Cutting Edge: CD1a Tetramers and Dextramers Identify Human Lipopeptide–Specific T Cells Ex Vivo Free

Humans express four CD1 Ag-presenting molecules, CD1a, CD1b, CD1c, and CD1d (1). Whereas the basic biology of CD1d and NKT cells has been extensively studied in mice lacking either CD1d or invariant NKT TCRs (2), CD1a proteins have been deleted from the murine genome (3), effectively limiting analysis to the study of a few human T cell clones propagated long-term in vitro (4, 5). The mycobacterial lipopeptide dideoxymycobactin (DDM) is the only well-characterized foreign Ag in the CD1a system and has served as a model to understand the specificity and mechanism of action of CD1a-restricted T cells in vitro (4, 6, 7). Building on the success of human MHC class I tetramers to quantitatively track fresh human T cells ex vivo (8), CD1 tetramers (9–16) capture lipid-reactive T cells as populations for study ex vivo. Dextramers rely on the same principle, but they use higher order multimers that allow detection of rare, lower affinity T cells (17). In this study, we investigated whether human TCRs directly bind to CD1a/lipopeptide to develop tetramers and dextramers that identify polyclonal T cell populations, which recognize foreign Ag bound to CD1a.

We produced soluble biotinylated CD1a monomers in lentivirus-transduced HEK293T cells at the National Institutes of Health Tetramer Core Facility (Emory University, Atlanta, GA) (14, 16).

DDM with defined stereochemistry (DDM 1S,3R or DDM 1R,3S) (7) was incubated with CD1a monomers and complexed with streptavidin coupled to allophycocyanin. Staining was optimized by analyzing T cells with CD1a tetramers generated after loading under varied conditions (Supplemental Fig. 1). Optimal staining was seen with DDM solubilized in DMSO, sonicated into 30-fold excess 50 mM sodium citrate plus 1% CHAPS at pH 6 for 2 min, incubated at 42°C for 1 h, and then added at 40-fold molar excess to CD1a monomers and incubated in a 37°C water bath for 1 h prior to neutralization to pH 7.4 with 2 μl Tris (pH 9).

CD1a tetramers were generated and validated by staining the DDM-reactive αβ T cell clone CD8-2 (18) and using methods similar to those reported for CD1b tetramers (14). We generated dextramers by diluting CD1a complexes to 0.1 mg/ml in PBS prior to incubation with 2 μl allophycocyanin-labeled dextramer backbone for 30 min at room temperature in the dark, followed by staining of T cells as previously reported (17). For dual staining with TCR Abs, TRBV3-1 (Beckman Coulter) was added after tetramer staining for the last 15 min of incubation. Cells were acquired on a FACSCanto II flow cytometer (Beckton Dickinson) and analyzed using FlowJo (Tree Star) software in the presence or absence of mAbs or recombinant TCRs.

The cDNAs of the α- and β-chains of the CD1b-restricted TCR LDN5 and the CD1a-restricted, DDM-specific TCR CD8-2 were produced as previously described (14, 19). Loaded CD1a-DDM was coupled to research-grade streptavidin-coated chips. Increasing concentrations of the CD8-2 TCR (0–533 μM) were injected over all flow cells. BIAevaluation version 3.1 software (Biacore) was used to fit the data to the 1:1 Langmuir binding model, and the equilibrium data were analyzed with the Prism program (GraphPad Software).

Work with human subjects was overseen by the Institutional Review Boards of the Lemuel Shattuck Hospital (00000786), Partners Healthcare (2002-P-000061), and the Harvard Committee on Microbiologic Safety (08-184). PBMCs were treated with dextramers at a 1:100 dilution for 15 min at room temperature followed by 15 min at 4°C, after which cells were washed and incubated with violet viability dye, CD3-FITC, CD14-PerCP-Cy5.5, and CD19-PerCP-Cy5.5 for 15 min at 4°C. Unfixed dextramer⁺ cells were sorted using a FACSAria flow cytometer and expanded by stimulation with anti-CD3 (30 ng/ml) in the presence of irradiated feeder cells and IL-2 (2 nM). After 3 wk, clones were analyzed for binding CD1a dextramers and were tested for DDM recognition in an ELISPOT assay (20).

We generated transmembrane-truncated soluble biotinylated CD1a proteins based on methods previously developed for MHC, CD1d, CD1b, and CD1c (8, 12, 14, 16). To test the folding and Ag-presenting function of this new CD1a construct, it was bound to streptavidin plates, treated with a synthetic DDM, and used to activate T cells (7). The synthetic Ag recapitulates the naturally occurring S or R stereochemistry among the four stereocenters in the natural DDM peptide, including S and R stereochemistry at positions 1 and 3 present in M. tuberculosis DDM (DDM 1S,3R) (Fig. 1A) (7). This lipopeptide activated the CD1a-restricted human T cell line CD8-2 in a dose-dependent manner. No activation was seen in response to DDM with the opposite stereochemical configuration at positions 1 and 3 (DDM 1R,3S) or a synthetic analog that deviated from the optimal natural DDM based on a fully saturated acyl chain and serine substituting for α-methyl serine (DDM analog; Fig. 1B). Thus, soluble CD1a monomers were properly folded and were sufficient to present a lipopeptide Ag to human T cells.

After optimizing loading conditions based on pH, time, temperature, and solvent variables (Supplemental Fig. 1), we observed that DDM (1S,3R)-treated CD1a tetramers selectively stain T cells expressing the clonotypic TRBV3-1 TCR, but not other T cells (Fig. 1C). T cells were not stained by CD1a tetramers that were not exposed to lipids or tetramers treated with the two nonantigenic DDMs that substantially mimic DDM (1S,3R). Thus, tetramer staining was dependent on the structure of the added lipopeptide and was specific for the clonotypic TCR that defines CD8-2. These data established a working CD1a tetramer and strongly supported the model of direct binding of an αβ TCR to CD1a. To further increase the avidity of interaction, we developed CD1a dextramers that are composed of 10–14 CD1a monomers on a flexible, fluorescently labeled dextran backbone (17). We observed bright, selective, and highly reproducible staining of the TRBV3-1 subset of CD8-2 T cells (Fig. 1D) with DDM (1S,3R)-loaded dextramers. Thus, dextramers provided a second reagent for probing the interaction of CD1a with T cells.

CD1a-DDM tetramer and dextramer binding to TRBV3-1⁺ T cells strongly implicated a cognate model in which a ternary interaction of CD1a/lipid binds to the TCR. We performed experiments to directly test binding between lipopeptide/CD1a complexes and the clonotypic human αβ TCR versus all other surface receptors. Preincubation of dextramers with anti-CD1a Ab blocked staining to background (Fig. 2A), confirming that staining was mediated by CD1a. We generated soluble TCRs containing the TCR α- and β-chains from CD8-2 (αβ heterodimers encoded by the TRAV3-1 and TRBV3-1 variable regions) or analogous constructs from the CD1b-restricted T cell LDN5 (composed of the TRAV17 and TRBV4-1 variable regions) (Fig. 2B). Preincubation with soluble LDN5 TCR minimally impacted staining, whereas preincubation with soluble CD8-2 TCR blocked T cell staining to background levels (Fig. 2C). Thus, CD1a-DDM staining of cells is mediated by the αβ TCR.

We then measured the affinity of interaction between soluble transmembrane-truncated CD8-2 TCR and CD1a proteins alone or pretreated with DDM using surface plasmon resonance (Fig. 2D). No binding was seen to CD1a alone; the determined K_D was 95.78 ± 13.51 μM. This affinity is significantly lower than that of TCRs recognizing α-galactosyl ceramide-CD1d (<1 μM) and glucose monomycolate-CD1b (∼1 μM) (15)) but approximates the affinity of NKT TCRs recognizing β-linked glycolipids (21).

To determine whether DDM-reactive T cells exist as cell populations ex vivo, PBMCs from subjects with active tuberculosis or positive tuberculin skin tests were stained with dextramers treated with DDM (1S,3R) and then gated on CD3⁺CD14⁻CD19⁻ live lymphocytes (Supplemental Fig. 2A). Rare CD3⁺ cells were identified with the DDM-loaded dextramer among four subjects with mycobacterial exposure (Fig. 3, Supplemental Fig. 2B). This result suggested that DDM-reactive T cells are present in the blood of unrelated human donors.

To determine whether the rare dextramer⁺ cells recognize CD1a and DDM, T cell yields were increased through leukapheresis (subject C58) or ex vivo expansion using anti-CD3 Ab and IL-2 (subject A32). We then used DDM-treated dextramers to sort cells (Supplemental Fig. 2C), cloned them using limiting dilution, and tested them for reactivity in ELISPOT assays. Dextramer-based sorting generated many T cell clones that were brightly stained using DDM-loaded dextramers, and these clones secreted TNF-α (clones 3, 6, 9, 21) or IFN-γ (clones P1, P5, P7, P9) in response to CD1a-expressing APCs treated with DDM (Fig. 4). Three of the clones express CD8 and the other five express CD4. We therefore conclude that CD1a- and DDM-reactive T cells are present as populations among genetically diverse donors.

Overall, these results distinguish between indirect and cognate TCR interaction models, illustrating that direct CD1a/DDM/TCR interactions are the mechanism of T cell activation by CD1a and lipopeptide. These studies extend prior work describing polyclonal responses to lipoprotein mixtures (18) and lipopeptide-specific T cell clones (22) to prove that lipopeptide-reactive T cells exist in the in vivo state. More generally, generation of working CD1a multimers completes the set of tools needed to study human CD1-restricted T cells, allowing determination of the phenotype and function of T cells recognizing CD1a bound to self and foreign Ags.

S.J. works for Immudex, the company that makes the dextramers used in this study to discover lipopeptide-specific T cells. The other authors have no financial conflicts of interest.