mAb targeting the γδ TCR have been used for γδ T-cell depletion with varying success. Although the depletion-capacity of the anti-γδ TCR mAb clone GL3 has been disputed repeatedly, many groups continue to use γδ T-cell depletion protocols involving the mAb clone UC7-13D5 and find significant biological effects. We show here that treatment with both GL3 and UC7-13D5 antibodies does not deplete γδ T cells in vivo, but rather leads to TCR internalization and thereby generates “invisible” γδ T cells. We addressed this issue using anti-γδ TCR mAb injections into WT mice as well as into reporter TCR delta locus-histone 2B enhanced GFP knock-in mice, in which γδ T cells can be detected based on an intrinsic green fluorescence. Importantly, the use of TCR delta locus-histone 2B enhanced GFP mice provided here for the first time direct evidence that the “depleted” γδ T cells were actually still present. Our results show further that GL3 and UC7-13D5 mAb are in part cross-competing for the same epitope. Assessed by activation markers, we observed in vitro and in vivo activation of γδ T cells through mAb. We conclude that γδ T-cell depletion experiments must be evaluated with caution and discuss the implications for future studies on the physiological functions of γδ T cells.
γδ T cells constitute a distinct lineage of T lymphocytes. Many T cells reside in mucosal tissues in close contact with epithelial cells. There, they often represent the majority of the T-cell pool and have tissue-specific oligoclonal TCR repertoires. Circulating γδ T cells show a more diverse TCR repertoire, but make up only 0.5–3% of all T cells in the blood and in secondary lymphoid organs such as lymph nodes and spleen. Both mucosal and circulating γδ T cells seem to play specific roles in the regulation and execution of host responses to many pathogens 1–3. Recently, γδ T cells have received an increasing attention as main producers of the pro-inflammatory cytokine IL17A 4–7.
In contrast to the progress made in understanding their counterparts of the αβ T-cell lineage, the biological role of γδ T cells is only slowly emerging. Research on γδ T cells must cope with various obstacles: (i) currently, information on γδ TCR ligands is still very limited; (ii) in the mouse model and in the human system there is no other γδ T-cell specific marker than the γδ TCR itself; and (iii) the functional phenotype of TCR delta locus (Tcrd)-deficient (Tcrd–/–) mice 8 may be concealed by opportunistic αβ T cells that occupy the niches of genuine γδ T cells and might partially take over their functions 9. However, a popular approach to overcome the latter issue is to deplete γδ T cells in WT mice with mAb directed against γδ TCR instead of using the Tcrd–/– model. To our knowledge, four different “classical” mAb directed against epitopes in the constant part of the mouse γδ TCR have been generated, namely the mAb UC7-13D5 10, 3A10 11, GL3 12 and GL4 12. Interestingly, all anti-TCR γδ mAb are Armenian hamster IgG.
To date, these mAb have been used for γδ T-cell depletion with varying success. Although the depletion-capacity of GL3 has been disputed repeatedly 13, 14, many groups continue to use γδ T-cell depletion protocols involving UC7-13D5 15–19. In most of the published studies, effective γδ T-cell depletion with UC7-13D5 was confirmed by surface staining using conjugated GL3 mAb. However, those two anti-γδ TCR mAb may still compete for the same epitope and therefore such validation may be questionable. In this study, we sought to settle the issue once and for all by using reporter knock-in mice in which γδ T cells can be readily detected based on their intrinsic green fluorescence and thus independent of TCR expression and TCR accessibility. We applied anti-γδ TCR mAb at different time points, doses and inoculation routes and analyzed the intra- and extracellular TCR γδ expression as well as the depletion efficiency of mucosal and circulating γδ T cells directly ex vivo or after in vitro culture.
Efficiency of mAb-mediated γδ T-cell depletion in WT mice
To determine the in vivo effects of GL3 and UC7-13D5 mAb, they were purified from hybridoma supernatants and various amounts were injected into WT C57BL/6 mice. We used doses ranging from 100 to 400 μg per 6–10-wk-old mouse and after 3 days we analyzed the expression of T-cell markers on splenocytes and intestinal intraepithelial lymphocytes (IEL) by flow cytometry. The presence of γδ T cells was assessed by staining with directly labeled GL3 mAb. Independent of the mAb dose and inoculation route, we observed in all conditions a strong reduction in circulating and mucosal γδ T cells as assessed by TCR γδ staining. Therefore, we used a dose of 200 μg i.p. injection in all subsequent experiments. Figure 1A shows a representative result for IEL from mice that received 200 μg of UC7-13D5 mAb i.p. This result is in accordance with published data. In mice, injected i.p. with 200 μg of GL3 mAb, the apparent reduction in detectable γδ T cells was even more pronounced (Fig. 1B). However, this finding does not necessarily mean that all γδ T cells were depleted, as circulating unlabeled GL3 may still be present and have simply blocked the binding of the same fluorescently labeled antibodies. In 6–10-wk-old WT C57BL/6 mice, the IEL compartment typically contains 40% αβ T cells and 45% γδ T cells. Conspicuously, 2 days after mAb injection we found no significant reduction in the frequency of TCR-β– IEL (46.0 versus 51.8%, respectively; Fig. 1B). Of the γδ IEL, approximately 75% are expressing CD8-αα homodimers. As shown in Fig. 1C, the proportion of CD8-α+ cells among the CD8-β– TCR-β– gated IEL was not diminished after mAb treatment (55.2 versus 67.9%, respectively), suggesting that CD8-αα+ γδ T cells may still be present although not detectable by GL3 staining. Next, we analyzed the effects of GL3 and UC7-13D5 mAb injection into WT C57BL/6 mice over a period of 14 days. At all time points, γδ T cells remained undetectable by staining with directly labeled GL3 mAb in IEL preparations (Fig. 2A) and in splenocytes (data not shown). However, at later time points the frequency of TCR-β– IEL decreased gradually from 40 to 50% in untreated mice to less than 30% 14 days after mAb injections (Fig. 2B), suggesting that the total amount of γδ T cells was in fact partially diminished. The proportion of CD8-αα+ cells among the TCR-β– IEL was also reduced at day 14 after mAb injections (Fig. 2C), indicating a potential loss of γδ T cells in the IEL compartment. It is conceivable that such late effects of mAb administrations are not mediated by complement lysis, but most likely through activation-induced γδ T-cell death. In all our preparations, the absolute numbers of recovered IEL did not show a significant trend to fewer cell counts after mAb treatment. Using our protocols, we constantly obtained 3–6 million IEL per small intestine, although extrapolations from IEL counts in histological sections would predict a much higher number of CD3+ IEL. Furthermore, compensatory expansion of specific TCR-β+ subsets may add to the complexity of the situation 20. It is thus difficult to draw solid conclusions on the effective loss or increase of IEL subsets from these data.
Analysis of Tcrd-H2BEGFP reporter mice
The above studies suggested that mAb γδ T-cell depletion is very inefficient, if it occurred at all. In order to elucidate this issue in a different approach, we next applied anti-TCR γδ mAb to Tcrd-histone 2B enhanced GFP (H2BEGFP) reporter mice. This knock-in comprises an H2B–eGFP fusion protein into the Tcrd locus under the control of an internal ribosomal entry site element, thereby leaving the Tcrd gene fully functional. In this model, γδ T cells can be identified by their intrinsic intra-nuclear fluorescence independent of TCR γδ expression 21. Notably, we found no reduction in γδ T cells as detected using EGFP after injection of GL3 or UC7-13D5 (Fig. 3A), although γδ T cells in mAb-treated animals could no longer be detected with conjugated GL3. Owing to the GFP expression in γδ T cells, we were able to further characterize the effects of the different mAb on the cells. Reduction in surface TCR γδ complexes as assessed by anti-CD3 staining was always more pronounced in GL3-treated mice than in UC7-13D5-treated mice when compared with mice that received mock injections (Fig. 3A). Similar results were obtained for splenic γδ T cells (data not shown). However, the surface TCR γδ complex reduction in mAb-treated mice was rather due to TCR internalization than TCR degradation as shown by comparison of intra- and extracellular CD3 levels (Fig. 3A), where the latter remained unaffected by mAb injections. In contrast to extracellular staining, intracellular staining of γδ T cells with secondary (anti-Armenian hamster) mAb against anti-TCR γδ mAb gave similar signal intensities in GL3- and UC7-13D5-injected mice (Fig. 3B), thus pointing at a more pronounced TCR internalization in vivo in GL3-treated mice. The suggestion that the γδ TCR after GL3 treatment is internalized but still present was further supported by the detection of strong fluorescent signals in γδ T cells after the in vivo application of GL3 mAb directly conjugated with Cy5 (Fig. 3C). Taken together, our analysis of Tcrd-H2BEGFP reporter mice demonstrated that anti-γδ TCR mAb injections did not reduce γδ T-cell numbers, but rather rendered the γδ T cells invisible due to γδ TCR internalization. However, GL3 was more effective than UC7-13D5 in inducing a downregulation of the TCR complex from the cell surface.
To test whether the seeming discrepancy of surface γδ TCR levels as assessed by either GL3 or anti-CD3ε surface staining (Fig. 3A) could be explained by at least partial epitope masking of the two mAb, we performed symmetrical experiments to determine the level of cross-inhibition of the two mAb. We assumed that injection of 200 μg of mAb per mouse resulted in effective concentrations of not more than 10 μg/mL in vivo. We found that in vitro pre-incubation of γδ T cells with concentrations starting from 1 μg/mL of GL3 decreased the MFI of subsequent staining with 10 mg/mL of the corresponding UC7-13D5 mAb (Fig. 4A). However, pre-incubation of γδ T cells with as much as 100 μg/mL of UC7-13D5 only slightly decreased the MFI of subsequent staining with 10 mg/mL of the corresponding GL3 mAb (Fig. 4B). At the same time, considerable TCR internalization through GL3 or UC7-13D5 pre-incubation was not observed as assessed by CD3 staining (Fig. 4A and B). We conclude that GL3 and UC7-13D5 likely recognize the same epitope, whereby GL3 shows a drastically higher affinity.
Activation of γδ T cells with mAb
TCR downregulation is a hallmark of T-cell activation. In γδ T cells, this is much more pronounced than in αβ T cells and can lead to a complete disappearance of TCR complexes from the cell surface of activated γδ T cells . We therefore sought to further analyze the effects of mAb directed against components of the TCR γδ complex in Tcrd-H2BEGFP reporter mice. Twenty-four hours in vitro cultivation of sorted γδ T cells on plate-bound GL3, UC7-13D5 or anti-CD3ε uniformly induced T-cell activation as indicated by the upregulation of the activation markers CD69 and CD44 (Fig. 5A). The same activation was not seen when soluble CD3ε, GL3 or UC7-13D5 mAb were added directly into the medium for 24 h (data not shown), suggesting that γδ TCR triggering depends on receptor clustering. When injected in vivo, the constant parts of the mAb GL3 and UC7-13D5 may be bound by Fc-receptors on interacting cells and thus be able to activate γδ T cells through their TCR. In fact, we observed upregulation of the activation marker CD44 on γδ T cells when GL3 or UC7-13D5 mAb were present in vivo (Fig. 5B). However, upregulation of CD69 was seen only in GL3-treated γδ IEL (Fig. 5B). In vivo activation of γδ T cells with GL3 has been suggested already by Kaufmann et al. 13. Although the authors could not directly detect GL3-activated γδ T cells, they could show reappearance of the TCR γδ complex after a few days of in vitro culture 13. In our Tcrd-H2BEGFP reporter system, we could confirm those data and found the same results for UC7-13D5-injected mice (Fig. 6). Collectively, these data indicate that at the time of analysis, the injected mAb were still present in the circulation of the mice. In conclusion, our data provide strong evidence that injection of either GL3 or UC7-13D5 mAb does not deplete but renders γδ T cells undetectable by anti-γδ TCR staining (invisible γδ T cells) and at the same time may induce γδ T-cell activation.
In this study, we make the point that neither GL3 nor UC7-13D5 mAb injection into the circulation of mice leads to γδ T-cell depletion. This clear result was obtained for circulating and mucosal γδ T cells in a wide range of experimental conditions that cover typical protocols employed for “γδ T-cell depletion”. Importantly, the use of Tcrd-H2BEGFP mice provided here for the first time a direct proof that the “depleted” γδ T cells were actually still present. This is a particular pity, since γδ T-cell depletion avoids the limitations of the Tcrd knock-out model in which opportunistic αβ T cells can occupy γδ T-cell niches and may partially substitute the functions of genuine γδ T cells 9. Nevertheless, many recent studies found interesting effects after administration of anti-γδ TCR mAb, mainly using the UC7-13D5 hybridoma 15–19. Moreover, several studies found opposing effects of γδ T-cell depletion and adoptive transfer of γδ T cells 17, 22 as well as similar experimental outcomes in Tcrd–/– and γδ T-cell-depleted mice 17, 23, strongly suggesting functional relevance of the mAb administration. However, based on our data it is clear that the effects reported were not due to the absence of γδ T cells. Therefore, the most likely scenario would be a “functional depletion” of γδ T cells through a block of TCR signaling with UC7-13D5 mAb or through GL3-mediated TCR downregulation. In this context it is conceivable that γδ T cells with no detectable TCR levels on their cell surface are poor responders to TCR-specific stimulation. An alternative, albeit unlikely, interpretation would be that already the depletion of only a minor fraction of γδ T cells, as observed in our time course experiments, is sufficient to produce significant effects on the experimental outcomes.
In addition, it is important to mention that the mAb studied here are pan γδ TCR reactive, and hence they target a variety of different γδ T-cell subsets. As different as these subsets may be, the potential effects of their stimulation may range from pro-inflammatory to regulatory/anti-inflammatory. For example, it was recently described that circulating Vγ1+ and Vγ4+ cells can play opposing roles in different mouse models for human pathologies 4, 24, 25. In this study, we analyzed the depletion potential of two commonly used anti-TCR γδ mAb in depth. It remains to be validated whether other mAb, in particular, Vγ/Vδ specific ones can serve as more potent and even subset specific depletion tools. In addition, two new anti-γδ TCR mAb clones named 3H2928 and 5K24 (Santa Cruz Biotechnology, CA) have recently been introduced. However, they are again Armenian hamster IgG, which somehow may be regarded as a discouraging sign for their potential depletion capacity.
In the field of γδ T-cell research, mAb-based strategies of γδ T-cell depletion in vivo have been termed “a poor man's Tcrd knock-out”, because it was never clear whether the γδ T cells were really depleted. Here we demonstrate that those depleted γδ T cells are in fact only “invisible”. Therefore, such regimens must be used with caution and should be complemented with additional experiments whenever possible. Future studies aiming at depleting γδ T cells to analyze their physiological functions should therefore consider other mAb or the generation of genetic models for inducible γδ T-cell ablation.
Materials and methods
Animals were bred at the central animal facility of Hannover Medical School under specific pathogen-free conditions, or purchased from Charles River (Germany). The Tcrd-H2BEGFP model has been described previously 21. Mice were handled in accordance with German and European directives. All mouse protocols were approved by the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit. Analyses used 6–10-wk-old mice.
The following antibodies and conjugates were used in this study: anti-CD45.2-PerCP-Cy5.5, anti-CD3-PE-Cy7 (145-2C11), anti-TCR-γδ-PE (clone GL3), anti-CD69-PE, anti-CD69-PerCP-Cy5.5 (BD Bioscience), anti-CD3-PE (145-2C11), anti-TCR-γδ-biotin (clone UC7-13D5), anti-Armenian hamster IgG-bio, anti-TCR-αβ-APC-Alexa750 and anti-CD44-Pacific Blue (eBioscience). Anti-CD8-α-Cy5 (clone 53–6.7) and anti-CD8-β-Pacific Orange (clone RmCD8-2) antibodies were provided by Elisabeth Kremmer (GSF München, Germany). Biotinylated antibodies were recognized by streptavidin coupled to PE-Cy7 (BD Bioscience). The GL3 clone was originally described in 12 and the UC7-13D5 clone in 10. Both hybridoma were kindly provided by Stefan H. Kaufmann (Max-Planck-Institute for Infection Biology, Berlin, Germany).
To obtain single-cell suspensions, spleen cells were meshed through a nylon mesh and washed with PBS supplemented with 3% FBS. Erythrocytes were removed by hypotonic lysis. For isolation of IEL, gut content and Peyer's patches were removed before intestines were opened longitudinally. Intestines were washed twice in cold PBS and once in cold PBS/5% FBS/5 mM EDTA, and incubated twice in 25 mL RPMI 1640 medium/5% FBS/5 mM EDTA at 37°C (spun down at 150 rpm). Supernatants were pooled, filtered through a nylon mesh, pelleted and resuspended in 40% Percoll (Amersham) in RPMI/5% FBS. This cell suspension was overlaid onto 70% Percoll in RPMI/5% FBS and centrifuged at 800g for 20 min. IEL were recovered from the interphase and washed twice in PBS/3% FBS. Cells were stained using the antibodies described above. For intracellular staining, surface epitopes were stained prior to fixation of the cells in PBS/3% FBS for 20 min on ice and then stained using a fixation/permeabilization kit (BD Pharmingen). The cells were analyzed on an LSRII (BD Biosciences) and the data evaluated using FlowJo-Software (Version 8.7.1, TreeStar).
γδ cells were sorted (FACS Aria, BD Bioscience) and cultured in T-cell medium (RPMI-1640) with L-glutamine (2 mM) supplemented with pyruvate (1 mM), 100 U/mL of penicillin, 100 μg/mL of streptomycin (all GIBCO, USA), IL-2 (300 IU/mL) and 10% FBS at 37°C, 95% humidity and 5% CO2, in 24 well or 96 welly round bottom plates. In some experiments, plastic dishes were pre-coated with PBS containing 5 μg/mL of anti-CD3, GL3 or UC7-13D5 for 90 min at 37°C. Cells were analyzed after 18–72 h.
C.K. is a scholar of the Deutsche Forschungsgemeinschaft (grant KO 3582/1-1). This work was supported by grants from the Deutsche Forschungsgemeinschaft PR727/2-1 (I.P.), the European Community ERG 046578 (I.P.) and from the Université franco-allemande G2R-FA-104-07 (I.P. and B.M.). We thank Jasmin Bölter for excellent mAb purification, Andreas Krueger and Markus Gräler for advice and for carefully reading the paper.
Conflict of interest: The authors declare no financial or commercial conflict of interest.