 and C15.7A-60, whereas C17.7B-25 variants were reconstituted along with C17.7A-73 or C17.7A-81 (Fig. 6).

Each transfectant was stimulated with plate-bound CD1c or CD1d monomers presenting HEK293-derived self-lipids, or no CD1 molecules, and CD69 expression was measured (Fig. 6A, 6D, 6G, 6J). CD3 expression of the mutant transfectants (Fig. 6B, 6E, 6H, 6K) and the maximal response of all transfectants tested (Fig. 6C, 6F, 6I, 6L) are shown for comparison. Stimulation of a Jurkat 76 transfectant expressing a clonotypic human iNKT cell receptor Cl.3010 (26) confirmed that functional CD1d molecules are present in this assay (Fig. 6M). Mutations at L97 and K104 of C15.7B-32 TCRs substantially reduced their CD1c-restricted reactivity, and this reduction was conserved when C15.-7B-32 was paired with the two different TCRα genes (Fig. 6A, 6D). In the case of C17.7B-25, T96A, V97A, and L98A, all demonstrated severe loss in CD1c reactivity when paired with C17.7A-81 (Fig. 6J). However, the importance of these residues were less pronounced when CD17.7B-25 was paired with C17.7A-73 (Fig. 6G), likely compensated by the strong CD1c reactivity of this TCRα. Nonetheless, these data demonstrated that particular CDR3β residues of TRBV4-1⁺ TCRs can play significant roles in CD1c recognition.

Lastly, the TRBV4-1 bias observed in the earlier experiments suggests that germline-encoded residues play a role in CD1c recognition, particularly at CDR1 and CDR2 sequences. As described earlier for CDR3β, we studied the functional importance of CDR1 and CDR2 residues by alanine scanning for the same set of TCR transfectants shown in Fig. 6. Among the alanine mutants of C15.7B-32 and C17.7B-25, R30A and Y51A consistently yielded the greatest loss in reactivity to CD1c, regardless of the CDR3β or TCRα sequences of the TCR (Fig. 7A, 7D, 7G, 7J). R30A and Y51A TCRs were expressed at comparable levels on the surface (Fig. 7B, 7E, 7H, 7K) and elicited similar maximal responses (Fig. 7C, 7F, 7I, 7L) as the other mutant and control transfectants. Importantly, Arg³⁰ and Tyr⁵¹ are present in only TRBV4-1 and, therefore, are unique to TRBV4-1 among the TRBV4 family (Fig. 7M). Consistently with this finding, no enrichment of TRBV4-3⁺ CD4 T cells was observed with CD1c⁺ aAPC stimulation (Fig. 1D, far right).

TCR gene bias in the CD1d-restricted iNKT cell receptor repertoire is well established (11). TCR bias is also observed in MR1-restricted mucosal-associated invariant T cells, where the TCRα-chain is largely limited to TRAV1-2-TRAJ33. Compared with iNKT cells, the mucosal-associated invariant TCRβ-chain repertoire is more oligoclonal but also biased toward TRBV6 and TRBV20 (39, 40). Until recently, it had not been known whether the repertoire of group I CD1-restricted TCRs also includes conserved molecular motifs. Two subsets of CD1b-restricted T cells have been identified that demonstrate TCR bias. Germline-encoded mycolyl lipid-reactive T cells preferentially encode TRAV1-2 rearranged to TRAJ9 in TCRα with a moderate bias toward TRBV6-2 in TCRβ genes (41). In contrast, LDN5-like T cells exhibit bias toward TRAV17 and TRBV4-1 (42). Both subsets were detected by multimeric CD1b loaded with the mycobacterial lipid glucose-6-O-monomycolate. In addition to these populations, our study has identified a TRBV4-1 bias in a subset of self-reactive CD1c-restricted T cells. In support of this finding, multiple autoreactive CD1c-restricted T cell clones that encode TRBV4 have been isolated previously (18, 36–38).

Because the nature of self-antigens presented by CD1c is not fully characterized, identifying T cells that respond to naturally presented endogenous lipids allows a more comprehensive analysis of autoreactive CD1c-restricted T cells. Similar strategies based on the same principle were used to estimate the frequency of autoreactive CD1-reactive T cells in the peripheral blood (18, 43). Our method of detecting changes in TRBV frequency after CD1c⁺ aAPC stimulation is limited by the availability of TRBV-specific mAbs. Therefore, there may have been other TRBV biases that were undetectable by our system. Notably, several mycobacterial-reactive CD1c-restricted T cell clones derived from different donors expressed TRBV7-9⁺ TCRs; these might represent another molecular motif of the CD1c-restricted T cells (44). Unfortunately, TRBV7-9–specific mAbs are not commercially available. Our approach also entailed pairing unique TCRα and TRBV4-1⁺ TCRβ genes from each donor to obtain a panel of CD1c-restricted TCRs. Accordingly, it is possible that not all of the TCR pairings determined in this study actually exist in vivo.

All of the CD1c⁺ aAPC-reactive TCRs we identified in Fig. 5A were significantly blocked by anti-CD1c mAb. However, a few clonotypic TCRs were only partially blocked. The incomplete blocking may be because of the potential high affinity of these particular TCRs for CD1c, or they are cross-reactive to another surface molecule uniquely expressed by the CD1c⁺ aAPC. Another explanation could be the different docking modes of the TCRs and the L161 mAb. Using alanine mutants of key solvent-exposed residues on α1 and α2 helices of CD1c, Roy et al. (44) determined where different TCRs docked onto the CD1c-phosphomycoketide complex. Interestingly, the requirement of α2 residues in the TCR–Ag interaction noticeably differed for several TCRs tested (44), suggestive of diverse docking orientations by CD1c-restricted TCRs. Therefore, it is possible that the insufficient blocking observed in this study could also be due to the binding of distal surfaces on the CD1c–lipid complex by the TCRs and L161 mAb.

Together with the LDN5-like CD1b-restricted T cell subset, our finding indicates that TRBV4-1 is a shared motif for both CD1b and CD1c restriction. Whether the CDR1 and CDR2 sequences of TRBV4-1 perform a similar role, that is, possibly recognizing conserved residues found in the α1 and α2 helices of both CD1b and CD1c, or whether they play unique roles in the context of different TCRα and CDR3β sequences remains to be elucidated. Two autoreactive CD1b-restricted TCRs have also recently been isolated by Van Rhijn et al. (45). Interestingly, one of the CD1b-restricted TCRs also encoded TRAV13-1 and TRBV4-1, similarly to the three C13.7A-18 (TRAV13-1⁺) transfectants, although CDR3 sequences were different. Our C13.7A-18 transfectants responded to stimulation with CD1c⁺ aAPCs, but not CD1b⁺ aAPCs (data not shown), possibly as a result of the differences in the lipids presented, the structures of CD1b and CD1c, the CDR3 sequences of the TCRs, or a combination thereof.

Some of the TRAV segments that formed CD1c-restricted TCRs studied in this article have also been detected in previously identified CD1c-restricted T cell clones, specifically TRAV9-2, TRAV13-1, TRAV22, and TRAV26-2 (18, 44). TRDV1 is also known to be associated with CD1c reactivity when it is rearranged to other TCRδ gene segments (17), but the C13.7A-27 gene encodes TRDV1 rearranged to TRAJ18 and TRAC. Molecular biases in the TCRα repertoire of CD1c-restricted T cells await further comprehensive analyses. From an evolutionary perspective, it may be efficient for the immune system to dedicate a certain set of germline-encoded TCR genes that are biased toward recognizing monomorphic targets, similarly to pattern recognition receptors. Indeed, TRAV1 may have coevolved with MR1, because the presence or absence of these two genes correlates perfectly across 15 species of mammals (46). However, no TRAV usage was shared among multicopied TCRα genes across the three donors tested. This discrepancy may have been a result of central selective pressures exerted by HLA class I and II, which are probably unique in each donor.

In summary, our study reports a novel molecular motif, TRBV4-1, associated with CD1c-restricted autoreactivity. The strength of self-reactivity of TRBV4-1⁺ TCRs was influenced by CDR3β sequences. Mutational mapping of CDR1β and CDR2β sequences revealed two unique residues, Arg³⁰ and Tyr⁵¹, which were crucial for TRBV4-1–mediated CD1c recognition. Therefore, this study advances the understanding of CD1c-restricted autoreactive T cells and their molecular properties underlying the recognition of CD1c.

Jurkat 76 cells were a generous gift from Dr. Heemskerk, Leiden University Medical Centre. We would like to acknowledge the National Institutes of Health Tetramer Core Facility for providing human CD1c and CD1d monomers.

The authors have no financial conflicts of interest.