ALBERT

Description: Background: Recent evidence demonstrated that the treatment of acute myeloid leukemia (AML) cells with daunorubicin (DNR) but not cytarabine (Ara-C) results in immunogenic cell death (ICD). In the clinical setting, chemotherapy including anthracyclines and Ara-C remains a gold standard for AML treatment. In the last decade, etoposide (Eto) and fludarabine (Flu) have been added to the standard treatment for AML to potentiate its therapeutic effect and have been tested in many trials. Very little data are available about the ability of these drugs to induce ICD. Methods: AML cells were treated with all four drugs. Calreticulin and heat shock protein 70/90 translocation, non-histone chromatin-binding protein high mobility group box 1 and adenosine triphosphate release were evaluated. The treated cells were pulsed into dendritic cells (DCs) and used for in vitro immunological tests. Results: Flu and Ara-C had no capacity to induce ICD-related events. Interestingly, Eto was comparable to DNR in inducing all ICD events, resulting in DC maturation. Moreover, Flu was significantly more potent in inducing suppressive T regulatory cells compared to other drugs. Conclusions: Our results indicate a novel and until now poorly investigated feature of antineoplastic drugs commonly used for AML treatment, based on their different immunogenic potential.

Description: Abstract 2190 Adenosine 5'-triphosphate (ATP) plays a pivotal role in several cellular processes, through specific cell membrane purinergic P2 receptors (P2Rs). During inflammation and tumor cell growth, cell necrosis causes the release of intracellular ATP into the extracellular space, thus increasing from low (1–10 nM) to high (5–10 mM) the concentration of extracellular ATP. For this reason, variations in the extracellular ATP concentration might activate/inhibit the immune system. Here we investigated the role of ATP on CD4+ T-cell functions. We first demonstrated the expression of P2Rs for extracellular nucleotides in human activated CD4+ T cells and regulatory T cells (Tregs) We then show that physiological concentrations of extracellular ATP (i.e. 1–50 nM) do not affect both activated CD4+ T cells and Tregs. Conversely, supraphysiological concentrations of ATP show a bimodal effect on activated CD4+ T cells. Whereas 250 nM of ATP stimulates proliferation, cytokine release, expression of adhesion molecules and adhesion, high ATP concentration (i.e. 1 mM) induces apoptosis and inhibits activated CD4+ T-cell functions. On the contrary, at the same high concentration, ATP enhances the proliferation, adhesion, migration and immunosuppressive ability of Tregs. Similar results are obtained when activated CD4+ T cells and Tregs are exposed to ATP released by necrotized leukemic blasts. The present results provide evidence that different concentrations of extracellular ATP modulate T cells according to their activation status. Therefore, high concentrations of ATP, compatible with fast-growing tumors or hyper-inflamed tissues, may have a key role in killing activated CD4+ T cells and in expanding Tregs. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 4298 Indoleamine 2,3-dioxygenase (IDO1) and indoleamine 2,3-dioxygenase-like (IDO2) are enzymes involved in the tryptophan catabolism along the kynurenine pathway. While it is established that IDO1-expressing dendritic cells (DCs) contribute to tolerance in a number of biological settings, little is known about the expression and function of IDO2 in DCs. Human DCs can be generated in vitro to obtain immunogenic antigen-presenting cells (APC), used as cellular vaccines. In the clinical setting, DCs are commonly matured with a cytokine cocktail (CC) which includes TNF-a, IL-1b, IL-6 and PGE2. In particular, PGE2 enhances APC function of DCs by increasing IL-12 production and facilitating DC migration to lymph nodes. However, PGE2 is also a strong IDO1 inducer, which by this route can also limit the anti-tumor activity of DC-based immunotherapies. Thus, understanding the roles of IDO1 and IDO2 in DCs may impact the development of vaccines or DC-based immunotherapies. In the present study, we fully characterized IDO1 and IDO2 expression and function in human monocyte-derived dendritic cells (Mo-DCs). Mo-DCs were generated from purified CD14+ monocytes after culture with GM-CSF and IL-4 and then matured with CD40L, LPS alone, LPS plus IFN-g and the CC. We observed that immature Mo-DCs had little if any expression of both IDO1 and IDO2, whereas mature Mo-DCs exhibited upregulation of both enzymes. Among the different maturation stimuli, CC was the most effective in upregulating IDO1 and IDO2, both at the message and protein levels. This effect was associated also with the highest kynurenine production. By means of IDO1 and IDO2 expression, mature Mo-DCs were inhibited in stimulating allogeneic T cell proliferation and generated a population of CD4+CD25+FOXP3+ Tregs which highly suppressed allogeneic and autologous T-cell proliferation. On the basis of evidence that IDO1 is preferentially inhibited by the L-isoform of 1 methyl-tryptophan (1-MT) and IDO2 by the D-isoform, we performed functional enzyme tests in presence of both isoforms. Notably, both isoforms exhibited inhibitory effects, although we observed a stronger effect of L-1-MT than with D-1-MT suggesting a greater contribution of IDO1 than IDO2. These results offer direct evidence that Mo-DCs express functional IDO1 and IDO2 proteins. During the maturation phase, Mo-DCs enhance their tolerogenic qualities, and in particular the capacity to induce Tregs, through the upregulation of both IDO1 and IDO2. Beside the critical role of IDO1 in enhancing the immunosuppressive capacity of DCs, we show, for the first time, that IDO2 is involved also. Our findings imply that, from a clinical standpoint, to improve the efficacy of DC-based vaccines mature DCs should be combined with molecules that can inhibit the activity of both IDO1 and IDO2. Disclosures: Metz: NewLink Genetics: Employment. Prendergast:New Link Genetics Corp: Consultancy, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties.

Description: Introduction: This is a Phase 1b monocentric clinical trial, aimed to evaluate the safety, feasibility and preliminary clinical efficacy of alloreactive natural killer (NK) cell infusions after immunosuppressive regimen in elderly acute myeloid leukemia (AML) patients, who achieved first complete remission (CR) after induction/consolidation chemotherapy. Methods: Seventeen AML patients, in first morphological CR (3 patients showed molecular disease) (median age 64 years, range 53-73) received highly purified CD56+CD3- NK cells from haploidentical KIR-ligand mismatched donors after fludarabine/cyclophosphamide (Flu/Cy) immunosuppressive chemotherapy, followed by interleukin-(IL-) 2. Results: No signs of NK cell-related toxicity, including graft-versus-host disease, were observed. Myelosuppression was moderate although 1 patient died due to overwhelming bacterial pneumonia and was censored for clinical follow-up. With a median follow-up of 22.5 months (range, 6-68 months), 9/16 evaluable patients (0.56) are alive disease-free, whereas 7/16 (0.44) relapsed with a median time to relapse of 9 months (range, 3-51 months). Among relapsed patients, 2 individuals showed a very prolonged CR phase of 24 and 51 months, respectively, in absence of any concomitant anti-leukemia treatments. Three patients, treated with molecular disease, achieved molecular CR lasting 9 and 4 months in 2 cases and 8+ months in the third patient. These results have been compared to that obtained in a cohort of patients, who achieved CR after induction/consolidation chemotherapy, but did not undergo NK cell immunotherapy due to the absence of a KIR-L mismatched donor (Fig. 1). Disease-free survival (DFS) in the control cohort was 0.39, indicating a trend towards better DFS under the NK therapy (0.56; p=0.07). To correlate the donor NK activity with the clinical response, donor NK cells, as evaluated both as the total number of infused NK cells and as the frequency of alloreactive NK clones, were assessed before NK cell infusion. No statistically significant correlation between the total number of infused NK cells and clinical response was observed. On the contrary, we observed a higher number of donor-derived alloreactive NK cell clones in the group of responders as compared to that of non-responders (Fig. 2A). Accordingly to statistical analysis, a threshold of 8 over 100 NK clones was chosen as the cut-off level discriminating responders versus non-responders (Fig. 2A). Importantly, the infusion of higher number of donor alloreactive NK cells (more than 8/100 clones) was associated with prolonged DFS (Fig.2B). In particular, for patients who received more than 8/100 alloreactive NK cell clones the probability of DFS was 0.81, whereas patients who received less than 8/100 clones had a DFS of 0.14. Such difference was statistically significant (p=0.03) and demonstrates that the number of alloreactive NK cell clones in the donor may discriminate patients receiving NK cell therapy as for response evaluation. Interestingly, the DFS of the group of patients who received less than 8/100 donor alloreactive NK cell clones was comparable to that of the historical control group (Fig. 2B). This finding suggests that the therapeutical effect of the whole procedure, which comprises chemotherapy (Flu/Cy) plus NK immunotherapy, mainly relies on the anti-leukemia activity of alloreactive NK cells. Noteworthy, the increased anti-leukemia effect in patients receiving the higher number of donor alloreactive NK cells was not associated with increased myelosuppression, as shown by comparable data of hematological recovery after NK cell infusion in responders and non-responders.Conclusions: infusion of purified NK cells is feasible in elderly patients with AML as post-CR consolidation strategy and donor NK alloreactivity has a predictive role on the clinical outcome of treated patients. Supported by: Italian Leukemia Association (BolognAIL), Section of Bologna and Fondazione Carisbo, Bologna, ELN, AIRC, PRIN, progetto Regione-Università 2010-12 (L. Bolondi), FP7 NGS-PTL project. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures No relevant conflicts of interest to declare.

Description: BACKGROUND: Overall survival of adult acute myeloid leukemia (AML) is still poor due to the lack of novel and effective therapies. In different malignancies including AML, some chemotherapy agents, such as daunorubicin (DNR) but not cytarabine (Ara-C), activate the immune response via the cross-priming of anti-tumor T cells by dendritic cells (DCs). Such process, known as immunogenic cell death (ICD), is characterized by intracellular and pericellular modifications of tumor cells, such as the cell surface translocation of calreticulin (CRT) and heat shock proteins 70/90 (HSPs 70/90), the extracellular release of ATP and pro-inflammatory factor HMGB1. Alongside with ICD, chemotherapy is known to induce inflammatory modifications within the tumor microenvironment, which may also elicit immunosuppressive pathways. In particular, DCs may be driven to acquire tolerogenic features, which may ultimately affect anti-tumor T-cell responses. In this study, we characterize ICD in AML to evaluate the involvement of some DC-related inhibitory pathways, such as the expression of indoleamine-2,3-dioxygenase 1 (IDO1) and the activation of PD-L1/PD-1 axis. METHODS: AML patients were analyzed at diagnosis.Before and after DNR-based chemotherapy, patient-derived T cells were extensively characterized by FACS and analyzed for their capacity to produce IFN-γ in response to autologous blasts. The AML cell line HL-60 and primary AML cells were then exposed, in vitro, to different drugs, including DNR and, as control drug, Ara-C. Dying cells were tested for the surface expression of CRT and HSPs 70/90, the release of HMGB1 and ATP. Functionally, immature DCs generated from healthy donors were pulsed with DNR-treated AML cells. Then, loaded DCs were tested for the expression of maturation-associated markers and of inhibitory pathways, such as IDO1 and PD-L1 and used to stimulate autologous CD3+ T cells. After co-culture, autologous healthy donor T cells were analyzed for IFN-g production, PD-1 expression and Tregs induction. A mouse model was set up to investigate in vivo the mechanism(s) underlying ICD in AML. The murine myelomonocytic leukemia cell line WEHI was transfected with luciferase PmeLUC probe, inoculated subcutaneously into BALB/c mice and used to measure in vivo ATP release after chemotherapy. Tumor-infiltrating T cells and DCs were characterized and correlated with ATP release. RESULTS: DNR treatment induced ICD-related modifications in both AML cell lines and primary blasts, including CRT, HSP70 and HSP90 exposure on cell surface, HMGB1 release from nucleus to cytoplasm and supernatant increase of ATP. Ex vivo, T-cell monitoring of DNR-treated AML patients displayed an increase in leukemia-specific IFN-g-producing CD4+ and CD8+ T cells in 20/28 evaluated patients. However, FACS analysis of CD8+ effector T cells emerging after chemotherapy showed a significant up-regulation of exhaustion marker such as LAG3 and PD-1, which paralleled with their reduced ability to produce active effector molecules, such as perforin and granzyme. Moreover, an increase of circulating Tregs was observed after DNR-based chemotherapy. In vitro, loading of chemotherapy-treated AML cells into DCs resulted not only in the induction of a maturation phenotype, but also in over-expression of inhibitory pathways, such as IDO1 and PD-L1. The silencing of IDO1 increased the capacity of DCs loaded with DNR-treated AML cells to induce leukemia-specific IFN-γ production by CD4+ and CD8+ T cells. In vivo, DNR therapy of mice inoculated with established murine AML cell line resulted in increased ATP release. Similarly to ex vivo and in vitro results, tumor-infiltrating DCs showed an increase in maturation status. Moreover, CD4+ and CD8+ T cells had increased IFN-γ production, but showed an exhausted phenotype. CONCLUSIONS: Our data confirm that chemotherapy-induced ICD may be active in AML and results in increased leukemia-specific T-cell immune response. However, a deep, ex vivo, in vitro and in vivo characterization of chemotherapy-induced T cells demonstrated an exhausted phenotype, which may be the result of the inhibitory pathways induction in DCs, such as IDO and PD-L1. The present data suggest that combination of chemotherapy with inhibitors of IDO1 and PD-L1 may represent an interesting approach to potentiate the immunogenic effect of chemotherapy, thus resulting in increased anti-leukemia immune response. Disclosures Cavo: Janssen-Cilag, Celgene, Amgen, BMS: Honoraria.

Description: Abstract 278 Indoleamine 2,3-dioxigenase (IDO) is the rate-limiting enzyme in tryptophan catabolism along the kynurenine pathway. IDO expression by different cell subsets inhibits T-cell activation, proliferation and survival and induces regulatory T cells (Tregs). Although human monocyte-derived dendritic cells have been shown to express IDO, little is known about its expression in other subsets of human DCs, including those generated from CD34+ hematopoietic progenitors (CD34+-derived DCs). In particular, no data are currently available for IDO expression in CD34+-derived DC subsets, such as dermal DCs and epidermal Langherans cells (LCs), which are physiologically resident DCs in the skin and act as the main sentinels against pathogens. In the present study, we performed a full characterization of IDO expression and function by human DCs generated from CD34+ cells. CD34+-derived DCs were generated from purified CD34+ cells after 7 days of culture with GM-CSF and TNF-alpha. In some experiments, day 7-DCs were sorted into different subsets and tested for IDO expression and function. Alternatively, day 7-DCs were cultured with GM-CSF and IL-4 and then matured with a cytokine cocktail of maturation stimuli (IL-1beta, TNF-alpha, IL-6, PGE2). After culture, DCs were analyzed for IDO expression by real-time PCR and western immunoblot, kynurenine production, inhibition of allogenic proliferation and Tregs induction. CD34+ cells did not express IDO mRNA, without significant differences among different progenitor cell sources (cord blood, mobilized peripheral blood, bone marrow). During DC differentiation, IDO mRNA expression was observed at day 7, but not at day 14. At day 14, maturation stimuli induced a marked up-regulation of IDO mRNA and protein, which resulted in increased kynurenine production and inhibition of T-cell proliferation. Flow cytometry analysis of day-7 cells revealed a double population, which comprised CD14+CD1a− dermal DCs and CD14−CD1a+ LCs. Interestingly, IDO mRNA and protein were observed only in dermal DCs, but not in LCs. IDO expression by dermal DCs resulted in increased production of kynurenine and in reduced allostimulatory capacity of T-cell proliferation. Dermal DCs were shown to induce a population of CD4+CD25+Foxp3+ which acted as Tregs by inhibiting allogeneic T-cell proliferation. This effect was abrogated by the addition of the IDO inhibitor 1-methyl tryptophan. Interestingly, dermal DCs and LCs showed a differential expression pattern of chemokine receptors. In particular, while both LCs and dermal DC expressed CCR7, CXCR4 expression segregated only in IDO-expressing dermal DCs. These data may correlate to different trafficking features of CD34+-derived DC subsets and may be associated to IDO expression. In conclusion, DC differentiation of CD34+ cells results in the expression of a functionally active IDO protein in dermal DCs, but not in LCs. Given the role of IDO in regulating immune tolerance, our data may suggest that within the complex skin microenvironment dermal DCs are intrinsically committed to function as regulatory DCs, whereas LCs are devoted to act as activating DCs. These data have implications for a better understanding of the development of the immune response during inflammation/infection. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 1047 Indoleamine 2,3-dioxygenase (IDO) enzymes (IDO1 and IDO2) critically regulate the rate-limiting step of tryptophan degradation along kynurenine pathway in several cellular subsets, including dendritic cells (DCs). In particular, it is known that IDO1 expression in DCs is induced by inflammatory mediators, especially PGE2, and enhances DCs tolerogenic capacity, such as the induction of T regulatory cells (Tregs). In absence of exogenous stimuli, IDO1 is driven to proteasomal degradation by SOCS3. Conversely, few and contrasting results are available about IDO2 expression and function in human DCs. We, then, besides IDO1, investigated the expression of IDO2 in human DC subsets, obtained from blood samples of healthy volunteers. Circulating myeloid and plasmacytoid DCs were purified and analyzed for the expression of both IDO enzymes. Our data demonstrate that myeloid DCs express higher levels of both IDO1 and IDO2 mRNA in comparison to plasmacytoid DCs. Moreover, PGE2 modulates IDO1 and IDO2 mRNA expression only in myeloid DCs and has no effect in plasmacytoid DCs. At protein level, IDO1 is expressed only in myeloid DCs and is modulated by PGE2, whereas IDO2 is expressed in both myeloid and plasmacytoid DCs and is not modulated by PGE2. Interestingly, when gene transcription or protein translation are inhibited, IDO2 protein expression persists independently from PGE2, while IDO1 expression requires PGE2 presence and needs continuous transcription and translation. Such discrepancy might be derived by a different binding of IDO1 and IDO2 to SOCS3, which in absence of pro-inflammatory stimuli is known to bind IDO1 and to drive it to proteasomal degradation. Indeed, SOCS3 immunoprecipitation and proteasomal inhibition experiments show that SOCS3 does not bind IDO2 which is not, therefore, driven to proteasomal degradation and is stably expressed. At the functional level, both myeloid and plasmacytoid DCs generate Tregs through an IDO-dependent mechanism, since the addition of IDO inhibitors 1-MT-D and 1-MT-L abrogate Tregs generation. More specifically, the silencing of IDO2 through its specific siRNA downregulates kynurenine production and Tregs generation, thus demonstrating that IDO2 is functionally active in DCs. In conclusion, our data demonstrate that, besides IDO1, human DCs also express IDO2, which contributes to tolerance by degrading tryptophan into kynurenine and by inducing Tregs. However, unlike IDO1, IDO2 expression is independent from inflammatory mediators, such as PGE2, and does not require continuous synthesis. These findings indicate IDO pathway as a critical check-point in the fine-tuning of DC-mediated tolerance induction. In particular, while IDO2 expression may confer a homeostatic tolerogenic capacity to DCs upon steady-state conditions, inflammation results in the induction of additional mechanisms of immune tolerance, such as the up-regulation of IDO1. Disclosures: Metz: New Link Genetics Corporation: Employment.

Description: Abstract 3231 The indoleamine 2,3-dioxygenases IDO1 and IDO2 participate in tryptophan catabolism along the kynurenine pathway and are specifically inhibited by 1-methyltryptophan (1-MT). IDO1 expression in dendritic cells (DCs) is upregulated by various maturation stimuli. In particular, PGE2 plays a crucial role in inducing IDO1-expressing DCs. While it is well established that IDO1-expressing DCs contribute to immunological tolerance by a number of mechanisms, little is known about the expression, function and regulation of IDO2 in DCs. Here we show that immature monocyte-derived DCs (Mo-DCs) express IDO2 which, like IDO1, is also upregulated upon maturation with a prostaglandin-E2 (PGE2)-containing cytokine cocktail. However, while IDO1 upregulation is strictly dependent on PGE2 during maturation, IDO2 expression level is not affected by PGE2. We show that IDO2 expressed by Mo-DCs is enzymatically active, producing L-kynurenine, and that this activity affects T-cell functions by inhibiting allogeneic and autologous T-cell proliferation and by inducing formation of a population of T regulatory cells (Tregs). Taken together, our findings establish different regulatory pathways for IDO2 and IDO1 in Mo-DCs, where IDO2 is functionally active, and they demonstrate that the tolerogenic function of Mo-DCs relies in part on PGE2-independent expression of IDO2. Disclosures: Metz: New Link Genetics Corporation: Employment. Prendergast:New Link Genetics Corporation: Employment.

Description: BACKGROUND: Overall survival of adult acute myeloid leukemia (AML) remains poor due to the lack of novel and effective therapies. The cancer cell death induced by some chemotherapeutic agents, especially anthracyclines, such as daunorubicin (DNR), named immunogenic cell death (ICD), is characterized by intra- and peri-cellular modifications, which favor the induction of anti-tumor T-cell immune response. Among them, the extracellular release of adenosine triphosphate (extracellular ATP, eATP) from dying tumor cells primes dendritic cells (DCs) by activating purinergic P2X7 receptors, thus eliciting the presentation of tumor antigens to T cell. DCs are key regulators of adaptive immunity, promoting or suppressing T-cell responses. One of the suppressive mechanisms involves the expression of indoleamine 2,3-dioxygenase 1 (IDO1), which plays a major role in the induction of T-cell tolerance through the expansion of regulatory T cells (Tregs). The present study aimed at evaluating the involvement of IDO-1 during ATP-driven ICD in AML. METHODS: AML patients were analyzed at diagnosis and after DNR-based chemotherapy. Ex vivo T cells were characterized by FACS and tested for their capacity to produce IFN-γ in response to autologous blasts. Then, CD8+IFN-γ-producing T cells were expanded and further characterized. ATP was used as an ICD representative model. In vitro, murine WEHI-3B and human HL-60 leukemic cell lines and primary blasts were tested for ATP release after DNR treatment. To in vivo investigate DNR-induced ICD, WEHI-3B cells stable transfected with luciferase PmeLUC were inoculated subcutaneously in BALB/c mice to measure ATP release directly from tumor mass. Tumor infiltrating DCs and T cells were characterized by FACS and immunohistochemistry after chemotherapy and plasma levels of cytokines were measured. In vitro DNR-treated AML cells were pulsed into immature DCs, previously generated from healthy donors. DCs maturation and IDO1 expression were examined (by FACS and western blot, respectively) and correlated with the presence of ATP in culture medium. IDO-driven Tregs induction was assessed. Finally, functional immunological tests were performed in vitro to test the ability of Tregs to inhibit leukemia antigen-specific IFN-γ production (FACS analysis) by ICD-activated T cells. RESULTS: After chemotherapy, 15/23 AML patients had an increase in leukemia-specific IFN-γ producing CD4+ and CD8+ T cells. Also an increase of Tregs was observed with a peak at day 21. CD8+ IFN-γ-producing T cells, which resulted in a skewing toward an effector memory phenotype, were activated and cytotoxic against autologous AML blasts but showed features of exhaustion and were defective in perforin production. In vitro and in vivo DNR induced ATP release from AML cells. In vivo the analysis of tumor-infiltrating T cells after treatment has shown an exhausted phenotype of cytotoxic CD8+ cells, increased IFN-γ+ Tregs and decreased TNF-α+ effector T cells. DNR treatment also increased in vivo plasma levels of cytokines IFN-γ, IL-1β, TNF-α, IL-12. Moreover, in DNR-treated mice we observed a significant increase of CD11c+ mature DCs which express IDO1 in tumor infiltrate. In vitro, loading of DNR-treated AML cells into DCs resulted in increased maturation, but also in IDO1 induction. Interestingly, extracellular ATP was directly involved in DCs maturation and IDO1 expression via purinergic receptor P2Y11. ICD-driven DCs were able to expand Tregs in an IDO-dependent manner. Finally, ICD both triggers a leukemia-specific IFN-γ production by CD8+T cells and induces Tregs, via IDO1-expressing DCs, which in turn inhibit leukemia-specific T cell. CONCLUSIONS: Overall, our data indicate that in AML chemotherapy-induced ICD has contrasting, and not fully elucidated, effects on T-cell immune response, resulting in the induction of leukemia-specific CTLs, albeit with defective features, and Tregs. In this scenario, the effects of ATP release from dying leukemia cells on DCs may be pivotal, as indicated by its capacity to concomitantly induce DC maturation and activation as well as tolerogenic function via IDO1. The combination of novel immunological drugs, such as IDO1 and/or checkpoint inhibitors, with conventional chemotherapy may represent an interesting approach to contrast tolerance induction and, then, fully exploit the immunogenic effect of chemotherapy. Disclosures Martinelli: Genentech: Consultancy; Amgen: Consultancy, Speakers Bureau; Celgene: Consultancy, Speakers Bureau; Novartis: Speakers Bureau; Ariad: Consultancy, Speakers Bureau; Roche: Consultancy, Speakers Bureau; Pfizer: Consultancy, Speakers Bureau; MSD: Consultancy; BMS: Speakers Bureau. Cavo:Celgene: Consultancy, Honoraria; Millennium: Consultancy, Honoraria; Bristol-Myers Squibb: Consultancy, Honoraria; Janssen-Cilag: Consultancy, Honoraria; Amgen: Consultancy, Honoraria.