Akhtar, Sharoon

The PRSS23 gene, encoding putative serine protease 23, has been associated with several diseases by correlative gene expression studies and genetic‐based comparative models. Our group has recently developed data implicating the expression of this gene as a potential driver of ovarian cancer malignancy. However, there have been no studies defining PRSS23 protein structure or demonstrating its hypothesized proteolytic function. We aim to elucidate PRSS23 protein structure and enzymatic activity. The first step to achieve this goal is to develop a robust expression system and purification protocol for recombinant PRSS23 protein. We hypothesize that developing a methodology to produce PRSS23 recombinant protein will enable us to determine its crystal structure, evaluate its biological and enzyme function, and ultimately target the enzyme therapeutically. We strategized to maximize chances of producing PRSS23 in its native form and with high yield by testing multiple systems and host species: human, yeast, and bacterial. We anticipate, based on the protein sequence, that PRSS23 is endogenously secreted and may be activated in the secretory pathway by furin, releasing the C‐terminal catalytic domain from a smaller N‐terminal domain. The protein is also predicted to be glycosylated. We generated expression plasmids for full‐length PRSS23, codon‐optimized for expression in human and Pichia pastoris expression systems, incorporating a His‐tag for detection and purification. In the human HEK 293F and HEK 293T cell systems, despite optimizing transfection protocols and testing multiple expression constructs, secretion signals, and promoters, the expression level remained very low and was judged unlikely to produce sufficient protein for structural and functional studies. Integration of codon‐optimized PRSS23 into the genome of P. pastoris produced integrants showing a spectrum of expression levels from very low to moderate. Experiments are ongoing to optimize expression conditions of the best clone and develop a purification workflow involving ammonium sulfate precipitation, immobilized nickel ion chromatography, and additional chromatographic steps as needed. In parallel, we cloned the catalytic domain of PRSS23, appropriately codon‐optimized and also containing a C‐terminal His‐tag, into an expression vector for Escherichia coli. High levels of protein were produced as insoluble inclusion bodies, and a protocol was developed for nickel ion affinity purification under denaturing conditions. Experiments are ongoing to identify suitable conditions for in vitro refolding of the PRSS23 catalytic domain. Additionally, we are evaluating a maltose binding protein fusion with the PRSS23 catalytic domain as an alternative strategy to improve soluble protein production in E. coli. Significant quantities of pure protein will be needed to enable crystal structure determination and enzymatic activity elucidation, to unravel the biological mechanism of action and identify endogenous substrates, and potentially lead to new therapeutic strategies targeting PRSS23. The expression systems and purification workflows developed here will be a step forward toward these long‐term goals.

Abstract Background: Waldenstršms macroglobulinemia (WM) is an indolent B-lymphoplasmacytic lymphoma, which is highly sensitive to 20S proteasome inhibitors such as bortezomib. However, clinical resistance to the drug can be observed early at onset of therapy (primary or inherent resistance) or develop over time, through repeated exposure to the drug (acquired resistance). Although prior investigations into the biological mechanisms associated with bortezomib resistance (BR) have identified aberrations at the genomic (PSMB5 mutation) and transcriptional level (loss of XPB-1), there is limited data on the epigenetic abnormalities that support resistance to 20S proteasome inhibition. With this in mind, we conducted a comprehensive methylation and mRNA expression analysis in WM cells focusing on the molecular patterns that characterize inherent BR (iBR) as well as acquired BR (aBR). Materials and Methods: Bortezomib sensitive (BS) human WM cell lines, BCWM.1 and MWCL-1, the iBR WM cell line RPCI-WM1 (developed from a bortezomib-refractory WM patient) as well as aBR subclones (BCWM.1/BR, MWCL-1/BR) were used in the study. Methylation status of WM cells was profiled by Illumina Infinium 450k methylation array. Methylation IDAT files were background subtracted using GenomeStudio v1.9. Beta mixture quantile dilation (BMIQ) method in R watermelon package was used to adjust type II probes into distribution characteristic of type I probes. For each gene, probes were partitioned into 6 regions using R IMA package and the ones belonging to the same region were averaged, including transcription start sites (TSS; both 200 bp and 1500 bp upstream), 5' untranslated region (UTR), and 3'UTR, the first exon, and genebody. Gene regions showing a significant adjusted p-value and a beta difference of >0.1 or <-0.1 were selected for differential hypermethylation and hypomethylation, respectively. Correlation with gene expression was conducted only on genes showing >1.5 fold change in expression. Overview of methodology is shown in Figure 1. Results: We first assessed the differential methylation pattern (TSS, 200bp upstream of gene body) of bortezomib-sensitive (BS, BCWM.1, IC50 6nM) and iBR cells, (RPCI-WM1). A total of 3,801 hypermethylated and 1,865 hypomethylated DNA regions were observed in iBR cells relative to BS cells (Figure 2A). In iBR cells, a negative correlation between hypermethylated DNA and corresponding gene expression was found in 1,039 genes (Spearman r = -0.12) (Figure 2B). In parallel, we compared the methylation pattern (TSS 200bp) of BS WM cells (BCWM.1 and MWCL-1) with their aBR derivatives, which display ~20 – 50 fold resistance to bortezomib. aBR subclones showed differential hypermethylation of 1,521 DNA regions and hypomethylation of 651 genomic regions relative to BS WM cells. These findings were not unexpected as aBR cell lines, being derivatives of BS WM cells, are clonally related to the former, whereas iBR cells are derived from a different clone. Meta-analysis of iBR and aBR followed by pathway enrichment (NextBio, Illumina Inc), revealed a significant clustering of commonly hypermethylated genes within the Validated Transcriptional Targets of TP63 isoforms Canonical Pathway (p<1.9-E5, Figure 2C) and the CEBPA binding site Geneset 2 (p<8.2-E9, Figure 2D); pathways whose genes are actively involved in suppression of cell growth and differentiation. Further validation and functional relevance of candidate gene targets from this analysis, which are relevant to WM pathobiology, are currently being investigated. Conclusion: This is the first study to examine and correlate the global methylation and gene expression profiles in WM tumor cells that are inherently wired to resist 20S proteasome inhibition as well as tumor cells, which have acquired resistance through continuous exposure to bortezomib. Although considerable differences were observed in the methylation profiles of WM cells displaying inherent vs. acquired resistance, a number of common genes emerged and pointed towards transcriptional repressors, whose inactivation through methylation has been linked to other cancers. Altogether, these findings uncover cellular systems in WM cells, which are affected by DNA methylation, subsequent mRNA suppression and which may be associated as a whole with both inherent and acquired forms of bortezomib resistance. Disclosures No relevant conflicts of interest to declare.

Abstract Background: Deubiquitinating enzyme (DUB) inhibitors are an emerging class of compounds, which are increasingly being regarded for their anti-cancer activity and therapeutic potential. The 19S regulatory particle of the proteasome contains two DUBs, USP14 and UCHL5, which are critical for the proper unfolding and deubiquitination of proteins prior to their entry into the 20S proteasome β-catalytic core. In B/plasma cell malignancies such as Waldenströms macroglobulinemia (WM) and multiple myeloma (MM), protein homeostasis and optimal proteasome functionality are paramount for tumor survival. This is evidenced by the success of 20S proteasome inhibition with bortezomib and carfilzomib, which have demonstrated remarkable clinical benefit in patients with WM or MM. As such, the proteasome degradation pathway represents a highly attractive and clinically validated, yet still largely unexplored therapeutic target. Here, we provide first evidence of the cytotoxic effects and molecular sequelae associated with a novel 19S proteasome DUB inhibitor, VLX1570, which selectively targets USP14 and UCHL5, in preclinical models of B/plasma cell cancers and their drug resistant tumor subclones. Materials: VLX1570 was obtained from Vivolux AB, Sweden. Bortezomib was purchased from Sellekchem, Houston TX. Human myeloma cell lines (OPM2, U266, KMS11), human WM cell lines (BCWM.1, MWCL-1 and RPCI-WM1), their respective bortezomib-resistant (BR) subclones as well as ibrutinib-resistant (IR) derivatives of BCWM.1 and RPCI-WM1 were used in experiments (total n=12). Peripheral blood mononuclear cells (PBMCs) from healthy donors were used to assess for cytotoxicity of VLX1570 in non-tumor cells. Results: A 72hr MTS assay first established sensitivity of plasma cell cancer models towards VLX1570, including drug resistant subclones (EC50 range, 20 – 90nM). Next we assessed whether loss of viability was due to apoptosis and observed dose-dependent annexin-V staining in ~50 – 70% of wild-type MM and WM cells, ~45 – 55% in corresponding BR models and ~50% in IR subclones when exposed to VLX1570 (100nm – 500nM) for 24hrs. VLX1570 was tested at similar concentrations in PBMCs and induced minimal cell death. To understand if VLX1570-induced apoptosis was mitochondrial mediated, we assessed mitochondrial outer membrane permeability (MOMP) in all plasma cell cancer models treated with the DUB inhibitor (100 – 500nM) and found a dose-dependent increase in MOMP (p<0.00025 – 0.000001). Activation of the mitochondrial/intrinsic apoptotic pathway was also evidenced by cleavage of caspase-9 and 3 on immunoblot analysis. As inhibition of protein deubiquitination results in accumulation of polyubiquitinated protein conjugates, we examined global ubiquitinated protein levels and found increasing amounts of high molecular weight poly-ubiquitinated protein in VLX1570 treated tumor cells in a dose-dependent manner. This prompted us to further delineate the potential cellular response mechanisms that relay the antitumor activity of VLX1570 in malignant plasma cells. Immunoblot analysis of drug sensitive and (bortezomib or ibrutinib) resistant MM and WM cell lines revealed several changes in ER stress response proteins, notably induction of HSP70 and splicing of XBP1; the latter event which was less prominent in BR cell lines treated with bortezomib. This suggested the ability of VLX1570 to overcome resistance through a rapid and potent proteotoxic effect, which overwhelmed the tumor cells stress response mechanisms beyond its capacity for survival. Finally, as abnormalities in the NFkB signaling pathway are implicated in malignant plasma cell survival, we found markedly decreased nuclear translocation of NFkB (p-P65) protein in tumor cells treated with VLX1570. Conclusions: The DUBs of the 19S proteasome cap offer an attractive target within the proteasome degradation system and whose inhibition by VLX1570 is able to circumvent mechanisms that support resistance to b-catalytic PI (bortezomib) or BTK inhibitor (ibrutinib) therapy. The studies undertaken thus far, coupled with more rigorous mechanistic and in vivo efficacy analyses currently underway, provides rational for the clinical development of DUB inhibitors for use in malignant B/plasma cell malignancies. Disclosures Martin: Janssen: Honoraria.

Abstract INTRODUCTION: Treatment of chronic lymphocytic leukemia (CLL) has expanded significantly with the approval of multiple small molecule inhibitors. This is of great significance for patients with adverse cytogenetic features who tend to respond poorly to standard chemo-immunotherapy (CIT). While single agent ibrutinib and venetoclax (V) have shown high rates of overall response, complete remission and minimal residual disease (MRD) eradication rates remain low. This argues for further testing and development of various combination strategies. The MURANO trial compared venetoclax and rituximab (VR) versus bendamustine and rituximab (BR) in patients with relapsed/refractory CLL reporting clear superiority of VR over BR. MRD rates in the bone marrow were reported to be 27.3% for VR versus 1.5% for BR. Given much higher rates of MRD eradication with combination of small molecule inhibitors and monoclonal antibodies (mAb) compared to standard CIT, we performed a comparative investigation into the direct and immune-mediated cytolytic effects of VR versus V + Obinutuzumab (O, type II anti-CD20 mAb) in primary CLL cells and B-lymphoid cell lines. METHODS: CD19+ B-cells were isolated from PBMCs of CLL patients (N=3). For all experiments using primary CLL cells, concentration of VR and VO was 3nM (V) and 10ug /ml (R, O), respectively. For cell lines, VR and VO was used at 5uM (V) and 10ug /ml (R, O), respectively. Apoptosis was determined by annexin-V/PI staining followed by flow cytometry. Antibody-dependent cell-mediated cytotoxicity (ADCC) induced by VR and VO was assessed in Calcein AM labeled CLL cells or cell lines co-cultured with healthy donor PBMCs (E:T ratio, 40:1); complement-dependent cytotoxicity (CDC) was measured using 10% serum from a healthy human donor. RESULTS: We assessed the ability of V+/-O or V+/-R to induce apoptotic cell death in the CD20+ BCWM.1 cell line (Waldenström's macroglobulinemia [WM] phenotype) and the MEC-1 cell line (B-PLL phenotype); with CD20- RPCI-WM1 (WM cells, negative control). Notably, Bcl-2 protein is expressed in all the aforementioned cell lines. We observed that single agent V, O and R induced ~30%, 61% and 13.64% annexin V/PI positivity in BCWM.1 cells, respectively. However, a significant degree of cell death was noted in VO-treated cells (~74%) compared to VR-treated cells (~40%) (p<0.01). Next, we examined for apoptosis in MEC1 cells and noted a similar trend; where the VO combination induced markedly more cell death (~71%) than VR (~57%). Contrastingly, in RPCI-WM1 cells neither single agent O or R could elicit >12% annexin V/PI positivity and where the addition of V increased apoptosis by only 3 - 4%. We also examined the apoptotic potential of VO or VR in tumor cells from low, intermediate and high-risk CLL patients. In low and intermediate-risk CLL cells from low and intermediate-risk patients, V alone induced ~30% cell death, which increased significantly with the addition of O (VO) to between 48 - 52%. Contrastingly, the combination of VR did not induce more than 29 - 32% apoptosis. In CLL cells from high-risk patient, we noted that exposure to single agent V induced ~ 28% cell death and in VO-treated cells, this number increased to 47%. We also examined for ADCC and CDC in the same cell lines and primary CLL cells. Despite considerable variability, single agent O and VO treatment of tumor cells resulted in greater ADCC than VR treatment. By contrast, in single agent R or VR-treated cells, more CDC was observed. CONCLUSION: Our preliminary investigation in VR- and VO-treated cell lines and primary CLL cells suggests the VO combination may be superior to VR in induction of direct tumor cell death. Mechanistic experiments underway will provide further insight and can aid in design of future VO-based clinical studies in CLL. Disclosures Ailawadhi: Pharmacyclics: Research Funding; Takeda: Consultancy; Celgene: Consultancy; Janssen: Consultancy; Amgen: Consultancy.

Background The ubiquitin proteasome system (UPS) maintains homeostasis in eukaryotic cells by degrading and recycling proteins associated with DNA transcription and cell cycle regulation. In tumor cells with a high protein turnover rate, particularly in immunoglobulin secreting cancers such as Waldenströms Macroglobulinemia (WM), proteasomes are critical for sustained cell proliferation. Proteasome inhibitors (PI) such as bortezomib and carfilzomib target the B5 subunit of the proteasome catalytic core and disrupt its activity leading to accumulation of polyubiquitinated proteins. Although highly effective in WM, persistent use of bortezomib eventually leads to emergence of a bortezomib resistant (BR) subclone. Point mutation in the PSMB5 gene (encoding the B5 subunit) leading to altered binding kinetics is identified as a potential underlying mechanism of PI resistance. Despite these adaptive modifications that support resistance to PI resistance, the UPS as a whole remains a relevant target with vulnerable sites that can be therapeutically exploited. The 19S cap of the proteasome contains deubiquitinating enzymes (DUBS) that function to deubiquitinate proteins prior to their entry into the proteasomal core. Using a novel DUB-inhibitor b-AP15, which blocks the UCHL5 and USP14 DUBS, D’Arcy et al demonstrated that disruption of these DUBS does not alter proteasomal activity yet induces robust tumor cell death in various cancers. Aim Investigation of b-AP15 in preclinical models of BR WM and characterization of associated downstream signaling interactions. Methods For these studies, we developed and used BR variants of the most commonly used WM cell lines (variants termed BCWM.1/BR, RPCI-WM1/BR and MWCL-1/BR). Proteasomal activity was measured in these cells using synthetic fluorogenic peptide substrates and apoptosis was measured by annexin-v/PI staining. For gene expression profiling (GEP), the Nanostring nCounter mRNA quantification assay was utilized. Results We treated all BR WM clones with either single agent bortezomib (10nM), carfilzomib (10nM), b-AP15 (100nM) or all three agents in combination and assessed their effect on chymotrypsin-like, caspase-like and trypsin-like proteasomal activity. b-AP15, either alone or in combination with the other PI, did not alter proteasomal activity, nor did it abrogate the inhibition of chymotrypsin-like activity of other PI. Although proteasomal activity was not disrupted, we observed an increase in high molecular weight ubiquitinated proteins on immunoblot analysis. We hypothesized that despite the BR cells resistance to B5-targeting PI, they would remain sensitive to UPS inhibition if targeted through DUB inhibition. We treated BR cells with b-AP15 at concentrations ranging from 50nM - 1.25uM and observed that b-AP15 induced robust WM cell death in a dose and time dependent manner. Apoptosis was more pronounced in MWCL-1/BR cells and optimal cell death (55%) was observed at 500nM; however a commitment towards cellular demise was evident by PARP-1 and caspase cleavage as early as 3 hrs. Use of a pan-caspase inhibitor resulted in rescuing of the BR WM cells from b-AP15 mediated death and demonstrated a caspase-dependent death process. In order to gain further mechanistic insight into the processes that regulates response to b-AP15, we performed GEP and immunoblot of BR WM cells treated with 50nM of b-AP15. We observed an increase in phospho-p38 protein, which belongs to the MAPK family and regulates IL8 secretion. When BR WM cells were treated with a p38 inhibitor in combination with b-AP15, a significant decrease in cell viability was noted relative to b-AP15 (100nM) alone. Furthermore, GEP and network analysis identified key regulator genes associated with p53, cell cycle and MAPK signaling. Summary We provide first evidence that inhibition of the DUBS, UCHL5 and USP14 with b-AP15 results in caspase-dependent apoptosis of BR WM cells. Our investigation uncovers important molecular interactions that may dictate the BR cells responsiveness to DUB-inhibition by b-AP15. These observations validate the pervasive role of the UPS and confirm alternate therapeutics targets using our unique BR WM models. We are grateful to the Waterfall Waldenström Macroglobulinemia Research Fund, the Leukemia and Lymphoma Society and the International Waldenström Macroglobulinemia Foundation for their continued support. Disclosures: Martin: Teva: Consultancy, Research Funding; Celgene: Consultancy, Research Funding; Genentech: Speakers Bureau; Millennium: Research Funding; Seattle Genetics: Consultancy, Speakers Bureau.

Abstract Background Waldenstršms macroglobulinemia (WM) is a rare and incurable indolent Non-Hodgkin's lymphoplasmacytic lymphoma. Rarity and limited availability of authentic preclinical models has been a major constraint in drug development for WM. We recently reported the establishment and characterization of a novel human WM cell line, RPCI-WM1. Thereafter we evaluated the tumorigenic potential of this unique WM cell line in a xenograft mouse model along with its ability to produce human IgM in vivo. Moreover, we assessed its potential as a possible model for enhanced understanding of WM cell biology in vivo and a mechanism to conduct functional studies and preclinical anti-WM drug evaluation. Aim The primary goal of this experiment was to develop and validate an in vivo WM model that can be used to investigate WM cell biology as well as drug evaluation. Methods The RPCI-WM1 cell line (1x106 cells) was injected both subcutaneously and via tail vein into female SCID mice. Engraftment was validated via detection of human IgM (ELISA) and tumor growth (as well as response to therapy) was monitored by weekly assessment of IgM levels and external tumor formation measurements. Results After subcutaneous injection (flank) into mice, the RPCI-WM1 cells developed localized tumors that secreted human IgM detectable in serum as early as 14 days. Progressive increase in IgM was measured in the serum and this corresponded with tumor growth in a time-dependent fashion. On average, tumor tissues grew to a median of 3 cm and IgM levels reached a median of 823 ng/mL by Day 30 post-implantation. We also assessed the practical utility of this xenograft mouse model for evaluation of therapeutics, using a compound X (propriety drug with demonstrable in vitro anti-WM effects). Eleven days post-implantation, the mice were randomized into 3 groups receiving either DMSO (n=6), 100 mg of compound X (n=7) or 200 mg/kg of compound X (n=7) every alternate day via intraperitoneal injection for 26 days. On Day 37, mice were sacrificed and tumor volume was measured. We observed a 48% reduction in tumor volume in the 100mg/kg cohort, with no additional anti-WM benefit noted in the 200mg/kg cohort. Human IgM secreted by the xenografted tumors was measured by ELISA and was significantly lower in mice treated with compound X versus the control group (Figure 1). We also conducted immunohistochemistry (IHC) analysis on tumor tissues obtained from 9 mice treated with compound X (both dose cohorts) or DMSO alone. IHC staining demonstrated changes in the known target proteins of compound X (such as BCL2, p53, Ki-67) between the treated and untreated mice. Summary Thus we demonstrate that the RPCI-WM1 cell line can be dependably used to induce tumor growth through xenotransplantation in SCID mice and this mimics the IgM secreting clinical disease model. This can then be used to serve as a platform for preclinical testing of novel anti-WM therapeutics and allow for subsequent correlative studies for pharmacodynamics purpose. Comprehensive data regarding this model its utility in evaluation of various biological questions such as those relevant to the tumor microenvironment as well therapeutic modeling will be presented. We are grateful to the Leukemia and Lymphoma Society and the International Waldenstršm Macroglobulinemia Foundation for their continued support. Disclosures: No relevant conflicts of interest to declare.

Abstract Background Multiple myeloma (MM) is the second most common hematologic malignancy in the U.S. and is characterized by increased immunoglobulin production and infiltration of the bone marrow by malignant plasma cells. Myeloma cell growth is supported by both the elements present in the bone marrow microenvironment as well as deregulation of internal cellular systems associated with proliferation and apoptosis. Defective programmed cell death by BCL2 or MCL1 upregulation is observed in >80% of myeloma cases and is associated with an aggressive clinical course. Remarkably, there is no approved drug with the ability to target BCL2 or MCL1. Various strategies to mitigate the effects of increased BCL2 functionality, via small molecule inhibitors or BCL2-specific antisense oligonucleotides have been previously examined in MM. These studies revealed that 1) apart from BCL2, additional anti-apoptotic members belonging to the BCL2 family (i.e. MCL1, BCL-xL) play integral roles in maintenance of the myeloma cell phenotype, 2) drugs, which target only BCL2 offer limited therapeutic advantage and 3) given that survival factors engaged in MM are multifactorial, perhaps isolated targeting of the BCL2 pathways may not be sufficient. We therefore hypothesized that the clinical development of drugs that target various antiapoptotic BCL2 family members will require a partnership with established anti-MM regimens whose activity can potentially be further enhanced with correction of the apoptotic response system. Aim Since lenalidomide and dexamethasone (LD) is an established therapy for plasma cell cancers, we investigated if the pan-BCL2 inhibitor AT-101 (BH3 mimetic), which has high binding specificity for BCL2, MCL1 and BCL-xL, can be an effective therapeutic partner to enhance anti-MM effects of LD. Methods Human MM (KMS11, U266, OPM2) and WM (BCWM.1) cell lines along with their corresponding bortezomib resistant (BR) clones, (KMS11/BR, U266/BR, OPM2/BR and BCWM.1/BR) which were developed in our laboratory were used in this study. For gene expression profiling, the Illumina HumanHT-12 v3 whole-genome gene expression array and Nanostring nCounter mRNA quantification assays (NanoString, Seattle, WA) were utilized. Statistical analysis was conducted using R-based packages and the MeV software (TIGR). Apoptosis was measured by annexin-v/PI staining, and mitochondrial membrane permeability (MOMP) was assessed using TMRM followed by flow cytometry. Protein profiles were ascertained by western blot. Results Gene expression and immunoblot analysis of six plasma cell cancer models showed upregulation of various BCL2 family members, notably MCL1 and BCL-xL. In a dose-dependent manner, AT-101 was able to downregulate BCL2 and MCL1 and induced apoptotic cell death in MM and WM cells in a dose dependent manner. Tumor cell death was associated with caspase and PARP-1 cleavage accompanied by an increase in MOMP. This cytotoxic effect and BCL2 downregulation were further potentiated when AT-101 was combined with lenalidomide/dexamethasone (LDA). Nanostring nCounter mRNA quantification and IPA analysis revealed differential changes in the CCNA, FRZB, FYN, IRF1, PTPN11 genes in LDA treated cells. IPA canonical pathway analysis demonstrated the p53 signaling and the cyclins and cell cycle regulation pathways to be the most significantly activated by LDA therapy. Summary In summary, we describe for the first time the cellular and molecular events associated with the use of AT-101 in combination with lenalidomide/dexamethasone in preclinical models of plasma cell malignancy. This study lays the rationale for engaging and handicapping the intrinsic apoptotic system through the pan-BCL2 targeting capabilities of AT-101; thus enhancing the anti-MM effects of lenalidomide/dexamethasone. Disclosures: Foran: Celgene: Research Funding.

Abstract Introduction: With approval of Daratumumab (Dara), targeting CD38 has become an attractive therapeutic option for various B cell cancers including WM. Using preclinical models of WM we investigated CD38 expression patterns, the impact of drug-resistance and the anti-WM effects of Dara, in vitro. An important aspect of our evaluation was the identification of optimal CD38 expression that correlates with maximal cytotoxic effects of Dara; such an approach has not been attempted previously in WM. Methods: WM cell lines (BCWM.1, MWCL.1 and RPCI.WM1) with isogenic proteasome inhibitor-resistant (CR, resistant to carfilzomib, bortezomib and ixazomib) and ibrutinib-resistant subclones (IR) were used. CD38 expression was measured by flow cytometry and a grading system was developed based on increasing CD38 MFI values with 1+= <50,000, 2+ = >50,000 and 3+ = >100,000. Daratumumab-induced specific lysis due to antibody dependent cell cytotoxicity (ADCC) or complement dependent cytotoxicity (CDC) was measured using the Calcein AM assay at an effector to target ratio (E:T) of 100:1. Peripheral blood mononuclear cells (PBMCs) and complement-containing serum were obtained from healthy human donors. Daratumumab, carfilzomib and ibrutinib were obtained from commercial sources. Results: Expression of CD38 (MFI) was highest in RPCI-WM1 (91295.6) > BCWM.1 (13865.3) > MWCL-1 (9804.19) (Figure 1A.). Intriguingly we observed an interesting change in CD38 expression with induction of drug resistance; all CR clones showed a decrease in MFI (RPCI-WM1/CR; øMFI 55960.5 > MWCL-1/CR; øMFI 3890.0 > BCWM.1/CR; øMFI 1381), while decreased expression was noted in all IR clones except one (MWCL-1/IR; øMFI 6272.96, BCWM.1/IR; øMFI 3076 vs. RPCI-WM1/IR; _MFI 461784.6) (Figure 1B). When we applied our CD38 expression grading system, RPCI-WM1/IR cells classified as Grade 3+, RPCI-WM1 and its CR clones as Grade 2+, while all other WM cells were Grade 1+. We noted that utilization of this grading system was helpful in correlating the extent of cytotoxicity to the level of CD38 expression (Table 1). We next treated WM cells with Dara and observed direct lysis in only RPCI-WM1/CR (43%) > BCWM.1/CR (19%) and RPCI-WM1 (11%) cells. However, when Dara-treated WM cells were cocultured with PBMCs, significantly increased ADCC was noted in RPCI-WM1 (90.3%), RPCI-WM1/IR (88.3%) and RPCI-WM1/CR (68.0%) cells (Figure 1C). Although CDC has been reported as a mechanism of mAb-mediated cell death in WM our evaluation of CDC as a potential effector mechanism in daratumumab-treated WM was negative as expected, highlighting ADCC rather than CDC as the primary cytotoxic mechanism. Conclusion: Our analysis of CD38 in WM preclinical models highlights several important findings: 1.) Expression of CD38 changes as WM tumor clones evolve with induction of drug-resistance; an effect that may also be cell-type specific as evidenced by increased CD38 expression in RPCI-WM1/IR cells, 2.) Daratumumab can exert a direct lethal effect in WM cells but this effect is variable and cell line dependent, 3.) Expression of CD38 is an important factor directing cytotoxicity of Dara however antigen density appears to be an important factor and a grading system can be beneficial for future preclinical and/or clinical studies, 4.) ADCC and not CDC appear to be the prominent mechanism of action and 5.) ADCC-mediated cytolytic activity of Dara is contingent on CD38 expression as evidenced by the most specific lysis being noted in CD38 Grade 2+ and 3+ expressing WM cells (Table 1.). These observations can help direct optimal design of clinical studies with Dara in WM where drug resistance and CD38 expression should be further explored as factors contributing to the clinical efficacy of this new agent. Disclosures Ailawadhi: Pharmacyclics: Consultancy; Novartis: Consultancy; Amgen Inc: Consultancy; Takeda Oncology: Consultancy. Ansell:BMS, Seattle Genetics, Merck, Celldex and Affimed: Research Funding.

Abstract Background: Aberrant B-cell receptor (BCR) and MYD88 signaling cooperatively drive Waldenstrom macroglobulinemia (WM) tumor cell survival. To sustain BCR-driven growth and maintain optimal cellular homeostasis, WM cells upregulate proteasome function. We have recently described the role of two distinct 19S proteasome-associated deubiquitinating enzymes (DUBs), USP14 and UCHL5, in maintaining WM-cell viability. Pharmacologic inhibition with b-AP15 (commercial compound) results in WM-cell death and accumulation of high molecular weight ubiquitinated protein conjugates; an event that is phenocopied by siRNA knockdown of USP14 and UCHL5. VLX1570 is a lead-optimized clinical grade analog of b-AP15 with enhanced solubility, stability and target residence time. We investigated the effect of VLX1570 on BCR components, WM cell survival proteins and its in vivo activity in a human xenograft model of WM. Materials: VLX1570 was obtained from Vivolux AB and bortezomib from Sellekchem. Human WM cell lines (BCWM.1, MWCL-1 and RPCI-WM1) and their drug-resistant derivatives were used (total n=12). Bone marrow aspirates from WM-patients and from healthy donors were obtained from the Mayo Clinic biospecimens core. PBMCs from healthy donors were used as comparators where needed. NOD/SCID mice were obtained from Taconic Labs and used for profiling the in vivo properties of VLX1570. Results: We first examined for USP14/UCHL5 protein expression by immunohistochemical analysis in primary tumor tissue from WM patients and compared it to healthy donor cells. WM patient-derived cells exhibited greater staining intensity of both DUBs. Preferential binding of VLX1570 to USP14 and UCHL5 in WM cells was probed in a WM cell using an enzyme-based competition assay (HA-Ub-Vs), which demonstrated time-dependent decrease in labeled USP14 and UCHL5 after treatment with VLX1570. In drug-resistant (bortezomib or ibrutinib-resistant) WM cells, VLX1570 induced >50% tumor-specific apoptosis and displayed synergistic activity with ibrutinib (median CI: 0.5). In addition to ER stress markers, we had previously observed that gene expression of BCR-signalosome components were reduced in VLX1570-treated WM cells (NanoString platform) and this was substantiated at the protein level where p-SYK, SYK and BTK decreased in a time-dependent manner. BCR end-effectors were probed in response to VLX1570 and a decrease in nuclear translocation of NFkB (p65) was noted. USP14 has been shown to control the availability of the chemokine receptor CXCR4, which regulates homing of WM cells to bone marrow. After VLX1570 treatment, we noted a significant decrease in CXCR4 surface receptor expression in all WM cell lines tested and primary WM cells. Lastly, to determine the in vivo anti-WM effects of VLX1570, we injected NOD/SCID mice (xenotransplanted with RPCI-WM1-Luc cells) with VLX1570 (I.P. route, 4.4mg/kg every other day) and measured tumor burden along with pharmacodynamic indicators. VLX1570-treated mice (n=7) demonstrated significantly lower tumor volumes by caliper measurement and bioluminescent signal (p=0.01 - 0.003) as well as reduced serum human-IgM, compared to vehicle-treated mice (n=7). Importantly, mice treated with VLX1570 survived longer vs. vehicle-treated mice (median survival, 62 vs. 44 days, respectively, p=0.0008). No unforeseen toxicities were noted and weight of the mice in vehicle vs. VLX1570-treated mice did not significantly change. IHC analysis in mice tumor tissues demonstrated reduced CXCR4, human-IgM, Ki67 and increased cleaved caspase-3 and ubiquitinated (k48-linked) protein level, the latter serving as a more specific indicator of USP14/UCHL5 inhibition. Conclusions: We provide substantial in vitro and in vivo evidence on the anti-WM effects of a novel first-in-class clinical grade USP14/UCHL5 inhibitor, VLX1570. In addition to anticipated induction of ER stress machinery, we describe for the first time VLX1570-mediated modulation of BCR components and WM cell survival proteins as well as demonstrate its robust activity in an aggressive human xenograft model of WM. These data provide the rationale for a clinical trial of single agent VLX1570 in patients with symptomatic WM. Disclosures Martin: Janssen: Consultancy, Honoraria; Acerta: Consultancy; Gilead: Consultancy; Celgene: Consultancy; Novartis: Consultancy; Bayer: Consultancy. Gullbo:Vivolux AB: Other: Shareholder. Linder:Vivolux AB: Other: Shareholder.