Recognition of Listeria Infection by Germline Elements of the Vγ1.1 Vδ6.3 TCR

More than 20 years ago, γδNKT cells were identified by different groups as a γδT population that shares many similarities with αβ–invariant NKT (αβiNKT) cells (1, 2). The role and function of γδNKT cells are still not clear. By contrast, αβiNKT have been well characterized. They are defined by the expression of a semi-invariant TCR that recognizes CD1d, a nonclassical MHC class I protein, loaded with self- and foreign lipids, and are known to have a broad immunomodulatory role (3). Although their TCRs are generated in the same way as conventional αβT cell TCRs, by somatic rearrangement of variable (V), diversity (D) and joining (J) gene segments (1), the αβiNKT TCR has restricted diversity. In mice, the αβiNKT TCR is formed by the Vα14Jα18 chain paired with a limited Vβ-chain repertoire (Vβ2, Vβ7, Vβ8.1, Vβ8.2, and Vβ8.3). In a similar way, the γδNKT cells also express a restricted TCR formed by Vγ1.1Jγ4Cγ4 and Vδ6.3Dδ2Jδ1Cδ1 in B6 mice (4) or Vδ6.4Dδ2Jδ1Cδ1 in DBA/2 (2, 5). γδNKT and αβiNKT populations have a highly similar gene expression profile, more similar to each other than αβiNKT with conventional αβ subsets (6). γδNKT and αβiNKT cell populations express many of the same cell-surface markers, such as NK1.1 and NKG2D, and a preactivated phenotype characterized by the high expression level of CD44 and low expression level of CD62L (2, 4). In more recent work, it was shown that Vγ1.1⁺ Vδ6.3⁺ γδNKT cells express PLZF (7), the same transcription factor that drives differentiation of αβiNKT (8, 9). In γδNKT cells, expression of PLZF depends on the strong TCR stimulation and drives the secretion of both IFN-γ and IL-4, a characteristic shared with αβNKT cells (7).

Despite having a well-characterized phenotype, the specific ligand for the γδNKT TCR has not been identified yet, limiting understanding of the function of these cells in the immune system. Several studies have reported that the γδNKT cell population can respond to L. monocytogenes (Listeria) infection (10–13). During early γδT cell studies, it was shown that Vδ6.3⁺ were the major subtype of γδT cells present in liver and thymus during Listeria infection (11). The Vγ1.1Vδ6.3 T cells adhered on macrophages isolated from Listeria-infected mice, and an anti-Vδ6.3 blocking Ab interfered with cell adhesion, indicating that the interaction between γδT cells and the macrophages depends on the TCR (12). In one study, a Vγ1.1Vδ6.3 hybridoma derived from Listeria-infected mice spleen was activated and secreted IFN-γ when cultured with peritoneal exudate cells obtained from Listeria-infected mice (10). In the same study, splenocytes enriched for γδT cells showed a cytotoxic effect on the peritoneal exudate cells, with the effect blocked by an anti-Vδ6.3 TCR blocking Ab (10). These results indicate that the interaction of Vγ1.1Vδ6.3 TCR with a ligand present in macrophages from Listeria-infected mice can result in a functional response. Using fluorescently labeled oligomers of a soluble version of a γδNKT TCR, carrying characteristic Vγ1.1 Vδ6.3 chains from clone butyrophilin (BTN) 19.8 [derived from the thymus of newborn B10 mice (14)], Aydintug et al. (13) were able to stain peritoneal-derived macrophages, indicating that the macrophages expressed a self-ligand for the TCR. Staining was enhanced when the mice were infected with Listeria (13). However, staining required higher-order oligomerization, with Ab and streptavidin used to form approximately octameric staining reagents, and staining was not observed with conventional tetramers (13). This suggests a very low affinity interaction between the TCR and the putative ligand present in the macrophages. Moreover, because the BTN 19.8 clone was isolated from newborn mice before induction of the machinery to generate junctional diversity, the BTN 19.8 TCR carries germline sequences, and the role of CDR3 diversity in ligand recognition was not investigated.

In the current study, we investigated the CDR3 sequence requirement of the Vγ1.1Vδ6.3 TCR to recognize macrophages and macrophages infected with Listeria. For this purpose, we expressed four different CDR3 γ variants of Vγ1.1Vδ6.3 TCRs in TCR-deficient hybridomas cell lines and showed that all four variants can induce cell activation measured as IL-2 secretion when cocultured with macrophages. The activation was enhanced when the macrophages were infected in vitro with Listeria. These results indicate that a functionally relevant γδNKT TCR self-ligand is present in macrophages. One out of the three TCR variants with an adult-derived Vγ1.1 CDR3 sequence showed a significantly weaker ability to activate the hybridomas when cultured in presence of the macrophages or macrophages infected with Listeria, suggesting that the CDR3γ sequence is involved in the γδNKT recognition of Listeria infection.

Total RNA was isolated from TCR-β^−/− Vγ1.1⁺Vδ6.3⁺ thymic γδT cells, and TCR-γ sequences were cloned as described by Yin et al. (15). Briefly, reverse transcription was performed with qScript cDNA SuperMix (Quantabio, Beverly, MA). TCR-Vγ region sequences were amplified from cDNA by PCR using previously described primers (16). Amplified sequences were cloned into the pCR2.1-TOPO TA vector using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA), and DNA isolated from individual clones was sequenced using both forward and reverse primers.

The γ and δ sequences cloned into the pCR2.1-TOPO vector did not include the transmembrane domain, and thus, for expression of TCR on the cell surface, the C region sequences were fully extended using overlapping PCR. Using T4 DNA polymerase, 160 and 140 nt of the missing Cγ4 and Cδ1 sequences, respectively, were synthesized. The following primers, shown with the 21-nt overlapping portions underlined, were used (all listed 5′ to 3′): Cγ4 forward, 5′-CTGGAAGATTGCATGAAAGGAAGAAAGGATATGTTGCAGCTTCAGGTCACCACCACCTATGCATTCTACACCTACCTCATCCTGTTCTTC-3′; Cγ4 reverse, 5′-GATCATCACAGGACATGGCTGCTCTTCTGAATAGACAGAATACAACGAAGGCCAAGTGGACCATGCTCTTGAAGAACAGGATGAGGTAGG-3′; Cδ1 forward, 5′-CAGAGCCTTGCTATGGCCCAAGAGTCACAGTTCACACTGAGAAGGTAAACATGATGTCCCTCACGGTGCTGGGCCTACGACTGC-3′; and Cδ1 reverse, 5′-TTAAAAGAATAACTTAACAGTCAAGAGAAAATTGATGGCAATGGTCTTGGCAAACAGCAGTCGTAGGCCCAGCACCG-3′

Amplification products were run on a 2% agarose gel and purified and then were used in a subsequent overlapping PCR with the pCR2.1-TOPO vector expressing each of the γ- and δ-chains and the following primers. The primers included restriction enzyme sites (shown as underlined) with the enzyme indicated in brackets: Cγ4 forward (XhoI), 5′-CCGCTCGAGATGCTGCTCCTGAGATGGCCCACC-3′; Cγ4 reverse (NotI), 5′-TTTTCCTTTTGCGGCCGCTTATGAACGTTGATCATCACAGGACATGGCTGCTCTTC-3′; Cδ1 forward (SalI), 5′-ACGCGTCGACATGCCTCCTCACAGCCTGTTCTGTG-3′; and Cδ1 reverse (NotI), 5′TTTTCCTTTTGCGGCCGCTTAAAAGAATAACTTAACAGTCAAGAGAAAATTGATGGC-3′.

Amplification products for γ- and δ-chains were cloned into murine stem cell virus–based vectors with resistance genes for zeocin and neomycin, respectively. These retroviral vectors, as well as the pCL-Eco accessory plasmid, were cotransfected using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA) into Phoenix cells. After 48 h, the transfected Phoenix cell supernatant was collected and used to transduced αβTCR-deficient T cell hybridoma 5KC-73.8.20 (kindly provided by Dr. E. Huseby, University of Massachusetts [UMass] Medical School, Worcester, MA) by spinfection at 1800 rpm for 2 h at room temperature (17). T cell hybridoma transfectants were selected using 0.5 mg/ml neomycin and 0.5 mg/ml of zeocin for 5 d.

T cell hybridoma transfectants (2 × 10⁶–4 × 10⁶) were treated with anti-mouse CD16/CD32 (2.4G2; Bio X Cell, Walpole, NH) to block Fc receptors and washed and stained with FITC anti-mouse TCR Vγ1.1/Cr4 clone 2.11 (BioLegend, Dedham, MA), PE anti-mouse Vδ6.3/2 clone 8F4H7B7 (BD Biosciences, San Jose, CA), and LIVE/DEAD Violet (Thermo Fisher Scientific) for 20 min at 4°C. Cells were acquired on an LSR II (BD Biosciences) flow cytometer, and data were analyzed using FlowJo software (version 9.7.5; Tree Star, Ashland, OR). For cell sorting, 7 × 10⁶–9 × 10⁶ cells were stained as described above, and FITC and PE double-positive cells were sorted on a FACSAria II (BD Biosciences).

Immortalized macrophages from wild-type C57BL/6 mice were generated using bone marrow–derived macrophages as previously described (18) and then immortalized by treatment with supernatant from NIH J2 Leukocyte cells, kindly provided by Dr. K. L. Rock (UMass Medical School, Worcester, MA) (19). Peritoneal cavity macrophages from wild-type C57BL/6 mice were elicited by injection of 3 ml 1% thioglycollate solution i.p., followed by peritoneal lavage and collection at 72 h. Immortalized macrophages or peritoneal cavity macrophages were plated in 96-well plates at 7 × 10⁴ or 1 × 10⁵ cells per well, respectively, in media without antibiotic and incubated at 37°C overnight. L. monocytogenes (10403s strain) was kindly provided by Dr. K. Fitzgerald (UMass Medical School, Worcester, MA). Listeria stock was thawed, and 200 μl was diluted in 10 ml of fresh media and incubated for 1.5 h in a shaking incubator at 37°C and 225 rpm. One milliliter of the bacteria culture was then centrifuged at 10,000 × g for 10 min at room temperature, and the pellet was resuspended in 1 ml PBS and was used to infect preplated macrophages at different multiplicities of infection (MOI). After addition of bacteria to each well, the plate was spun at 400 g for 5 min at room temperature and then incubated for 1 h at 37°C. Medium was then replaced with medium containing 100 μg/ml of gentamicin. After 3 h of incubation, the medium was removed, and 250 μl/well of medium alone or medium containing 1 × 10⁵ T cell hybridomas transfected with different γδTCRs (or untransfected or transfected with an αβTCR as controls) was added to the infected macrophages. After 16 h, IL-2 and IL-6 levels in the supernatant were tested by ELISA (BD Biosciences and Thermo Fisher Scientific, respectively).

In some experiments macrophages were not directly infected with Listeria but, instead, were treated with supernatant from an infected macrophage culture. In this case, macrophages were infected as described above and then incubated for 2 h with the media containing 100 μg/ml of gentamicin. Then, the medium was replaced with 180 μl of the medium used to culture hybridomas (Dulbecco's modified eagle medium, 1.2% mass-to-volume ratio [m/v] glucose, 21% v/v nonessential amino acids, 0.07% m/v gentamicin, 0.09% m/v penicillin G, 0.15% m/v streptomycin sulfate, and 0.005% 2-ME). After 16 h, the cells were spun and 100 μl of their supernatant was used to treat macrophages plated the night before in 96-well plates at 7 × 10⁴ cells per well, and 1 × 10⁵ T cell hybridomas expressing the different γδTCRs were added to a final volume of 250 μl/well. After 16 h, the IL-2 levels in the supernatant were tested by ELISA.

We wanted to express different variants of the Vγ1.1Vδ6.3 TCR to test their response to macrophages and macrophages infected with Listeria. We used a TCR Vδ6.3 sequence previously characterized for γδNKT cells isolated from the liver of wild-type mice (15). The TCR δ sequences observed in that study had restricted junctional diversity with only four different sequences observed, all using the same Dδ2 open reading frame and limited N nucleotides, all with the same overall CDR3 length. Out of the four sequences observed, we used the most highly represented one (88%) (Fig. 1A), arbitrarily called variant 10. To identify TCR Vγ1.1 γ sequences that could pair with the selected Vδ6.3 sequence, we characterized TCR Vγ1.1 sequences from thymus of adult βTCR knockout mice (Supplemental Fig. 1). Adult thymus γδNKT cells derive from fetal progenitors or from adult precursors, which remain as a thymic resident population (2, 20) and so exhibit heterogenous junctional diversity allowing us to characterize a range of TCR Vγ1.1 sequences. We chose five Vγ1.1 sequences, one with germline sequence called variant 10, and the other four selected based on differences in the CDR3 region, called variants 2, 3, 6, and 13 (Fig. 1B).

We transfected the five different γ-chains in combination with the δ-chain in TCR-deficient T cell hybridomas and tested which of these combinations could fold properly and form a heterodimeric TCR on the cell surface. For this purpose, we stained the transfected cells with fluorescent anti-Vγ1.1 and anti-Vδ6.3 Abs and measured the percentage of cells stained with both Abs by flow cytometry. When T cell hybridomas were transfected only with the δ-chain, they did not express TCR on the surface (Fig. 2A, top left panel). When T cell hybridomas were cotransfected with the Vδ6.3 variant 10 in combination with Vγ1.1 variants 2, 3, 6, or 10, most of the transfected cells bound both anti-Vγ1.1 and anti-Vδ6.3 Abs, indicating that they express TCR on the surface (Fig. 2). However, when Vδ6.3 variant 10 was cotransfected with Vγ1.1 variant 13 (Fig. 2A, bottom right panel), no TCR was detected, indicating that this combination could not fold properly and/or traffic to the cell surface. These results showed that out of all the possible γ/δ combinations tested, Vγ1.1 variants 2, 3, 6, and 10 paired with Vδ6.3 variant 10 (called 2-10, 3-10, 6-10, and 10-10 TCR variants, respectively) to form a heterodimeric TCR on the cell surface of transfected cells, and only Vγ1.1 variant 13 could not form a TCR with Vδ6.3 variant 10. It is possible that the valine-tryptophan to alanine-leucine substitution at the end of the Vγ1.1 sequence observed only in variant 13 could explain why this variant did not fold to form a TCR on the cell surface (Fig. 1B).

To test if the Vγ1.1Vδ6.3 TCRs expressed on the T cell hybridomas were functional and could trigger TCR signaling, we treated the cells with anti-Vγ1.1 clone 2.11 Ab, which was previously shown to activate the TCR (21). As an indicator of cell activation, we measured the level of IL-2 secreted to the medium using an ELISA. We used, as controls, hybridomas untransfected with any TCR (no TCR), hybridomas transfected with a control αβTCR, or hybridomas transfected with Vγ1.1 variant 13- Vδ6.3 variant 10 (13-10 TCR) that was shown not to form a TCR on the surface. We observed that the hybridomas expressing the four γδTCR variants were activated and secreted IL-2 when cultured in presence of the Vγ1.1 Ab (Fig. 3). But this Ab could not activate the hybridoma transfected with 13-10 TCR, control αβTCR, or hybridomas with no TCR transfected (Fig. 3). We also treated cells with an anti-CD3 Ab, which should activate cells expressing any type of TCR on the surface. As expected, this treatment was able to activate and induce IL-2 secretion from all the hybridomas expressing a TCR on the cell surface (2-10, 3-10, 6-10, 10-10, and αβTCR). These results indicate that the four pairs of Vγ1.1 and Vδ6.3 variants formed a functional TCR on the surface, all of them with different CDR3 regions, and one encoding a germline TCR (10-10).

Once we knew which Vγ1.1 and Vδ6.3 variants could pair and get expressed on the cell surface, and that those TCRs were functional, we sorted the cells expressing high levels of 2-10, 3-10, 6-10, and 10-10 TCR out of the total T cell hybridoma transfectants. After sorting the cells, they were assayed by flow cytometry, and more than 90% of the hybridomas were stained positive with an anti-Vδ6.3 Ab. We used the sorted cells for further experiments.

In a previous report was shown that a Vγ1.1Vδ6.3 TCR clone, BNT 19-8, obtained from newborn murine thymus (22) was able to stain peritoneal cavity macrophages when used as a fluorescently labeled octamer. Furthermore, the staining increased if the macrophages were isolated from a mice previously infected with Listeria (13, 16). Taking these results into account, we wanted to test if the Vγ1.1Vδ6.3 TCR variants could also recognize ligands on macrophages. For this purpose, we cocultured the Vγ1.1Vδ6.3 TCR⁺ hybridomas with immortalized macrophages and assayed TCR activation as the level of secreted IL-2. We observed a general increase in the secretion of IL-2 by the Vγ1.1Vδ6.3 TCR⁺ hybridomas when they were cultured in presence of uninfected immortalized macrophages compared with when they were cultured alone (Fig. 4A). Hybridomas expressing 2-10 TCR exhibited a relatively high basal level of IL-2 secretion when they were cultured alone, which was similar to the high levels of IL-2 in presence of uninfected immortalized macrophages, making these two values not significantly different. For all the other hybridomas expressing the 3-10, 6-10, and 10-10 TCR variants, we observed a significant increase in IL-2 secretion when they were cultured in presence of uninfected immortalized macrophages, compared with when they were cultured alone (Fig. 4A).

To test if the Vγ1.1Vδ6.3 TCR⁺ hybridomas would respond to Listeria infection, we cultured the hybridomas expressing different TCRs with immortalized macrophages previously infected with different doses of Listeria (MOI of 0.1, 0.5, 1, and 5). The hybridomas expressing the four variants of Vγ1.1Vδ6.3 TCR were activated significantly more strongly with macrophages infected with Listeria as compared with uninfected macrophages, in a dose-dependent manner (Fig. 4A). The activation was dependent on the γδTCR expression, because cells not expressing TCR (transfected with 13-10 or untransfected) or expressing an αβTCR did not secrete IL-2 when cultured with macrophages either uninfected or infected with Listeria (Fig. 4B). In some cases, we observed that immortalized macrophages infected with the highest dose of Listeria were not able to activate the Vγ1.1Vδ6.3 TCR⁺ further than uninfected immortalized macrophages. This result can be explained by our observation of increased cell death for macrophages infected at MOI of 5.

To confirm that the macrophages were infected by the Listeria, we monitored macrophage activation after treatment with different Listeria doses using the level of IL-6 secreted as a readout of macrophage response to infection. We observed that the levels of secreted cytokine increased with increasing doses of Listeria, indicating that the doses we were using were not saturating (Fig. 4C). These results indicate that the Vγ1.1Vδ6.3 TCR⁺ hybridomas could recognize a TCR self-ligand present in macrophages and that ligand was modified or upregulated after Listeria infection.

Previous work on a different γδ TCR (human Vγ4Vδ5 clone) showed that it recognized the endothelial protein C receptor (EPCR), but only on transformed cells or on cells infected with CMV, indicating that, in that case, the γδ T cell response required a particular cellular environment (23). We wanted to investigate whether the Vγ1.1Vδ6.3 TCR⁺ hybridomas were similarly dependent on transformation or whether they could also be activated by primary cells. Thus, we repeated the experiments shown in Fig. 4 using primary macrophages isolated from the peritoneal cavity instead of an immortalized macrophage cell line.

Hybridomas expressing the Vγ1.1Vδ6.3 TCRs secreted higher levels of IL-2 when cultured in presence of uninfected macrophages isolated from wild-type B6 peritoneal cavity than when cultured alone without macrophages (Fig. 5A). The amount of IL-2 secreted by the hybridomas was even higher when the macrophages were previously infected with Listeria (Fig. 5A). The activation of the hybridomas depended on the γδTCR expression, because cells not expressing a TCR or expressing an αβTCR did not secrete IL-2 when cultured with uninfected or infected macrophages from the peritoneal cavity (Fig. 5B). Levels of IL-6 secreted by the macrophages treated with the different Listeria doses reached a maximum after the second dose of Listeria (Fig. 5C), indicating that the Listeria doses we used had a stronger effect on infecting and/or activating the macrophages isolated from the peritoneal cavity than the immortalized macrophages (Fig. 4C). This could also explain our observation that macrophages infected with doses higher than MOI of 0.5 do not activate the hybridomas more strongly than uninfected macrophages, suggesting that these peritoneal-derived macrophages could be killed by Listeria at lower doses than the immortalized macrophages. From these results, we can conclude that Vγ1.1Vδ6.3 TCR with different CDR3 regions can recognize a ligand present in primary or transformed macrophages.

In the experiments described above, we observed that the hybridomas carrying different Vγ1.1Vδ6.3 TCR variants appeared to be activated to different extents when cultured with infected macrophages, but there was high variability experiment to experiment, and the hybridomas also were activated to different extents after anti-TCR Ab treatment. To compare the responses induced by the different TCRs, we normalized the data using the IL-2 secretion level of each Vγ1.1Vδ6.3 TCR⁺ hybridoma cultured with anti-Vγ1.1 Ab, which was measured in every experiment as a way to account for differences in the TCR expression level and TCR signaling among different hybridomas. When the level of IL-2 secreted by hybridomas cultured with macrophages was expressed as a percentage of the IL-2 secreted when cultured with the anti-Vγ1.1 Ab, we observed that the 3-10 TCR variant induced much weaker activation of the hybridomas than did the other three TCR variants (Fig. 6). This difference was observed for hybridomas treated with uninfected macrophages (Fig. 6), or with macrophages infected with four different doses of Listeria (MOI of 0.1, 0.5, 1, and 5). The variants 2-10, 6-10, and 10-10 were not substantially different to each other in terms of the normalized activation they can induce when expressed in the hybridomas. Because the hybridomas differ only in their CDR3γ sequences, we conclude that the CDR3 sequence of the γ-chain is likely to play a role in the Vγ1.1Vδ6.3 TCR recognition of its ligand.

We wanted to test whether Listeria infection directly induced macrophages to activate the Vγ1.1Vδ6.3 TCR⁺ hybridomas or, alternatively, whether infected macrophages secreted a soluble factor able to induce uninfected macrophages to activate the Vγ1.1Vδ6.3 TCR⁺ hybridomas. For this purpose, we performed an experiment in which we cultured the more sensitive Vγ1.1Vδ6.3 TCR⁺ hybridomas expressing the 2-10, 6-10, and 10-10 TCR variants with immortalized macrophages either uninfected or infected with different doses of Listeria, as before, and compared these results to Vγ1.1Vδ6.3 TCR⁺ hybridomas cocultured with immortalized macrophages treated with supernatants from separate cultures of immortalized macrophages infected with different doses of Listeria collected 16 h after the Listeria infection. As observed previously, the hybridomas expressing different Vγ1.1Vδ6.3 TCR variants were activated slightly by uninfected immortalized macrophages, with activation levels higher in a dose-dependent manner when the macrophages were infected with Listeria (Fig. 7, solid bars). When macrophages were treated with supernatants from Listeria-infected cultures, instead of being directly infected, and then used to activate the Vγ1.1Vδ6.3 hybridomas, activation levels overall were much lower, and no dose-dependent activation was observed (Fig. 7, striped bars). Hybridomas expressing the 6-10 and the 10-10 TCR variants (Fig. 7, middle and bottom panels, respectively) were not activated at all by immortalized macrophages treated with supernatant from Listeria-infected macrophage cultures. Hybridomas expressing the 2-10 TCR variant (Fig. 7, top panel) showed a low level of activation with the supernatant of the Listeria-infected macrophages (Fig. 7 top, striped gray bar), which, however, was comparable to the activation achieved when cultured with untreated uninfected macrophages in the same experiment (Fig. 7 top, solid gray bar). Taken together, these results indicate that the Vγ1.1Vδ6.3 TCR⁺ hybridomas are activated efficiently by macrophages directly infected with Listeria and are activated much less, if at all, by uninfected macrophages activated by soluble factor(s) released from Listeria-infected macrophages.

In this study, we investigated ligand recognition by different variants of the Vγ1.1Vδ6.3 TCR from γδNKT cells. All of the γδTCR variants we studied shared the same δ sequence and had different γ junctional regions. We observed that TCR-deficient hybridomas transfected with the various Vγ1.1Vδ6.3 TCR became activated when cultured in presence of macrophages and that this activation increased when the macrophages were previously infected with Listeria. Activation was specific to the Vγ1.1Vδ6.3 TCR, because hybridomas not expressing any TCR or expressing an αβTCR did not show activation when cultured in presence of macrophages. These results expand on aspects of previous work on the Vγ1.1⁺Vδ6.3⁺ T cell population and its role in recognizing Listeria infection (10–13).

In previous work, it was shown that Vγ1.1Vδ6.3 TCR could recognize macrophages from Listeria-infected mice (10, 12, 13). However, it was not clear if the macrophages needed to be directly infected or if the Listeria- infected environment induced changes in the macrophages that allowed for Vγ1.1Vδ6.3 TCR recognition. In our studies, we infected the macrophages in vitro with cultured Listeria, using the level of IL-6 secretion to monitor the degree of macrophage infection. We found that direct infection of the macrophages with Listeria was necessary for efficient activation of the Vγ1.1Vδ6.3 TCR⁺ hybridomas, as macrophages treated with the supernatants of infected macrophages were not able to activate the Vγ1.1Vδ6.3 TCR⁺ hybridomas, ruling out an important role in this in vitro system for soluble factors secreted by the Listeria-infected macrophages in upregulating or generating the Vγ1.1Vδ6.3 TCR ligand(s). These do not, however, rule out the possibility that other cell populations or soluble factors could be involved in vivo.

Several potential mechanisms could be envisioned to explain how uninfected macrophages could induce activation on the Vγ1.1Vδ6.3 TCR⁺ hybridomas, with activation increased when the macrophages were infected with Listeria. One possibility is that the Vγ1.1Vδ6.3 TCR recognizes a self-ligand that is expressed at low levels on resting macrophages, with surface expression of the self-ligand increased after Listeria infection. This mechanism has been observed for murine G8 and KN6 γδTCR clones that recognize T10/T22 (24, 25), which are MHC class I protein–like molecules that do not bind or present Ags (26). The expression level of T10 and T22 increases in activated lymphocytes, and it has been suggested that this modulates the activation of γδT cells expressing the G8 and the KN6 receptors (27). A second possibility is that this TCR recognizes a self-ligand that binds a Listeria-derived molecule upon infection, causing an increase in the affinity between the TCR and the ligand. This type of mechanism has been observed for γδT cells recognizing CD1d/lipid complexes, in which the TCR affinity for CD1d increases when it is presenting a specific lipid derived from bacteria (28–30), and it also appears to be the mechanism for a human Vγ9Vδ2 TCR that recognizes an allosteric change on the A1 isoform of the BTN 3 in response to intracellular accumulation of phosphoantigens in infected or transformed cells (31, 32). A third possibility involves both TCR recognition of a self-ligand present in uninfected macrophages and a second ligand induced by Listeria infection that either favors TCR recognition of the self-ligand or promotes activation of the hybridomas by interacting with a different receptor on these cells. This appears to be the mechanism for the human LES Vγ4Vδ5 TCR that recognizes the CD1d-like lipid-binding protein EPCR in certain tumor cell lines and in CMV-infected fibroblasts (23). Further studies would be required to determine whether Vγ1.1Vδ6.3 TCR recognizes Listeria infection by upregulation or modification of a self-ligand, induction of additional ligand(s), or another mechanism.

Previous work showing that the Vγ1.1Vδ6.3 TCR has a role in recognizing Listeria infection (10–13) did not address the role of the junctional regions in ligand recognition. Previous studies analyzed only the variable gene usage, but not CDR3 sequence dependence for γδT cells responding to Listeria infection (11, 12), or investigated only individual Vγ1.1Vδ6.3 TCR sequences expressed on hybridomas or as soluble oligomers (10, 13). Thus, is was not known if all Vγ1.1Vδ6.3 TCRs can recognize macrophages from Listeria-infected animals, or whether some particular CDR3 sequences are required. In this study, we showed that the four Vγ1.1Vδ6.3 TCR variants, formed by the combination of a germline-encoded Vδ6.3-Dδ2-Jδ1 chain with four Vγ1.1-Jγ4 variants with different junctional regions, were able to recognize macrophages and macrophages infected with Listeria. However, not all variants were equally activated in presence of the macrophages. One of the variants, called the 3-10 variant, which had a serine at the beginning of the Jγ4, showed a significantly weaker activation in response to macrophages infected with Listeria than did the other three variants, which all have a glutamic acid in that position. It is possible that differences in the CDR3γ sequence among the different Vγ1.1Vδ6.3 TCR variants studied in this study translate into different TCR affinities for the ligand. This was also shown for mouse γδTCRs that specifically recognized T22 molecules for which differences in their CDR3δ region translated to different affinities for T22 (33, 34). We conclude that the CDR3γ is likely to be involved in the ligand recognition of these cells. Furthermore, one of the TCR variants (10-10) was completely germline derived, suggesting that both fetal and adult-derived Vγ1.1Vδ6.3 TCRs can recognize the same Listeria-modulated ligand.

Crystal structures determined for three other γδTCRs in complex with their ligands show a limited role for the CDR3γ in ligand recognition (28, 30, 35). These γδTCRs dock differently on their ligand but all sit with a tilted angle on top of the ligand, resulting in an interaction mode that involves mainly the δ-chain. This interaction mode is different from the conventional diagonal αβTCR/MHC interaction, in which both CDR3α- and CDR3β-chains typically interact with the MHC-peptide binding surface (36). For the murine G8 γδTCR in complex with T22, the main interaction is between CDR3δ loop and the empty cleft of the T22 protein, followed by the interactions involving residues from the CDR1δ and CDR2δ loops. The CDR3γ loop it is also involved in the T22 recognition, but it only contributes 11% of the buried surface area (35). For the human Vγ5Vδ1 TCR in complex with CD1d/αGalCer (30), CDR1δ and the CDR3δ contributed most of the TCR ligand interaction, whereas only one residue from the γ-chain contacted CD1d. For the human Vγ4Vδ1 TCR in complex with CD1d/sulfatide (28), CDR1δ, CDR2δ, and CDR3δ loops were evenly involved in the interactions with CD1d and the sulfatide, whereas the γ-chain did not show any interactions with the ligand. Further studies will be required to establish how the Vγ1.1Vδ6.3 TCR recognizes its ligand, and the participation of other parts of Vγ1.1 and Vδ6.3 chains in addition to the CDR3γ loop suggested in this study.

The findings reported in this study support the idea that γδNKT TCRs formed by a Vγ1.1Vδ6.3 TCR recognize a self-ligand present in macrophages that is upregulated or modified on macrophages infected by Listeria, causing an increase on the γδNKT TCR activation. The TCRs that recognize this ligand can originate from fetal, as well as from adult, γδT cell development, with the CDR3γ region likely involved in ligand recognition.

We thank Eric Huseby (UMass Medical School) for providing the TCR-deficient T cell hybridoma, 5KC-73.8.20, and a control αβTCR transfectant; Kenneth Rock (UMass Medical School) for providing immortalized C57BL/6 macrophages; Katherine Fitzgerald (UMass Medical School) for providing the Listeria stock; and anonymous reviewers for helpful suggestions.

The authors have no financial conflicts of interest.