ALBERT

Description: Abstract 2385 The t(4;14) group in multiple myeloma (MM) is associated with a significantly impaired prognosis based on the overexpression of MMSET, a histone methyltransferase and epigenetic modifier that is thought to play a major role in myeloma genesis. We have previously shown that t(4;14) myelomas are associated with a specific pattern of DNA hypermethylation using Illumina HumanMethylation27 BeadChip arrays. These results were interpreted as being consistent with that MMSET overexpression leads to specific changes in the epigenetic architecture of MM. However, the detailed mechanism underlying this remains unclear partly due to the low resolution of methylation array technology used. In order to address this we have performed high-resolution genome-wide analyses of DNA methylation for t(4;14) and t(11;14) samples from patients with myeloma, plasma cell leukemia (PCL) and myeloma cell lines (HMCL) using methyl binding domain next generation sequencing (MBD-seq) in order to define DNA methylation patterns specific for t(4;14) MM, as well as to analyse methylation changes accompanying the progression from MM to PCL at a high-resolution, genome-wide scale. DNA from 6 myeloma patients at diagnosis [3 MM with t(4;14), 3 MM with t(11;14)], 6 patients with PCL [3 PCL with t(4;14), 3 PCL with t(11;14)] and 6 HMCL [3 with t(4;14), 3 non-t(4;14)] was fragmented. Methylated DNA fractions were captured using biotin-labelled methyl-binding domain 2 (MBD2) protein. Captured sequences bound to MBD were washed and eluted using elution buffers with increasing salt concentrations. Eluted fragments were purified and sequenced on an Illumina GAIIx, using 1.5 lanes per patient sample, generating 36 bp single-end reads. On average, 1.4 Gbases of reads were generated per sample. Sequences were aligned and de-duplicated using stampy and bwa algorithms. The reference genome (build hg19) was divided into overlapping bins of 200 bp (termed probes) and short read coverage per bin was normalised to per million reads aligned. Differentially methylated regions were defined by comparing normalised reads per probe between the t(4;14) and the t(11;14) groups for MM, PCL and HMCL groups. We first compared probe values that were higher in all three t(4;14) MM samples compared to the three t(11;14) samples. About 16500 probe values were higher in t(4,14) cases compared the t(11;14) group, whereas only 470 probes values were higher in all t(11;14) cases compared to t(4;14) cases. This confirms our previous observation that t(4;14) MM cases are characterised by pronounced hypermethylation. Of the 16500 probes values higher in t(4;14), about 9500 probes mapped to gene bodies and 600 to gene promoters, affecting in total about 1600 genes, indicating that gene or gene regulatory sequence hypermethylation is a common feature in t(4;14). Gene set enrichment analyses of these genes demonstrated highly significant enrichment of KEGG pathways ‘pathways in cancer’, ‘cell adhesion molecules’, the GO term ‘cell development’, among others, and an overrepresentation of probes mapping to chromosomal regions on chromosome 1q. When comparing the progression from MM to PCL, about 2600 genomic probe values were higher in all 3 t(11;14) PCL vs all 3 t(11;14) MM and 1600 probes in all 3 t(4;14) PCL vs MM, indicating that hypermethylation from MM to PCL is more pronounced in t(11;14) than in t(4;14). Very few differences in probe values were present when comparing all 6 MM (both t(4;14) and t(11;14)) with the 6 PCL samples, indicating that the epigenetic mechanisms involved in progression from MM to PCL might be different between the cytogenetic subgroups. Enrichment of methylated sequences was strong for both translocation groups when comparing PCLs with HMCLs, demonstrating that the epigenetic architecture of HMCLs differs significantly even from late-stage patient tumour material. This genome-wide methylation analysis provides us with candidate genes that are likely to be directly or indirectly epigenetically modified by MMSET. We are integrating this methylation data with gene expression data to identify expression-methylation correlations. Furthermore, additional experiments using MMSET knockout models will be used to further filter MMSET-specific effects on genome wide methylation. Finally, we go on to define epigenetic markers that could serve as biomarkers for future epigenetic therapies targeting epigenetic modifiers in t(4;14) myeloma. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction A significant proportion of myeloma patients relapse early and show short survival with current therapies. Molecular diagnostic tools are needed to identify these high risk patients at diagnosis to stratify treatment and offer the prospect of improving outcomes. Two validated molecular approaches for risk prediction are widely used: 1) molecular genetic risk profiling [e.g. del(17p), t(4;14)] 2) gene expression (GEP) risk profiling, [e.g. EMC92 (Kuiper et al., Leukemia 2012)]. We profiled patients from a large multicentric UK National trial using both approaches for integrated risk stratification. Methods A representative group of 221 newly diagnosed, transplant eligible patients (median age 64 years) treated on the UK NCRI Myeloma XI trial were molecularly profiled. DNA and RNA were extracted from immunomagnetically CD138-sorted bone marrow plasma cells. Molecular genetic profiles, including t(4;14), t(14;16), Del(17p), Gain(1q) were generated using MLPA (MRC Holland) and a TC-classification based qRT-PCR assay (Boyle EM, et al., Gen Chrom Canc 2015, Kaiser MF, et al., Leukemia 2013). GEP risk status as per EMC92 was profiled on a diagnostic Affymetrix platform using the U133plus2.0-based, CE-marked MMprofiler (SkylineDx) which generates a standardised EMC92 risk score, called 'SKY92'. Progression-free (PFS) and overall survival (OS) were measured from initial randomization and median follow-up for the analysed group was 36 months. Statistical analyses were performed using R 3.3.0 and the 'survival' package. Results were confirmed in an independent dataset, MRC Myeloma IX, for which median follow-up was 82.7 months. Results Of the 221 analysed patients, 116 were found to carry an established genetic high risk lesion [t(4;14), t(14;16), del(17p) or gain(1q)]. We and others have recently demonstrated that adverse lesions have an additive effect and that co-occurrence of ≥2 high risk lesions is specifically associated with adverse outcome (Boyd KD et al, Leukemia 2011). 39/221 patients (17.6%) were identified as genetic high risk with ≥2 risk lesions (termed HR2). By GEP, 53/221 patients (24.0%) were identified as SKY92 high risk. Genetic and GEP high risk co-occurred in 22 patients (10.0%), 31 patients (14.0%) were high risk only by GEP and 17 patients (7.7%) by genetics only. SKY92 high risk status was associated with significantly shorter PFS (median 17.1 vs. 34.3 months; P

Description: Background The t(4;14) translocation in multiple myeloma is associated with aggressive disease behavior and shorter survival despite recent advances in therapy. The histone methyltransferase MMSET/WHSC1 is overexpressed as a consequence of the translocation, but the detailed epigenetic downstream effects of this overexpression are unknown. Our group identified a specific DNA hypermethylation pattern in t(4;14) cases, indicating a strong contribution of epigenetic changes to the aggressive biology of this subtype. We have here performed genome wide methyl-binding domain protein capture and next generation sequencing (MBD-seq) to characterize the epigenetic perturbations in t(4;14) myeloma in contrast to t(11;14) cases, a subgroup with favorable clinical behavior. Patient bone marrow myeloma cells at presentation and at progression to plasma cell leukemia were analysed to identify epigenetic mechanisms characteristic for early and late stage disease. Methods MBD-seq data were generated from bone marrow myeloma samples of 3 t(4;14) and 3 t(11;14) cases at both presentation and plasma cell leukemia (PCL) stages. Data were analyzed with the MEDIPS package and RPM values for 500 bp bins were calculated. Due to copy number sensitivity of the method, matched Affymetrix SNP genotyping array data were used to correct for focal genomic gains and losses. Pairwise contrasts were calculated and results were filtered for statistical significance (p20) to generate lists of differentially methylated regions (DMRs). Publicly available ENCODE datasets such as chromatin state segmentation for the B-lymphocytic GM12878 cell line were used to identify epigenetic mechanisms involved in MMSET activity and disease progression. Matched gene expression data have been generated for presentation myeloma cases and exome sequencing has been performed for a t(4;14) and a t(11;14) myeloma case. Results Copy number corrected MBD-seq results confirmed the previous observation of a specific DNA hypermethylation pattern in t(4;14) versus t(11;14) cases (2729 hyper vs. 582 hypomethylated DMRs). Comparison of t(4;14) specific DMRs with publicly available datasets (GREAT, Bejerano Lab, Stanford) revealed genome wide significant overlap with the top 50 overexpressed genes of the “MS” cluster consisting of t(4;14) cases (binomial FDR q-value: 1.28e-9), confirming the functional epigenetic involvement in t(4;14) biology. There was also significant overlap with the module Genes in the AKT1 Pathway (q-value: 6.62e-8). Comparison of results with ENCODE Chromatin states data for B lymphocytes (GM12878) demonstrated significant overrepresentation of “weak enhancer” and “weak promoter” regions in t(4;14) hypermethylated DMRs. In contrast, hypomethylated DMRs in t(4;14) were significantly enriched for “transcriptional transition” and “transcriptional elongation” (MCFDR p-value=0.0 for both), with overrepresentation of histone marks H3K20me1 and H3K36me3 in GM12878. Hypomethylated regions mapped to MUM1/IRF4 and to several epigenetic modifiers such as CTCF, highlighting a functionally important, specific set of hypomethylated genes in t(4;14). Progression to PCL was accompanied by significant preferential hypermethylation of regulatory enhancers and promoter regions and of the “polycomb group repressed” chromatin state both in t(4;14) and t(11;14) groups. Hypomethylation occurred in regions of “heterochromatin” state in t(4;14) MM to PCL progression and significantly more than expected in repetitive regions accompanying t(11;14) progression. Comparing t(4;14) and t(11;14) PCL demonstrated significant t(4;14) hypermethylation of regulatory regions (enhancers, promoters) and significant relative preferential t(4;14) hypomethylation of the “transcriptional elongation” chromatin state, indicating some maintained functional epigenetic programming by MMSET in t(4;14) PCL. Conclusion We have characterized the genome wide epigenetic programming of DNA methylation in MMSET overexpressing t(4;14) myeloma and PCL identifying key regulatory pathways involved in t(4;14) biology. We have also identified focal regulatory sequences, enhancers and promoters, which are affected specifically in the t(4;14) which may be linked to the biology of the this disease subtype. Disclosures: No relevant conflicts of interest to declare.

Description: Key Points Epigenetic inactivation of tumor suppressor genes is associated with an unfavorable prognosis in multiple myeloma. Drug response and microenvironment interaction pathways are affected by epigenetic inactivation, linking tumor biology to prognosis.

Description: Introduction Epigenetic dysregulation is a hallmark of cancer and has significant impact on disease biology. The epigenetic structure of myeloma is heterogeneous and we previously demonstrated that gene specific DNA methylation changes are associated with outcome, using low-resolution arrays. We now performed a high-resolution genome wide DNA methylation analysis of a larger group of patients from a UK national phase III study to further define the role of epigenetic modifications in disease behaviour and outcome. Patients and Methods Highly purified (〉95%) CD138+ myeloma bone marrow cells from 465 newly diagnosed patients enrolled in the UK NCRI Myeloma XI study were analysed. The extracted DNA was bisulfite-converted using the EZ DNA methylation kit (Zymo) and hybridized to Infinium HumanMethylation450 BeadChip arrays. Raw data was processed using the R Bioconductor package "minfi". SNP containing probes and probes on the sex chromosomes were removed. 464 samples and 441293 probes were retained following inspection of quality control metrics. Beta values were summarized across functional genomic units or differentially methylated regions (DMRs) that included: gene bodies, promoters, insulators, CpG-islands and enhancers. K-means was applied to each DMR to cluster patients into 2 groups (high or low methylation) per region. Filters were applied to define a clinically meaningful minimum group size and methylation differences between the groups. Overall survival (OS) and progression free survival (PFS) were assessed by a Cox proportional hazards regression model fitted to each DMR with a time-dependent covariate of the trial pathway. Pathway analyses were performed using GREAT (Stanford University) and GSEA (Broad Institute). Results We identified 589 differentially methylated regions that were significantly associated with PFS and OS when using a cut-off of P

Description: Background: The main genetic features of myeloma identified so far have been the presence of balanced translocations at the immunoglobulin heavy chain (IGH) region and copy number abnormalities. Novel methodologies such as massively parallel sequencing have begun to describe the pattern of tumour acquired mutations detected at presentation but their biological and clinical relevance has not yet been fully established. Methods: Whole exome sequencing was performed on 463 presentation patients enrolled into the large UK, phase III, open label, Myeloma XI trial. DNA was extracted from germline DNA and CD138+ plasma cells and whole exome sequencing was performed using SureSelect (Agilent). In addition to capturing the exome, extra baits were added covering the IGH, IGK, IGL and MYC loci in order to determine the breakpoints associated with translocations in these genes. Tumour and germline DNA were sequenced to a median of 60x and data processed to generate copy number, acquired variants and translocation breakpoints in the tumour. Progression-free and overall survival was measured from initial randomization and median follow up for this analysis was 25 months. These combined data allow us to examine the effect of translocations on the mutational spectra in myeloma and determine any associations with progression-free or overall survival. Results: We identified 15 significantly mutated genes comprising IRF4, KRAS, NRAS, MAX, HIST1H1E, RB1, EGR1, TP53, TRAF3, FAM46C, DIS3, BRAF, LTB, CYLD and FGFR3. By analysing the correlation between mutations and cytogenetic events using a probabilistic approach, we describe the co-segregation of t(11;14) with CCND1 mutations (Corr 0.28,BF=1.5x106 (Bayes Factor)) and t(4;14) with FGFR3 (Corr=0.40, BF=1.12x1014) and PRKD2 mutations (Corr=0.23, BF=3507). The mutational spectrum is dominated by mutations in the RAS (43%) and NF-κB (17%) pathway, however they are prognostically neutral. We describe for the first time in myeloma mutations in genes such as CCND1 and DNA repair pathway alterations (TP53, ATM, ATR and ZFHX4 mutations) that are associated with a negative impact on survival in contrast to those in IRF4 and EGR1 that are associated with a favourable overall-survival. By combining these novel risk factors with the previously described adverse cytogenetic features and ISS we were able to demonstrate in a multivariate analysis the independent prognostic relevance of copy number and structural abnormalities (CNSA) such as del(17p), t(4;14), amp(1q), hyperdiploidy and MYC translocations and mutations in genes such as ATM/ATR, ZFHX4, TP53 and CCND1. We demonstrate that the more adverse features a patient had the worse his outcome was for both PFS (one lesion: HR=1.6, p=0.0012, 2 lesions HR=3.3, p

Description: Introduction High expression of the H3K27 histone methyltransferase EZH2 mRNA in myeloma (MM) patient samples is associated with molecular features of high risk disease, including increased proliferation, and adverse outcomes (1). Mutations or deletions in the H3K27 demethylase KDM6A are associated with similar findings (2) and would be expected to have the same epigenetic effect, increasing H3K27me3 levels, a mark associated with repression of gene expression. We, therefore, sought to identify the role EZH2 plays in controlling myeloma cell proliferation. Methods A panel of MM cell lines and primary patient samples (CD138 selected from bone marrow with consent) representing a variety of different MM molecular subgroups were used. Cell viability (WST-1), cell cycle (PI) and apoptosis (AnnexinV/PI, Caspase-Glo 3/7) assays were performed. Affymetrix gene expression arrays followed by validation with RT-PCR were used to identify patterns of gene expression change with EZH2i. Western blotting confirmed changes at the protein level and Chip-PCR was performed using a validated antibody and isotype control to identify H3K27me3 changes at the relevant gene promotors. Affymetrix gene expression data for 1213 patients enrolled in the Total Therapy studies were used to investigate the relevance of our findings in myeloma patient samples. Results We confirmed a reduction in viability following EZH2i using two chemically distinct, specific small molecule inhibitors (EPZ005687 and UNC1999) and the negative control compound UNC2400. There was a reduction in viability in 6/8 cell lines and 5/6 patient samples. Response to inhibition was not related to molecular subgroup or the presence of high-risk molecular features including del17p. Global levels of H3K27me3 measured by Western blot were reduced in all cell lines regardless of response to EZH2i. In responding cell lines EZH2i induced cell cycle arrest at G1/S followed by induction of apoptosis. Gene expression arrays performed using mRNA from KMS11 and KMM1 cell lines highlighted a change in expression of cell cycle control genes associated with EZH2i. This finding was validated using qRT-PCR, which demonstrated upregulation of the cyclin dependent kinase inhibitors CDKN2B, CDKN1A or both. These findings were confirmed at the protein level by Western blotting. Chip-PCR experiment using cell lysates from KMS11 cells following incubation with EZH2i over 6 days identified changes in H3K27me3 at the promoter and transcriptional start site (PROM/TSS) regions of the CDKN2B and CDKN1A genes. The most specific changes occurred at the CDKN1A PROM/TSS, which were more heavily marked with H3K27me3 at baseline compared to a region approx. 5KB upstream. Given these results, which suggest that CDKN1A expression may be controlled by changes in H3K27me3, we explored the effect of CDKN1A mRNA expression in our patient datasets. We found the expression of EZH2 and CDKN1A to be inversely correlated (R=-0.170, p

Description: Digital image analysis and multivariate data analysis were used in this study to identify a set of leaf and fruit morphometric traits to discriminate white mulberry (Morus alba L.) cultivars. The trial was conducted using three- to five-year-old potted cuttings of several white mulberry cultivars. 32 leaf morphometric descriptors were recorded in 2011 and 2012 from 11 mulberry cultivars using image analysis of scanned leaves, whereas six fruit descriptors were recorded in 2011 from nine mulberry cultivars. Linear discriminant analysis (LDA) was used to identify a subset of measured variables that could discriminate the cultivars in trial. Biplot analysis, followed by cluster analysis, was performed on the discriminant variables to investigate any possible cultivar grouping based on similar morphometric traits. LDA was able to discriminate the 11 cultivars with a canonical function, which included 13 leaf descriptors. Using those 13 descriptors, the Biplot showed that over 84% of the variability could be explained by the first three factors. Clustering of standardized biplot coordinates recognized three groups: the first including ‘Korinne’ and ‘Miura’ with similar leaf angles and apical tooth size; the second including ‘Cattaneo’, ‘Florio’, ‘Kokusò-21’, ‘Kokusò-27’, and ‘Kokusò Rosso’ with similar leaf size and shape; and the third including ‘Ichinose’, ‘Kayrio’, ‘Morettiana’, and ‘Restelli’, with similar leaf margin. Fruit descriptors were fewer and measured on fewer cultivars, yielding smaller discriminatory power than leaf descriptors. Use of leaf morphometric descriptors, along with image and multivariate analysis, proved to be effective for discriminating mulberry cultivars and showed promise for the implementation of a simple and inexpensive characterization and classification tool.