ALBERT

Description: Tumor cells that survive initial courses of chemotherapy may do so by acquiring a multidrug-resistant phenotype. This particular mechanism of drug resistance may also confer resistance to physiological effectors of apoptosis that could potentially reduce the efficacy of immune therapies that use these pathways of cell death. We have previously demonstrated high efficacy for a cytokine-based tumor cell vaccine in a murine MPC11 myeloma model. In the present study, the effects of this vaccination were compared in MPC11 cells and their isogenic sublines selected for mdr1/P-glycoprotein (Pgp)-mediated multidrug resistance (MDR). Immunization with MPC11 cells expressing granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-12 (IL-12) led to long-lasting protection of mice against subcutaneous (sc) challenge with both parental cells or their MDR variants. Similarly, immunization with GM-CSF/IL-12–transfected MDR sublines caused rejection of transplantation of both parental cells and the MDR sublines. Whereas MPC11 cells and their MDR variants were resistant to APO-1/CD95/Fas ligand, the immunization generated potent granzyme B/perforin-secreting cytotoxic T lymphocytes (CTLs) that were similarly effective against both parental and isogenic MDR cells. We conclude that MDR mediated bymdr1/Pgp did not interfere with lysis by pore-forming CTLs. Immunotherapy based on pore-forming CTLs may be an attractive approach to the treatment of drug-resistant myeloma.

Description: Abstract 2806 Poster Board II-782 Hematopoietic stroma provides a sanctuary for multiple myeloma (MM) cells and promotes resistance to immune control mediated by APO2L/TRAIL in part by increasing synthesis of the anti-apoptotic protein c-FLIP (J Immunol, 180: 3; 2008). Bortezomib, a reversible 26S proteosome inhibitor, sensitizes cancer cells to APO2L/TRAIL induced apoptosis in a variety of malignancies including MM. In addition, it has been well established that bortezomib interferes with the protective interaction between cancer cells and the bone marrow stroma. In this study we tested whether bortezomib can reverse the APO2L/TRAIL environmental mediated-immune resistant (EM-IR). Using MM cell lines (RPMI 8226 and U266) or CD138+ positive selected cells from MM patient's bone marrow, we found that exposure to HS5 stroma cells by direct adhesion or in a transwell system induced resistance to bortezomib (10nM for 24 hours) and that pretreatment with bortezomib (10nM for 20 hours) effectively overcomes APO2L/TRAIL resistance (10ng/mL for 4 additional hours). Conditioned medium made from a 14-day culture of MM patient's bone marrow stroma induced resistance to bortezomib and pretreatment with bortezomib sensitized cells to APO2L/TRAIL induced cell death as shown with HS5 cells. Bortezomib showed not to be cytotoxic to HS5 stroma cells and only MCP-2-3 and IL-10 levels were altered in the stroma-MM milieu within cytokines measured by ELISA array. IL-6, a cytokine shown to induce APO2L/TRAIL mediated resistance, remained unchanged with bortezomib treatment. We found that bortezomib increased expression of TRAIL receptors (DR5, DCR1 and DCR2), but such expression did not predict for sensitivity to apoptosis as DR4 demonstrated to be the receptor responsible for activation of APO2L/TRAIL. Soluble factor(s) released by HS5 stroma increased expression of c-FLIP and induced STAT-3 and ERK phosphorylation in myeloma cells. However, only c-FLIP protein expression was effectively reduced by bortezomib. c-FLIP quantitative-PCR found that HS5 increased transcription only after 20 hours bortezomib treatment and significantly reduces soluble factor(s) induced c-FLIP transcription. In a transwell assay, HS5 stroma cells induced NF-κB activation and the addition of bortezomib diminished its activation at 20 hours. These findings provide the rationale to combine bortezomib and APO2L/TRAIL to disrupt the influence of the stroma microenvironment on myeloma cells. Disclosures: Off Label Use: sirolimus for graft-versus-host disease.

Description: Proteasome inhibitors (PI) such as bortezomib and carfilzomib are critical components of anti-multiple myeloma (MM) therapy, yet all MM patients eventually develop refractory disease. We developed a non-biased method to identify and validate dysregulated pathways associated with PI-resistance in myeloma by combining RNAseq data from 522 MM patient specimens obtained from our Total Cancer Care/M2Gen/ORIEN network at Moffitt Cancer Center with paired ex vivo sensitivity to PIs and kinase inhibitors (KI). Dimensionality reduction analysis (t-SNE) and Fuzzy C-means was used to identify 422 clusters of genes that co-express in individual patients, and Gene Set Enrichment Analysis (GSEA) was used to identify clusters with gene expression patterns that correlated with PI sensitivity. Using publicly curated databases and in silico integrative analyses, we built protein-protein interaction networks to identify putative transcription factors, corresponding master regulators (kinases), and candidate KIs to promote PI sensitization. This systems biology approach identified a Chk1-Cdk1-Plk1 circuit associated with PI-resistance and also found 21 additional kinases (of 501 expressed in our cohort's kinome) that could be targeted to re-sensitize PI-resistant MM, which we confirmed in cell lines, specimens from relapsed patients, and two in vivo models. A panel of paired isogenic PI-resistant and sensitive MM cell lines were differentially screened to find kinases associated with PI-resistance using activity-based protein profiling (ABPP) and KI activity measured by high-throughput viability assay. The MM cell lines 8226 and U266, along with their drug resistant counterparts 8226-B25 and U266-PR, were grown in mono-culture for 24h and lysates were enriched for ATP binding proteins by affinity purification versus a chemical probe. Tryptic peptides were measured using discovery proteomics (nano-UPLC and QExactive Plus mass spectrometer) to identify 85 kinases out of a total of 715 proteins in 8226-B25 MM cells and 35 kinases out of a total of 688 proteins in U266-PR MM cells that were preferentially enriched by 2-fold change compared to parental cell lines. Twenty-four kinases were commonly activated among PI-resistant cell line pairs and were screened in PI-resistant myeloma lines using a label-free, high throughput viability assay that simulates the tumor microenvironment. Three KIs targeting Plk1 (volasertib and GSK461364) and Cdk1/5 (dinaciclib) consistently maintained LD50s in the low-nanomolar range and induced caspase-3 activation in four PI-resistant MM cell lines: 8226-B25, U266-PR, ANBL-6-V10R, and Kas6-V10R. Twenty-four kinases each were identified by RNAseq/ex vivo PI sensitivity of MM specimens and ABPP of PI-resistant/sensitive MM cell line pairs. Of these, 7 kinases were identified by both methods: Cdk1, Chk1, Plk1, ILK, Syk, PKA, and p70S6K. Several KIs targeting Cdk1, Plk1, ILK, DNAPK, Syk, MKK7, Nek2, and mTOR identified in patient specimen or cell-line screens showed single agent activity in MM patient bone marrow specimens purified by a CD138 affinity column. Among these, inhibitors to Cdk1, ILK, mTOR, and Plk1 showed the most activity in patient specimens with an average 96h LD50 of 25 nM (n=56), 2.4 uM (n=42), 2.7 uM (n=57) and 3.8 uM (n=53), respectively. Six KIs targeting Plk1, ILK, Syk, MKK7, Nek2 and MARK3 were synergistic with carfilzomib in 20 patient specimens and maintained or improved ex vivo activity in relapsed refractory MM (RRMM) specimens. Volasertib, which targets Plk1, was the most synergistic with carfilzomib of all KIs tested in patient specimens and was further validated in two in vivo models: a NSG/U266 xenograft model of PI resistance and the syngeneic C57BL/6-KaLwRij/5TGM1 immunocompetent model. Volasertib significantly increased survival and reduced tumor burden in both models as a single agent, and was more effective versus PI-resistant tumors compared to PI-sensitive counterparts. Our pharmaco-proteomic screen, coupled with rich gene expression data from patients identified Plk1 as a target critical to MM survival in the context of acquired PI resistance and represents a unique workflow to find tumor vulnerabilities that arise during therapy. We anticipate that these data will also produce a critical path for the personalized allocation of therapy to maximize efficacy and minimize the use of ineffective therapies in RRMM. Disclosures No relevant conflicts of interest to declare.