Direct Costimulatory Effect of TLR3 Ligand Poly(I:C) on Human γδ T Lymphocytes¹

A small fraction of peripheral blood CD3⁺ T cells expresses the TCRγδ instead of the conventional TCRαβ. γδ T cells recognize Ags without requirement for Ag processing and independently of classical MHC molecules (1, 2). In the blood of adult humans, a majority (50–95%) of γδ T cells expresses Vγ9Vδ2 TCR. These cells primarily recognize phosphorylated intermediates of the nonmevalonate pathway of the bacterial isoprenoid biosynthesis pathway (phosphoantigens) (3, 4) and thus sense microbial infection at a very early stage. After ligand recognition, γδ T cells rapidly release cytokines such as IFN-γ and TNF-α, thereby activating innate immune cells and facilitating adaptive immune responses by αβ T cells (1, 5). Therefore, it is assumed that γδ T cells may serve as protection against infections as a first line of defense before Ag-specific αβ T cells expand. Moreover, γδ T cells play a not precisely defined role during antiviral immunity. Increased numbers of γδ T cells are present in the peripheral blood of EBV-infected (Vδ2) or HIV-infected (Vδ1) individuals and CMV-infected patients (Vδ1) after kidney transplantation (6, 7, 8).

TLRs are pattern recognition receptors involved in the innate immune response to infection, which also contribute to the regulation of adaptive immune responses. To date, TLR have been considered to be expressed mainly in APC, including monocytes, macrophages, and dendritic cells (DCs).³ In APC, TLR ligands induce the up-regulation of costimulatory molecules, including CD80, CD86, and CD40 and the production of proinflammatory cytokines, thereby indirectly costimulating T cell activation (9, 10, 11, 12, 13). Recent studies suggested that γδ T lymphocytes are also stimulated indirectly via TLR-mediated activation of immature myeloid DC (via TLR3) or plasmacytoid DC (via TLR9) (14, 15).

TLR3 recognizes viral dsRNA and a synthetic analog, polyinosinic-polycytidylic acid (poly(I:C)), which both lead to the production of type I and type II IFN via NF-κB activation and IFN regulatory factor-3 (16, 17, 18, 19, 20). In addition to TLR3, the RNA helicase retinoic acid-inducible gene I (RIG-I) has recently been shown to interact with cytoplasmic dsRNA, leading to the activation of NF-κB and IFN regulatory factor-3 (21, 22). Furthermore, cellular mRNA functions as a host-derived TLR3 ligand (23). TLR3 was shown to be expressed in myeloid BDCA-1⁺ (CD1c⁺)/CD11c⁺ DC, NK cells, fibroblasts, and intestinal epithelial cells and more recently in human T lymphocytes using quantitative PCR (24, 25, 26, 27). Although TLR3 function and signal transduction have been well studied in nonlymphoid cells, the functional relevance of TLR3 in T cells is enigmatic.

In the present study we focused on the expression of TLR3 in αβ and γδ T cells and their functional responsiveness after stimulation with poly(I:C). Our results show that TCR-activated γδ T cells, but not αβ T cells, produce large amounts of IFN-γ when costimulated with poly(I:C).

γδ T cell clones and lines were established and cultured with occasional restimulation, as previously described (28). Vδ1 γδ T cell clones and lines were cultured in 96-well microculture plates coated with 0.5 μg/ml anti-CD3 mAb OKT3, and Vδ2 γδ T cell clones and lines were stimulated with optimal concentrations of bromohydrin pyrophosphate (BrHPP; 200 nM; provided by Innate Pharma), 0.5 μg/ml anti-CD3 mAb, or anti-Vγ9 mAb 7A5 (29). Both γδ T cell subsets were stimulated in the absence or the presence of different concentrations of poly(I:C) (Calbiochem/Merck or Amersham Biosciences) 10–12 days after restimulation in serum-free X-VIVO 15 medium (Cambrex BioScience). Proliferation was measured by uptake of [³H]TdR during the last 8 h of a 3-day culture period in the presence of 50 U/ml rIL-2 (Chiron). IFN-γ secretion was determined in the absence of IL-2 after 12 h using a commercially available ELISA kit (Quantikine; R&D Systems) according to the manufacturer’s instructions.

γδ T cells, αβ T cells, and CD14⁺ monocytes/macrophages were positively separated from freshly isolated PBMC using the MACS system (Miltenyi Biotec). To avoid examination of in vivo (pre)activated γδ T cells, care was taken to select blood donors in whom γδ T cells accounted for <6%. Briefly, PBMC were pretreated with Fc-blocking reagents (Miltenyi Biotec) to avoid unspecific binding of Abs to FcR-bearing cells. To isolate γδ T cells, PBMC were stained with haptenated anti-TCRγδ mAb, followed by anti-hapten microbeads-FITC; for αβ T cells, PBMC were incubated with anti-TCRαβ mAb BMA031, followed by goat anti-mouse IgG microbeads, and for isolation of CD14⁺ monocytes/macrophages, PBMC were incubated with CD14 microbeads. The purity of the positively selected cells was ≥98%. CD56⁺ NK cells were purified by negative selection procedures using NK Cell Isolation Kit II (Miltenyi Biotec). Immature DCs were generated from the adherent cell fraction of PBMC. Adherent cells were cultured in serum-free X-VIVO 15 medium (Cambrex BioScience) supplemented with IL-4 (1000 U/ml; R&D Systems) and GM-CSF (1000 U/ml; R&D Systems) for 4–5 days.

The following mAb were used to analyze the purity of the freshly isolated lymphocytes and monocytes/macrophages: anti-CD3, pan anti-TCRγδ, pan anti-TCRαβ, anti-CD56, anti-CD14, anti-CD11c, and anti-CD1c (all from BD Biosciences). To determine the CD69 expression on γδ T cells within PBMC, PE-labeled CD69 mAb and FITC-labeled pan anti-TCRγδ mAb (BD Biosciences) were used. The positively separated γδ T cells, which were labeled with haptenated anti-TCRγδ mAb, followed by FITC-conjugated anti-hapten microbeads during separation, were stained only with PE-labeled CD69 mAb. PE-conjugated anti-TLR3 mAb TLR3.7 (Biosciences) was used for surface and intracellular staining. For intracellular detection, cells were washed, fixed, and permeabilized with the Cytofix/Cytoperm kit (BD Pharmingen) following the manufacturer’s instructions. Thereafter, cells were stained with mAb TLR3.7 or isotype control, and measured on a FACScan or FACSCalibur flow cytometer (BD Biosciences) using CellQuest software. For laser scanning microscopy, established γδ T cell clones were allowed to adhere to poly-l-lysine-coated glass slides. The cells were fixed and permeabilized with 100% methanol (Merck), then washed extensively three times with PBS. After blocking with 0.5% BSA, cells were stained with unconjugated anti-TLR3 mAb TLR3.7 or isotype control for 1 h, followed by the second step Ab Alexa 546-conjugated goat anti-mouse (Invitrogen Life Technologies, Inc.) for 1 h. The stained cells were visualized at ×630 magnification with a confocal laser scanning microscope (Zeiss).

Positively isolated, highly purified γδ or αβ T cells (4 × 10⁶/well) were cultured in wells coated with rabbit anti-mouse Ig (to cross-link the TCR) in the presence or in the absence of 50 μg/ml poly(I:C) without IL-2 in serum-free X-VIVO 15 medium. Isolated T cells were also stimulated with poly(I:C) alone or via TCR cross-linking in the presence of 1 μg/ml soluble anti-CD28 mAb (BD Pharmingen). Supernatants were collected after 24 h and were stored at −20°C until use. Supernatants were screened for cytokines and chemokines using the RayBio Human Cytokine Ab Array VI and 6.1 Map, and VII and 7.1 Map (Hoelzel Diagnostic), which allow simultaneous detection of 2 × 60 cytokines and chemokines. Signals were detected by chemiluminescence, followed by semiquantitative analysis with AIDA software (Ray-Test). To determine the intensity, local background was subtracted from each value and normalized against the positive controls of each membrane. Normalization was performed by Excel calculation. To confirm and quantify the results of the cytokine array, IFN-γ was also measured by intracellular flow cytometry and ELISA (Quantikine; R&D Systems). For analysis of intracellular IFN-γ, 1 × 10⁶/well positively isolated γδ or αβ T cells were stimulated for 24 h with immobilized rabbit anti-mouse Ig with or without poly(I:C). During the last 4 h of stimulation, 3 μM monensin was added to the cultures. The cells were harvested, washed, fixed, and permeabilized using the Cytofix/Cytoperm kit according to the manufacturer’s instructions. Subsequently, the cells were washed and stained with PE-labeled mouse anti-human IFN-γ or isotype-matched Ig control (BD Pharmingen) for 25 min. After washing steps, cells were analyzed on a FACSCalibur flow cytometer.

RNA was isolated using the NucleoSpin RNA II kit (Macherey & Nagel). cDNA was synthesized from 1 μg of total RNA in a 20-μl reaction volume using random hexamers as primer (first-strand cDNA synthesis kit; Amersham Biosciences). PCR was performed with Platinum SYBR Green quantitative PCR SuperMix UDG (Invitrogen Life Technologies) on an iCycler (Bio-Rad) with the following cycling conditions: 2 min at 50°C; 2 min at 95°C; and 45 cycles of 15 s at 95°C, 15 s at 62°C, and 30 s at 72°C. Two microliters of cDNA was used in each amplification reaction. Primers were designed using GeneFisher software for TLR3 (forward, 5′-ACAGCCAGCTGTCCACCA-3′; reverse, 5′-TCCATGTTAAGGTGCTCCAA-3′) and for RNA polymerase II as the housekeeping gene for normalization (forward, 5′-GCACCACGTCCAATGACAT-3′; reverse, 5′-GTGCGGCTGCTTCCATAA-3′). Primer specificity was confirmed by melting curve analysis. No unspecific products were observed. Serial dilutions of plasmids containing the cloned PCR products were used to obtain a standard curve.

Student’s t test (paired data) was used to analyze the statistical significance of differences.

In our first set of experiments we investigated whether the TLR3 ligand poly(I:C) has a direct effect on γδ T cells. We observed a reproducible and significant increase in proliferation (as measured by [³H]TdR uptake) in the absence of APC when established Vδ2 or Vδ1 γδ T cell clones and lines were cultured in the presence of poly(I:C) alone or in combination with a TCR stimulus such as anti-Vγ9 mAb for Vδ2 T cells and anti-CD3 for Vδ1 T cells (Fig. 1,a). Comparable results were obtained with BrHPP or anti-CD3 mAb for Vδ2 T cell clones and lines instead of anti-Vγ9 mAb (data not shown). The proliferation in response to TCR stimulus alone was not significantly increased in the majority of γδ T cell clones and lines due to the partially overlapping effect of activation-induced cell death after TCR triggering (29). Furthermore, IFN-γ production by these γδ T cell clones and lines stimulated with BrHPP or anti-Vγ9 mAb for Vδ2 T cell, or anti-CD3 mAb for Vδ1 T cells was enhanced by poly(I:C) (Fig. 1 b and data not shown). These results suggested that γδ T cells, like NK cells (18, 30), might indeed directly respond to poly(I:C) in addition to the previously described indirect effects mediated via myeloid DC. Therefore, we investigated in detail the expression of TLR3 in freshly isolated resting and short-term activated (12–24 h) γδ T cells and activated γδ T cell clones and lines.

TLR3 expression was previously found in total T lymphocytes by Northern blot analysis and RT-PCR (24, 25, 26, 27), but no data are available for purified human γδ T cells. We observed the expression of TLR3 by RT-PCR in four established γδ T cell clones and in six newly generated γδ T cell lines expressing various Vγ- and Vδ-chains (data not shown). Next, we investigated TLR3 mRNA by quantitative RT-PCR in freshly isolated, highly purified (>98%) γδ T cells in comparison with other subsets of human PBMC (Fig. 2,a). No other potentially TLR3-expressing cells (neither CD56⁺ NK cells nor CD11c⁺ CD1c⁺ myeloid DCs) were detected in the purified γδ or αβ T cells. In line with published data, TLR3 expression was nearly absent in freshly isolated monocytes/macrophages and was prominent in NK cells. Surprisingly, highly purified γδ and αβ T lymphocytes gave strong TLR3 signals. In all four tested blood samples, however, TLR3 mRNA expression was 1.5–4 times higher in γδ T cells compared with αβ T cells (Fig. 2). TLR3 expression was confirmed in cell sorter-purified (99.9% purity) γδ T cells and αβ T cells (Fig. 2,b, ▵). Next, we analyzed TLR3 expression at the protein level. We detected TLR3 protein in γδ T cells by flow cytometry and confocal laser scanning microscopy. In line with the results of Matsumoto et al. (17), we found that TLR3 is expressed intracellularly, but not on the cell surface, in monocyte-derived immature DC (Fig. 3,a). We observed the same pattern in freshly isolated (resting) γδ and αβ T lymphocytes, where TLR3 was predominantly localized intracellularly and not on the cell surface (Fig. 3,a). Peripheral blood γδ T cells preferentially express Vγ9 Vδ2, whereas Vδ1-expressing cells are under-represented. However, all γδ T cells expressed TLR3 independently of the Vγ/Vδ-chain in the six tested donors (data not shown). Additionally, we investigated TLR3 expression in several Vδ1 and Vδ2 γδ and αβ T cell clones and T cell lines by confocal laser scanning microscopy and flow cytometry, again with positive results. Intracellular TLR3 expression of one representative Vδ1 γδ T cell clone is shown in Fig. 3,a (upper right, laser scanning microscopy). Additionally, we observed modest TLR3 expression on the surface of all tested γδ T cell clones and lines (Fig. 3,b). These results unambiguously demonstrate that resting T lymphocytes and T cell clones and lines, including γδ T cells, express TLR3 intracellularly, whereas only γδ T cell clones and lines display weak cell surface expression. For this reason, we decided to examine the potential up-regulation of cell surface TLR3 on resting γδ T cells after stimulation. We observed an up-regulation of TLR3 on the surface of positively isolated γδ T cells after TCR cross-linking, which was increased even more 24 h after TCR cross-linking in the presence of poly(I:C) (Fig. 3,c). We obtained comparable results when γδ T cells were analyzed within PBMC stimulated with BrHPP in the presence or the absence of poly(I:C) (Fig. 3 c). This prompted us to investigate possible direct effects of TLR3 ligation on freshly isolated, highly purified γδ T cells.

Highly purified, positively selected γδ and αβ T cells from three healthy donors were stimulated by TCR cross-linking (via rabbit anti-mouse Ig) in the presence or the absence of 50 μg/ml poly(I:C) or with poly(I:C) alone. CD69 expression and secretion of cytokines and chemokines in the supernatants were analyzed after 24 h (Figs. 4 and 5). In contrast to NK cells and myeloid immature DCs, freshly isolated γδ T lymphocytes did not respond to poly(I:C) alone in the absence of APC. As shown in Fig. 4, we compared CD69 expression on highly purified γδ T cells and γδ T cells within PBMC. Similar to the findings reported by Kunzmann et al. (14), we observed CD69 up-regulation on γδ T cells by poly(I:C) within PBMC stimulated, or not, by BrHPP (Fig. 4,a). Additionally, CD69 expression on highly purified γδ T cells was increased when cells were cultured for 24 h in medium alone (possibly due to the effects of positive magnetic separation) and was further up-regulated after TCR cross-linking or TCR cross-linking in the presence of poly(I:C)), whereas poly(I:C) alone did not enhance CD69 on purified γδ T cells (Fig. 4,b). In these experiments, cytokine and chemokine productions were measured in parallel after 24 h by a human cytokine Ab array. The results showed that poly(I:C) significantly increased the TCR-mediated IFN-γ production of positively selected, highly purified γδ T lymphocytes, whereas poly(I:C) alone did not induce IFN-γ production (Fig. 5,a). The costimulatory effect of poly(I:C) on IFN-γ production was confirmed by intracellular flow cytometry (Fig. 5,b) and ELISA (Fig. 6,a). Additionally, several other cytokines (TNF-α, IL-1β, IL-6, and GM-CSF) and chemokines (RANTES and MCP-1) were moderately modulated by poly(I:C) (Fig. 5,a). To selectively activate Vγ9Vδ2 T cells, highly purified γδ T cells were stimulated with BrHPP. Again, poly(I:C) drastically enhanced BrHPP-stimulated IFN-γ production (Fig. 6,a). Furthermore, the addition of anti-TLR3 Ab to these cultures partially inhibited IFN-γ production from 20–30% in three donors (Fig. 6,b). In striking contrast to γδ T cells, poly(I:C) did not stimulate an IFN-γ response in αβ T cells, neither alone nor in combination with a TCR stimulus. As expected, positively selected and TCR cross-linked αβ T cells produced a variety of cytokines and chemokines when costimulated with soluble anti-CD28 mAb (Fig. 5,a). The productions of cytokines and chemokines, such as IL-2, IFN-γ, TNF-β, IL-13, and RANTES, were not modified when poly(I:C) was additionally added, whereas moderate effects were noted for IL-10, TNF-α, and IL-6 after TCR cross-linking with poly(I:C) in the absence as well as the presence of soluble anti-CD28 mAb (Fig. 5 b and data not shown).

In this study we demonstrate that human γδ T cells express TLR3, the pattern recognition receptor for dsRNA. TLR3 is known to be expressed in immature myeloid DC, NK cells, fibroblasts, intestinal epithelial cells, and, to some extent in T cells (16, 17, 18, 24, 25, 26, 27). We observed that highly purified γδ T cells expressed similar levels of TLR3 mRNA as NK cells. In contrast to the results reported by Kunzmann et al. (14), we found that freshly isolated human γδ T cells as well as established γδ T cell lines and clones express TLR3 protein. The reason for this discrepancy is not clear, but might be related to the use of different Abs. We tested different commercially available anti-TLR3 mAbs. Only the anti-TLR3.7 mAb, which is suitable for cell surface and intracellular staining (31), yielded reliable results when directly fluorochrome-conjugated mAb were used. We did not detect TLR3 on the cell surface of resting T cells, but did find TLR3 intracellularly in γδ and also αβ T cells. After short-term stimulation, TLR3 was up-regulated on the surface of γδ T cells. Moderate expression of TLR3 was also detectable on the surface of γδ T cell clones and lines. This provides a basis for poly(I:C) ligand binding to TLR3 and subsequent signaling events. Moreover, we were able to partially inhibit the costimulatory effect of poly(I:C) by adding neutralizing anti-TLR3 mAb, which presumably antagonizes TLR3 on the cell surface. Furthermore, we have demonstrated a direct costimulatory effect of the TLR3 surrogate ligand poly(I:C) on the IFN-γ production of freshly isolated γδ, but not αβ T cells. In the studies by Kunzmann et al. (14), poly(I:C) alone did not induce effects on purified γδ T cells and T cell clones, and the authors concluded that the costimulation was mediated by TLR3 expressing immature myeloid DC via the production of type I IFN. In agreement with these results, we observed that poly(I:C) alone did not induce any effect on freshly isolated γδ T cells, in line with the absence of TLR3 surface expression on unstimulated (resting) γδ T cells. Our data indicate that γδ T cells have to be activated via TCR to up-regulate TLR3. This fit well with the recent study by Brandes et al. (32), which demonstrated that resting isolated γδ T cells have to be activated to perform APC function, as shown by them, or to become directly sensitive to costimulatory effects, as shown by us. Furthermore, we observed that highly purified γδ T cells secrete several other cytokines and chemokines (e.g., TNF-α, GM-CSF, IL-1β, IL-6, RANTES, and MCP-1) after TCR cross-linking, which seems to be modulated in the presence of poly(I:C). We noticed constitutive expression of RANTES, MCP-1, and IL-6 in γδ T cells cultured in medium alone as described previously for RANTES (33), whereas MCP-1 was not detected by Cipriani et al. (34) with ELISA using PBMC cultured in isopentenyl pyrophosphate plus IL-2. To confirm and detect marginal changes in expression levels, additional experiments with assays more sensitive than the human cytokine Ab array are required. In this study we have focused on IFN-γ production, which is the most prominent cytokine produced after TCR cross-linking and poly(I:C) stimulation of freshly isolated γδ T cells. Poly(I:C) significantly increased TCR-stimulated IFN-γ production by purified γδ T cells in the absence of other TLR3-expressing cells (no detectable CD1c⁺CD11c⁺ DC) and without added IL-2. IL-2 is dispensable when expression of cytokines and chemokines is analyzed in γδ T cells after short-term stimulation; however, IL-2 is essential for the proliferation of γδ T cells. Lamont et al. (35) reported that enhanced STAT4 DNA binding and increased IFN-γ production are triggered by IL-2 in human Vγ9Vδ2 T cells. Freshly isolated human peripheral blood γδ T cells are not able to produce IL-2 (36), which explains the low level of IFN-γ after TCR stimulation in the absence of IL-2 (Fig. 5,a, upper part). The addition of poly(I:C) might overcome the requirement for IL-2 in this regard. In contrast to resting γδ T cells, activated γδ T cell clones and lines are able to produce IL-2 and produce higher amounts of IFN-γ upon TCR stimulation alone. In this study the addition of poly(I:C) caused only a moderate enhancement of IFN-γ secretion due to the preactivated state of the γδ T cell clones (Fig. 1,b). In contrast to freshly isolated γδ T cells, poly(I:C) did not costimulate IFN-γ production in TLR3-expressing αβ T cells. As expected, αβ T cells required costimulatory signals, such as anti-CD28, to produce high amounts of cytokines and chemokines, which were not modified when poly(I:C) was also added (Fig. 5 b and data not shown). Poly(I:C), instead of soluble anti-CD28, moderately enhanced some cytokines (e.g., TNF-α, IL-6, and IL-10), an observation that requires additional analysis and quantification by other methods.

Increasing evidence indicates that certain T cells can indeed express functional TLRs. TLR2, TLR5, and TLR7/8 ligands have been recently found to directly costimulate memory CD4⁺ T cells (37, 38), and TLR8 ligands have been shown to reverse suppression by regulatory T cells (39). These studies failed to detect an effect of poly(I:C) on CD4⁺ αβ T cells (37, 38). In accordance, we did not detect a costimulatory effect of poly(I:C) on IFN-γ-producing αβ T cells. However, the previous studies did not examine whether CD4⁺ or perhaps CD8⁺ αβ T cells produced TNF-α or IL-6 after stimulation via TCR and poly(I:C). The moderately enhanced cytokine production of αβ T cells after TCR stimulation in the presence of poly(I:C) in our experiments might be due to the responsiveness of only a subset of αβ T cells, e.g., memory cells. In line with such an assumption, memory CD4⁺ T cells are known to be more sensitive to TLR-mediated activation than naive CD4⁺ T cells (38). Moreover, it might also be possible that TLR3 promotes αβ T cell survival, as described by Gellman et al. (40) for activated CD4⁺ αβ T cells in the mouse.

RIG-I was recently identified as another receptor for poly(I:C) and therefore for dsRNA (21, 22). RIG-I belongs to the antiviral host response mechanisms. It detects dsRNA, which accumulates in the cytoplasm in virus-infected cells. For our experiments, we cannot exclude that poly(I:C) stimulates an intracellular host response within γδ T cells. However, the incubation with poly(I:C) alone did not result in activation of γδ T cells detectable with the readout systems used in our study.

As we showed, γδ T cells express TLR3, which is an important determinant of cellular responses to external dsRNA, and therefore senses virus infection in other cells. Similar to human NK cells, γδ T cells respond to poly(I:C) with up-regulation of CD69 as well as increased IFN-γ production (13, 18, 30); however, γδ T cells need a second stimulus provided via TCR to efficiently support antiviral immunity. As discussed by Caron et al. (38), non-T cells are more sensitive to pathogen-associated molecular pattern-mediated activation, whereas pathogen-associated molecular pattern activation of T cells would take place during massive entry of microorganisms. Our data clearly suggest that γδ T cells emerge as another cell population to guard and eliminate viral infections.

It has long been observed that γδ T cells are involved in antiviral defense reactions; however, their precise role is still elusive (1, 6, 7, 8). In CMV-infected patients after kidney transplantation, Vδ2-negative γδ T cells show a long-lasting expansion and display a strong reactivity against CMV-infected cells (8, 41). This reactivity requires TCR engagement. Also, this subset of γδ T cells has been reported to recognize stress-induced MHC class I-related chains A and B on epithelial cells via their TCR (42). The additional signaling via TLR3 in the case of a virus infection might provide specific information for the cells to initiate the appropriate effector functions. In this context, it is very interesting that TLR9 and TLR3 are mainly involved in antiviral immunity against murine CMV infection. As demonstrated by Tabeta et al. (43), TLR3^−/− and TLR9^CpG1/CpG1 mice are susceptible to murine CMV infection, which resulted in impaired type-I IFN production. Their data suggest that both TLR pathways together are necessary for overall protection against murine CMV infection (43).

In summary, our results suggest that integrated signals from TLR3 and TCR together induce a strong antiviral effector function in γδ T cells and thus support a decisive role of γδ T cells in early defense against viral infection. Most likely, such a mechanism is not restricted to human γδ T cells, because bovine γδ T cells were also recently found to express TLR3 (44). Together, these results support the idea that γδ T cells form a link between the innate and adaptive immune systems (1, 45). Furthermore, our study supports the emerging concept that certain subsets of human T lymphocytes react directly to certain TLR ligands (37, 38, 39, 40).

We gratefully acknowledge the technical assistance of Ina Martens, Hoa Ly, Katrin Köbsch, and Parvin Davarnia. We also thank Michaela Kaniess for sorting the γδ T cells, and Innate Pharma for the gift of BrHPP.

The authors have no financial conflict of interest.