We investigated the interactions between human monocyte-derived dendritic cells (DCs) and Ag-activated circulating TCR-γδ-expressing lymphocytes (Vδ2). Coculture of immature DCs (iDCs) with peripheral blood Vδ2 T cells activated with either pyrophosphomonoesters (isopentenyl pyrophosphate; IPP) or aminobiphosphonates (pamidronate; PAM) led to a significant up-modulation of CD86 and MHC class I molecules and to the acquisition of functional features typical of activated DCs. DC activation induced by both IPP- and PAM-stimulated γδ T cells was mostly mediated by TNF-α and IFN-γ secreted by activated lymphocytes. However, the effect of PAM-activated γδ T cells, but not that of IPP-activated cells, required cell-to-cell contact. Reciprocally, activation of Vδ2 T cells by PAM, but not by IPP, was dependent on cell contact with iDCs. In fact, when PAM-stimulated DC-γδ T cell cocultures were separated by a semipermeable membrane or treated with blocking anti-CD86 Abs, induction of CD25 and CD69 as well as IFN-γ and TNF-α secretion by Vδ2 cells were strongly reduced. These results demonstrate for the first time a bidirectional activating interaction between iDCs and PAM-stimulated γδ T lymphocytes, thus suggesting a potential adjuvant role of this early cross-talk in the therapeutic activity of aminobiphosphonate drugs.
Human T cells expressing the TCR γδ represent a unique lymphocyte population with an unusual tissue distribution. These cells are present in lymphoid tissues as well as in the skin- and intestine-associated lymphoid organs (1). In particular, the γδ T cell subset Vγ9/Vδ2 represents the majority of peripheral blood γδ T lymphocytes and is involved in the immune response against intracellular pathogens and hematological malignancies (2, 3, 4). These lymphocytes circulate in the lymph nodes as naive cells, in the blood as effector/memory cells, and in inflamed tissues as terminally differentiated effectors (5).
Unlike classical αβ T cells, Vδ2 T cells are endowed with the capacity to interact with low molecular mass phosphate-containing nonprocessed Ags, such as pyrophosphomonoesters (6) and alkylamines (7). Recognition of these antigenic compounds is TCR mediated but does not require presentation by conventional MHC molecules (8, 9). Recently, another group of nonpeptide compounds, the aminobiphosphonates, currently used as therapeutic molecules in cancer therapy, was shown to activate Vδ2 T cells both in vitro and in vivo (10, 11). A number of studies have clearly shown that the antitumor effect of aminobiphosphonates, including pamidronate (PAM),3 mainly results from their capacity to activate γδ T cells, as assessed by induction of their proliferation and Th1 cytokine production as well as stimulation of cytotoxic activity against selected tumor cells (10, 11).
In contrast to pyrophosphomonoesters, aminobiphosphonate-induced activation and clustering of human peripheral blood Vδ2 T cells is strictly dependent on the presence of adherent cells of the monocyte lineage (12). Moreover, it has been recently shown that γδ T cell activation by aminobiphosphonates is achieved through the accumulation of mevalonate pathway metabolites, such as isopentenyl pyrophosphate (IPP), which are subsequently recognized by the Vγ9Vδ2 TCR (13).
Dendritic cells (DCs) are professional APCs exhibiting the unique capacity to initiate primary immune responses (14). These cells reside in peripheral tissues in an immature state, where they are adapted to capture and accumulate Ags, thus acting like “immunological sensors.” Immature DCs (iDCs) typically respond to pathogen exposure by undergoing considerable morphological and functional changes, collectively called maturation, that occur while they migrate from the peripheral tissues into the draining lymph nodes. Migration of DCs to lymphoid tissues, a process orchestrated by chemotactic signals and tightly regulated expression of the cognate receptors, results in the efficient presentation of optimally processed Ags to T cells. These specialized DC functions permit the rapid generation and maintenance of specific immune responses to invading pathogens, largely irrespectively of access site (15, 16).
In this study, we analyzed the interactions between human Vδ2 T cells and monocyte-derived iDCs. We report that both aminobiphosphonate (PAM)- and pyrophosphomonoester (IPP)-activated, but not resting, Vδ2 T lymphocytes specifically stimulate iDCs to undergo maturation. Both these phosphoantigens contribute to DC maturation by triggering TNF-α and IFN-γ secretion even though they trigger Vδ2 T cell activation by different mechanisms. Notably, interaction of Vδ2 T cells with iDCs is a prerequisite for their activation by PAM and requires CD86 expression. This study represents the first evidence for a cross-talk between iDCs and γδ T cells exploited to achieve reciprocal full activation. We suggest that this reciprocal activating interaction may play an adjuvant role in the therapeutic activity of aminobiphosphonate drugs.
Materials and Methods
IPP and PAM were obtained from Sigma-Aldrich and Novartis Pharmaceuticals, respectively. FITC- and PE-conjugated mAbs used for flow cytometry as well as blocking anti-CD86 (clone IT2.2, IgG2b) and control IgG2b Abs were obtained from BD Pharmingen. Blocking anti-TNF-α and anti-IFN-γ Abs were purchased from R&D Systems, whereas Escherichia coli LPS, FITC-dextran, and mevastatin were obtained from Sigma-Aldrich. GM-CSF and IL-4 were kindly provided by Schering-Plough.
Generation of DCs, isolation of γδ T cells, and cell culture conditions
Human monocytes were isolated from peripheral blood of healthy donors. Briefly, PBMC were separated by Ficoll-Paque (Sigma-Aldrich) density centrifugation and subsequently subjected to counterflow centrifugal elutriation (17, 18). Because elutriated monocytes contained a small percentage of contaminant cells (1–2% T cells and ≤0.5% B and NK cells), they were further purified by depleting the nonmonocyte populations to ≤1% by immunomagnetic beads negative selection (Miltenyi Biotec) as previously described (17). The resulting monocyte populations (99% pure) were routinely checked for the presence of contaminant cells as well as for the expression of monocyte-specific markers by FACS. CD1a+CD14− iDCs were generated by culturing monocytes for 7 days in flat-bottom 24-well plates at the density of 106 cells/ml (0.5 ml/well) in RPMI 1640 supplemented with 10% FBS, GM-CSF (50 ng/ml), and IL-4 (500 U/ml). Fresh medium plus cytokines were added every 3 days. Mature DCs were obtained by treatment of iDCs with 200 ng/ml E. coli LPS for 24 h.
γδ T cells were separated from total cryopreserved autologous lymphocytes cultured for a few hours in complete medium and then subjected to positive selection with immunomagnetic beads (Miltenyi Biotec) according to the manufacturer’s instructions. Briefly, elutriated PBLs were incubated with apten-conjugated anti-γδ TCR Abs for 15 min at 4°C, washed, and then subsequently incubated with FITC-conjugated anti-apten immunomagnetic beads. Positively selected populations contained >95% viable γδ T cells as assessed by flow cytometry.
After an overnight culture in complete medium, either total PBLs or purified γδ T cells were washed, suspended in RPMI plus 10% FBS at the density of 106 cells/ml, and added to DC cultures (0.5 ml/well). DC/γδ cocultures were left untreated or stimulated with the nonpeptide phosphate Ags IPP (2 μg/ml) or PAM (10 μg/ml) and then analyzed 48 h later. In some experiments, γδ T lymphocytes were seeded, at the same density, in the absence of DCs and stimulated with the specific Ags as described above. In parallel, DCs and γδ T cells were separated by a semipermeable membrane through culture in trans-well chambers (Falcon). In some experiments, DC/γδ T cell cultures were stimulated with Ags in the presence of blocking mAb to CD86 (5 μg/ml) or control IgG2b mAb, anti-TNF-α, and anti-IFN-γ mAbs (10 μg/ml) or after a 1-h pretreatment with mevastatin (25 μM). Some DC cultures were stimulated with LPS (200 ng/ml) for 24 h as positive control of maturation.
Phenotypic analysis of DCs and γδ T cells
Untreated as well as Ag-stimulated DC/γδ T cell cultures were stained with FITC-conjugated mAbs to CD1a, CD14, CD80, CD86, CD40, HLA-DR, HLA-ABC, and CD83. Briefly, 2 × 105 cells were preincubated with PBS containing 10% human AB serum for 30 min on ice to block unspecific Ig binding, then stained with the specific mAbs or their isotype-matched control mAb for 30 min on ice, washed, and analyzed by a FACS flow cytometer (BD Biosciences).
To analyze γδ T cell activation, cells were collected 48 h after stimulation with PAM or IPP and double-stained with anti-Vδ2-FITC and anti-CD25-PE mAbs or with anti-Vδ2-FITC and anti-CD69-PerCP mAbs. Double-stained cells were then washed and analyzed by flow cytometry.
To investigate the expression of HLA-DR on both DCs and γδ T cells after sorting, γδ T cells were positively selected from the cocultures 48 h after their stimulation with PAM. Both lymphocytes and negatively selected DCs were stained with FITC-conjugated anti-HLA-DR or isotype-matched control mAbs as described above and analyzed by flow cytometry.
iDCs were cocultured with Ag-activated γδ T cells as described above. Total PBLs and γδ-depleted PBLs were used as controls. After 48 h, cells were washed and incubated with dextran-FITC at a concentration of 0.05 mg/ml for 40 min at 37°C. Some cultures were kept at 0°C as negative controls. Cells were then extensively washed and FITC-dextran fluorescence was determined by flow cytometry.
Seven-day cultured iDCs were cocultured with control or Ag-stimulated γδ T cells at a 1:1 ratio. Forty-eight hours later, cells were washed, irradiated (6000 rad), and then seeded in mixed cultures with allogeneic elutriated PBLs for 5 days. In parallel, γδ T cells were positively selected from irradiated PAM-stimulated cocultures. After selection, both γδ T cells and γδ T cell-depleted DCs were seeded in mixed cultures with allogeneic PBLs and cultured as described above. During the final 16 h of culture, cells were pulsed with 0.5 μCi of [3H]thymidine (sp. act., 5 Ci/mmol; Amersham Biosciences) and counted as described previously (17). Results are shown as mean cpm of triplicate cultures. [3H]Thymidine incorporation in negative controls (either stimulator DCs or responder T cells alone) was always below 800 cpm.
Determination of cytokine levels
Culture medium from DC/γδ T cell cocultures was harvested at different time points after Ag stimulation and subsequently analyzed for IL-12, TNF-α, and IL-10 secretion by ELISA (R&D Systems; detection limits, 7.8, 15.6, and 7.8 pg/ml, respectively).
Intracellular staining of TNF-α and IFN-γ
γδ T cells stimulated with phosphoantigens, in the presence or in the absence of iDCs, were treated with brefeldin (Sigma-Aldrich; 10 μg/ml) during the last 5 h of culture. Cells were collected, fixed with paraformaldehyde, and then permeated with 70% ethanol. Permeated cells were double-stained for 30 min at 4°C with PE-conjugated mAbs to TNF-α or IFN-γ and FITC-conjugated anti-TCR Vδ2 mAb. Double-stained cells were analyzed by flow cytometry.
Coculture of iDCs with phosphoantigen-stimulated γδ T lymphocytes up-regulates the expression of CD86 and MHC class I molecules
We analyzed the ability of phosphoantigen-activated human PBLs to modulate the phenotype and functional activities of iDCs. To this aim, 7-day-cultured iDCs were exposed to autologous PBLs at a 1:1 ratio and treated with PAM, a phosphoantigen known for its capacity to indirectly activate Vγ9/Vδ2 T lymphocytes (12, 13). As shown in Fig. 1, exposure of iDCs to PAM-stimulated PBLs strongly up-modulated the expression of the costimulatory molecule CD86, as both percentage of positive cells and mean fluorescence intensity. Moreover, a significant up-modulation of the fluorescence intensity of MHC class I molecules was observed under the same experimental conditions (Fig. 1). The effect of PAM-activated lymphocytes on CD86 and MHC class I molecule expression was comparable with that observed in DCs stimulated to mature by classical stimuli, such as LPS (Fig. 1). In contrast, as shown in Fig. 2, no significant changes were observed in the expression level of other surface molecules on PAM stimulation of iDC-PBL cocultures, whereas, as expected, a significant induction of CD80, CD86, CD40, CD83, and MHC class I and II, together with a slight reduction in CD1a expression, was observed in DC stimulated with LPS (data not shown). CD86 and MHC class I up-modulation was specifically induced by PAM-activated PBLs, because direct exposure of iDCs either to the Ag alone or to unstimulated PBLs did not result in any change in their expression. Similar results were obtained on coculture of iDCs with IPP-activated γδ T cells (data not shown).
CD86 and MHC class I molecule up-regulation is specifically induced by PAM-activated γδ T cells
To establish whether CD86 and MHC class I up-modulation correlated with the phosphoantigen-induced specific activation of γδ T cells constituting the PBL populations, Vδ2 T cells were purified or depleted from the PBL pool by immunomagnetic selection. Purified Vδ2 T cells as well as Vδ2-depleted PBLs were cocultured with iDCs at a 1:1 ratio and stimulated with PAM. As shown in Fig. 3, coculture of iDCs with PAM-activated purified γδ T cells induced CD86 and MHC class I up-modulation at levels comparable with those observed with activated total PBLs. Conversely, depletion of γδ T cells from the PBL pool completely abrogated the capacity of these cells to up-modulate CD86 and MHC class I expression. Similar results were obtained on coculture of iDCs with IPP-activated purified γδ T cells (data not shown). Coculture of iDCs with phosphoantigen-activated purified γδ T cells at a ratio up to 10:1 still induced up-modulation of CD86 and MHC class I molecules (data not shown). Moreover, the magnitude of the effect exerted by purified γδ T cells was not dependent on the frequency of circulating γδ T cells observed among donors (data not shown).
γδ T cell-exposed DCs acquire the functional characteristics of mature cells
To investigate whether the phenotypic changes observed in iDCs cocultured with PAM-activated γδ T cells paralleled modifications in their functional activities, experiments aimed at evaluating Ag uptake capacity, cytokine secretion, and induction of T cell-proliferative responses were performed. The ability of DCs to capture Ags is an important feature of iDCs and is known to decline on DC maturation (14). The endocytic profile of iDCs was compared with that of DCs matured on LPS stimulation or cocultured with unstimulated PBLs, purified Vδ2 T cells, and Vδ2 T cell-depleted PBLs. As expected, iDCs exhibited a high capacity to take up FITC-labeled dextran, but this activity was almost completely lost on maturation induced by LPS (Fig. 4,A). DCs cocultured with PAM-stimulated total PBLs or purified γδ T cells showed a marked decrease in their ability to take up FITC-labeled dextran, whereas no significant changes were observed on coculture with unstimulated or γδ T cell-depleted PBLs. Likewise, exposure of iDCs to PAM-activated Vδ2 T cells led to IL-12 and TNF-α production, although to a lower extent with respect to LPS-stimulated DCs (Fig. 4,B). Conversely, no IL-10 was detected in the coculture medium (data not shown). Finally, experiments were conducted to investigate whether IL-12 producing DCs, generated on exposure to PAM-activated γδ T cells, exhibited the capability to stimulate proliferation of allogeneic lymphocytes. To this aim, iDCs cocultured with purified Vδ2 T cells (1:1 ratio), either untreated or stimulated with PAM, were irradiated and seeded as stimulators in MLR assay with allogeneic PBLs. As shown in Fig. 4,C, DCs cocultured with PAM-stimulated γδ T cells significantly increased their capacity to stimulate proliferation of allogeneic lymphocytes with respect to DCs cocultured with unstimulated γδ T cells. The allostimulatory capacity of γδ T cell-activated DCs was comparable with that exhibited by irradiated LPS-matured DCs (data not shown). To exclude the possibility that the observed proliferation induced by the DC/γδ T cell coculture could be due to the up-regulation of MHC class II expression on PAM-activated Vδ2 T cells, these latter were purified from the coculture by positive selection with magnetic beads 48 h after PAM stimulation. Both γδ T cells and γδ T cell-depleted DCs were analyzed for MHC class II surface expression as well as for their ability to stimulate allogeneic αβ T cell proliferation. Although >90% of PAM-activated γδ T cells expressed comparable levels of HLA-DR on their surface as activated DCs (data not shown), only these latter represented the major contributors to the allostimulatory activity exerted by the PAM-stimulated cocultures, whereas the stimulatory potential exhibited by γδ T cells was negligible (Fig. 4 C). Similar results were obtained with IPP-stimulated γδ T cells despite the induction of a clear-cut HLA-DR up-modulation (data not shown).
Phenotypic and functional changes induced in DCs on interaction with PAM-activated, but not IPP-activated, γδ T cells require cell-to-cell contact
γδ T cells have been shown to produce high amounts of TNF-α and IFN-γ on activation by phosphoantigens (19). To investigate whether DC activation induced by PAM- and IPP-stimulated γδ T cells was somehow related to the release of these immune mediators or required cell-to-cell contact, iDCs and γδ T cells were separated by a semipermeable membrane and cultured in Transwellchambers. As shown in Fig. 5,A, no up-modulation of CD86 was induced in DCs when coculture with PAM-activated γδ T cells occurred in Transwell chambers. Likewise, no up-modulation of MHC class I molecules was observed (data not shown). As expected, the effect of IPP-activated γδ T cells on CD86 up-modulation was not significantly reduced under the same experimental conditions (Fig. 5,A). Experiments were then conducted to evaluate whether the effect of PAM-activated γδ T cells on Ag uptake capacity as well as IL-12 production was abrogated when the two cell populations were physically separated. In keeping with the observed phenotypic effect, Transwell coculture of DCs with PAM-stimulated γδ T cells did not lead to any significant reduction in the Ag uptake capacity of DCs, whereas a marked reduction was observed on IPP stimulation (Fig. 5,B). Moreover, no IL-12 secretion was detected in DCs cultured with PAM-stimulated γδ T cells in Transwell chambers (Fig. 5 C).
DC activation induced by PAM-stimulated γδ T cells is abrogated by blocking mevalonate metabolite production
It has been recently demonstrated that γδ T cell activation induced by PAM involves the dysregulation of the mevalonate biosynthesis pathway, leading to a specific accumulation of endogenous IPP, ultimately responsible for the TCR stimulation (13). Blocking this accumulation, γδ T cell activation induced by PAM is completely impaired (13). To further evaluate the molecular basis underlying the differential effect of PAM vs IPP, the role of the mevalonate pathway in the γδ T cell-mediated DC activation was investigated by using mevastatin, a specific inhibitor of the hydroxymethylglutaryl-CoA reductase, a crucial enzyme for IPP production. As shown in Fig. 6, mevastatin pretreatment of DC/γδ T cell cocultures completely abrogated CD86 and MHC class I induction by PAM-stimulated γδ T cells. Moreover, IL-12 production induced by PAM-stimulated γδ T cells was also completely inhibited in these experimental conditions (data not shown). In contrast, mevastatin did not affect DC maturation induced by LPS, thus excluding any toxic effect. In keeping with previously published data (13), mevastatin pretreatment completely inhibited Vδ2 T cell activation, as assessed by the analysis of the T cell activation markers CD25 and CD69 (data not shown).
iDCs are strictly required for PAM-mediated γδ T cell activation, whereas they are dispensable for the IPP-mediated activation: role of cell-to-cell contact and CD86 expression
Although it is generally assumed that γδ T cells are activated by phosphoantigens through the TCR in the absence of APC, a recent study showed that PAM-mediated γδ T cell activation requires the presence of monocyte-like cells (12). On the basis of these observations, we thus investigated whether a reciprocal activating interaction between DCs and γδ T cells could exist. To this aim, we analyzed the expression of T lymphocyte activation markers in our DC/γδ T cell cocultures. As shown in Fig. 7,A, PAM stimulation of purified γδ T cells in the absence of DCs did not lead to any up-modulation of typical T cell activation markers, such as CD25 and CD69, on Vδ2+ cells. A strong induction of the expression of these surface markers was observed in Vδ2 cells on coculture with iDCs. Similar results were obtained when total PBLs activated with PAM were used (data not shown). On the contrary, IPP stimulation was per se sufficient to induce Vδ2 T cell activation, which was not significantly modified in the presence of iDCs (Fig. 7,A). To further investigate the role of DCs in PAM-induced Vδ2 T cell activation, we evaluated the possible contribution of soluble mediators and cell contact in this phenomenon. As shown in Fig. 7,B, abrogation of physical contact by a semipermeable membrane completely inhibited Vδ2 T cell activation by PAM. On the basis of our previous observation that CD86 plays a role in the γδ T cell-mediated DC activation (Fig. 5), we conducted experiments aimed at evaluating whether this costimulatory molecule could also be involved in the activation of γδ T cells by iDCs. Therefore, we analyzed the expression of CD25 and CD69 in Vδ2 T cells cocultured with iDCs and exposed to PAM in the presence or in the absence of a blocking mAb to CD86. As shown in Fig. 7 B, a marked reduction in the expression of both activation markers was detected in cultures exposed to PAM in the presence of anti-CD86 mAb as compared with cell cultures left untreated or treated with a control mAb. Moreover, a marked decrease in TNF-α production was also found in the culture medium of PAM-stimulated DC-γδ T cell cocultures (data not shown).
To further evaluate the role of DCs in the secretion of inflammatory cytokines by activated Vδ2 T cells, the intracellular accumulation of IFN-γ and TNF-α was measured in purified γδ T lymphocyte cultures exposed to the Ags in the presence or in the absence of DCs. As shown in Fig. 7 C, in the absence of DCs, IFN-γ and TNF-α intracellular accumulation was detected only in Vδ2 T cells stimulated with IPP, but not with PAM. Similar results were obtained when total PBLs activated with PAM were used (data not shown). In the presence of DCs, a very low proportion of unstimulated γδ T cells produced TNF-α, whereas IFN-γ was not produced under these experimental conditions. However, on stimulation with both phosphoantigens, a consistent increase in the percentage of IFN-γ- and TNF-α-producing Vδ2 cells was found.
Differential contribution of TNF-α and IFN-γ secreted on γδ T cell activation in the reciprocal activating interaction between DCs and γδ T cells
To address the question of whether the secretion of TNF-α and IFN-γ by phosphoantigen-activated γδ T cells was responsible for the induction of DC maturation, the biological activity of these cytokines was neutralized by means of specific mAbs. iDCs were cocultured with PAM- or IPP-stimulated γδ T cells in the presence or in the absence of neutralizing mAbs to TNF-α or IFN-γ and subsequently subject to phenotypic analysis. Neutralization of TNF-α activity almost completely abrogated the up-regulation of CD86 generally observed on coculture of iDCs with PAM- and IPP-activated γδ T cells, whereas only a partial effect was observed in the presence of a mAb to IFN-γ (Fig. 8). No effect on CD86 surface expression was detected in the presence of isotype-matched control Abs (data not shown).
In this study, we investigated the interactions between DCs and activated γδ T cells and describe, for the first time, the existence of a reciprocal activating interaction between these cell types. We show that coculture of iDCs with γδ T cells activated by two phosphoantigens, PAM or IPP, ultimately leads to remarkable DC phenotypic and functional changes characterized by a marked up-regulation of CD86 and MHC class I molecules as well as the acquisition of functional activities typical of mature/activated DCs. This phenomenon is strictly linked to the presence of activated γδ T cells in that their removal from the PBL pool completely abrogates DC activation. Conversely, highly purified Vδ2 T cells are per se sufficient to induce full DC activation. The modalities through which DC activation is achieved are completely different for the two Ags. In fact, we show that whereas PAM-stimulated T cells require physical contact with iDCs to induce full activation, no physical contact is needed with IPP-activated cells. In both cases, however, cytokines released in the coculture medium, mostly TNF-α but also IFN-γ, were found to play an important role in this phenomenon, given that their neutralization by specific Abs reduces, although to a different extent, the capacity of γδ T cells to activate DCs, independently of the phosphoantigen used.
PAM belongs to a new class of synthetic compounds, the aminobiphosphonates, originally developed as therapeutic drugs for osteoporosis, and nowadays used for cancer therapy. The antitumor activity of these compounds is apparently linked to their capacity to activate γδ T cells, both in vitro and in vivo (10, 11). On activation, these cells acquire a potent cytotoxic activity toward a wide spectrum of tumors in vitro (20). In contrast to IPP, PAM-induced activation and clustering of circulating human γδ T cells is strictly dependent on the presence of adherent cells of the monocytic lineage (12). Moreover, it has been recently reported that stimulation of tumor cell lines with PAM leads to IPP accumulation and subsequent γδ T cell activation (13). The accumulation of phosphorylated metabolites, such as IPP, and activation of Vδ2 T cells was found to be regulated by the mevalonate pathway because its blocking through specific inhibitors completely abolished these effects (13). In keeping with these findings, we show that DC activation induced by PAM-activated γδ T cells was completely abrogated by blocking the mevalonate pathway, indicating that a similar mechanism of regulation is also operative in iDCs, thus providing an explanation for the differential effect observed with the two types of phosphoantiagens.
Overall, these observations prompted us to evaluate the existence of a reciprocal activating interaction between DCs and PAM-stimulated γδ T cells. In keeping with this hypothesis, we observed that coculture of iDCs with PAM-stimulated γδ T cells leads to a marked up-modulation of lymphocyte activation markers (i.e., CD25 and CD69), as well as TNF-α and IFN-γ production, which is not observed when γδ T cells are stimulated with PAM in the absence of iDCs. Moreover, we provide evidence that this reciprocal activation requires cell-to-cell contact and is partially dependent on CD86 expression. Accordingly, Fikri et al. (21) showed that activation of bovine WC1+ γδ T cells by bacterial superantigens is dependent on the presence of either monocytes or DCs. Furthermore, γδ T lymphocyte proliferation is markedly inhibited by CD86 blocking. Our finding that CD86 blocking only partially impairs DC-mediated γδ T cell activation suggests that other DC molecules are likely involved in DC-dependent PAM recognition by γδ T cells. In this regard, a role for LFA-1/ICAM-1 interaction between γδ T cells and PAM-presenting tumor cells has been recently proposed (22).
In this study, we also report that the reciprocal activating interaction between iDCs and γδ T cells occurs only when these latter are activated by PAM. In keeping with previous reports, we show that IPP-mediated γδ T cell activation does not require the presence of accessory cells. On the basis of these results and of previous observations showing that γδ T cell activation by aminobiphosphonates depends on IPP accumulation in accessory cells (13), it is likely that DCs can play a dual role in this phenomenon. In fact, iDCs could either provide a costimulatory signal for γδ T lymphocytes (23) under conditions in which PAM-induced intracellular IPP accumulation is very low or, in contrast, iDCs could represent indispensable accessory cells actively involved in PAM uptake. This hypothesis is supported, at least in part, by our results showing that, in the absence of iDCs, PAM stimulation of γδ T cells does not result in any functional activation (Fig. 7).
Cognate interactions are a common feature of the Ag-specific immune response, where intimate cell contacts govern critical events such as Ag presentation and delivery of T cell help to CTLs and B cells. The interaction of iDCs with naive T lymphocytes represents the crucial step through which full activation of DCs and the development of a protective immune response are achieved. With respect to the interaction between DCs and αβ T cells activated by bacterial superantigens, the induction of costimulatory molecules and MHC Ags on iDCs has been reported in mouse and human DC subsets. In particular, the maturation of murine splenic CD11c+ DCs is induced in response to staphylococcal enterotoxin B in a αβ T cell-dependent manner (24), whereas αβ T cells stimulated with staphylococcal enterotoxin A have been shown to induce activation of human blood DCs (25). However, superantigens have also been described to directly induce activation of human monocyte-derived DCs in the absence of T lymphocytes (26). Although it has been reported that the coordinated control of innate immunity is mostly mediated by cytokine loops, recently, a number of studies highlighted the importance of DC interactions with NK cells, an important effector cell type of innate immunity, in the regulation of DC maturation (27, 28, 29). Moreover, Gerosa et al. (29) reported the existence of a cross-talk between iDCs and resting NK cells, leading to cell activation on microbial stimulus encounter. Conversely, only a few studies have previously reported some functional interactions between DC and γδ T cell subsets (30, 31, 32). In particular, Leslie et al. (30) showed induction of DC maturation on coculture with nonactivated Vδ1+ T cell clones. This interaction was found to be dependent on CD1 expression. Moreover, Ismaili et al. (31) recently reported that circulating γδ T cells, either resting or activated by a synthetic analogue of IPP, can induce DC maturation. However, none of these studies has investigated the mechanisms of action of different classes of phosphoantigens or described functional reciprocal interactions between γδ T cells and DCs. To the best of our knowledge, this is the first study showing a bidirectional activating interaction between Vδ2 T cells and DCs. We provide evidence that this cross-talk is restricted to γδ T cells activated by specific phosphoantigens (PAM but not IPP), and is likely analogous to that existing between activated αβ T lymphocytes or NK cells and DCs, allowing the reciprocal transfer of an activation state.
In light of our results as well as of previously published reports, we can envisage the following scenario. Interaction between γδ T cells and iDCs, locally recruited on tissue damage or microbial infection, might lead to the activation of γδ T lymphocytes, triggering cytokine secretion. TNF-α and IFN-γ would in turn act on iDCs by stimulating their functional maturation. Activated DCs might contribute to the development of specific immune responses toward microbes or tumors through IL-12 secretion and αβ T lymphocyte activation (33).
It has been hypothesized that aminobiphosphonates exhibit their antitumor activity either directly, by inducing cell cycle arrest in tumor cells (34), or through activation of γδ T cells, leading to both tumor-specific cytotoxic response and IFN-γ-mediated inhibition of tumor growth (11). The newly described γδ T cell-mediated DC activation in response to aminobiphosphonates would argue in favor of a wider spectrum of action for these compounds, unraveling their potential capacity of inducing specific immune responses against malignancies. On the other side, activated IL-12-producing DCs might stimulate γδ T cell activation and expansion as well as the recruitment of other γδ T cells, thus contributing to the amplification of the immune response. In this regard, it has been shown that IL-12 strongly enhances endothelial transmigration of Vδ2 T cells toward peripheral tissues (35). Moreover, Lopez et al. (36) recently demonstrated a role for IL-12 in protecting γδ T cells from programmed cell death induced by mitogenic stimulation, thus favoring their expansion and antitumor activity. The broad cross-reactivity of circulating γδ T cells and the possibility to improve their expansion in response to phosphorylated metabolites might provide a useful tool to manipulate the activation state of this cell population for therapeutic and/or vaccination exploitation in both malignancies and infectious diseases. The use of aminobiphosphonates as activating Ag would limit the activation of γδ T cells at specific districts, where iDCs and resting γδ T cells coexist, thus avoiding uncontrolled immune activation and cytotoxic responses. Therefore, aminobiphosphonates could act as useful vaccine adjuvants capable of linking innate to adaptive immunity by rapidly activating γδ T cells in the presence of DCs and potentially inducing αβ T cell-mediated protective immunity.
We thank Cristina Purificato for her excellent technical help in cell preparation and Stefano Billi and Anna Maria Fattapposta for graphical and editorial assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by European Commission Grant QLK2-CT-2001-02103 (to S.G.).
Abbreviations used in this paper: PAM, pamidronate; DC, dendritic cell; IPP, isopentenyl pyrophosphate; iDC, immature DC.