Vγ2Vδ2 T-cell receptor–mediated recognition of aminobisphosphonates

Bisphosphonates are antiresorptive drugs effectively used to treat osteoporosis, multiple myeloma, and metastatic breast and prostate cancer, in which patients have excessive bone resorption, leading to pathologic fractures. Chemically, bisphosphonates are synthetic analogs of endogenous pyrophosphate.1,2 Their side-chain composition contributes to the relative potencies of bisphosphonates.3 One mechanism whereby these drugs inhibit bone resorption is by directly inhibiting osteoclast function.4 However, indirect mechanisms also may exist. Two recent studies have shown that aminobisphosphonates can expand the most abundant population of γδ T cells in the human peripheral blood, Vγ2Vδ2 T cells (alternate nomenclature, Vγ9Vδ2), in vivo and in vitro, raising the possibility that aminobisphosphonates may induce γδ T cells to mediate antitumor and antiresorptive activity.5,6 However, these studies did not address the role of the T-cell receptor (TCR) in aminobisphosphonate recognition and whether these drugs in soluble form activate γδ T cells to kill tumor targets subsequently or sensitize these targets to γδ T-cell–mediated lysis by binding to the target cell surface.

The structural similarities between aminobisphosphonates and defined alkylamine and prenyl pyrophosphate γδ T-cell antigens prompted us to more closely investigate γδ T-cell–mediated recognition of aminobisphosphonates. Using TCR transfer studies, we show here that reactivity to aminobisphosphonates requires expression of the Vγ2Vδ2 TCR and that tumor cells and monocytes can be sensitized for Vγ2Vδ2 T-cell–mediated lysis by pretreatment with risedronate, an aminobisphosphonate. These data suggest a new mechanism whereby treatment of multiple myeloma and osteoporosis with aminobisphosphonates may decrease tumor burden and bone resorption.

The Vγ2Vδ2 T-cell lines and clone isoamyl 5.C7 were isolated from human peripheral blood mononuclear cells (PBMCs) as previously described.7 Fresh monocytes were derived from PBMCs obtained from healthy volunteers by Ficoll-Paque density centrifugation, followed by plastic adherence and harvesting with a rubber policeman.

DBS43, Vγ2Vδ2 and 27/3.62, Vγ1Vδ1 TCR transfectants, respectively, were made as described8 by transfecting TCR⁻ J.RT3-T3.5 cells with TCRγ and TCRδ complementary DNAs.

Risedronate (Actonel) and etidronate (Didronel) were obtained from Procter and Gamble (Cincinnati, OH), alendronate (Fosamax) from Merck (West Point, PA), and pamidronate (Aredia) from Novartis (Basel, Switzerland). Myeloma cell line U266B1 was obtained from American Type Culture Collection (Rockville, MD).

Stimulation of γδ T-cell lines or of the Vγ2Vδ2 TCR transfectant was performed as previously described.7 The SD of the triplicate determination was less then 10% of the mean. Interferon (IFN)-γ and tumor necrosis factor (TNF)-α were analyzed by enzyme-linked immunosorbent assay (Pharmingen, San Diego, CA) after 24 hours of antigenic stimulation.

Three-color flow cytometric analyses for intracellular cytokine were performed using Cytofix/Cytoperm Plus (with GolgiStop) (Pharmingen) to determine IFN-γ and TNF-α production of γδ T cells at the single-cell level. GolgiStop was added for the last 4 hours during antigenic stimulation to the γδ T-cell lines for 17 hours. Cells were stained for surface expression of TCRγδ by alexafluor (AF)-conjugated anti-γδ TCR (TCRδ1) and Cy-conjugated anti-CD3. Cells were fixed and permeabilized with Cytofix/Cytoperm solution and were stained with either anti–IFN-γ or anti–TNF-α or an isotype-matched monoclonal antibody (mAb) (Pharmingen) conjugated to phycoerythrin. Samples were analyzed in a FACScan flow cytometer using FlowJo software (Becton Dickinson, Palo Alto, CA).

Cytotoxic activity was performed as described, using the classical ⁵¹Cr release assay.7 

To examine the effect of antigen specificity on the Vγ2Vδ2 T-cell–mediated cytokine responses to bisphosphonates with different side chains, we performed cytokine release assays using a Vγ2Vδ2 T-cell line as an effector population and 3 aminobisphosphonates (risedronate, alendronate, pamidronate) or a bisphosphonate (etidronate) as antigens. Compared with pamidronate, 10- to 100-fold less alendronate and 100- to 1000-fold less risedronate was needed to achieve a half-maximal IL-2 (Figure 1A) or IFN-γ (Figure 1B) release from the γδ T-cell lines. There was no significant IL-2 or IFN-γ release in the absence of antigens or with etidronate stimulation, even with concentrations much higher than physiologic levels, suggesting that the amino group is necessary for antigenic activity. Fresh PBMCs exposed to these aminobisphosphonates showed a similar pattern of reactivity when γδ T-cell expansion was analyzed after incubation for 10 days, with γδ T cells expanding up to 12-fold in response to as little as 20 nM aminobisphosphonate (Figure 1C). Single-cell analysis of cytokine production by intracellular staining of IFN-γ and TNF-α, using 3-color flow cytometric analyses to gate on γδ T cells, confirmed that γδ T cells were the source of cytokine in these experiments (Figure1D).

The potent stimulatory activity of risedronate was somewhat surprising because compounds with ring structures that are not catabolized have previously been shown to be poorly stimulatory for γδ T cells.9,10 However, the antigenic activity of aminobisphosphonates in our study correlated with their antiresorptive potency (risedronate > alendronate > pamidronate) and with their ability to inhibit the activity of farnesyl pyrophosphate synthase, an enzyme that catalyzes a rate-limiting step in the mevalonate pathway of cholesterol synthesis.11 These data confirm and extend previously published data showing that the aminobisphosphonates alendronate, pamidronate, and ibandronate induce γδ T-cell expansion, while the bisphosphonates etidronate and clodronate do not.6 

To examine if Vγ2Vδ2 T-cell recognition of aminobisphosphonates was mediated by their TCR, either Vγ2Vδ2 or Vγ1Vδ1 TCRs were reconstituted in a TCR-deficient Jurkat T-cell line by transfection with appropriate TCR γ and δ complementary DNAs, and the transfected cells were tested for their ability to release IL-2 in response to bisphosphonates. Both Vγ2Vδ2 TCR–reconstituted Jurkat cells and Vγ1Vδ1 TCR–reconstituted Jurkat cells produced IL-2 in response to anti-CD3 mAb in a dose-dependent manner, suggesting that the reconstituted TCRs could transduce signals in both transfectants. However, only Vγ2Vδ2⁺ and not the Vγ1Vδ1⁺ TCR transfectants produced IL-2 in response to risedronate and alendronate in a dose-dependent manner (Figure 1D). In a separate experiment, we observed a very low dose-dependent response by the Vγ2Vδ2⁺ TCR transfectant in response to pamidronate, suggesting that recognition of pamidronate is Vγ2Vδ2 TCR–dependent (data not shown). However, because these TCR transfectants are up to 100-fold less sensitive than native γδ T cells in IL-2 release assays, we believe that the response to pamidronate, the weakest aminobisphosphonate studied here, was below the detection level of the assay in most experiments. Taken together, these data show that recognition of aminobisphosphonates is Vγ2Vδ2 TCR–dependent.

Using a Vγ2Vδ2 T-cell clone or a risedronate-derived Vγ2Vδ2 T-cell line as effectors, we examined the sensitivity to lysis of an Epstein-Barr virus (EBV)–transformed lymphoblastoid cell line (LCL), U266B1 myeloma cells, or fresh monocytes either pretreated with 10 μM risedronate, and then washed extensively, or in the continuous presence of 10 μM risedronate in a 4-hour cytotoxicity assay. When exposed to risedronate, target cells were lysed 3- to 30-fold more efficiently by either the Vγ2Vδ2 T-cell clone (isoamyl 5.C7) or γδ T-cell line (risedronate-derived) as compared with mock-treated targets. Surprisingly, pretreatment of the target cells with risedronate, followed by extensive washing, resulted in up to a 2-fold increase in sensitivity to lysis as compared with continuous antigen exposure (Figure 2A-E). No difference in toxicity or spontaneous release of label was observed. This result was unexpected because previous studies have shown that alkylamine and prenyl pyrophosphate antigens cannot be pulsed onto targets and must be present throughout the assay period to stimulate γδ T cells.9,12 The most likely explanation for this comparative decrease in cytolysis in the continuous presence of soluble antigen is that γδ T cells can kill each other under these conditions and that soluble antigen can compete with target cell–associated antigen, resulting in decreased target cell lysis (data not shown). These results suggest that risedronate, perhaps due to its high potency, in contrast to lower-potency alkylamine and prenyl pyrophosphate Vγ2Vδ2 T-cell antigens, can stably associate with target cell surfaces at levels high enough to render these targets extremely sensitive to lysis by Vγ2Vδ2 T cells.

Treatment with aminobisphosphonates of cancer metastatic to bone results in a decrease in metastatic events and pathological fractures.13 Aminobisphosphonates also are used widely for the treatment and prevention of osteoporosis. Aminobisphosphonates are preferentially bound to skeletal sites of bone resorption, resulting in concentrations in the millimolar range, much higher than the nanomolar concentrations found in blood.14 Our data showing that blood monocytes and myeloma cells can be sensitized to γδ T-cell–mediated lysis by pretreatment with micromolar concentrations of risedronate (Figure 2) suggest that osteoclasts and myeloma cells that resorb bone containing aminobisphosphonates could be subject to elimination by cytotoxic Vγ2Vδ2 T cells present in the bone marrow. Alternatively, these aminobisphosphonate-laden myeloma cells and osteoclasts may induce γδ T cells to secrete IFN-γ, a known inhibitor of tumor cell growth and bone resorption.15Thus, VγVδ2 T cells, via TCR-dependent recognition of tumor- and osteoclast-bound aminobisphosphonates, may be key mediators of the antitumor and antiosteoporotic effects associated with aminobisphosphonate treatment.

We thank Dr Michael B. Brenner for helpful discussions and Lin Li for technical assistance.