ALBERT

Description: Tumor cells exploit immune checkpoints to withstand immune recognition and effector cells' onslaught. Pre-clinical findings, corroborated by initial results of clinical studies, indicate that immune checkpoint blockade is a promising strategy to harness anti-tumor immune responses and improve the clinical outcome of patients with hematological malignancies. Multiple myeloma (MM) is a prototypic disease in which immune checkpoints significantly contribute to the immune suppressive contexture that myeloma cells establish in the bone marrow (BM) in cooperation with regulatory T cells (Tregs), myeloid derived suppressor cells (MDSC), and BM stromal cells (BMSC). Vγ9Vδ2 T cells are among the immune effector cells strategically victimized by suppressor cells. These are non-conventional T cells halfway between innate and adaptive immunity with a natural inclination to react against malignant B cells, including myeloma cells. Vγ9Vδ2 T cells are equipped with a peculiar array of receptors for stress-induced self-ligands and a unique TCR-dependent recognition ability of phosphoantigens (pAgs) generated in the mevalonate (Mev) pathway. Recently, we have shown that BM Vγ9Vδ2 T cells are anergic to pAg stimulation and that the programmed death 1(PD-1)/programmed death ligand 1 (PD-L1) immune checkpoint pair contributes to their dysfunction. This is an early event already detectable in individuals with monoclonal gammopathy of undetermined significance (MGUS) and not fully reverted even when MM patients achieve clinical remission after autologous stem cell transplantation (auto-SCT). Anti-PD-1 treatment partially recovers the ability of BM Vγ9Vδ2 T cells to proliferate and exert cytotoxic activity after pAg stimulation, but early studies based on single-agent PD-1 blockade have fallen short of clinical expectations in MM. Thus, several strategies are under consideration to implement the clinical efficacy of immune checkpoint blockade like the association with lenalidomide and/or concurrent tumor vaccination. Our results indicate that TIM-3 is significantly upregulated in BM Vγ9Vδ2 T cells from MM patients at diagnosis. We have previously shown that pAg stimulation of PD-1+ BM Vγ9Vδ2 T cells further increase PD-1 expression and preliminary data suggest that this stimulation also increases TIM-3 expression. Interestingly, TIM-3 up-regulation is even more pronounced than PD-1 up-regulation in BM Vγ9Vδ2 T cells and it occurs also in peripheral blood (PB) Vγ9Vδ2 T cells from anergic MM patients. We have recently shown that pAg reactivity of BM Vγ9Vδ2 T cells from MM at diagnosis can be partially recovered by PD-1 blockade. Our results reveal that TIM-3 blockade is also able to partially recover pAg-induced Vγ9Vδ2 T-cell proliferation. The best recovery is obtained when pAg stimulation is carried out in the presence of concurrent PD-1 and TIM-3 blockade. BM Vγ9Vδ2 T cells from MM patients who are in remission (MM-rem) after auto-SCT are still PD1+ and anergic to pAg stimulation. Remarkably, percentages of PD-L1+ MDSC in the BM of MM-rem are also unchanged compared to MM at diagnosis (MM-dia) and MM in relapse (MM-rel). These data indicate that the immune suppressive contexture is still operative at the tumor site even when most of myeloma cells have been cleared by chemotherapy. Interestingly, chemoresistant residual myeloma cells after auto-SCT have been reported to be PD-L1+, and circulating exhausted PD-1+ CD8+ T cells have been described in the PB after auto-SCT. This may explain why our previous idiotype vaccination studies in MM patients have failed. We have initiated to investigate the effect of immune checkpoint blockade in different phases of the disease and preliminary results suggest that the functional outcome of PD-1 blockade can be very different according to the disease status: the most signifcant recovery of Vγ9Vδ2 T-cell proliferation is observed after PD-1 blockade in MM-rem, while the anergy of Vγ9Vδ2 T cells from MM-rel is totally refractory to immune checkpoint blockade. In conclusion, our results suggest that recovery of pAg reactivity by PB Vγ9Vδ2 T cells is a reliable biomarker to predict or assess the clinical efficacy of immune checkpoint in vivo and provide scientific groundwork to optimize anti-PD1 treatment as single agent or in combination with other antibodies (i.e, anti-TIM-3) to maximize the efficacy of immune checkpoint blockade according to the disease status. Disclosures Boccadoro: Amgen: Honoraria, Research Funding; SANOFI: Honoraria, Research Funding; BMS: Honoraria, Research Funding; Mundipharma: Research Funding; Janssen: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; CELGENE: Honoraria, Research Funding; Abbivie: Honoraria. Massaia:Roche: Other: advisory board, research support; Janssen: Other: advisory board; Gilead: Other: advisory board.

Description: The results of gene expression profiling (GEP) and immunohistochemical studies indicate that survival is worsened by macrophages (MΦ) in the tumor microenvironment of various B-cell lymphomas including follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL). Tumor-associated macrophages (TAMs) are known to be different from other types of MΦ, but the effects of TAMs that worsen prognosis in B-cell lymphoma are essentially unknown, as are the mechanisms of these effects. Here, we determined the phenotype and effects of TAMs on tumor survival, proliferation, and drug resistance in B-cell lymphomas and evaluated strategies to reverse their effects. As compared to peripheral blood monocytes (Mo) from normal donors (ND), Mo from FL patients were differentiated less into M1 MΦ (defined as CD68+CD163loCD206loCD86hi) by culture with CSF-1 for 5 days followed by IFN-g + LPS for 2 days more. In contrast, Mo from FL patients and ND were differentiated similarly into M2 MΦ (defined as CD68+CD163hiCD206hiCD86lo) by culture with CSF-1 followed by IL-4. Consistent with this, MΦ gene signatures from FL tumors were more similar to previously-described signatures of M2 rather than M1 MΦ (Martinez et al, J Immunol, 2006, 177(10):7303-11). In co-culture, primary FL tumor cells and lymphoma cell lines (including RL, a transformed FL cell line; Granta 519, a mantle cell lymphoma (MCL) cell line; and Raji, a Burkitt lymphoma cell line) induced differentiation of Mo into MΦ. Differentiation could be prevented by CS4 monoclonal antibody (mAb), a fully human IgG1 anti-human CSF-1R mAb (ImClone/Eli Lilly), but not isotype control Ab. Elevated levels of CSF-1 in culture supernatants after addition of CS4 mAb and real-time PCR of tumor cells suggested secretion of CSF-1 by lymphoma cells. Spontaneous apoptosis of primary FL and MCL tumor cells, determined by Annexin V and propidium iodide staining, was significantly reduced by co-culture with ND Mo (p

Description: The role of Vγ9Vδ2 T cells in chronic lymphocytic leukemia (CLL) is unexplored, although these cells have a natural inclination to react against B-cell malignancies. Proliferation induced by zoledronic acid was used as a surrogate of γδ TCR-dependent stimulation to functionally interrogate Vγ9Vδ2 T cells in 106 untreated CLL patients. This assay permitted the identification of responder and low-responder (LR) patients. The LR status was associated with greater baseline counts of Vγ9Vδ2 T cells and to the expansion of the effector memory and terminally differentiated effector memory subsets. The tumor immunoglobulin heavy chain variable region was more frequently unmutated in CLL cells of LR patients, and the mevalonate pathway, which generates Vγ9Vδ2 TCR ligands, was more active in unmutated CLL cells. In addition, greater numbers of circulating regulatory T cells were detected in LR patients. In multivariate analysis, the LR condition was an independent predictor of shorter time-to-first treatment. Accordingly, the time-to-first treatment was significantly shorter in patients with greater baseline numbers of total Vγ9Vδ2 T cells and effector memory and terminally differentiated effector memory subpopulations. These results unveil a clinically relevant in vivo relationship between the mevalonate pathway activity of CLL cells and dys-functional Vγ9Vδ2 T cells.

Description: Abstract 2250 Poster Board II-227 Background: Allogeneic hematopoietic cell transplantation (allo-HCT) is a potentially curative treatment for multiple myeloma (MM) due to its graft-versus-myeloma effect. The use of reduced-intensity conditioning regimens with lower transplant related mortality has extended applicability of allo-HCT but is associated with a high risk of relapse. Active immunization of donors with a tumor-specific antigen prior to stem cell harvest and recipients following transplantation may enhance graft-versus myeloma effect without increasing the risk of graft-versus-host-disease (GVHD). Here, we evaluated the magnitude and quality of the vaccine-induced T-cell immunity in the donors and its transfer to recipients. Methods: Patient-specific idiotype (Id) protein was isolated from the plasma of the MM patient, conjugated to keyhole limpet hemocyanin (Id-KLH) and administered with GM-CSF as an adjuvant. Donors received 3 vaccine doses subcutaneously at 10, 8, and 4 weeks before the stem cell harvest. MM pts received 3 booster subcutaneous injections at 3, 4, and 6 months after allo-HCT following reduced-intensity conditioning regimen (Fludarabine/Cyclophosphamide) and wash out of GVHD prophylaxis. Th1 (IL-2, IFN-γ, TNF-β, and GM-CSF) and Th2 (IL-4, IL-5, IL-10, and IL-13) cellular immune responses against Id and the carrier molecule KLH were determined by multiplex cytokine assay on a Luminex platform. Th1 responses (IL-2, IFN-γ, TNF-β) were further quantified and characterized at the single cell level by intracellular cytokine assay (ICCA) in CD4+ T cells from unfractionated cryopreserved pre and postvaccine PBMC. Multicolor intracellular flow cytometry analysis was also used to determine the frequency of circulating regulatory T cells (Tregs), identified as CD4+CD25+CD127lowFOXP3+, at various time points in both donors and recipients. Results: Ten MM pts (median age: 54 years; IgG = 8, IgA = 2) and their respective donors were enrolled in this study. All 10 donors completed vaccinations without significant toxicity. Following vaccination, KLH-specific Th1/Th2 responses were observed in all 10 donors by multiplex cytokine assay and confirmed by intracellular cytokine assay (TNFβ and IL-2). Id-specific Th1 and Th2 responses were noted in 5 and 6 donors, respectively. All 10 MM pts underwent reduced intensity allo-HCT and 9 of these completed their post-transplant vaccinations. KLH-specific Th1 responses were detected in 6/7 evaluable recipients prior to post-transplant immunization and in 7/7 following booster immunizations by ICCA. Phenotypic characterization revealed that the transferred donor-derived KLH-specific CD4+TNF-a producing T cells mainly consisted of intermediate or late effectors (CD62L-CD27+ or CD62L-CD27-). Id-specific Th1 and Th2 responses were measured in 5 and 3 pts, respectively, prior to post-transplant booster injections. Two pts had Id-specific Th1 responses (IFNγ or TNFβ) detectable by ICCA after in vitro expansion from post-transplant immunizations PBMC with a frequency of 2.1% and 2.5% respectively. Since the use of GM-CSF has been associated with increased Tregs in mouse models, we determined the effects of Id-KLH+GM-CSF vaccination on the number of Tregs in the peripheral blood. The percentage of circulating Tregs did not change significantly following vaccination in the donors. In contrast, we observed a progressive increase in the number of Tregs in recipients reaching statistical significance at 7 months post transplant as compared to day 100 (median values 12.8% and 6.1 respectively, p

Description: Vγ9/Vδ2 (γδ) T cells play a major role in innate immunity against microbes, stressed, and tumor cells. They represent less than 5% of peripheral blood lymphocytes but can be activated and expanded in vitro by aminobisphosphonates (ABP)–treated monocytes. The aim of this work was to determine whether ABP-treated dendritic cells (DCs) can also activate γδ T cells and regulate immune responses mediated by conventional αβ T cells. Highly purified immature (iDC) and mature DC (mDC) were generated from peripheral blood monocytes of healthy donors and incubated with zoledronic acid (Zol) for 24 hours. Zol-treated iDC and mDC retained their immunostimulatory properties and induced the vigorous expansion of central memory and effector memory γδ T cells. γδ T cells displayed antitumor activity and appropriate cell surface antigens to target secondary lymphoid organs and exert costimulatory activity. Antigen-specific MHC-restricted immune responses, mediated by conventional αβ T cells, were improved by the concurrent γδ T-cell activation. In conclusion, large numbers of γδ T cells with effector and costimulatory activities are rapidly generated by Zol-treated iDC/mDC. This strategy is worthy of further investigation to improve adoptive cell therapy and vaccine interventions against tumors and infections.

Description: Abstract 3878 Introduction: In this study, a serological proteome analysis (SERPA) was applied for the first time to identify novel tumor-associated antigens (Ags) capable of eliciting humoral immune responses in patients with chronic lymphocytic leukemia (CLL). SERPA has been demonstrated to be a valuable method to identify tumor associated Ags in several human solid and hematological malignancies. The identification and characterization of circulating antibodies (Abs) and corresponding Ags in CLL can provide useful information to understand cell transformation, predict clinical outcome, and develop immune-based interventions. Methods: SERPA was performed in 21 untreated patients. Proteins extracted from purified CLL cells were separated by 2-D electrophoresis (2-DE) to obtain proteomic maps which were blotted with corresponding sera by Western Blot to reveal Ab-based reactivity with autologous proteins. To verify the CLL specificity of Abs recognition, 7 out of 21 maps were also probed with sera collected from 7 healthy donors (HD). For identification, Ag spots in WB were aligned with proteins in 2-DE maps. The protein spots corresponding to the assigned Ags were excised from the gel, trypsin digested and analyzed by peptide mass fingerprint by MALDITOF Mass Spectrometry (MS) with the software MASCOT. T cells from 6 CLL patients and 3 HD were stimulated with autologous ENOA-pulsed and control dendritic cells (DC) and evaluated by IFNγ ELISPOT assay. Ags surface expression was analyzed by flow cytometry. Statistical correlations were performed using t-test, Mann-Withney rank sum test and χ2-test. Results: Sixteen out of 21 CLL sera (76%) were immunoreactive and produced a total number of 45 Ag spots, whereas HD sera produced only 3 spots (p

Description: Abstract 3881 Background: The mutational status of tumor immunoglobulin heavy chain variable region (IGHV) is a reliable prognosticator in chronic lymphocytic leukemia (CLL): patients with unmutated (UM) IGHV have a worse prognosis than patients with mutated (M) IGHV. The tumor microenvironment actively supports the survival of CLL cells and confers a multidrug resistance (MDR) phenotype to CLL cells. MDR is due to the over-expression of membrane transporters, like P-glycoprotein (Pgp), which actively extrudes several anticancer drugs. Pgp is under the positive control of the transcription factor Hypoxia-Inducible-Factor-1-alfa (HIF-1α) which is activated by isoprenylated Ras/Rho-dependent downstream signaling pathways. Ras and Rho isoprenylation are regulated by the mevalonate (Mev) pathway activity suggesting that this pathway can be exploited as a metabolic checkpoint to regulate chemresistance. Aim: The aim of this study was twofold: 1) to investigate the correlation between chemoresistance and the activity of the Mev pathway and Ras/Rho-A downstream signaling pathways in purified M and UM CLL cells under basal conditions and after incubation with stromal cells; 2) to evaluate the chemosensitizing effects of agents specifically targeting the Mev pathway and downstream signaling pathways under the same culture conditions. Methods: M and UM CLL cells were cultured in the presence and in the absence of murine stromal cells (M210B4) and exposed to Zoledronic acid (ZA) (1 μmol/L), Simvastatine (Sim) (1 μmol/L), ERK1/2 kinase inhibitor PD98059 (10 μmol/L), HIF-1α inhibitor YC-1 (10 μmol/L) and Doxorubicine (Doxo) (1 μmol/L). The Mev pathway activity was measured by cells radiolabelling with [14C]-mevalonic acid and thin layer chromatography. Ras, ERK1/2 and Akt activity were detected by Western blot. Rho, Rho Kinase and HIF-1α activity were assessed by ELISA. Mdr1 expression was measured by Real Time-PCR. PgP activity was evaluated by measuring Doxo intracellular accumulation. Doxo cytotoxicity was assessed by annexin V and propidium iodide staining. Results: The Mev pathway is significantly more active in UM than in M CLL cells. This hypermetabolic activity translates into a higher activation of Ras/Akt and Rho/Rho kinase signaling pathways and higher expression of the phosphorylated active form of HIF-1α. HIF-1α activation positively regulates mdr1 gene expression in UM CLL cells leading to a more effective Doxo extrusion and therefore better survival upon Doxo exposure. M210B4 stromal cells further protect UM CLL cells from Doxo induced cell death by upregulating Mev pathway activity, HIF-1α/mdr1/PgP axis activation, and Doxo extrusion. Targeting the Mev pathway of UM cells with ZA and Mev reduces the basal activity of HIF-1α/mdr1/PgP axis and significantly increases Doxo retention and cytotoxicity. Similar effects are obtained with PD85 and YC1–10 which are specific inhibitors of the downstream molecules ERK-1/2 and HIF-1α, respectively. All these agents are able to overcome the protective effect exerted by stromal cells by significantly increasing PgP activity and Doxo-induced cell death. Conclusions: Our data demonstrate that the Ras- and Rho-dependent HIF-1α/mdr1/PgP axis is more active and associated with higher levels of MDR in UM compared with M CLL cells. Targeting the Mev pathway and/or downstream signalling pathways is a promising strategy to circumvent basal and stroma-mediated chemoresistance especially in UM CLL cells. Disclosures: Massaia: Novartis Farma S.p.A: Honoraria, Research Funding.

Description: Abstract 724 Background: Programmed death (PD)–1 is an inhibitory receptor that impairs the function of activated T-cells and natural killer (NK) cells when engaged by its ligands PD-L1 or PD-L2. We have previously demonstrated that PD-1 is markedly up-regulated in intratumoral and peripheral blood CD4+ and CD8+ T cells in patients with follicular lymphoma (FL), a finding associated with impaired T-cell function, suggesting that PD-1 blockade may improve FL immune control. CT-011, a humanized anti PD-1 monoclonal antibody, was previously studied in a phase I clinical trial in patients with advanced hematological malignancies. CT-011 was well tolerated and induced sustained elevations of CD4+ T cells in the peripheral blood. More importantly, apparent clinical benefit was observed in six patients, including one patient with FL who had large tumor masses that achieved a durable complete remission lasting 〉14 months. Here, we studied the in vitro and in vivo effects of CT-011 on T-cell and/or NK-cell immune responses against human B-cell lymphoma and the hypothesis that CT-011 may improve tumor control when combined with rituximab, a chimeric anti-CD20 monoclonal antibody for the treatment of human FL. Materials and Methods: To determine the effects of CT-011 on antitumor T cells, intratumoral T cells were isolated from primary FL tumor samples, and cultured with or without autologous tumor cells in the presence or absence of CT-011 or isotype control antibody (50 μg/ml each) for 5 days, and tested for proliferation by 3H thymidine incorporation assay. To determine the effects of CT-011 on NK cells, peripheral blood mononuclear cells (PBMCs) derived from normal donors or patients with FL were cultured in the presence or absence of CT-011 (50 μg/ml) with or without IL-2 for 96 hours and analyzed for expression of various activating receptors including CD16, CD32, CD64, Fas ligand, NKG2D, NKp30, NKp44, and NKp46. The in vivo effects of CT-011 were tested in two B-cell lymphoma xenograft models. Ramos and RL lymphoma tumor cells were injected subcutaneously into nude and SCID mice, respectively, and CT-011 (10 μg/mouse) was injected weekly with or without rituximab starting approximately 7–10 days after tumor inoculation. Results: We observed that CT-011 significantly increased the proliferation of intratumoral T cells in response to autologous tumor cells compared with isotype control antibody. Treatment with CT-011 enhanced the expression of Fas ligand, CD32, CD64, and NKp30 on human NK cells in the presence of IL-2 as compared with PBMCs treated with IL-2 alone or media control. In the RL lymphoma xenograft model in SCID mice, treatment with CT-011 significantly delayed tumor growth (P≤0.05) and improved survival (P≤0.01) compared with control mice injected with saline. In a Ramos lymphoma xenograft model in nude mice, treatment with CT-011 and rituximab eradicated established tumors in a significant proportion of mice (P≤0.05) and markedly improved survival compared with rituximab alone or saline. Conclusions: Taken together, these studies suggest that blockade of PD-1 with CT-011 enhances the function of anti-tumor T-cells and augments the expression of activating receptors on NK cells. Treatment with CT-011 led to improved tumor control against human B-cell lymphoma in xenograft models and the combined use of CT-011 and rituximab was more effective that rituximab alone. These results provide the rationale to test the combination of CT-011 with rituximab in patients with B-cell lymphoma, given that the combination is likely to be complementary and may even be synergistic, leading to enhanced clinical efficacy without increasing toxicity. The development of such approaches that activate both the innate (NK-cells) and adaptive (T-cells) immune systems is likely to minimize the emergence of immune escape variants and improve clinical outcome in patients with lymphoma. A clinical trial evaluating CT-011 in combination with rituximab is planned in patients with relapsed FL. Disclosures: Rodionov: Cure Tech Ltd.: Employment. Rotem-Yehudar:Cure Tech Ltd.: Employment.

Description: BACKGROUND: Treatment of fludarabine-resistant chronic lymphocytic leukemia (CLL) patients is an unmet clinical need. Fludarabine resistance in CLL depends on intrinsic molecular features of the tumor cells, and on bidirectional interactions occurring between CLL cells and stromal cells (SC) of the tumor microenvironment. One of the main players of SC-induced fludarabine resistance is the CXCL12/CXCR4 axis. CXCR4 is a G protein-coupled receptor constitutively expressed on CLL cells. The binding of CXCR4 with CXCL12 activates the Ras/ERK1-2/Akt and the RhoA-dependent signalling pathways. To be active transducers Ras and RhoA need to undergo a post-translational modification (i.e. isoprenylation) by means of small molecules produced by the mevalonate (Mev) pathway. We have recently demonstrated that the Mev pathway is more active in IGHV unmutated than in mutated CLL cells, and is amenable to pharmacological manipulation by statins (i.e. simvastatin [Sim]). It is currently unknown whether the Mev pathway and its pharmacological targeting are implicated in the modulation of the CXCL12/CXCR4 axis and in the SC-induced fludarabine resistance of CLL cells. AIM: The aim of this study was to investigate the reciprocal interactions between the Mev pathway and the CXCL12/CXCR4 axis, in order to identify potential targets to counteract the constitutive and SC-induced fludarabine resistance of CLL cells. METHODS: Immuno-magnetically purified patient-derived CLL cells were cultured alone or with murine SC (M2-10B4 cell line). In selected experiments, cell cultures were exposed to human recombinant CXCL12 (100 μg/ml), CXCR4 inhibitor AMD3100 (5 μg/ml), fludarabine (F-ara-A, 10 μM), Sim (1 μM), ERK1-2 kinase inhibitor PD98059 (10 μM), RhoA kinase inhibitor Y276 (10 μM), HIF-1α inhibitor YC-1 (10 μM). The activity of the Mev pathway was measured by the quantification of metabolites [i.e. cholesterol and farnesyl pyrophosphate (FPP)] produced by CLL cells after 24 h incubation with 1 μCi of [3H]acetate. Ras and RhoA activities were evaluated measuring their GTP-bound fraction, taken as an index of the G-protein activation, respectively by pull-down assay and by an ELISA based assay. ERK1-2 and HIF-1α phosphorylation were evaluated by Western Blot. RhoA kinase, Akt and HIF-1α activities were measured with specific immunoassay kit. The amount of CXCL12 in culture supernatants was assessed by ELISA assay. Cell viability was determined by Annexin-V/propidium Iodide immunostaining and flow cytometry analysis. RESULTS: Co-culture with SC upregulated the Mev pathway activity of CLL cells, as shown by the increased production of cholesterol and FPP. This SC-induced increase in the Mev pathway activity was followed by the activation of the downstream Ras/ERK1-2 and RhoA/RhoA kinase signalling, the upregulation of the pro-survival factor Akt, and an increase in the transcriptional activity of HIF-1α. These biological and molecular effects were identically observed when CLL cells were exposed to recombinant CXCL12, and were completely abrogated by the CXCR4 antagonist AMD3100, thus showing the key role of the CXCL12/CXCR4 axis in the SC-induced modulation of the Mev pathway and the downstream Ras/ERK1-2 and RhoA/RhoA kinase signalling. On the other hand, blocking the Mev pathway by Sim and targeting ERK1-2 kinases, RhoA kinase and HIF-1α by specific small-molecule inhibitors significantly reduced the constitutive activity and the SC-induced upregulation of the Ras/ERK1-2 and RhoA/RhoA kinase signal transduction. A regulatory role of the Mev pathway on the CXCL12/CXCR4 axis was observed not only in CLL cells but also in SC, as shown by the significant reduction in CXCL12 secretion by SC exposed to Sim. In the last set of experiments, we found that the inhibition of the Mev pathway by Sim potentiated the direct cytotoxic effect of fludarabine against CLL cells. Even more importantly, both Sim and the downstream HIF-1α inhibitor YC-1 were capable of counteracting the SC-mediated protection of CLL cells from fludarabine-induced cytotoxicity. CONCLUSIONS: Our data demonstrate that the Mev pathway has a regulatory role on the CXCL12/CXCR4 axis and on the SC-mediated protective effects toward spontaneous and fludarabine-induced CLL cell death. The upstream inhibition of the Mev pathway and the downstream targeting of HIF-1α are promising strategies to circumvent fludarabine resistance in CLL. Disclosures Boccadoro: Janssen-Cilag: Consultancy, Membership on an entity's Board of Directors or advisory committees; Onyx Pharmaceuticals: Consultancy, Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees; Sanofi: Consultancy, Membership on an entity's Board of Directors or advisory committees.

Description: Abstract 3602 The clinical course of chronic lymphocytic leukemia (CLL) depends on the intrinsic tumor cell features but also on the interactions between tumor cells, local microenvironment and host immunity. The interplay between CLL cells and conventional αβ T cells has already been investigated in details, whereas very little is known about the role of Vgamma9Vdelta2 T cells. These cells have intrinsic antitumor properties which can be further enhanced by stimulation with aminobisphosphonates such as zoledronic acid (ZA) via monocytes or other antigen-presenting cells. ZA targets the mevalonate (Mev) pathway and induces the intracellular accumulation and release of intermediate metabolites, like isopentenyl pyrophosphate (IPP), which are very similar to the natural ligands of Vgamma9Vdelta2 T cells. In this study we have performed a phenotypic and functional analysis of Vgamma9Vdelta2 T cells in 93 untreated CLL patients and correlated the results with intrinsic CLL cell features and clinical outcome. Stimulation of peripheral blood mononuclear cells with ZA induced Vgamma9Vdelta2 T cell proliferation in 47/93 patients (responders, R), but not in 46/93 patients (non-responders, NR). Vgamma9Vdelta2 T-cell subset distribution of R CLL was similar to healthy donors, whereas effector memory (EM) and terminally differentiated effector memory (TEMRA) cells predominated in NR CLL. A significant association was found between the R/NR status of Vgamma9Vdelta2 T cells and the mutational status of the tumor IGHV genes: 77% of R patients were M, whereas 70% of UM patients were NR. To test the hypothesis that this difference reflected a different activity of the Mev pathway in M and UM CLL cells, we performed a bioinformatic elaboration of data obtained from gene expression profiling of CLL cells and a biochemical quantification of the Mev pathway in CLL cells including the intermediate metabolites farnesyl pyrophosphate (FPP) and IPP, and the final product cholesterol. The Mev pathway was significantly more active in UM than in M CLL cells, suggesting that the IPP overproduction by UM CLL cells is responsible for a chronic stimulation of Vgamma9Vdelta2 T cells leading to their differentiation into EM and TEMRA cells. This biochemical-driven immunoediting process has clinical implications. After a median follow-up of 46 months from diagnosis, univariate analysis identified R status as a predictor of reduced TFT (NR: 59 months vs R: not reached; p=0.01). The R/NR status also allows to further identify two subsets in UM CLL patients (R-UM and NR-UM) with significantly different TFT (NR-UM: 32 months vs R-UM: not reached; p=0.02). In multivariate analysis, Binet stage (P=0.027), IGHV mutational status (p=0.016), R/NR status (p=0.007), high and low-risk cytogenetic abnormalities (p