Faculty Bibliography

Background: The prognosis of multiple myeloma (MM) patients who are refractory to all currently available therapies, including anti-CD38 monoclonal antibodies, remains poor, indicating that there is an unmet need for novel treatment strategies. Although data on G protein-coupled receptor 5D (GPRC5D) protein expression is scarce, RNA levels are selectively elevated in MM cell lines and primary MM cells. This makes GPRC5D an attractive target for the novel anti-myeloma JNJ-7564 bispecific antibody.
Method(s): We determined GPRC5D protein expression levels on MM cell lines and bone marrow (BM) mononuclear cells (MNCs) derived from healthy donors (HD) and MM patients by flow cytometry. We also analyzed the impact of differential GPRC5D gene expression in purified CD138+ MM cells derived from patients who participated in 5 large randomized clinical trials (HOVON65, MRC-IX, TT2, TT3 and APEX), on overall survival (OS) and progression free survival (PFS). Furthermore, the activity of a new GPRC5DxCD3 bispecific antibody (JNJ-7564) was evaluated in MM cell lines and patient-derived whole BM samples after 48 hours of incubation (concentrations 0.00064-4mug/ml), as well as biomarkers for in vitro JNJ-7564 response. At baseline, MNCs were characterized for the composition of T-cell subsets.
Result(s): GPRC5D protein expression was significantly higher on MM cells compared to HD plasma cells and immune cells. Gene expression levels of GPRC5D were highly variable in MM, but were significantly higher in patients with t(4;14) or gain 1q. There was no association with OS or PFS in the clinical trials (n=1421). JNJ-7564 effectively killed GPRC5D+ MM cell lines (MM1.S, UM9 and RPMI-8226) in a dose-dependent manner, using HD or patient-derived blood MNCs as effector cells. Co-incubation with patient-derived BM stromal cells resulted in a modest impairment of killing capacity in 2 out of 3 cell lines. In MM patient samples (n=29), the mean lysis of MM cells with 4.0mug/mL JNJ-7564 was 57% (range: -8-97%), while NK-cell and T-cell frequencies were not affected. JNJ-7564 was also active in samples from extensively pretreated and daratumumab (DARA)-refractory patients (n=8; mean lysis with 4.0mug/mL: 59%; range: 22-97%). JNJ-7564-mediated MM cell lysis was associated with activation (CD25+) and degranulation (CD107a+) of CD4+ and CD8+ T-cells. Maximum kill was correlated with GPRC5D expression on MM cells (Spearman r=0.40, p=0.029), % of regulatory T-cells (r=-0.41, p=0.023) and % of PD-1+ T-cells (r=-0.43, p=0.033).
Conclusion(s): GPRC5D is highly and selectively expressed on MM cells, which makes it an attractive therapeutic target. The GPRC5DxCD3 bispecific antibody, JNJ-7564, effectively lysed MM cell lines and primary MM cells in samples derived from both newly diagnosed and heavily pretreated patients, including DARA-refractory patients. Altogether, this strengthens the preclinical rationale for an ongoing phase 1 study with JNJ-7564 in relapsed/refractory MM. Keywords: bispecific antibody immunotherapy Multiple myeloma Tracks: Multiple Myeloma Novel Agents
Copyright

Multiple myeloma (MM) newly diagnosed patients have a wide range of prognoses, which are traditionally been defined by key translocations and microarray-based gene expression profiling (GEP). In diseases that can differ greatly in prognosis based on initiating genomic events, such as MM, GEP is reliably used to establish risk. Several prognostic scores based on microarray data have been developed to identify high-risk MM patients, which include EMC-92, GEP-70, GEP-17, GEP-80, IFM-15, MRC-IX-6, and GPI-50. Microarrays are inexpensive and give consistently reproducible results, but with the wide adoption of next-generation sequencing, reduced costs, and improved algorithms, RNA-Seq is long overdue to replace microarrays. The goal of this study was to show the efficacy of replacing microarray expression data with RNA-Seq data in well-established MM prognostic risk-scores. We took 167 MM samples that had both RNA-Seq and microarray data and recapitulated each of the prognostic scores. Microarray probes used in each of the signatures were mapped back to genes. In many instances there was not a one-to-one conversion and resulted in fewer genes in the corresponding RNA-Seq score. The new cutoff was achieved in two-ways: 1) we ranked the samples by risk score and empirically determine the best RNA-Seq cutoff, and 2) using the running log rank tests on progression free survival (PFS) and overall survival (OS) separately in the training set and identify optimal cutoffs. The agreement between the calls generated with microarray and RNA-Seq take the microarray based call as being the truth set. GEP70 risk scores from RNA-Seq correlated with the Microarray risk scores with an R-squared value of 0.854. Sensitivity and specificity of the RNA-Seq risk scores had values of 0.83 and 0.93 respectively. The RNA-Seq based scores were validated on the CoMMpass dataset of 615 MM samples with RNA-Seq data with corresponding clinical data. RNA-Seq GEP70 scores classified 30% of the patients in the validation dataset as high-risk. Survival analysis showed that the RNA-Seq based scores was able to classifying samples as high-risk as demonstrated with a significant difference in PFS and OS in the validation dataset (p-value<0.001, p-value=0.002). It is extremely important to accurately identify high-risk patients early so that they can receive the most aggressive treatment and enter into appropriate clinical trials. This preliminary work demonstrated that with adjusted cutoffs, RNA-Seq expression data can be used in place of microarray data to classify patients as either high-risk or low-risk using established clinically relevant risk signatures. Keywords: Gene expression profiling Risk stratification RNA-Seq Tracks: Multiple Myeloma Genomics
Copyright

Background: Multiple myeloma (MM) has significant spatial and temporal clonal heterogeneity suggesting therapeutic agents with different mechanisms of action, delivered in combination or sequentially, are required to maximize the depth of response and improve outcomes. The UK NCRI Myeloma XI phase III randomized trial compared induction with the second generation proteasome inhibitor carfilzomib and lenalidomide containing quadruplet, KCRD, vs a response-adapted approach of sequential triplet therapies in newly diagnosed transplant eligible patients.
Method(s): 1056 patients were randomized between KCRD (28 day cycles of carfilzomib (K) 36mg/m2 IV d1-2,8-9,15-16, cyclophosphamide (C) 500mg PO d1,8, lenalidomide (R) 25mg PO d1-21, dexamethasone (D) 40mg PO d1-4,8-9,15-16) and immunomodulatory drug (IMiD) triplet CTD/CRD prior to ASCT. Patients with a suboptimal response to CTD/CRD underwent response-adapted intensification randomization to a proteasome inhibitor (bortezomib, CVD) containing triplet or no CVD. A maintenance randomization at 3 months post ASCT compared lenalidomide to observation. Molecular high-risk (HiR) was classified by t(4;14), t(14;16), t(14;20), del(17p) or gain(1q) with ultra-high risk (UHiR) the presence of >1 lesions.
Result(s): KCRD was associated with a significantly longer PFS than IMiD triplet therapy (HR 0.63, 95%CI 0.51, 0.76, 3yr PFS KCRD 64.5% vs CTD/CRD 50.3%, p<0.0001). PFS2 was also significantly improved with KCRD (HR 0.75, 95% CI 0.56, 0.99, 3yr PFS2 KCRD 81.8% vs CTD/CRD 75.1%). Deeper response rates were seen in patients treated with KCRD vs CTD/CRD pre and post-transplant (p<0.0001). All regimens were well tolerated with no significant additional toxicity due to the quadruplet regimen. A higher proportion of patients receiving KCRD induction were able to undergo ASCT than those who received response-adapted induction and in an analysis restricted to those who had completed ASCT, KCRD induction was still associated with a significantly longer PFS. There was no significant heterogeneity in PFS outcome between molecular risk groups with a benefit for KCRD over triplets in all. In patients receiving KCRD there was no difference in response rate at the end of initial induction by risk group but UHiR disease was associated with significantly shorter PFS than both SR and HiR, whilst there was no difference in outcome between patients with HiR (one adverse lesion only) and SR. An exploratory analysis compared the patients receiving KCRD to patients in the CTD/CRD arm who had received the optimum response-adapted approach (i.e. excluding those with a suboptimal response randomized to no CVD). KCRD was associated with significantly longer PFS than using a response adapted sequential triplet approach (HR 0.64, 95% CI 0.52, 0.78, p<0.0001).
Conclusion(s): KCRD was well tolerated with deep responses pre- and post-transplant and a significant PFS benefit compared to triplet therapy across all risk groups. Keywords: carfilzomib carfilzomib-lenalidomide-dexamethasone Lenalidomide Tracks: Treatment of Newly Diagnosed Myeloma Transplant Eligible
Copyright

Deletions in Chromosome 17p (del17p) are associated with poor outcome in Multiple Myeloma (MM). We compiled a cohort of 139 whole-exome and whole-genome sequenced del17p newly diagnosed patient tumors and correlated the size of the deletion, as calculated by controlFreec, against progression-free and overall survival. Large (> 10.9MB) deletions in chromosome 17p that included TP53 region were significantly associated with poorer overall survival (mOS = 28.8 months vs. mOS = 45.0 months, pval= 0.047). Further, patients with large del17p deletions were more likely to be bi-allelic TP53 with high clonal cell fraction (CCF) for deletion of TP53 (p < 0.05). We identified a gene expression signature composed of genes in the 17p region that were associated with high CCF and large TP53 deletions. Compared to low CCF tumors, gene expression pathways uniquely downregulated to large-deletion 17p tumors included mitotic checkpoint (p < 0.05) and spindle formation (p < 0.05) pathways. This analysis suggests that biallelic inactivation or large TP53 deletion are involved in early events associated with myelomagenesis and in-part drive high-risk phenotype. Further clinical data validation in relapsed setting and experimental validation of TP53-driven pathways and targets in isogenic CRISPR and KD models are ongoing and Results: from these analyses will be presented. Keywords: Genetic profiling genomic architecture Overall survival Tracks: Multiple Myeloma Genomics
Copyright

Background: Multiple myeloma (MM) is a heterogeneous disease with variable outcomes. In the past several years, serum proteins, cytogenetics and gene expression profiling (GEP) have been used to predict outcomes of MM patients. Patients identified by the prognostic GEP70 gene signature as high risk (HR) have more aggressive disease and shorter survival compared to patients identified as low risk (LR). We have analyzed cfDNA levels in LR and HR patients and determined its correlation to GEP70 risk score. In a subset of patients, we performed low-pass whole genome sequencing (LP-WGS) and targeted sequencing to determine genomic alterations in cfDNA.
Method(s): GEP was performed on CD138+ plasma cells from bone marrow using Affymetrix U133 Plus 2.0 arrays. Low-risk (n=38) and high-risk MM patients (n=44) as determined by GEP70 score were selected. Plasma cells in bone marrow aspirate were ascertained for tumor burden by flow cytometry. cfDNA was extracted from plasma using QIAamp circulating nucleic acid kit (Qiagen) and was measured using Qubit fluorometer. In a subset of patients, plasma cfDNA, matched tumor DNA and matched white blood cell genomic DNA were sequenced using a targeted panel covering key driver genes and immunoglobulin regions involved in translocations in myeloma. Targeted sequencing was performed on NextSeq500 (Illumina) to a depth of 400-600x. LP-WGS was performed at 0.1X coverage. Sequencing data were analyzed using Strelka and ichor.
Result(s): Total cfDNA (ng/ml plasma) was significantly higher in the HR group compared to the LR group, median cfDNA LR=18.76 ng/ml range 0.2-140 ng/ml plasma vs. HR=33 ng/ml plasma range 7.3-726.66 ng/ml; p=0.02. cfDNA levels among different GEP subgroups did not reach significance, however patients in the PR subgroup had higher cfDNA in plasma compared to other subgroups. Ranked cfDNA levels correlated with GEP risk score r=0.32, p=0.0029 (Spearman's test). cfDNA levels also correlated with tumor burden r=0.41, p<0.0001 (Pearson's test). Additionally, LP-WGS analysis performed in a subset of patients showed that circulating tumor DNA (ctDNA) fraction correlated strongly with GEP70 risk score (Spearman r=0.83, p=0.0012). Monitoring of cfDNA levels in patients before and after chemotherapy showed an increase in cfDNA levels between 3-5 days after chemotherapy and fell to baseline levels a week later. Variant allele frequencies of NRAS and KRAS mutations were higher immediately post-chemotherapy compared to baseline in 3/4 patients.
Conclusion(s): cfDNA levels correlate with GEP70 risk score. In the small subset of sequenced patients, ctDNA fraction also correlated with GEP70 risk score. Molecular monitoring of myeloma using cfDNA can capture the mutational landscape in myeloma and can be a potential prognostic biomarker for MM disease. Keywords: cell-free DNA Gene expression profiling Whole genome sequencing Tracks: Multiple Myeloma Genomics
Copyright

MRD is a powerful predictor of survival outcomes in multiple myeloma but treatment regimens show increasing duration and complexity for both TE and TNE populations. For this reason, sequential assessments are preferable to single timepoints. This has been evaluated in the Myeloma XI trial, which was a phase 3 trial with three potential randomisations to determine induction therapy, consolidation therapy and maintenance in both TE and TNE patients. Bone marrow aspirates were obtained after induction, post consolidation (if given), post ASCT for TE patients and 6 months post maintenance randomisation. MRD was assessed using flow cytometry (minimum sensitivity 0.004%). This analysis represents a subset of 1630 samples that represent all patient groups and therapeutic timepoints. Overall MRD-negativity post-induction was 164/722 (22.7%). 70.1% of patients randomised to CCRD were MRD-negative compared to 19.6% after CTD or RCD in TE patients and 12.7% after CTDa and 19.2% after RCDa (p <0.001) in TNE patients. Levels of residual disease in MRD-positive patients were lower following CCRD induction; median 0.08% of leucocytes (range 0.001-9.5%), compared to medians of 0.31%, 0.23%, 0.38% and 0.42% following CTD, RCD, CTDa and RCDa respectively. MRD-negativity was demonstrated in 16/54 (29.7%) of patients who received bortezimib-based consolidation following a suboptimal response to induction. Post ASCT, 257/413 (62%) of samples were MRD-negative, with a significant increase in MRD-negativity rate relative to post induction seen with CTD (59.2% vs 19.6% post induction) and RCD (53.6% vs to 19.6% post induction). The level of conversion from MRD-positive post induction to MRD-negative post-ASCT was lower in the CCRD cohort (83.3% MRD-negative post ASCT compared to 70.1% following induction). Amongst the patients who underwent maintenance randomisation, MRD-negativity was demonstrated in 238/413 (57.6%). Conversion to MRD-negativity was seen in 32% of MRD-positive patients on lenalidomide maintenance, whilst those patients who became MRD-positive during maintenance had poor outcome. For those patients that remained MRD-positive, lower levels of residual disease were noted in those receiving maintenance (median 0.15% on maintenance vs 0.39%, p=0.04). Sequential MRD monitoring allows for the assessment of individual components of complex myeloma therapeutic strategies. It enables comparison of induction regimens, possibly identifying patients that may not require ASCT. Consolidation or maintenance strategies can also be assessed. This data suggests a potential future role for flow cytometric MRD assessment in individual patient management. A full dataset of >4000 samples and mature outcome data will be presented at the meeting. Keywords: Flow Cytometry Minimal residual disease Tracks: Myeloma Response Assessment including MRD
Copyright

Introduction /Background: Multiple myeloma (MM) is a malignancy that is characterized by clonal proliferation of neoplastic plasma cells. Epigenomic abnormalities contribute to the pathogenesis of MM. Bromodomain 4 (BRD4), a member of the bromodomain and extra terminal (BET) family, binds to acetylated histones during M/G1 transition in the cell cycle promoting progression to S phase. Other histone modifiers include the acetyltransferases cyclic AMP response element binding protein-binding protein (CBP) and the E1A interacting protein of 300 kDa (EP300). Disruption of these epigenetic processes leads to altered gene function and tumorigenesis, and contributes to the pathogenesis of MM. A number of studies have shown that targeting these individual classes of proteins has anti-tumor activity in MM cell lines (MMCLs), as well as other cancers. Here we present the first data exploring the anti-tumor activity of the novel dual inhibitors of BET and CBP/EP300 in MM.
Method(s): 16 MMCLs were exposed to increasing concentrations of dual BET/CBP inhibitors NEO2734 and NEO1132, single BETis JQ1, OTX015, iBET-762 and iBET-151, and a single CBP/EP300 inhibitor CPI-637 for 72 hrs. Cell viability, cell cycle and protein levels were analysed in MMCLs following exposure to compounds.
Result(s): 16 MMCLs were exposed to NEO2734 and NEO1132. The compounds showed anti-tumor activity with a median IC50 of 140 nM (95% C.I., 62-284 nM) and 402 nM (95% C.I., 153-817 nM) respectively, with NEO2734 being significantly more potent that NEO1132 (P<0.001). As a comparison, all MMCLs were exposed to the BETis JQ1, OTX015, iBET-762 and iBET-151, and to a CBP/EP300 inhibitor CPI-637, the median IC50 values of these compounds were 83 nM (95% C.I., 40-155 nM), 250 nM (95% C.I., 110-497 nM), 360 nM (95% C.I., 178-912 nM), 687 nM (95% C.I., 313-1316 nM) and 1970 mM (95% C.I., 606-5250 nM) respectively. The novel dual inhibitor NEO2734 was more potent than the single BET inhibitors iBET-762 (P = 0.0084), iBET-151 (P = <0.0001) and the CBP/EP300 inhibitor CPI-637 (P = <0.0001), and showed no significant difference from the highly potent JQ1 and OTX015 inhibitors (P = 0.067 and P = 0.056 respectively). When comparing the IC50 across the different MMCLs JQ1 ranks as consistently the most potent followed by NEO2734 for each of the MMCLs tested. Sensitivity to compounds differs for each MMCL with some being more sensitive than others. Like the single BET and CBP/EP300 inhibitors, NEO2734 and NEO1132 induce a G1 cell cycle arrest in sensitive MMCLs at 24 hrs. We also observed a marked decreased in the levels of proteins regulated by BET and CBP/EP300, including c-Myc and IRF4, following compound exposure indicating an effect on the regulatory domains.
Conclusion(s): The dual BET and CBP/EP300 inhibitor NEO2734 shows a strong anti-tumor activity and is consistently highly active against all MMCLs, being as potent as JQ1 and more so than other single BET and CBP/EP300 inhibitors. Keywords: BET/BRD4 CBP/EP300 Dual inhibitor Tracks: Multiple Myeloma Novel Agents
Copyright

The iliac crest is the sampling site for minimal residual disease (MRD) monitoring in multiple myeloma (MM). However, the disease distribution is often heterogeneous, and imaging can be used to complement MRD detection at a single site. We have investigated patients in complete remission (CR) during first-line or salvage therapy for whom MRD flow cytometry and the two imaging modalities positron emission tomography (PET) and diffusion-weighted magnetic resonance imaging (DW-MRI) were performed at the onset of CR. Residual focal lesions (FLs), detectable in 24% of first-line patients, were associated with short progression-free survival (PFS), with DW-MRI detecting disease in more patients. In some patients, FLs were only PET positive, indicating that the two approaches are complementary. Combining MRD and imaging improved prediction of outcome, with double-negative and double-positive features defining groups with excellent and dismal PFS, respectively. FLs were a rare event (12%) in first-line MRD-negative CR patients. In contrast, patients achieving an MRD-negative CR during salvage therapy frequently had FLs (50%). Multi-region sequencing and imaging in an MRD-negative patient showed persistence of spatially separated clones. In conclusion, we show that DW-MRI is a promising tool for monitoring residual disease that complements PET and should be combined with MRD.

The emergence of treatment resistant sub-clones is a key feature of relapse in multiple myeloma. Therapeutic attempts to extend remission and prevent relapse include the maximisation of response and use of maintenance therapy. We used whole exome sequencing to study the genetics of paired presentation and relapse samples from 56 newly diagnosed patients, following induction therapy, randomised to receive either lenalidomide maintenance or observation as part of the Myeloma XI trial. Patients included were considered high risk, relapsing within 30 months of maintenance randomisation. Patients achieving a complete response had predominantly branching evolutionary patterns leading to relapse, characterised by a greater mutational burden, an altered mutational profile, bi-allelic inactivation of tumour suppressor genes, and acquired structural aberrations. Conversely, in patients achieving a partial response the evolutionary features were predominantly stable with a similar mutational and structural profile. There were no significant differences between patients relapsing after maintenance lenalidomide vs observation. This study shows that the depth of response is a key determinant of the evolutionary patterns seen at relapse.

FRAX is a commonly used tool to evaluate patient fracture risk based on individual patient models that integrate the risks associated with clinical risk factors with or without bone mineral density (BMD) at the femoral neck. Retrospectively, factors identified by the FRAX scoring algorithm were used to predict the risk for vertebral compression fractures at baseline in newly diagnosed multiple myeloma patients. The data were derived from myeloma patients enrolled in Total Therapy Protocols (TT4 & TT5) between 8/2008 and 9/2017. FRAX scores were calculated and baseline PET and MRI imaging obtained. Univariate and multivariate logistic regression analyses determined the association between FRAX components and the existence of vertebral compression fractures, both pathologic and osteoporotic. The patient population had a median age of 61 years (43-76), 37% female, and 87% white. The median major osteoporotic score (MOS) and Hip fracture scores (HFS) for TT4 patients (low-risk myeloma) were 5.6 and 0.5, respectively, while median MOS and HFS for TT5 (high risk myeloma) patients were 6.2 and 0.7, respectively. The odds ratio for fracture at diagnosis in patients with elevated MOS (>2), and HFS (>4.5) was significant OR (1.48, 95% confidence interval (1.35,1.62)) and OR (1.61, 95% confidence interval (1.42, 1.81)), respectively. In sum, an elevated baseline FRAX score was highly predictive of baseline vertebral fractures in MM patients at presentation. In addition, patients with higher FRAX scores had significantly shorter survival in the low-risk (TT4) group but this survival effect was not seen in the high-risk (TT5) group. These findings suggest that FRAX assessment of baseline fracture risk is beneficial in MM patients to identify an individual patients' risk of vertebral fracture.