Search

Circulating myeloma tumor cells (CTCs) as defined by the presence of peripheral blood (PB) clonal plasma cells (PCs) are a powerful prognostic marker in multiple myeloma (MM). However, the biological features of CTCs and their pathophysiological role in MM remains unexplored. Here, we investigate the phenotypic, cytogenetic, and functional characteristics as well as the circadian distribution of CTCs vs paired bone marrow (BM) clonal PCs from MM patients. Our results show that CTCs typically represent a unique subpopulation of all BM clonal PCs, characterized by downregulation (P < .05) of integrins (CD11a/CD11c/CD29/CD49d/CD49e), adhesion (CD33/CD56/CD117/CD138), and activation molecules (CD28/CD38/CD81). Fluorescence in situ hybridization analysis of fluorescence-activated cell sorter-sorted CTCs also unraveled different cytogenetic profiles vs paired BM clonal PCs. Moreover, CTCs were mostly quiescent and associated with higher clonogenic potential when cocultured with BM stromal cells. Most interestingly, CTCs showed a circadian distribution which fluctuates in a similar pattern to that of CD34(+) cells, and opposite to stromal cell-derived factor 1 plasma levels and corresponding surface expression of CXC chemokine receptor 4 on clonal PCs, suggesting that in MM, CTCs may egress to PB to colonize/metastasize other sites in the BM during the patients' resting period.

Background: Venetoclax is a BCL-2 inhibitor particularly effective in patients with multiple myeloma (MM) harboring the t(11;14). However, resistance to venetoclax has been linked to MCL-1 overexpression. On the other hand, it is wellknown that MM cells depend on MCL-1 rather than BCL-2 for survival, and this dependence has recently been reported to be enhanced by the tumor-associated microenvironment. Therefore, the combination of venetoclax with the potent MCL-1 inhibitor S63845 arises as a promising and novel approach for the treatment of MM. Aims: To evaluate the efficacy and mechanism of action of S63845 alone and in combination with venetoclax in absence and presence of the bone marrow tumor microenvironment in preclinical in vitro, ex vivo and in vivo models of MM. Methods: S63845 was provided by an agreement with Servier and Novartis. In vitro activity of S63845 and venetoclax alone and in combination was evaluated by bioluminescence on a MM cell line expressing luciferase (MM.1S-luc) in absence and presence of mesenchymal stromal cells isolated from bone marrow aspirates of MM patients (pMSCs). MM.1S cells cultured in absence or presence of pMSCs were analyzed for MCL-1 and BCL-2 protein levels by Western blot. Interactions between these anti-apoptotic proteins with the pro-apoptotic protein BIM were assessed by immunoprecipitation assays. The efficacy of S63845 and venetoclax alone and in combination was also evaluated ex vivo in MM cells and normal lymphocytes from MM patients. Finally, a disseminated MM model in BRG mice was used for in vivo studies. Results: S63845 and venetoclax showed a strong antimyeloma dose-dependent effect on MM.1S-luc cells co-cultured with pMSCs. However, whereas the presence of tumor-associated MSCs increased the IC50 value of venetoclax in MM.1S-luc cells from 6.2 to 9.8 mM, it reduced that of S63845 from 94.1 to 81 nM, suggesting a mild sensitization to this drug in the context of the microenvironment. Neither S63845 nor venetoclax affected pMSC viability even at high concentrations by MTT assay. The co-culture with the BM stromal microenvironment increased MCL-1 expression on untreated MM.1S cells in two out of four experiments performed with MSCs from different MM patients, whereas it surprisingly induced a decrease on BCL-2 levels in all of them. Treatment with S63845 completely blocked MCL-1 binding to BIM, both in the absence or presence of pMSCs but did not induce the compensatory increase of BCL-2/BIM complexes observed in MM.1S cells in monoculture. Venetoclax also completely blocked the binding of BCL-2 to BIM in MM.1S alone or in co-culture, and induced a similar compensatory increase of MCL-1/BIM complexes in both situations. Importantly, the double combination S63845 + venetoclax was significantly superior to both drugs in monotherapy in killing MM.1S-luc cells co-cultured in the presence of the stromal microenvironment. BIM immunoprecipitation assays showed that the double combination was able to counteract the compensatory upregulation of MCL-1 bound to BIM observed on MM.1S cells treated with venetoclax and to entirely disrupt BCL-2/BIM complexes, both in the absence and presence of pMSCs. Furthermore, S63845 + venetoclax increased the percentage of apoptotic MM plasma cells from three MM patients with respect to single treatments with moderate toxicity detected on normal lymphocytes, suggesting the existence of a therapeutic window for the double combination. Finally, the combination of S63845 + venetoclax clearly delayed tumor growth as compared with the agents in monotherapy in a disseminated model of MM with statistically significant differences from day 19 of treatment. This in vivo effect translated into a significatively improved survival for mice treated with the double combination (median 60 days) vs control mice (median 32 days; log-rank test P=0.045). Conclusion: Our preclinical data demonstrate the potent activity of the combination of venetoclax with S63845 in MM even in presence of the stromal associated-tumor microenvironment, and provides the rationale for the clinical development of this combination in relapsed or refractory MM patients. This project was supported by Novartis Pharmaceuticals and by the Spanish , ISCIII-FIS PI15/00067 and PI15/02156, GRS 1604/A/17 and CRMRTC de Castilla y León. Predoctoral grant to EMA by Consejería de Educación de Castilla y León. Disclosures Schoumacher: Servier: Employment. Banquet:Servier: Employment. Kraus-Berthier:servier: Employment. Kloos:Servier: Employment; Novartis: Other: Partnership. Halilovic:Novartis: Employment, Equity Ownership. Maacke:Novartis: Employment. Mateos:Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; GSK: Consultancy, Membership on an entity's Board of Directors or advisory committees; Amgen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Abbvie: Consultancy, Membership on an entity's Board of Directors or advisory committees; Janssen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; GSK: Consultancy, Membership on an entity's Board of Directors or advisory committees; Amgen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Takeda: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees. Ocio:AbbVie: Consultancy; Novartis: Consultancy, Honoraria; BMS: Consultancy; Seattle Genetics: Consultancy; Janssen: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Pharmamar: Consultancy; Sanofi: Research Funding; Amgen: Consultancy, Honoraria, Research Funding; Mundipharma: Research Funding; Celgene: Consultancy, Honoraria, Research Funding; Array Pharmaceuticals: Research Funding.

It is well known that the efficiency of Herpes simplex virus thymidine kinase gene/ganciclovir (HSV-tk/GCV) therapy is improved by the bystander effect, which mainly relies on gap junctional intercellular communication (GJIC). Malignant gliomas communicate poorly through gap junctions, consequently, agents with the ability to increase GJIC are good candidates to improve the efficiency of this therapy. Since we previously showed that the inhibition of ATP-sensitive potassium (KATP) channels promoted by tolbutamide increased GJIC in rat C6 glioma cells, we have investigated whether tolbutamide could increase the bystander effect in HSV-tk/GCV therapy against human glioma cells. We found that tolbutamide increased GJIC in U373 human glioma cells, an effect that was due to the up-regulation of connexin43, a protein that forms gap junctions channels. More interestingly, our results show that tolbutamide increased the efficiency of HSV-tk/GCV in co-cultures containing U373 cells and U373 cells transfected with HSV-tk. This effect was impaired in the presence of carbenoxolone, an inhibitor of GJIC. Furthermore, tolbutamide did not enhance the bystander effect in connexin43-silenced co-cultures. Together our results reveal that the inhibition of KATP channels promoted by tolbutamide enhances the bystander effect in HSV-tk/GCV therapy by increasing connexin43-mediated gap junctional intercellular communication in U373 human glioma cells.

Introduction: Filanesib (ARRY-520) is a novel inhibitor of the "kinesin spindle protein" (KSP), which has demonstrated efficacy in heavily pretreated patients with refractory MM, (Lonial et al, ASH 2013). Our preliminary studies demonstrated synergy with standard anti-MM agents, especially with pomalidomide and dexamethasone. This set the stage for a recently activated trial being run by the Spanish MM group investigating FPD in relapsed MM patients. In this abstract we investigate the mechanisms underlying the synergy of the combination. Methods: In vitro action of FPD was evaluated in MM cell lines by MTT assay, bioluminescence, Annexin V staining, cell cycle profile analysis and TMRE staining by flow cytometry. Synergy was quantified with the Calcusyn software. In vivo efficacy was assessed in a subcutaneous plasmacytoma model of MM1S in CB17-SCID mice (The Jackson Laboratory, Bar Harbor, ME, USA). The mechanism of action was analyzed by Western blot, flow cytometry, genomic techniques, immunohistochemistry and immunofluorescence techniques. Results: The triple combination of FPD resulted in clear synergy in multiple myeloma cell lines (MM1S, OPM2, and RPMI8226) with combination indices between 0.4-0.7, and abrogated the effect of the soluble cytokines IL-6 and IGF-I and the protective effect of the adhesion of plasma cells to BMSCs, HS-5 and TERT cells. FPD caused cell cycle arrest in G2/M and specific apoptosis of cells arrested in these proliferative phases (with apoptosis percentage of 5, 23, 58 and 88 for control, poma+dexa, filanesib and FPD, respectively) demonstrated by flow cytometry with DRAQ5 and Annexin-V. Thus, FPD and filanesib in monotherapy treatments induced a similar effect on the cell cycle profile (arrest in G2/M) with a concordant increase of cyclin B1 and phosphorylated Histone H3. Although a secondary increase of KSP protein levels would be expected, pomalidomide and dexamethasone induced a decrease of the levels of this protein, which was still present in the triple combination (FPD). This fact could be contributing to the potentiation observed with the combination. Attending to apoptosis mechanism, proapoptotic stimulus from the extrinsic and intrinsic apoptotic pathways were promoted by pomalidomide and dexamethasone and filanesib, and converged in the triple combination. In this regard, a decrease of MCL-1 (antiapoptotic protein) and a significant increase of the proapoptotic BCL2 family members of the intrinsic pathway like NOXA and BIMEL BIML, BIMS(this last one being the most potent proapoptotic isoform), tBID (extrinsic pathway) and Bax protein were observed. We confirmed that all these proteins were translocated into the mitochondria, resulting in a decrease of the mitochondrial membrane potential by TMRE, increase of permeability and a release of cytochrome C and AIF. These results were confirmed in vivo in a model of subcutaneous plasmacytoma in small (70 mm3) and large (2000 mm3) tumors. In this model we observed a significant reduction of tumor growth, which was correlated with a statistically significant improvement in survival. Changes induced by FPD in the gene expression profile were concordant with the in vitro results as several overexpressed genes belonging to the previous pathways were identified, such as spindle assembly checkpoint (CENP-E and CENP-F) and apoptosis (BCL2L11, gene that codifies BIM protein). Furthermore, IHC of tumors treated with FPD showed more apoptosis by TUNEL and a significant increase of monopolar spindles (2, 0, 53 and 140 per 10 high-power fields, for control, poma+dexa, filanesib and FPD, respectively). Conclusions: The synergy observed with filanesib in combination with pomalidomide and dexamethasone is the result of several coincidental mechanisms: a potentiation of the KSP inhibition with a subsequent increase in monopolar spindle formation and a simultaneous activation of the intrinsic and extrinsic pathways of apoptosis. In this regard, NOXA, BIM, BAX and tBID are probably the central players that, through different mechanisms, inhibit antiapoptotic proteins (MCL-1, BCL2 and BCL-XL) and promote mitochondrial outer membrane permeabilization and the release of apoptogenic factors such us cytochrome C and AIF. This work was funded in part by the company Array BioPharma. Disclosures Tunquist: Array BioPharma: Employment. Mateos:Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees; Amgen: Consultancy, Membership on an entity's Board of Directors or advisory committees; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees; Takeda: Consultancy, Membership on an entity's Board of Directors or advisory committees; Onyx: Consultancy; Janssen: Consultancy, Membership on an entity's Board of Directors or advisory committees; BMS: Consultancy. Ocio:Jassen: Honoraria; Celgene: Honoraria, Research Funding; Pharmamar: Consultancy, Research Funding; MSD: Research Funding; Novartis: Consultancy, Research Funding; Mundipharma: Consultancy, Research Funding; Bristol Myers Squibb: Consultancy; Amgen/Onyx: Consultancy, Honoraria, Research Funding; Array BioPharma: Consultancy, Research Funding.

Background and objectives. PIM kinases (PIM1, PIM2, PIM3) are proteins known to be overexpressed in several hematological malignancies. In particular, in chronic lymphocytic leukemia (CLL) they are involved in cell survival, resistance to apoptosis (especially PIM2 and PIM3) and interactions with the microenvironment (PIM1). The aim of this study was dual: I) to evaluate the preclinical efficacy of PIM447, a pan PIM kinase inhibitor, in CLL and to study potential synergies with other drugs; and II) to evaluate the expression of PIM-kinases in different stages of the disease and correlate it with the prognosis and the sensitivity to the drug. Methods. Peripheral blood samples from untreated patients with different stages of the disease (monoclonal B lymphocytosis (MBL), stable CLL not requiring treatment (sCLL), and active CLL requiring treatment (aCLL)) were collected after informed consent. The ex vivo efficacy of PIM447 was analyzed by flow cytometry with annexin V in these samples. Moreover, PIM447 efficacy was also analyzed in two cell lines (MEC-1 and JVM-2) by MTT assay. Synergy with other drugs effective in CLL (bendamustine and fludarabine) was evluated with the calcusyn software. Protein levels of PIM Kinase proteins were evaluated by capillary electrophoresis immunoassay (WESTM ProteinSimple) in monoclonal B cells purified by CD19 selection with anti-CD19 magnetic microbeads and the autoMacs Cell separator (both from Miltenyi Biotec) from a subset of patients. Results. The pan PIM inhibitor, PIM447 was active in both cell lines tested, MEC-1 (IC50 5μM) and JVM2 (IC50 7μM), and also in monoclonal B cells from freshly isolated patients samples (sCLL=11; aCLL=5), with no difference in sensitivity between the different stages of the disease (IC50 of 4,8 μM and 4,7 μM for sCLL and aCLL respectively). There was a clear therapeutic window as treatment with PIM447 at doses toxic for monoclonal B cells, preserved T lymphocytes (figure 1) (median % of apoptosis for B cells and T lymphocytes respectively of 23 vs 20 at 5μM and 87 vs 35 at 10 μM). Moreover, PIM447 demonstrated to potentiate the activity of both bendamustine and fludarabine, being especially synergistic with this last one (combination index 0.1-0.6). A second objective was to analyze PIM2 protein expression by western blot in monoclonal B cells from these samples and correlate it with clinical and biological features. Up to now, it has been evaluated in 18 samples (MBL=4; sCLL=8; aCLL=6,). All of them expressed PIM-2. Expression levels of this protein were significantly higher in active CLL as compared with indolent stages of the disease (p=0,012). Patients with an unmutated IGHV status also displayed higher levels of PIM2 (p=0,01). Finally, samples with high PIM2 levels were slightly more resistant to PIM447 as compared with samples with lower protein levels (IC50 of 7,7 μM vs 5 μM, respectively). We are currently completing the analysis of the PIM2 levels of remaining samples and we are also measuring the levels of PIM1 protein, what will be available at the meeting. Conclusions: PIM-Kinase inhibition with PIM447 is effective in vitro in CLL cell lines and ex vivo in samples from patients. It synergizes with other agents especially fludarabine. PIM2 protein levels correlated with the clinical activity of CLL and with the mutational state of IGHV. Although all patients appear sensitive ex vivo to PIM447, further work is required to define PIM2 expression as a marker of sensitivity. Figure 1. Figure 1. Disclosures Ocio: Array BioPharma: Consultancy, Research Funding; Celgene: Honoraria, Research Funding; Amgen/Onyx: Consultancy, Honoraria, Research Funding; Bristol Myers Squibb: Consultancy; Mundipharma: Consultancy, Research Funding; Novartis: Consultancy, Research Funding; MSD: Research Funding; Pharmamar: Consultancy, Research Funding; Jassen: Honoraria.

Introduction. EDO-S101 is a hybrid molecule of bendamustine plus vorinostat, new in its class. Our group has previously demonstrated that EDO-S101 is effective in vitro in MM cell lines independently of p53 state, and also in a murine plasmacytoma model where it decreases tumor growth and prolongs survival with respect to bendamustine and/or vorinostat treatment. The objective of this work was to gain further insights into the efficacy of EDO-S101, its mechanism of action and its combination with other drugs used in MM. Methods. The mechanism of action was assessed by western blot, comet assay, immunohistochemistry, and flow cytometry. Homologous recombination (HR) efficiency was calculated using chromosomally integrated green fluorescent protein reporter construct-based assay. The efficacy of different combinations was studied in vitro (HMCLs), in vivo (murine plasmacytoma model CB-17 SCID mice) and ex vivo (cells from patients). Results. In addition to the activity of EDO-S101 in MM cell lines we demonstrated that it was active ex vivo in cells isolated from 7 MM patients, with median IC50 of 5 µM (ranging from 1,8 to 8 µM), some of them previously exposed and resistant to alkylators such as melphalan. Interestingly, EDO-S101 could also overcome alkylators-resistance in vitro, as it was active in melphalan resistant cells (U266-LR7 and RPMI8226-LR5). EDO-S01 was also effective in the presence of factors that confer proliferative advantage to plasma cells, like IL-6, IGF or co-culture with mesenchimal cells hMSC-TERT. Regarding its mechanism of action, we found that the apoptosis induced by EDO-S101 was caspase-independent but calpain-dependent, since PD150606, an inhibitor of this protein could overcome EDO-S101-induced apoptosis, whereas the caspase inhibitor Z-VAD -FMK did not. This data was consistent with the finding that under treatment with EDO-S101, MM1S cells showed AIF (apoptotic inducing factor) translocation from the mitochondria into the nucleus. Interestingly, the release of this pro-apoptotic protein from the mitochondria could be mediated by calpains, as it has been described in literature. We subsequently demonstrated that EDO-S101 causes DNA damage, as revealed by the phosphorylation and subsequent activation of several components of the DNA Damage Response (DDR) such as ATM, H2AX, chk1, chk2 or p53, and the induction of DNA fragmentation, that was detected by the comet assay. EDO-S101 was also found to induce cell cycle arrest in different phases depending on the dose and cell line. It has previously been suggested that DACi may impair DNA repair by inhibiting homologous recombination (HR), a pathway related with genomic instability and progression, very active in MM. Therefore we next evaluated the efficiency of HR using a reported construct that was chromosomally integrated in two MM cell lines, JJN3 and U266. Treatment with EDO-S101 significantly reduced the efficiency of HR in both cell lines, by 50% and 20% of untreated controls respectively. Finally, we tested potential combinations with other antimyeloma agents like lenalidomide and thalidomide; and also with proteasome inhibitors (bortezomib, carfilzomib and oprozomib). EDO-S101 potentiated the activity of all these agents, but the most synergistic combination was that including Bortezomib + Dexamethasone (CI 0,4). This combination was also evaluated in vivo, where it significantly decreased tumor growth and prolonged survival compared to agents in monotherapy and in double combinations. We are currently deepening into the mechanism of action of this combination. Conclusions. EDO S101 is active ex vivo in cells isolated from patients and is able to overcome resistance to alkylators. It induces caspase-independent apoptosis, and cell cycle arrest in MM cell lines. These effects are due to the potent DNA damage which is enhanced by HR impairment induced by the hybrid molecule. Moreover, the combination with bortezomib and dexamethasone is especially attractive to be taken into the clinical setting. Disclosures Mehrling: 4Mundipharma-EDO GmbH, Basel, Switzerland: Employment. Mateos:Takeda: Consultancy; Janssen-Cilag: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Onyx: Consultancy.

PO-302 Amiloride induces potent antitumoral activity in multiple myeloma through the reactivation of mutant p53 E. Rojas, I. Misiewicz-Krzeminska, S.H. Garcia, T. Paino, L.A. Corchete, I. Isidro, M.V. Mateos, E. Ocio, N.C. Gutierrez Hematology Departament, University Hospital, IBSAL IBMCC (USAL-CSIC) Funding: “Instituto de Salud Carlos III” (PI13/00111), “Gerencia Regional de Salud” (BIO/SA57/13), “Asociacion Espanola contra el cancer” (AECC: GCB120981SAN). Background: The antihypertensive agent, amiloride, has been demonstrated to inhibit tumor cell proliferation in different experimental models. Here, we investigate the antimyeloma effect of amiloride in myeloma cell lines and primary samples, and the potential mechanisms involved. Methods: Seven human myeloma cell lines (NCI-H929, JJN3, KMS12-BM, KMS12-PE, U266, MM1S and RPMI-8226) and 6 bone marrow (BM) samples from MM patients were used for the experiments. Cell viability, apoptosis and cell cycle analyses were carried. The measuring of mitochondrial membrane potential (m) was evaluated by flow cytometry analyses. The gene expression was quantified by Taqman assay qRT-PCR. Synergy with other antimyeloma agents was calculated with the “CalcuSyn” software program. Results: To test the antiproliferative effect of amiloride, myeloma cells were treated with increasing concentrations of the compound (0.1mM-1mM) for 24, 48 and 72 hours. Amiloride displayed potent antimyeloma activity in the panel of 7 cell lines. A substantial induction of apoptotic cell death (greater than 40% at 24 h and 60% at 48 h) and cell cycle blockade was observed in all MM cell lines. The apoptosis induced by amiloride was not related to the TP53 mutational status, since amiloride was still able to induce cell death in p53 mutated cell lines, such as U-266, KMS12-BM and JJN3. Moreover, this drug caused a notable decrease of m. Mechanistic studies showed that apoptosis triggered by amiloride was associated with the overexpression of BAX, BAK1, BBC3, TNFRSF10B, FAS, CDKN1B and CDKN1A, even in those MM cell lines in which p53 gene was mutated. The cell death was reduced in MM cell lines under the treatments with p53 inhibitors (pifithrin, PFT and PFT), even in p53 mutated cell lines. These results suggest the reactivation of mutant p53 in amiloride-induced apoptosis in myeloma cells. The double combinations of amiloride with bortezomib, dexamethasone and melphalan were synergistic in NCI-H929, JJN3, RMPI and MM1S cell lines. In vitro studies confirmed the antimyeloma efficacy of amiloride in patient samples and revealed significant lower cytotoxicity in the remaining cell subpopulations compared to tumor plasma cells. Conclusions: Our results demonstrate a potent and selective antimyeloma effect of amiloride independent of p53 status. The present data support the investigation of amiloride as a treatment option for MM patients, either alone or in combination.

Introduction: Multiple myeloma (MM) is characterized by the presence of complex karyotypes and chromosome instability, suggesting that cell cycle checkpoints are defective. Filanesib (ARRY-520) is a highly selective, targeted inhibitor of kinesin spindle proteins (KSP), which are required to establish mitotic spindle bipolarity, driving centrosome separation. Filanesib monotherapy has demonstrated clinical activity in heavily pretreated MM patients (Lonial et al, ASH 2013). In this work we aimed to explore the preclinical activity of filanesib alone and in combination with IMiDs. Methods: In vitro activity of filanesib alone and in combination with IMiDs (thalidomide, lenalidomide and pomalidomide) was evaluated in MM cell lines by MTT assay and Annexin V, Propidium Iodide and DiOC6 analysis by flow cytometry, Western Blot and immunofluorescence. Synergy was quantified with combination indices (CI) by Calcusyn software. In vivo efficacy was assessed in a subcutaneous plasmacytoma model of MM1S in CB17-SCID mice. Results: Filanesib demonstrated significant activity in a broad panel of 11 MM cell lines, with 48-hour IC50 values ranging between 0.3 and 5 nM. Interestingly, the highest activity was observed in drug-resistant cell lines such as OPM-2 and RPMI-LR5. We next evaluated whether the cell death MoA was dependent on apoptosis or blockade of proliferation. Time response experiments performed in three different cell lines with different sensitivity to filanesib (OPM-2, MM1S and U266) showed accumulation of cells in G2/M, followed by loss of the mitochondrial membrane potential and activation of apoptosis. Accordingly, Western blot analysis demonstrated an activation of the mitotic checkpoint indicated by an increase in Cyclin B1, and activation of apoptosis with PARP, and caspase-3 and -7 cleavage. Furthermore, filanesib activity was very rapid as 15 minutes of exposure was sufficient to exert all of the apoptotic and cell cycle effects observed at 48 hours. Immunofluorescence microscopy using alpha-tubulin demonstrated that filanesib induces monopolar spindle formation. The sensitivity of MM to filanesib has been previously correlated with the cell-dependency of the anti-apoptotic protein Mcl-1. We, therefore, studied the basal levels of six Bcl-2 family members (Bcl-2, Bcl-XL, Mcl-1, Bax, Bak, Bad) in the 11 cell lines and observed a correlation between the basal levels of these proteins and drug sensitivity. In particular, we confirmed the relationship with the anti-apoptotic proteins Mcl-1 and Bcl-2 and demonstrated that conversely to what is observed with proteasome inhibitors, cells with high basal levels of pro-apoptotic Bax or Bak were more resistant to filanesib. Moreover, treatment with filanesib induced a clear decrease of Mcl-1 in the 3 cell lines analyzed (U266, MM1S and OPM-2) that coincided with a decrease of Bcl-2 in the most sensitive cell line, OPM-2. Finally, we evaluated the activity of filanesib in combination with IMiDs and dexamethasone. In vitro studies showed a synergistic effect of filanesib with dexamethasone (CI: 0.21), and with all IMiDs, being most pronounced with pomalidomide (CI: 0.09). Of note, this triple combination demonstrated the highest synergistic activity (CI: 0.06). These results were confirmed in vivo where the triple combination of filanesib, dexamethasone, and pomalidomide was also synergistic, with a significant reduction of tumor growth of up to 50 days, which correlated with a statistically significant survival improvement. Mechanistic studies on the combination are ongoing. Conclusions: Our results demonstrate the potent, rapid activity of filanesib which induces a cell cycle blockade through the inhibition of KSP, leading to apoptosis in MM. It was identified that this activity is dependent on the anti-apoptotic protein Mcl-1 and pro-apoptotic proteins Bax and Bak. Furthermore, in animal xenograft studies, filanesib exhibited robust synergism in combination with dexamethasone and any IMiD, pomalidomide being the most synergistic. These data are the rationale for the clinical trial "Pomdefil" which uses this combination in patients with relapsed refractory MM. The trial will begin soon as a collaboration of the Spanish Myeloma Group (GEM). This work was supported in part by Array BioPharma. Disclosures Humphries: Array Biopharma: Employment. Tunquist:Array Biopharma: Employment. Mateos:Array Biopharma: Honoraria. Ocio:Array Biopharma: Honoraria, Research Funding.

Background Alkylating histone deacetylase inhibitors (HDACi) enhance the anticancer efficacy of alkylators by increasing chromatin accessibility and also down regulating DNA repair. EDO-S101 is a first-in-class fusion molecule that combines DNA damaging effect of bendamustine with the pan-HDACi vorinostat. Objectives To study the bi-functional properties of EDO-S101 as an alkylating agent and a pan-HDACi in various in vitro and in vivo xenograft models of hematological malignancies. Methods In vitro inhibition of HDAC Class I and II enzymes by EDO-S101 and vorinostat was tested using an recombinant human enzymatic assay (BPS Bioscience, Enzo Life Science) and in vivo in rat peripheral blood mononuclear cells (PBMCs). The degree of inhibition was measured 1 hour following a single dose of 10–50 mg/kg i.v. and duration of inhibition over 24 hours after a single i.v. dose of EDO-S101 of 25 mg/kg. HDAC inhibition, alkylation and apoptotic activity were evaluated in vitro in myeloid (HL60 AML cell line) and lymphoid cell lines, including Daudi Burkitt’s lymphoma (BL) and a panel of 6 MM cell lines (MM1S, MM1R, RPMI-8226, RPMI-LR5, U266, U266-LR7). In vivo intra-tumor effects were analyzed after short courses of treatment with EDO-S101 in MM1S human plasmacytoma (PC) and BL xenograft models. Changes in pathway activation, protein expression and activities influencing the cell cycle were measured by Western blot and immunohistochemistry. Anti-tumor activity in vitro was measured by MTT and in vivo using a caliper to assess tumor size at regular intervals. Results In vitro, EDO-S101’s pan-HDACi activity, at nanomolar concentrations in Class I and II recombinant enzymes, was similar to vorinostat. In vivo, in intact rat PBMCs, HDAC inhibition was maximal at 1 hour after a single dose of 10 mg/kg i.v.–the dose where antitumor activity starts. HDAC inhibition did not increase with doses up to 50 mg/kg, recovery began within 3 hours and was nearly complete at 16 hours. In the AML HL60 cell line in vitro, hyperacetylation of lysine residues K9, K14, K23 and K56 on histone 3 was found after exposure to 2–4 µM of EDO-S101. Histone 3 and 4 hyperacetylation was also demonstrated in MM cell lines at 1–5 µM concentrations. In xenograft models of human plasmacytoma and BL, EDO-S101 induced histone 3 hyperacetylation, indicating an HDACi effect in vivo. Alkylating activity was demonstrated in vitro in HL60 and MM cell lines by DNA cross-linking and double strand break formation in the comet assay by immunofluorescence. In vivo, in xenograft models of human plasmacytoma (60 mg/kg d 1, 8, 15) and BL (40 and 80mg/kg d1) exposure to EDO-S101 caused a strong DNA-repair response shown by activation of pH2AX and p53 (PC and BL) followed by an increase of DNA damage check point proteins pCHK1 (PC) and even more prominent pCHK2 (PC and BL). The kinetics of this effect, studied in vivo in BL tumors, showed that the pH2AX response fell at Day 8 after dosing while the p53 response lasted, particularly in the group treated with 80mg/kg. In Daudi-bearing mice tumors, p-ATR was completely suppressed at Day 8 after treatment, which was not clear in the PC tumors. EDO-S101 triggered apoptosis in vitro and in vivo, resulting in strong antitumor activity in HL60, Daudi and the panel of six MM cell lines. Initial in vitro experiments in HL60 cells showed an activation of the intrinsic pathway of apoptosis with cleavage of caspases 3, 9 and PARP and a marked reduction of anti-apoptotic proteins XIAP and Mcl-1. In the MM cell line, MM1S activation of the intrinsic and extrinsic pathways of apoptosis (C 8, 9, 3, 7 and PARP cleavage) was seen with a loss of mitochondrial membrane potential by DiOC6. Tumors of human plasmacytoma and BL in vivo were rapidly shrinking or completely eradicated after i.v. administration of EDO-S101. A decrease in proliferation (Ki67) and slight PARP cleavage was found in the tumor tissue (PC), and evidence of activation of apoptosis by cleavage of caspases 7 and 9 at Day 4 and caspase 8 and PARP at Day 8 after treatment in BL tumors. The level of caspase 3, different to MM, remained unchanged. Importantly, EDO-S101 induced a rapid and dose-dependent strong decrease of XIAP and Mcl-1 which lasted until Day 8. Conclusions This study demonstrates the bi-functional mechanism of ED0-S101 in both myeloid and lymphoid hematological malignancies. The data support the clinical investigation of EDO-S101 in treating hematological malignancies. Disclosures Ocio: Mundipharma: Honoraria, Research Funding. Mehrling:Mundipharma: Employment.

MRD monitoring is emerging as a powerful prognostic biomarker in MM. In addition to its utility to redefine the depth of response achieved after therapy, the MRD clone represents, from a biological point of view, a very small fraction of tumor cells that are truly chemoresistant, potentially quiescent (not producing M-protein), and able to recapitulate the initial tumor burden at relapse. Thus, the MRD clone represents a unique model to understand chemoresistance (in otherwise serological responding patients) and the characteristics of eventual MM clonogenic cells; however, the MRD clone has never been characterized at the biological level. Here, we compared the immunophenotypic and genomic profiles of paired baseline vs MRD clonal PCs from a total of 40 newly-diagnosed elderly MM patients enrolled in the GEM2010 trial: sequential VMP (9 cycles) followed by Rd (9 cycles) vs alternating VMP with Rd (18 cycles). MRD monitoring was performed by multidimensional flow cytometry (MFC) at cycles 9 and 18. First, we focused on the immunophenotypic expression profiling (iPEP) of patient-paired baseline vs MRD clonal PCs through 23-color MFC, combining backbone markers for the identification of clonal PCs with additional integrins, adhesion, activation and maturation molecules. Our results show that the iPEP from MRD clonal PCs at cycle 9 differed from paired baseline tumor cells by significantly (P Using patient-specific aberrant phenotypes, we then sorted clonal PCs (purity ≥97%) by FACS for subsequent genomic studies. Patient-specific paired comparison of baseline vs MRD clonal PCs (n=11) genomic profile was performed by high density Cytoscan750K array. The pattern of copy number abnormalities (CNA; only those with minimum of 25 consecutive imbalanced markers/segment and minimum 100 Kb length were considered) significantly varied from baseline to the chemoresistant MRD PC clone (accounting for whole chromosome, chromosomal arm or interstitial imbalances). Individual patient analysis showed one case in which both PC clones showed exactly the same 25 CNA; a second case in which MRD cells showed +1p36.21, -9p23, -17q11.2 and +17q25.3 in addition to 10 CNAs already present at baseline; 4 cases in which MRD cells maintained initial chromosomal imbalances but lacked specific CNA (range: 1-34) present at diagnosis; and 5 patients in which MRD cells displayed none of the CNA (range: 7-20) detected at baseline. It should be noted that markedly subclonal heterogeneity was detected at diagnosis. Accordingly, in one out of the 5 latter cases baseline but not MRD clonal PCs had 1q+ and 13q- which were further confirmed by FISH (in 85% and 58% of clonal PCs, respectively); this suggests that MRD PCs were present at the baseline state as a minor sublclone with respect to the initial tumor bulk. On the other hand, copy number neutral loss of heterozygosity (larger than 3Mb) was slightly more frequent in MRD as compared to patient paired baseline clonal PCs (13 vs 8, respectively). Only in 5/40 patients sufficient RNA was extracted from baseline and paired MRD FACS-sorted clonal PCs to perform gene expression profiling (GEP; HumanGene 1.0ST). Although the numbers are small, it should be noted that GEP of baseline vs. patient-paired MRD clonal PCs mostly overlapped, with only 19 genes found to be down-regulated after chemotherapy (SAM Excel add-in with a FDR q-value In summary, this integrated phenotypic and genomic analysis suggests that from the baseline MM tumor bulk, primarily chemoresistant MRD PCs may be those potentially closer to the BM niche (with initial increased expression of integrins and adhesion molecules), as well as a more ancestral subclone (with less CNA) in a significant proportion of patients. Disclosures: No relevant conflicts of interest to declare.

Introduction Multiple myeloma (MM) is characterized by the accumulation of malignant plasma cells in the bone marrow (BM) and is closely associated with osteolytic lesions, in part due to an increase in the bone-resorptive activity and number of osteoclasts (OCs). The activation of survival pathways in myeloma cells could be the cause of treatment failure rendering the disease incurable. Pim kinases are a family of survival serine/threonine kinases composed of three members (Pim1, Pim2 and Pim3) that are overexpressed in MM cells and may have a role in MM pathogenesis. However, little is known about the role of Pim kinases in OCs and its involvement in myeloma bone disease. Here, we have evaluated the preclinical activity of a new pan-Pim kinase inhibitor, LGH447, on MM cells and OCs. Cell lines, primary samples, material and methods LGH447 was provided by Novartis Pharmaceuticals. The human MM cell lines MM1S, MM1R, RPMI-8226 (or RPMI-8226-luc), RPMI-LR5, MM144, NCI-H929, OPM-2, U266, U266-Dox4 and U266-LR7 were employed. PBMCs from healthy volunteers were used to generate OCs, whereas primary mesenchymal stromal cells (MSCs) were obtained from bone marrow aspirates of MM patients. Cell viability was studied using MTT colorimetric assay or bioluminescence. Apoptosis was measured by annexin-V staining. For cell cycle analysis, propidium iodide staining was used. OC formation was assessed by enumeration of multinucleated (≥3) TRAP-positive cells and OC resorption was assessed on calcium-coated slides. Immunoblotting, quantitative PCR and immunofluorescence were used to further investigate the mechanism of action of LGH447. Results All MM cell lines expressed the three isoforms of Pim kinases with higher levels of Pim2. The dose-response curves to LGH447 after a 48 hour treatment revealed two groups of MM cell lines with regard to sensitivity to this drug: high sensitive, with IC50 values ranging from 0.2 to 3.3 µM (MM1S, MM1R, RPMI-8226, MM144, U266 and NCI-H929); and low sensitive, with IC50 values >7 µM (OPM-2, RPMI-LR5, U266-Dox4 and U266-LR7). Our results indicated that LGH447 promoted apoptosis in myeloma cells as shown by the increase in annexin-V positive cells and by the cleavage of initiator (caspases 8 and 9) and effector caspases (caspases 3 and 7) and of PARP. LGH447 also blocked the cell cycle in MM cells as demonstrated by the increase in G0-G1 and the decrease in S-G2-M phases. Importantly, LGH447 was also able to overcome the growth advantage conferred to RPMI-8226-luc cells by co-culture with MSCs or OCs. Regarding the mechanisms involved in these effects, LGH447 inhibited the mTOR pathway, demonstrated by a decreased phosphorylation of the downstream mTOR effectors, 4EBP1 and S6 in residues Thr37/46 and Ser235/236, respectively. Interestingly, LGH447 also inhibited OC formation and resorption activity. LGH447 treatment of human pre-OCs diminished the expression of key molecules involved in OC differentiation (p-Erk1/2 and NFATc1) and function [CAII (carbonic anhidrase II), CLCN7 (chloride channel 7), ATP6V1A (vacuolar-H+-ATPase catalytic subunit A1) and MMP9 (matrix metalloproteinase 9)] and also disrupted the F-actin ring necessary for OC effective resorption. Conclusion Overall, our results demonstrate that both MM cells and OCs are targets of the pan-Pim kinase inhibitor, LGH447. Therefore, the inhibition of Pim kinases could potentially provide a dual benefit in myeloma patients as a consequence of cytotoxic effects exerted on MM cells and an anti-resorptive activity on bone. This work was supported by funding from the Fundación Española de Hematología y Hemoterapia (AG-G), Centro en Red de Medicina Regenerativa y Terapia Celular de Castilla y León, the RTICC-Hematology Group (RD12/0036/0058), Spanish FIS (PI12/02591) and the Junta de Castilla y León, Gerencia Regional de Salud (GRS 862/A/13). Disclosures: Off Label Use: LGH447 is a pan-Pim kinase inhibitor (Novartis Pharmaceuticals). It has been used for pre-clinical studies in multiple myeloma.

The identification of subclones in multiple myeloma (MM) could have clinical implications since the complete eradication of all clones is required to prolong patients’ survival. Hence, there is an unmet need to fully characterize such subclones. In the present study, we started by investigating using multidimensional (23-color) flow cytometry (MFC) the presence of ≥2 distinct bone marrow PC subclones in 116 newly-diagnosed (MM) patients included in the Spanish GEM2010 trial: sequential VMP (9 cycles) followed by Rd (9 cycles) vs alternating VMP with Rd (18 cycles). By principal component analysis (PCA) using the Infinicyt software, the immunophenotypic expression profile (based on the simultaneous evaluation of 23 antigens in single-clonal-PCs) showed the presence of two or more subclones in 35/116 (30%) newly-diagnosed MM patients. The most powerful antigens for the identification of subclonal heterogeneity were CD56, CD27, CD81, CD20, CD33, β7, CD138, CD49e, CD54, HLADR, CD19, CD44 and CXCR4. Then, we investigated the genetic profile of distinct phenotypic subclones (sorted by FACS with purity ≥97%) through high-density Cytoscan750K (n=5) arrays and FISH (n=8). In 2/5 (40%) patients, specific copy number abnormalities (CNA; only those with minimum of 25 consecutive imbalanced markers per segment and minimum 100 Kb length were considered) were present in one clone [del(13q), del(14q), del(20p), +(20q), +(21q), or del(22q)] but absent in the other. Similar results were observed by FISH while investigating those probes used as part of the clinical workout. Noteworthy, in three cases del(17p13) was detected in only one subclone, whereas del(13q14) was restricted to one subclone in two patients. To assess the clonogenic potential of distinct phenotypic subclones detected in the whole malignant PC compartment, these were sorted by MFC and seeded in coculture with the human mesenchymal cell line hTERT (with Methocult® supplemented with lymphocyte conditioned medium (10%) plus 20 ng/ml IL-6 plus 20 ng/ml IGF-1). In 2 out of 7 patients (29%) we observed different clonogenic potential between two patient-specific subclones (patient 1, 208 vs 0 colonies; patient 2, 5 vs 1 colony). Finally, we aimed to investigate if distinct phenotypic subclones display different chemoresistance. To address this question, we compared the immunophenotypic expression profile of paired baseline vs MRD clonal PC in 9 patients enrolled in the Spanish GEM2010 trial in which ≥2 subclones were identified at diagnosis. Our results show a phenomenon of clonal restriction in 7/9 (78%) patients after chemotherapy, indicating that among the initial tumor bulk, only specific subclones are primary chemoresistant (i.e.: MRD) in patients otherwise achieving a serological response. As an example, in one patient showing at diagnosis four clonal PC subsets (CD33-/CD117-, CD33+/CD117-, CD33+/CD117+ and CD33+/CD117+), only one subclone (CD33-/CD117-) persisted after 9 cycles of alterning Rd-VMP. In conclusion, our results show that distinct phenotypic subclones translate into a different cytogenetic profile and clonogenic potential. Since the balance between different subclones may change over the course of treatment, so may change the extent of specific cytogenetic abnormalities and the patient risk. Thus, careful monitoring of specific subclones may have implications towards an adapted therapy to overcome chemoresistant cells. This work was supported by funding from the RTICC-Hematology Group (RD12/0036/0058), the Asociación Española Contra el Cáncer (AECC) (GCB120981SAN) and Multiple Myeloma Research Foundation (MMRF) 2012 Research Fellow Award (B Paiva) Disclosures: No relevant conflicts of interest to declare.

Abstract 4014 Introduction: Bone destruction, a hallmark of multiple myeloma (MM), arises as a consequence of the interactions between MM cells and the bone marrow microenvironment, which lead to an increase in the bone-resorptive activity and number of osteoclasts (OC) and a reduction of the bone-forming activity and differentiation of osteoblasts (OB). MLN9708, which hydrolyzes to pharmacologically active MLN2238 in aqueous solution, is an investigational proteasome inhibitor (PI) with demonstrated preclinical anti-myeloma activity. However, it is currently not known whether MLN9708, may have a beneficial effect on myeloma-associated bone disease. Here, we have conducted in vitro and in vivo studies to evaluate its ability to promote osteogenic differentiation and to inhibit OC formation and function in the myeloma setting. Patient samples, material and methods: The human MM cell lines RPMI-8226 and MM.1S (or RPMI-8226-luc and MM.1S-luc) together with the mesenchymal stem hMSC-TERT cell line were employed. Also, MSCs from BM samples of healthy donors and MM patients were used in OB differentiation studies, whereas PBMCs from healthy volunteers were used to generate OCs. NOD.SCID.IL2Rγ−/− mice were used in the in vivo model of disseminated human MM. MLN2238 and bortezomib (Velcade) were provided by Millennium Pharmaceuticals, Inc. OB differentiation from MSCs and OB function were investigated by measurement of ALP activity, quantitative mineralization, luciferase reporter assays, siRNA gene silencing and real time RT-PCR. The effect of the new PI on OC formation was assessed by enumeration of multinucleated (≥3) TRAP-positive cells. Measurement of resorbed area, immunofluorescence and flow cytometry were used to further investigate the effect of MLN2238 on OC function. In our in vivo model, bioluminescence imaging, micro-CT analysis and serum levels of Igλ and bone markers were determined. Results: Physiologic concentrations of MLN2238 were able to stimulate the osteogenic differentiation of MSCs from both myeloma patients and healthy donors in vitro to an extent comparable to bortezomib; this was assessed by increased levels of ALP activity, higher expression of bone formation markers (Runx2, osterix, osteopontin and osteocalcin) and augmented matrix mineralization. The enhanced OB formation and function induced by MLN2238 was at least partly due to induction of T-cell factor 4 (TCF4) transcriptional activity, as well as to activation of the unfolded protein response. A similar range of MLN2238 doses also markedly inhibited OC formation and resorption from human progenitors. Similarly to that described with bortezomib, MLN2238 treatment of human pre-OCs prevented RANKL-induced NF-κB activation, disrupted the integrity of the F-actin ring and also reduced the expression of the αVβ3 integrin, thus contributing to inhibition of OC function. MLN2238 was also able to overcome the growth advantage conferred to MM.1S-luc cells by co-culture with MSCs or OCs. Oral administration of MLN2238 in a mouse model of disseminated human MM decreased human RPMI-8226-luc tumor burden as assessed by diminished bioluminescence signal and decreased serum levels of Igλ secreted by RPMI-8226-luc cells. In addition, MLN2238 prevented tumor-associated bone loss with significant increases in femoral trabecular bone parameters as compared to vehicle control animals. Serum markers of bone turnover showed that MLN2238 inhibited bone resorption (decreased levels of CTX) while enhancing bone formation (increased levels of P1NP). Conclusion: MLN2238 in vitro was capable of promoting osteoblastogenesis and OB activity as well as of inhibiting OC formation and function to an extent similar to bortezomib. In a disseminated human MM mouse model, orally administered MLN2238 showed anti-resorptive and bone-anabolic effects in addition to its anti-tumor properties. Given the thus far available data on the preclinical safety and favorable pharmacologic properties of MLN2238, it is conceivable that MLN9708, the clinical formulation of this proteasome inhibitor, may also achieve bone benefits in myeloma patients. Disclosures: Berger: Millennium Pharmaceuticals, Inc.: Employment. San-Miguel:Millennium Pharmaceuticals, Inc.: Consultancy.

Abstract 726 Circulating Tumor Cells (CTCs) are present in the vast majority of MM patients, but the biologic basis for the movement of CTCs from the bone marrow (BM) to the PB circulation is not clear. Are all BM myelomatous plasma cells (mPCs) capable to egress into PB, or only a specific sub-clone? Do CTCs have stem cell-like features and are enriched by clonogenic cells? Notably, hematopoietic stem cells (HSCs) also circulate in PB, under a specific circadian rhythm orchestrated by the central clock promoting the regeneration of the stem cell niche in other sites of the BM during the resting period; does circadian rhythms also affect CTCs? In the present study we investigated the phenotypic, functional and circadian characteristics of CTCs from symptomatic MM patients (at diagnosis or relapse), comparing them with patient' paired BM mPCs. The immunophenotypic characterization (n=13) was performed using 8 color multiparameter flow cytometry (MFC). CTCs and BM mPCs were sorted by MFC according to (patient specific) aberrant phenotypes, and clonogenic growth (n=6) was evaluated by platting the same number of cells in co-culture with the human BM stromal cell line (hTERT) at 10:1. Colonies consisting of more than 40 cells were scored at day 14. Proliferation (n=10) was measured using MFC with the DRAQ5 dye. The circadian rhythm of CTCs and HSCs in the PB of MM patients at relapse (n=5) was investigated by MFC starting at 16:00pm and quantifying both cell populations every 4h up to 12:00am next day (when patients' initiated treatment). Plasma was simultaneously collected and the levels of SDF-1 were measured using a quantitative ELISA assay. Among 20 antigens analyzed, CTCs differed (P≤.05) from paired BM mPCs by a down-regulation of the mean fluorescence intensity of the integrins CD11a, CD11c, CD41a, CD49d and CD49e (all these molecules known as mediators of cell adhesion and motility) (Figure 1A). Similarly, expression of the adhesion molecules CD38 and CD56 was significantly decreased in CTCs, and so was the stem-cell factor receptor CD117. CD81, a tetraspanin involved in B-cell activation and proliferation was also down-regulated in CTCs. In turn, no differences were noted (P>.05) for maturation-related markers such as CD19, CD20, CD27, CD45, CD79b, and CD138, as well as CXCR4. We then investigated the proliferation index of CTCs vs. BM mPCs (Figure 1B). Our results show a marked reduction of PCs in S-phase in the fraction of CTCs as compared to their counterpart in the BM (mean values of 0.07% vs. 1.5%; P=.005). In turn, colony formation assays showed that when co-cultured with human BM stromal cells, CTCs have a 3-fold enhanced clonogenic capacity as compared to paired BM mPCs. Finally, we investigated whether the number of CTCs followed a specific circadian rhythm similarly to that reported for HSCs. Strikingly, we founded that CTCs exhibited circadian fluctuations (Figure 1C), peaking at 4:00am (288 cells/μL) and reaching a low point at 16:00pm (183 cells/μL). This pattern was also found for normal HSCs (Figure 1D), with mean numbers peaking at 4:00am (5.0 cells/μL) and reaching a nadir at 16:00pm (2.4 cells/μL). Conversely, median SDF1 levels fluctuated in anti-phase with CTCs and HSCs, peaking at 16:00pm (1892 pg/mL) and dropping at 12:00am (1198 pg/mL). In summary, our results show that CTCs represent a sub-clone of BM mPCs characterized by a unique profile of integrin and adhesion molecules, but without sign of being at a different phenotypic maturation stage. CTCs are mostly quiescent, but show enhanced clonogenic potential as compared to BM mPCs when co-cultured with stromal cells. Similarly to HSCs, CTCs show a circadian rhythm triggered by rhythmic expression of SDF1 which may indicate that as the former, CTCs egress to PB during patients' resting period to colonize other sites in the BM. Figure 1. Figure 1. Disclosures: No relevant conflicts of interest to declare.

Abstract 3003 During the last years, several novel drugs combinations have changed the outcome of Multiple Myeloma (MM) patients; nevertheless, relapse is still the rule for most of these patients, and, therefore, other drugs with novel mechanisms of action are required. The current research in this topic is usually based on the addition of novel drugs to standard of care combinations in order to try to increase their efficacy. Bortezomib + steroids has turned out to be one of these standards, both for induction therapy before high dose melphalan and also for non-trasplant candidates. Zalypsis is a novel marine compound with remarkable preclinical activity in MM, which is currently being explored in a phase I/II trial. We explored the preclinical efficacy of the triple combination of zalypsis, bortezomib and dexamethasone in different models and investigated its mechanism of action. This combination showed high potency when compared to the respective single agents and double combinations, and, in fact, the analysis by the Calcusyn software, evidenced combination indexes in the highly synergistic range (CI 0.2–0.6). This effect was observed not only in the MM1S cell line, but also in freshly isolated cells from MM patients and, most importantly, in a subcutaneous xenograft of a human plasmocytoma in SCID mice. In this last model, we used suboptimal doses of bortezomib and dexamethasone (0.1 mg/Kg and 1 mg/Kg respectively ip, 5 days per week indefinitely) combined with an active dose of zalypsis (0,75 mg/Kg iv weekly for three total doses). Probably due to the low doses used, the double combination of bortezomib + dexamethasone did not show much effect as compared to the control group, and neither bortezomib nor dexamethasone alone could significantly improve the efficacy of Zalypsis. Nevertheless, when the three drugs were combined together, the tumor volume drastically shrunk with differences highly statistically significant (p These results impelled us to investigate the mechanisms relying under the synergy observed. The triple combination induced caspase dependent apoptosis, demonstrated by Annexin V induction and cleavage of caspases 3, 7, 8, 9 and also PARP. By contrast, the activity on the cell cycle was not evident for this combination, indicating a preferential apoptotic effect and not a substantial antiproliferative one. Of note, two mechanisms seem to be specifically associated to the combination of the three agents: 1. A significant release of Endonuclease-G and Cytochrome-C from the mitochondria into the cytosol was observed, suggesting a role for the intrinsic pathway of apoptosis; and 2. One of the most significant changes induced by the combination was an increase in the levels of AIF (apoptosis inducing factor) in the cytosol and nucleus, which is probably indicating the involvement of caspase independent mechanisms. In line with these results, pretreatment with ZVAD was only able to partially rescue from apoptosis, also indicating the potential activation of a dual mechanism: dependent and independent of caspases. Finally, we had previously shown that Zalypsis acts, at least partially, through the induction of DNA damage, based on a p53 dependent mechanism. In this study, the addition of bortezomib and dexamethasone, further increased the H2AX phosphorylation induced by Zalypsis, suggesting a potentiation of the DNA damaging effect; and the combination was also associated with a clear increase in the levels of p53 protein by WB as compared with the agents in monotherapy. The study of the mechanism of action is currently being completed and final results will be presented at the meeting. In summary, the combination of zalypsis + bortezomib + dexamethasone shows high synergy in MM through the activation of different pathways, and these results provide the rationale for its use in relapsed/refractory MM patients. Disclosures: Cuevas: Pharmamar: Employment. Avilés:Pharmamar: Employment. Pandiella:Pharmamar: Research Funding. San Miguel:Pharmamar: Research Funding; Millenium: Consultancy, Research Funding.

Abstract 3792 Poster Board III-728 Background p53 mutation is the most frequent single genetic abnormality found in therapy related AML and is also quite frequent in de novo AML with an incidence between 10% and 25%; Moreover patients with p53 mutations characteristically present complex karyotypes and complicated chromosome rearrangements, leading to an adverse prognosis and resistance to conventional chemotherapeutic agents. Consequently, there is a need for novel antileukemic drugs that could work on both p53-wild type and p53-mutated AML patients. In this study we have evaluated the activity of Zalypsis, a novel alkaloid from marine origin, in several AML cell lines and patients samples with different p53 status. Methods The efficacy of Zalypsis was analyzed in four AML cell lines (HL60, HEL, MV4.11 and KG1) and in cells from fresh bone marrow samples from ten newly diagnosed AML patients. The cytotoxicity was analyzed by means of MTT assay and with a multiparametric flow cytometry (MFC) technique that allowed us to study the efficacy of the drug in both mature and immature blast cell compartments. Proteomic and genomic changes induced after treatment with Zalypsis were analyzed in two AML cell lines with different p53 status (HEL and HL60) by Western-Blot and gene expression profile studies. Results Zalypsis showed a very potent antileukemic activity in the four AML cell lines tested, with IC50s at 48 hours below 1 nM. When compared with the in vitro activity of conventional antileukemic agents, such as cytarabine, doxorubicine or fludarabine, Zalypsis turned out to be 10-100 times more potent. It also showed remarkable ex vivo potency in freshly isolated blasts from ten AML patients. In these patients samples, we had the opportunity to separately analyze by multiparametric flow cytometry, the activity of Zalypsis in the different blasts populations, and we constantly observed similar activity in the most mature blast population and in the most immature one (CD34+, CD38- Lineage-), which is thought to include the leukemic stem cells and which is usually more resistant to conventional chemotherapy. Regarding toxicity, Zalypsis preserved the CD34+ non tumoral hematopoietic progenitor cells. The combination of this novel drug with conventional antileukemic agents indicated that Zalypsis is a good partner for combination with all of them, being the combination with cytarabine and daunorubicin particularly attractive for the clinic. Interestingly, as already said, all cell lines were extremely sensitive to Zalypsis, independently of the p53 status, as some of them displayed high basal levels of this protein by Western-Blot and had mutated p53 (HEL and MV 4.11), whereas its protein expression in the remaining cell lines was low, with KG1 bearing a mutation of the gene and HL60 a p53 deletion. Nevertheless, these last cells with low levels of the protein were a bit more sensitive to the drug, pointing out to a role of p53 in Zalypsis induced cell death. This was in agreement with a clear induction of DNA double strand breaks after treatment with Zalypsis, which was evidenced by an increase in phospho Histone H2AX, phospho CHK1 and phospho CHK2 after treatment of both HEL and HL60 with this compound. This was followed by the overexpression of p53 in the AML cell line bearing low basal levels of this protein (HL60). These results were further confirmed in the gene expression profile studies, in which treatment of AML cells with Zalypsis resulted in an important deregulation of genes involved in DNA damage response, including genes implicated in the ATM repair pathway and other mRNAs related to DNA repair, such as TLK2, ATR, ATMIN, CHEK2, RAD5, RAD52, RAD54L, BRCA1, BRCA2 and GADD45B among others. This response ended up in a clear induction of apoptosis as assessed by annexin V studies, caspase and PARP cleavage with partial rescue of cell death with the caspase inhibitor ZVAD-FMK, DNA laddering and loss of mitochondrial membrane potential analyzed by DiOC6 with the subsequent release of cytochrome C and AIF into the cytosol. Conclusion The potent and selective antileukemic effect observed with Zalypsis in AML cell lines and in patient's samples through a DNA damage response, together with its activity in both p53 mutated- and deleted-cells, provides the rationale for the investigation of Zalypsis in clinical trials for patients with AML. Disclosures: Galmarini: Pharmamar: Employment. Avilés:Pharmamar: Employment. Cuevas:Pharmamar: Employment. Pandiella:Pharmamar: Research Funding. San-Miguel:Pharmamar: Research Funding.

Abstract 3834 Poster Board III-770 Introduction Multiple myeloma (MM), an hematological malignancy of terminally differentiated plasma cells, is characterized by the presence of bone disease, caused by increased osteoclast (OC) activity and differentiation as well as a reduction in osteoblast (OB) number and function. Dasatinib (BMS-354825) is an oral multitargeted tyrosin-kinase inhibitor approved for chronic myeloid leukemia which is also under clinical investigation in several other types of tumors. Preclinical data suggests that dasatinib can also be of value in MM based on its effects on myelomatous plasma cells and angiogenesis. In this study, we have further investigated the effects of dasatinib on in vitro OB differentiation and function, as well as on OC formation and resorption activity. Materials and methods For studies on OB differentiation and function, human mensenchymal stem cells (hMSC) from bone marrow samples of healthy donors and MM patients were used. Alternatively, the mesenchymal hMSC-TERT, the osteoblast-like (MG-63) and multiple myeloma (MM.1S) cell lines were employed. Dasatinib mechanism of action was investigated by Western blotting, PKH67/Annexin V/7 aminoactinomycin D staining, real time RT-PCR, alkaline phosphatase (ALP) activity and quantitative mineralization assays. Receptor activator of nuclear factor κ B ligand (RANKL) and osteoprotegerin (OPG) levels in conditioned media were measured by ELISA. OCs were generated by culture of peripheral blood mononuclear cells from healthy volunteers in medium containing macrophage colony stimulating factor and RANKL. The effect of dasatinib on osteoclastogenesis was assessed by enumeration of multinucleated (≥3) tartrate resistant acid phosphatase-positive cells, whereas bone resorption was calculated by the resorbed area on calcium-coated slides. Results We found that dasatinib inhibited platelet derived growth factor (PDGF)-stimulated activation of PDGF receptor β (PDGFRβ) and c-Src in hMSC-TERT and MG-63 cell lines, both tyrosin kinases implicated in OB remodelation processes. Inhibition of PDGFRβ and c-Src signalling correlated with diminished proliferation of the same cell lines without affecting cell viavility as assessed by PKH67/Annexin V/7 aminoactinomycin D staining. Also, treatment of human osteoprogenitor cells with low dasatinib concentrations (2 - 5 nM) promoted OB differentiation since ALP activity at day 7 and gene expression levels of bone formation markers (Runx2/Cbfa1, ALP, and COLIA1) at day 7 and 14 in the osteoblastic differentiation process, were elevated; besides, dasatinib also increased mineral nodular formation as per quantification of alizarin red staining. Finally, treatment with dasatinib decreased the RANKL/OPG ratio in conditioned media from co-cultures of MG-63 and MM.1S cell lines. Similar range of dasatinib concentrations markedly inhibited OC formation, both at the initial and late stages of differentiation from hemopoietic progenitor cells. Finally, a significant reduction of OC resorptive activity of a calcium-coated substrate was observed. Conclusion Our results indicate that dasatinib favours both OB differentiation and activity and markedly impairs osteoclastogenesis and OC resorption, thus providing rationale for its use to improve bone lesions in MM patients and other bone pathologies. This work was supported by grants from Ministerio de Ciencia e Innovación - ISCIII (PI081825); Fundación de Investigación Médica Mutua Madrileña AP27262008, and Centro en Red de Medicina Regenerativa y Terapia Celular de Castilla y León 07-09, Consejería Sanidad JCyL-ISCIII. Disclosures: Garzon: Bristol-Myers Squibb Company: Employment.

Introduction: Filanesib (ARRY-520) is a highly selective inhibitor of kinesin spindle protein (KSP), a microtubule motor protein active in proliferating cells, which plays an essential role in assembly and maintaining of the bipolar spindle. In cells arrested by KSP inhibition, Mcl-1 is rapidly depleted resulting in cell death, and consequently cells that are dependent on this pro-survival protein, such as myeloma cells, are particularly sensitive to filanesib. We investigated the mechanisms underlying the antimyeloma effect of this agent, focusing on other Bcl-2 family members. Methods: In vitro action of filanesib, alone and in combination with calpain inhibitor PD150606, was evaluated in multiple myeloma (MM) cell lines by MTT assay, Annexin V staining and cell cycle profile analysis by flow cytometry. MM cells were transiently transfected with non- targeting control short interfering RNA (NT-siRNA), Bax siRNA ON TARGET plus SMART pool siRNA using the cell line Nucleofector Kit V. Expression levels of different proteins were analyzed by Western-Blot. Results: We previously showed that all 11 MM cell lines tested were sensitive to filanesib and that sensitivity to this agent correlated with Bax levels. For these experiments, we focused on 3 cell lines with different Bax expression and sensitivity to filanesib: OPM-2 and MM1S (sensitive and high Bax levels) and U266 (less sensitive and low Bax levels). Treatment of MM1S with this agent triggered the translocation to the mitochondria of several proapoptotic Bcl-2 family members such as Noxa, Bim and Bax with several downstream effects. The mitochondrial translocation and activation of Noxa is key in the degradation of Mcl-1 by mediating the translocation of this protein from the cytosol to the mitochondria promoting its degradation. Regarding Bim, filanesib also induced the early (12-24 hours) expression of several proapoptotic isoforms of Bim that also translocated to the mytochondria. As previously reported, Bax is the top determinant of sensitivity to filanesib. In the present study, remarkably, once translocated into the mitochondria, Bax was also cleaved into the very potent proapoptotic 18 kDa fragment. All these events triggered the mitochondrial release of cytochrome C (caspase dependent apoptosis) and AIF (caspase independent apoptosis). In order to confirm the role of Bax in filanesib-induced apoptosis, we knocked-down Bax in MM1S by using small interfering RNA. This approach clearly decreased the sensitivity of these cells to filanesib, as treatment with 10 nM for 24 hours induced only 26% apoptosis in the siRNA-Bax cells as compared with 50% in the non-targeted cells (as compared with 58% vs 61% for bortezomib). Furthermore, treatment with filanesib also induced cleavage of effector caspases (3 and 7) in all cell lines studied (OPM2, MM1S, U266). PARP was also cleaved in these cells, but it was previous to caspase activation in the most sensitive cell lines (OPM2, MM1S) suggesting a caspase-independent mechanism of apoptosis. This was confirmed by pre-treatment with the pan caspase-inhibitor Z-VAD-FMK, which did not rescue OPM2 and MM1S cells from apoptosis. Interestingly, one potential mechanism that could link both effects is the activity of calpain, a cysteine protease involved in caspase-independent apoptosis. This protein is a well-known caspase-independent way of processing PARP into the 60 kDa fragment, and has also been described as being responsible for Bax cleavage into the 18-kDa fragment. Consistent with this hypothesis, pre-treatment with the calpain inhibitor PD150606 clearly reduced the activity of filanesib in these cells (35 % to 70 % of survival) as assessed by MTT. Finally, consistent with the previous hypothesis, the less sensitive U266 cell line contained undetectable Bax protein suggesting that filanesib was not able to trigger caspase-independent apoptosis. However, a secondary caspase dependent apoptosis mechanism was confirmed as the pan-caspase inhibitor ZVAD-FMK was able to almost completely abrogate the activity of filanesib. Conclusions: Our results show that filanesib primarily initiates apoptosis by activating Bax in a caspase-independent manner, probably via calpain, a powerful accelerator of the apoptotic process. In addition, Noxa and BIM appear to be crucial for modulating Mcl-1 proteasomal degradation and Bax activation. This work was funded in part by the company Array BioPharma. Disclosures Tunquist: Array BioPharma: Employment. Mateos:Takeda: Consultancy; Onyx: Consultancy; Janssen-Cilag: Consultancy, Honoraria; Celgene: Consultancy, Honoraria. Ocio:Mundipharma: Consultancy, Research Funding; Bristol Myers Squibb: Consultancy; Novartis: Consultancy, Research Funding; MSD: Research Funding; Amgen/Onyx: Consultancy, Honoraria, Research Funding; Array BioPharma: Consultancy, Research Funding; Celgene: Consultancy, Honoraria; Pharmamar: Consultancy, Research Funding; Janssen: Honoraria.

Abstract 4077 Several murine models resembling multiple myeloma have been developed in the last years: the subcutaneous plasmocytoma model and the disseminated model, based on the iv or intracardiac injection of MM cells, have been widely used to test the efficacy of novel agents. Nevertheless, it is well known the crucial role that the interaction between MM plasma cells and the bone marrow microenvironment plays in MM pathogenesis as well as in the proliferation of MM cells and drug resistance; however, unfortunately, most murine models do not recapitulate these interactions. In order to overcome this caveat, other models such as the SCIDhu or the SCIDrab have been developed. However, in these models, the induced microenvironmental milleu is derived from a fetus and a rabbit respectively, which can be relatively different from that found in an adult patient with MM. We have developed a novel model of a human disseminated MM with the presence of human adult microenvironment. For this purpose, CB17-SCID mice, which lack B and T lymphocytes, were anesthetized and a human bone chip obtained from a healthy adult patient undergoing a hip replacement due to a non-malignant fracture, was subcutaneously implanted in the back of the mice. In the same day 3×106 MM1S-Luc cells were iv injected through the tail vein. Two groups were also included as controls in the study, one with MM1S-Luc cells without the bone chip, and the other with only the human bone chip (without MM cells). Luciferase activity was measured three times per week using a Xenogen-IVIS 50 system after ip injection of 0,15 mg/Kg luciferin. Mice were sacrificed when they developed signs associated with advanced disease, such as body weight loss or back limb paralysis. Although the group of mice receiving MM1S-Luc without bone, did develop a disseminated myeloma, this was much more precocious in the group of mice with a human bone subcutaneous implant. In fact, in mice with human bone and MM1S-Luc cells, the luciferase activity was detectable a median of 9 days after the injection of the cells whereas it was necessary 16 days for obtaining the same effect in mice without the bone chip. This is pointing out that the bone chip in these mice induces a favourable environment for myeloma growth and could be appropriate model to study the interactions of MM cells with human stroma. The increase of tumour growth rate also correlated with a decrease in survival. Median survival (CI 95%) of 64 (47-80) days for mice without bone vs 48 (45-50) days for mice with human bone (Log Rank=0.004). Regarding the sites of infiltration, luminescence images showed a disseminated involvement in all skeletal regions, and this was confirmed by histopathology, as there was a general infiltration of spine, femora, tibia and cranium of these mice. By contrast we did not find infiltration of extra-skeletal regions such as heart, lung or spleen in any mice and only unclear infiltrates were observed in the kidney and liver of some mice. In order to validate the suitability of this model for testing the in vivo antimyeloma activity of novel drugs in the presence of a human microenvironment, 6 mice bearing this novel model were randomized to receive: bortezomib (0.5 mg/Kg ip five days per week) + dexamethasone (0.5 mg/Kg 2 days per week) or vehicle control. A second group of the same number of mice with MM1S-Luc but without the bone chip were randomized in the same way and used as control. Bortezomib + dex significantly decreased tumour growth in both groups of mice without any clear differences in activity between both models, suggesting that this combination overcomes the influence of the microenvironment. In summary we have developed a murine model of MM that may allows us to 1) gain insights into the role of the BM microenvironment interactions in MM pathogenesis. 2) Evaluate the in vivo activity of different anti-MM agents in the presence of human microenvironment. 3) Potentially, this model could also contribute to analyze the effect of different agents in the myeloma bone disease. Disclosures: No relevant conflicts of interest to declare.