Search

Understanding the molecular and phenotypic heterogeneity of cancer is a prerequisite for effective treatment. For chronic lymphocytic leukemia (CLL), recurrent genetic driver events have been extensively cataloged, but this does not suffice to explain the disease’s diverse course. Here, we performed RNA-sequencing on 184 CLL patient samples. Unsupervised analysis revealed two major, orthogonal axes of gene expression variation: the first one represented the mutational status of the immunoglobulin heavy variable (IGHV) genes, and concomitantly, the three-group stratification of CLL by global DNA methylation. The second axis aligned with trisomy 12 status and affected chemokine, MAPK and mTOR signaling. We discovered nonadditive effects (epistasis) of IGHV mutation status and trisomy 12 on multiple phenotypes, including the expression of 893 genes. Multiple types of epistasis were observed, including synergy, buffering, suppression and inversion, suggesting that molecular understanding of disease heterogeneity requires studying such genetic events not only individually but in combination. We detected strong differentially expressed gene signatures associated with major gene mutations and copy-number aberrations including SF3B1, BRAF and TP53, as well as del(17)(p13), del(13)(q14) and del(11)(q22.3) beyond dosage effect. Our study reveals previously underappreciated gene expression signatures for the major molecular subtypes in CLL and the presence of epistasis between them.

Despite several potent targeted treatments for chronic lymphocytic leukemia (CLL), the clinical challenge of treating drug-resistant disease is emerging. In this study, we discovered that the dual-specific phosphatases DUSP1 and DUSP6 are required to negatively regulate Mitogen-activated protein kinases (MAPK) and thus counterbalance excessive MAPK activity to prevent apoptosis in CLL. We show that DUSP1 and DUSP6 are widely expressed in CLL and high expression of DUSP6 in CLL correlates with a poor clinical prognosis, which may reflect high levels of MAPK activity. Importantly, genetic deletion of the inhibitory phosphatase DUSP1 or DUSP6 and blocking DUSP1/6 function using a small molecule are toxic for CLL cells in vitro and in vivo. Analyzing downstream effects using global phospho-proteome approaches, we observed that acute activation of MAPK signaling by DUSP1/6 inhibition induces DNA damage response and thereby apoptotic cell death in CLL cells. This cell death is mediated by CHK kinases and can function independent of p53 and ATM, both effectors of DNA damage response, which are frequently deleted in CLL. Finally, we observed that DUSP1/6 inhibition is particularly effective against treatment-resistant CLL and therefore suggest transient DUSP1/6 inhibition as a promising novel treatment concept to eliminate drug-resistant CLL cells.

Drug combinations that target critical pathways are a mainstay of cancer care. To improve current approaches to combination treatment of chronic lymphocytic leukemia (CLL) and gain insights into the underlying biology, we studied the effect of 352 drug combination pairs in multiple concentrations by analysing ex vivo drug response of 52 primary CLL samples, which were characterized by “omics” profiling. Known synergistic interactions were confirmed for B-cell receptor (BCR) inhibitors with Bcl-2 inhibitors and with chemotherapeutic drugs, suggesting that this approach can identify clinically useful combinations. Moreover, we uncovered synergistic interactions between BCR inhibitors and afatinib, which we attribute to BCR activation by afatinib through BLK upstream of BTK and PI3K. Combinations of multiple inhibitors of BCR components (e.g., BTK, PI3K, SYK) had effects similar to the single agents. While PI3K and BTK inhibitors produced overall similar effects in combinations with other drugs, we uncovered a larger response heterogeneity of combinations including PI3K inhibitors, predominantly in CLL with mutated IGHV, which we attribute to the target’s position within the BCR-signaling pathway. Taken together, our study shows that drug combination effects can be effectively queried in primary cancer cells, which could aid discovery, triage and clinical development of drug combinations.

Several gene expression profiles with a strong correlation with patient outcomes have been previously described in chronic lymphocytic leukemia (CLL), although their applicability as biomarkers in clinical practice has been particularly limited. Here we describe the training and validation of a gene expression signature for predicting early progression in patients with CLL based on the analysis of 200 genes related to microenvironment signaling on the NanoString platform. In the training cohort (n = 154), the CLL15 assay containing a 15-gene signature was associated with the time to first treatment (TtFT) (hazard ratio [HR], 2.83; 95% CI, 2.17-3.68; P < .001). The prognostic value of the CLL15 score (HR, 1.71; 95% CI, 1.15-2.52; P = .007) was further confirmed in an external independent validation cohort (n = 112). Notably, the CLL15 score improved the prognostic capacity over IGHV mutational status and the International Prognostic Score for asymptomatic early-stage (IPS-E) CLL. In multivariate analysis, the CLL15 score (HR, 1.83; 95% CI, 1.32-2.56; P < .001) and the IPS-E CLL (HR, 2.23; 95% CI, 1.59-3.12; P < .001) were independently associated with TtFT. The newly developed and validated CLL15 assay successfully translated previous gene signatures such as the microenvironment signaling into a new gene expression–based assay with prognostic implications in CLL., This work was supported by research funding from the Asociacion Espa ´ nola Contra el C ˜ ancer grant [5U01CA157581- 05, ECRIN-M3 - A29370] and in part by the Instituto de Salud Carlos III, Fondo de Investigaciones Sanitarias [PI17/00950, M.C., PI17/00943, F.B, PI18/01392, P.A.], and the Spanish Ministry of Economy and Competitiveness [CIBERONC-CB16/12/00233], the Education Council or Health Council of the Junta de Castilla y Leon [GRS 2036/A/19], and Gilead Sciences [GLD15/00348]. ´ This work was supported by research funding from the Asociacion´ Espanola Contra el C ˜ ancer grant [5U01CA157581-05, ECRIN-M3 ´ - A29370] and in part by the Instituto de Salud Carlos III, Fondo de Investigaciones Sanitarias [PI17/00950, M.C., PI17/00943, F.B, PI18/01392, P.A.], and the Spanish Ministry of Economy and Competitiveness [CIBERONC-CB16/12/00233], the Education Council or Health Council of the Junta de Castilla y Leon [GRS ´ 2036/A/19], Gilead Sciences [GLD15/00348] and Gilead Fellowships [GLD16/00144, GLD18/00047, F.B.], and Fundacio la ´Marato de TV3 [201905-30-31 F.B]. All Spanish funding was ´ cosponsored by the European Union FEDER program “Una manera de hacer Europa”. M.C. holds a contract from Ministerio de Ciencia, Innovacion y Universidades [RYC-2012-2018].

The genomic landscape of mantle cell lymphoma (MCL) includes frequent alterations of TP53, ATM, CCND1 and KMT2D. Thus far, the mutational landscape provides little information for treatment stratification. We characterized a cohort of MCL by targeted next generation sequencing and discovered SAMHD1 as a novel recurrently mutated gene (8.5% of investigated cases, 4/47 samples). Furthermore, we provide evidence of in vitro resistance of SAMHD1 mutated patient-derived MCL cells to cytarabine and fludarabine.

Alterations in global DNA methylation patterns are a major hallmark of cancer and represent attractive biomarkers for personalized risk stratification. Chronic lymphocytic leukemia (CLL) risk stratification studies typically focus on time to first treatment (TTFT), time to progression (TTP) after treatment, and overall survival (OS). Whereas TTFT risk stratification remains similar over time, TTP and OS have changed dramatically with the introduction of targeted therapies, such as the Bruton tyrosine kinase inhibitor ibrutinib. We have shown that genome-wide DNA methylation patterns in CLL are strongly associated with phenotypic differentiation and patient outcomes. Here, we developed a novel assay, termed methylation-iPLEX (Me-iPLEX), for high-throughput quantification of targeted panels of single cytosine guanine dinucleotides from multiple independent loci. Me-iPLEX was used to classify CLL samples into 1 of 3 known epigenetic subtypes (epitypes). We examined the impact of epitype in 1286 CLL patients from 4 independent cohorts representing a comprehensive view of CLL disease course and therapies. We found that epitype significantly predicted TTFT and OS among newly diagnosed CLL patients. Additionally, epitype predicted TTP and OS with 2 common CLL therapies: chemoimmunotherapy and ibrutinib. Epitype retained significance after stratifying by biologically related biomarkers, immunoglobulin heavy chain mutational status, and ZAP70 expression, as well as other common prognostic markers. Furthermore, among several biological traits enriched between epitypes, we found highly biased immunogenetic features, including IGLV3-21 usage in the poorly characterized intermediate-programmed CLL epitype. In summary, Me-iPLEX is an elegant method to assess epigenetic signatures, including robust classification of CLL epitypes that independently stratify patient risk at diagnosis and time of treatment.

Aberrant CXCR4 activity has been implicated in lymphoma pathogenesis, disease progression, and resistance to therapies. Using a mouse model with a gain-of-function CXCR4 mutation (CXCR4C1013G) that hyperactivates CXCR4 signaling, we identified CXCR4 as a crucial activator of multiple key oncogenic pathways. CXCR4 hyperactivation resulted in an expansion of transitional B1 lymphocytes, which represent the precursors of chronic lymphocytic leukemia (CLL). Indeed, CXCR4 hyperactivation led to a significant acceleration of disease onset and a more aggressive phenotype in the murine Eµ-TCL1 CLL model. Hyperactivated CXCR4 signaling cooperated with TCL1 to cause a distinct oncogenic transcriptional program in B cells, characterized by PLK1/FOXM1-associated pathways. In accordance, Eµ-TCL1;CXCR4C1013G B cells enriched a transcriptional signature from patients with Richter’s syndrome, an aggressive transformation of CLL. Notably, MYC activation in aggressive lymphoma was associated with increased CXCR4 expression. In line with this finding, additional hyperactive CXCR4 signaling in the Eµ-Myc mouse, a model of aggressive B-cell cancer, did not impact survival. In summary, we here identify CXCR4 hyperactivation as a co-driver of an aggressive lymphoma phenotype.

Many functional consequences of mutations on tumor phenotypes in chronic lymphocytic leukemia (CLL) are unknown. This may be in part due to a scarcity of information on the proteome of CLL. We profiled the proteome of 117 CLL patient samples with data-independent acquisition mass spectrometry and integrated the results with genomic, transcriptomic, ex vivo drug response, and clinical outcome data. We found trisomy 12, IGHV mutational status, mutated SF3B1, trisomy 19, del(17)(p13), del(11)(q22.3), mutated DDX3X and MED12 to influence protein expression (false discovery rate [FDR] = 5%). Trisomy 12 and IGHV status were the major determinants of protein expression variation in CLL as shown by principal-component analysis (1055 and 542 differentially expressed proteins, FDR = 5%). Gene set enrichment analyses of CLL with trisomy 12 implicated B-cell receptor (BCR)/phosphatidylinositol 3-kinase (PI3K)/AKT signaling as a tumor driver. These findings were supported by analyses of protein abundance buffering and protein complex formation, which identified limited protein abundance buffering and an upregulated protein complex involved in BCR, AKT, MAPK, and PI3K signaling in trisomy 12 CLL. A survey of proteins associated with trisomy 12/IGHV-independent drug response linked STAT2 protein expression with response to kinase inhibitors, including Bruton tyrosine kinase and mitogen-activated protein kinase kinase (MEK) inhibitors. STAT2 was upregulated in unmutated IGHV CLL and trisomy 12 CLL and required for chemokine/cytokine signaling (interferon response). This study highlights the importance of protein abundance data as a nonredundant layer of information in tumor biology and provides a protein expression reference map for CLL.

Epigenetic changes during B cell differentiation generates distinct DNA methylation signatures specific for B cell subsets, including memory B cells (MBCs) and plasma cells (PCs). Waldenström's macroglobulinemia (WM) is a complex B cell malignancy uniquely comprised of a mixture of lymphocytic and plasmacytic phenotypes. Here we integrated genome-wide DNA methylation, transcriptome, mutation and other phenotypic features of tumor cells from 35 MYD88-mutated WM patients in relation to normal plasma and B cell subsets. We discovered that WM patients naturally segregate into two groups according to DNA methylation patterns, related to normal MBC and PC profiles, and reminiscent of other memory and plasma cell-derived malignancies. Concurrent analysis of DNA methylation changes in normal and WM development were used to capture tumor-specific events, highlighting a selective reprogramming of enhancer regions in MBC-like WM and repressed and heterochromatic regions in PC-like WM. MBC-like WM hypomethylation was enriched in motifs belonging to PU.1, TCF3 and OCT2 transcription factors and involved elevated MYD88/TLR pathway activity. PC-like WM displayed marked global hypomethylation and selective overexpression of histone genes. Finally, WM subtypes exhibited differential genetic, phenotypic and clinical features. MBC-like WM harbored significantly more clonal CXCR4 mutations (P=0.015), deletion 13q (P=0.006), splenomegaly (P=0.02) and thrombocytopenia (P=0.004), while PC-like WM harbored more deletion 6q (P=0.012), gain 6p (P=0.033), had increased frequencies of IGHV3 genes (P=0.002), CD38 surface expression (P=4.1e-5), and plasmacytic differentiation features (P=0.008). Together our findings illustrate a novel approach to subclassify WM patients using patterns of DNA methylation and reveal divergent molecular signatures among WM patients.

We report a systematic analysis of the biological and clinical implications of DNA methylation variability in five categories of B-cell tumors derived from B cells spanning the entire maturation spectrum. We used 2056 primary samples including training and validation series and show that 88% of the human DNA methylome is dynamically modulated under normal and neoplastic conditions. B-cell tumors display both epigenetic imprints of their cellular origin and de novo, disease-specific epigenetic alterations that in part are related to differential transcription factor binding. These differential methylation patterns were used by a machine-learning approach to create a diagnostic algorithm that accurately classifies 14 B-cell tumor entities and subtypes with different clinical management. Beyond this, we identified extensive patient-specific epigenetic variability targeting constitutively silenced chromatin regions, a phenomenon we could relate to the proliferative history of normal and neoplastic B cells. We observed that, depending on the maturation stage of the tumor cell of origin, mitotic activity leaves different imprints into the DNA methylome. Subsequently, we constructed a novel DNA methylation-based mitotic clock called epiCMIT (epigenetically-determined Cumulative MIToses), whose lapse magnitude represents a strong independent prognostic variable within specific B-cell tumor subtypes and is associated with particular driver genetic alterations. Our findings reveal DNA methylation as a holistic tracker of B-cell tumor developmental history, with implications in the differential diagnosis and prediction of the outcome of the patients.

Tumors accumulate high levels of mutant p53 (mutp53), which contributes to mutp53 gain-of-function properties. The mechanisms that underlie such excessive accumulation are not fully understood. To discover regulators of mutp53 protein accumulation, we performed a large-scale RNA interference screen in a Burkitt lymphoma cell line model. We identified transformation/transcription domain-associated protein (TRRAP), a constituent of several histone acetyltransferase complexes, as a critical positive regulator of both mutp53 and wild-type p53 levels. TRRAP silencing attenuated p53 accumulation in lymphoma and colon cancer models, whereas TRRAP overexpression increased mutp53 levels, suggesting a role for TRRAP across cancer entities and p53 mutations. Through clustered regularly interspaced short palindromic repeats (CRISPR)-Cas9 screening, we identified a 109-amino-acid region in the N-terminal HEAT repeat region of TRRAP that was crucial for mutp53 stabilization and cell proliferation. Mass spectrometric analysis of the mutp53 interactome indicated that TRRAP silencing caused degradation of mutp53 via the MDM2-proteasome axis. This suggests that TRRAP is vital for maintaining mutp53 levels by shielding it against the natural p53 degradation machinery. To identify drugs that alleviated p53 accumulation similarly to TRRAP silencing, we performed a small-molecule drug screen and found that inhibition of histone deacetylases (HDACs), specifically HDAC1/2/3, decreased p53 levels to a comparable extent. In summary, here we identify TRRAP as a key regulator of p53 levels and link acetylation-modifying complexes to p53 protein stability. Our findings may provide clues for therapeutic targeting of mutp53 in lymphoma and other cancers.

Background: Chronic lymphocytic leukemia (CLL) is the most common adult leukemia in the western world and shows a very heterogeneous clinical course. While the genetic landscape of CLL has been well characterized during recent years it can only partially explain the underlying biology of this heterogeneity. Proteogenomics could offer a valuable tool to fill this gap and improve the understanding of CLL biology. Methods: Here, we performed a large proteogenomic analysis (n=263) of three clinically annotated CLL cohorts: For the discovery cohort (Germany_1: n=68) we performed in-depth HiRIEF LC-MS based proteomics (more than 9000 proteins quantified) alongside genome-, transcriptome and ex-vivo drug response-profiling with 43 clinically established drugs. The proteome of two additional validation cohorts (Germany_2: n=44, Sweden_1: n=89), were characterized by data-independent acquisition (DIA) mass spectrometry. Results: To connect the CLL genotype with the molecular phenotype, we investigated associations between recurrent genetic alterations of CLL, mRNA expression and protein abundance. We found that trisomy 12, IGHV status and SF3B1 mutations had the greatest impact on protein abundances. CLL specific recurrent chromosomal deletions and gains (trisomy 12, del17p, del13q, del11q, gain8q24) consistently impacted on gene expression and protein abundance through gene dosage effects. We explored functional consequences of these gene dosage effects and found that the additional copy of chromosome 12 increased the abundance of central B-cell receptor (BCR) protein complexes through cis- and trans-effects, which could explain the increased response of trisomy 12 patient samples to BCR inhibition. Somatic mutations of TP53, ATM and XPO1 were associated with less, but specific and biologically-relevant protein abundance changes. p53 for instance, was the most upregulated protein in TP53 mutated samples, owing to the known stabilisation of mutant p53. This effect was not detectable on transcript level. ATM and XPO1 protein abundances were significantly lower in ATM and XPO1 mutated cases, indicating loss-of-function phenotypes of these mutations. To understand global similarities and differences between CLL patients on the proteomic level, we performed unsupervised clustering and identified clinically meaningful subgroups. Unsupervised clustering of the proteomics data identified six subgroups with contrasting clinical behaviour. TP53 mutations, IGHV status, trisomy 12 and their interactions explained five subgroups. These results show that quantitative mass spectrometry-based proteomics distinguished clinically relevant subgroups of CLL. Most importantly, we identified a previously unappreciated subgroup of CLL, comprising 20% of all cases, which could be uncovered by proteomic profiling and showed no association with frequent genetic or transcriptional alterations. This new CLL subgroup was characterized by accelerated disease progression, SF3B1 mutation-independent splicing alterations, metabolomic reprogramming and increased vulnerability to inhibitors of metabolic enzymes and the proteasome. Surprisingly, major BCR signaling proteins were downregulated in this subgroup, suggesting less dependence on BCR activity. In accordance with this observation, an unsupervised analysis revealed that low levels of many BCR signaling proteins (e.g. PLCG2 and PIK3CD) were associated with short time to next treatment. The existence of this subgroup could be confirmed in the validation cohorts. Finally, we performed an unsupervised multi-omics factor analysis (MOFA) across all omics data sets in parallel. This unsupervised analysis confirmed the existence of the above identified CLL subgroups and an important role of SF3B1 mutation-independent splicing alterations in CLL. Conclusion: Our integrative multi-omics analysis provides the first comprehensive overview of the interplay between genetic variants, the transcriptome, and the proteome, along with functional consequences for drug response and clinical outcome in CLL. Importantly, we identified a new subgroup with accelerated disease progression, a distinct proteomic signature and a clinically exploitable drug sensitivity profile. Figure Disclosures Mueller-Tidow: BiolineRx: Research Funding; Daiichi Sankyo: Research Funding; Pfizer: Membership on an entity's Board of Directors or advisory committees, Research Funding; BMBF: Research Funding; Wilhelm-Sander-Stiftung: Research Funding; Jose-Carreras-Siftung: Research Funding; Bayer AG: Research Funding; Deutsche Krebshilfe: Research Funding; Deutsche Forschungsgemeinschaft: Research Funding; Janssen-Cilag Gmbh: Membership on an entity's Board of Directors or advisory committees. Dreger:Neovii: Research Funding; Roche: Consultancy, Speakers Bureau; Riemser: Consultancy, Research Funding, Speakers Bureau; Novartis: Consultancy, Speakers Bureau; Janssen: Consultancy; Gilead: Consultancy, Speakers Bureau; AstraZeneca: Consultancy; AbbVie: Consultancy, Speakers Bureau. Stilgenbauer:Pharmacyclics: Consultancy, Honoraria, Other, Research Funding; Novartis: Consultancy, Honoraria, Other, Research Funding; Mundipharma: Consultancy, Honoraria, Other, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Other: travel support, Research Funding; GlaxoSmithKline: Consultancy, Honoraria, Other: travel support, Research Funding; Gilead: Consultancy, Honoraria, Other: travel support, Research Funding; Genzyme: Consultancy, Honoraria, Other: travel support, Research Funding; Genentech: Consultancy, Honoraria, Other: travel support, Research Funding; F. Hoffmann-LaRoche: Consultancy, Honoraria, Other: travel support, Research Funding; Celgene: Consultancy, Honoraria, Other: travel support, Research Funding; Boehringer-Ingelheim: Consultancy, Honoraria, Other: travel support, Research Funding; Amgen: Consultancy, Honoraria, Other: travel support, Research Funding; AbbVie: Consultancy, Honoraria, Other: travel support, Research Funding. Tausch:Roche: Consultancy, Honoraria, Research Funding; AbbVie: Consultancy, Honoraria, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Research Funding. Dietrich:Roche: Membership on an entity's Board of Directors or advisory committees; Janssen: Membership on an entity's Board of Directors or advisory committees, Research Funding; Celgene: Membership on an entity's Board of Directors or advisory committees; KITE: Membership on an entity's Board of Directors or advisory committees.

Introduction: B-cell receptor (BCR) signaling is crucial for chronic lymphocytic leukemia (CLL) biology. IGLV3-21-expressing B-cells may acquire a single point mutation (R110) that triggers autonomous BCR signaling conferring aggressive behavior. Epigenetic studies have defined three CLL subtypes based on methylation signatures reminiscent of pre- and post-germinal center B-cells named naïve-like (n-CLL), intermediate (i-CLL) and memory-like CLL (m-CLL) with different biological features. i-CLL carry a borderline IGHV mutational load and a significant higher usage of IGHV3-21/IGLV3-21. The integration of these factors might translate into novel insights in CLL pathogenesis with implications on the proposed stratification of the patients. Aim: To determine the clinical and biological features of the IGLV3-21R110 in CLL in the light of the epigenetic subtypes and immunogenetic, genomic and transcriptomic landscapes of the tumors. Methods: We characterized the immunoglobulin (IG) gene of 584 CLL cases from whole-genome/exome and RNA sequencing using our recently developed algorithm IgCaller (Nadeu et al., Nat. Commun. 2020) and MiXCR, respectively. The genomic makeup of the tumors was obtained from whole-genome/exome sequencing while RNA sequencing data for 369 cases was used for gene expression analyses. Expression levels of WNT5A and WNT5B were verified by quantitative PCR with reverse transcriptase. Primary end points were time to first treatment (TTFT) and overall survival (OS) calculated from the date of diagnosis. All patients gave written informed consent. The study was approved by the Ethics Committee of the Hospital Clínic of Barcelona. Results: The IGLV3-21R110 was detected in 6.5% of cases being similarly distributed between mutated (6.5%) and unmutated (6.6%) IGHV cases (P=0.56). In contrast, the IGLV3-21R110 was found in 30/79 (38%) i-CLL compared to only 5/291 (1.7%) m-CLL and 1/189 (0.5%) n-CLL (P Although composite regarding IGHV mutational status, IGLV3-21R110 i-CLL transcriptomically resembled naïve-like/unmutated IGHV CLL and had a specific expression signature of 64 genes with overexpression of WNT5A and WNT5B as hallmarks. No differences were observed in the expression profile of subset #2 and non-subset #2 IGLV3-21R110 i-CLL tumors. On the other hand, i-CLL lacking the IGLV3-21R110 phenotypically mirrored memory-like/mutated IGHV cases. In relation to prognosis, IGLV3-21R110 i-CLL had a short TTFT and OS similar to n-CLL/unmutated IGHV cases whereas non-IGLV3-21R110 i-CLL had a good prognosis similar to memory-like/mutated IGHV. Therefore, i-CLL cases, which have been associated with an intermediate prognosis between m-CLL and n-CLL in previous studies, can be divided in two subgroups of cases with opposed clinical evolutions based on the IGLV3-21R110. Indeed, the IGLV3-21R110 and n-CLL subtype retained independent prognostic value in multivariate analyses while the i-CLL lost its prognostic prediction both for TTFT and OS. The prognostic value of the IGLV3-21R110 was also independent of the IGHV mutational status. In terms of applicability in the clinics, all n-CLL cases were classified as unmutated IGHV and 98% of m-CLL were mutated IGHV. Thus, either a complete IG characterization (IGHV mutational status and IGLV3-21R110) or the integration of the n-CLL subtype and IGLV3-21R110 identified virtually the same subset of patients with aggressive disease. Conclusions: The IGLV3-21R110 defines a CLL subset with borderline IGHV mutations, specific driver alterations, a gene expression signature including WNT5A/B overexpression, and an unfavorable prognosis independent of the IGHV mutational status and epigenetic subtypes. Our findings support the identification of IGLV3-21R110 CLL as a particular subgroup of the disease with relevance in the risk stratification of the patients. Disclosures Nadeu: Janssen: Honoraria. Campo:NIH: Consultancy, Other: Co-inventor on a patent related to the MCL35 assay filed at the National Institutes of Health, United States of America..

Chronic lymphocytic leukemia cells have an altered energy metabolism compared to normal B cells. While there is a growing understanding of the molecular heterogeneity of the disease, the extent of metabolic heterogeneity and its relation to molecular heterogeneity has not been systematically studied. Here, we assessed 11 bioenergetic features, primarily reflecting cell oxidative phosphorylation and glycolytic activity, in leukemic cells from 140 chronic lymphocytic leukemia patients using metabolic flux analysis. We examined these bioenergetic features for relationships with molecular profiles (including genetic aberrations, transcriptome and methylome profiles) of the tumors, their ex vivo responses to a panel of 63 compounds, and with clinical data. We observed that leukemic cells with mutated immunoglobulin variable heavy-chain show significantly lower glycolytic activity than cells with unmutated immunoglobulin variable heavy-chain. Accordingly, several key glycolytic genes (PFKP, PGAM1 and PGK1) were found to be down-regulated in samples harboring mutated immunoglobulin variable heavy-chain. In addition, 8q24 copy number gains, 8p12 deletions, 13q14 deletions and ATM mutations were identified as determinants of cellular respiration. The metabolic state of leukemic cells was associated with drug sensitivity; in particular, higher glycolytic activity was linked to increased resistance towards several drugs including rotenone, navitoclax, and orlistat. In addition, we found glycolytic capacity and glycolytic reserve to be predictors of overall survival (P

Introduction: Genomic studies of chronic lymphocytic leukemia (CLL) have uncovered >80 potential driver mutations. The vast majority of these mutations affect coding regions, and just two potential drivers have been identified in non-coding elements. Aim: To describe the biological and clinical impact of a recurrent A>C mutation at the third base of the small nuclear RNA U1, the non-coding component of the spliceosome involved in the recognition of the 5' splice site (5'SS). Methods: Whole-genome sequencing (WGS) and RNA-seq from 318 CLL patients were used to identify and characterize a highly recurrent A>C point mutation occurring at position 3 of the U1 snRNA gene (g.3A>C mutation). The U1 wild-type and mutant forms were introduced into three CLL cell lines (JVM3, HG3, MEC1) to validate in vitro the predicted effect of this alteration. We screened two independent cohorts including a total of 1,314 CLL patients for the presence of the mutation using the rhAmp SNP genotyping assay, and integrated the U1 mutational status with well-known driver alterations, IGHV and epigenetic subgroups, and clinical parameters. Results: The U1 mutation was found in 8/78 (10.3%) CLL cases analyzed by WGS. Given its role in 5'SS recognition by base-pairing, we reasoned that this mutation was likely to alter the splicing and expression patterns of CLL. We were able to confirm widespread specific alterations in the transcriptome by comparing RNA-seq data between wild-type and g.3A>C mutated samples. Applying this knowledge to an algorithm aimed to infer the U1 mutational status from expression data, we were able to identify 4 mutated cases among 240 additional cases that had RNA-seq but no WGS. In total, 12/318 (3.8%) CLL patients analyzed by WGS and/or RNA-seq harbored this mutation. This g.3A>C U1 mutation changes the preferential A-U base-pairing between U1 and 5'SS to C-G base-pairing, creating novel splice junctions and altering the splicing pattern of 3,193 introns in 1,519 genes. In addition to altered splicing, 869 genes were differentially expressed between mutated and wild-type cases. We identified specific cancer genes (e.g. MSI2, POLD1, or CD44) and pathways (B-cell receptor signaling, promotion of apoptosis, telomere maintenance, among others) altered by the U1 mutation. To confirm a causal link between this mutation and splicing changes, we introduced exogenous U1 genes with or without the mutation into three cell lines. Subsequent RNA-seq of these cell lines recapitulated the altered splicing and expression patterns observed in CLL patients. We next screened for the presence of the U1 mutation 1,057 patients (cohort 1) using the rhAmp assay and it was found in 30 (2.8%) cases. The distribution of the mutation was similar in Binet stages and CLL vs monoclonal B-cell lymphocytosis. However, the U1 mutation was almost always found in IGHV unmutated CLL (29/30, p=9.0e-11) and within the naïve-like CLL epigenetic subgroup (p=3.7e-7). None of the U1 mutated cases had mutations in the SF3B1 splicing factor. Considering only pre-treatment CLL samples, U1 mutation was associated with a shorter time to first treatment independently of the Binet stage, IGHV mutational status, epigenetic subgroups, and mutations in the well-known CLL drivers SF3B1, NOTCH1, ATMor TP53. In cohort 2 (n=257), this mutation was found in 13 (5.1%) patients, confirming its enrichment in IGHV unmutated cases, naïve-like epigenetic subgroup, and splicing modulation. Despite the relatively small number of pre-treatment samples carrying the U1 mutation (7/178) and short follow-up of the patients (median 2.6 years), the effect of this mutation on time to first treatment in cohort 2 was compatible with the one observed in cohort 1. Finally, we screened for the U1 mutation a cohort of diffuse large B-cell lymphoma (n=108), mantle cell lymphoma (n=101), follicular lymphoma (n=87), splenic marginal zone lymphoma (n=12), acute myeloid leukemia (n=52), and myelodysplastic syndrome (n=67). The mutation was not present in any of the samples analyzed. Conclusions: Here we have reported that the third base of the small nuclear RNA U1 is recurrently mutated in CLL, proved its effect in splicing and gene expression, and shown that this mutation is independently associated with faster disease progression. The g.3A>C U1 mutation represents a novel non-coding driver alteration in CLL with potential clinical and therapeutic implications. Disclosures Ramirez Payer: GILEAD SCIENCES: Research Funding. Terol:Astra Zeneca: Consultancy; Gilead: Research Funding; Abbvie: Consultancy; Janssen: Consultancy, Research Funding; Roche: Consultancy. Lopez-Guillermo:Celgene: Consultancy, Research Funding; Janssen: Research Funding; Roche: Consultancy, Research Funding; Gilead: Consultancy, Research Funding.

Introduction: Mature normal and tumor B cells express a unique rearranged immunoglobulin (IG) gene that can be used as a marker of the clonal expansion of the cell. Somatic hypermutation (SHM) in the V(D)J region of IG genes are acquired in the germinal center and are a surrogate imprint of the cell of origin of lymphoid neoplasms. In chronic lymphocytic leukemia (CLL) and mantle cell lymphoma (MCL), the identification of SHM distinguishes subtypes of tumors with different clinical and biological behavior. Although still less studied, cases carrying highly similar IG sequences (i.e. stereotyped IG), specific light chain (LC) rearrangements, and the presence of class switch recombination (CSR) of the constant region of the heavy chains further sub-classify patients into potentially distinct clinico-biological subgroups. The analysis of the rearranged IG gene is currently performed by specific Sanger sequencing (SSeq) or next-generation sequencing protocols. Whole-genome sequencing (WGS) of B-cell neoplasms should store the information to reconstruct the entire rearranged IG gene [heavy (IGH) and kappa or lambda (IGK, IGL) LC]. However, the high genomic complexity and homology of these regions have prevented the analysis of the rearranged IG genes in WGS using standard bioinformatics pipelines. Aim: To assess the use of WGS data to fully characterize the rearranged IG gene in B-cell neoplasms. Methods: We developed IgCaller, a fast, easy-to-run program that uses already aligned WGS data to dissect the rearranged IGH V(D)J genes, IGK and IGL VJ genes, and the presence of constant heavy chain CSR. IgCaller also determines the homology of the rearranged sequences compared to the patient's germ line or reference genome. We demonstrated the accuracy of IgCaller using WGS data of 331 B-cell neoplasms [240 CLL (152 cohort 1 (C1)-CLL; 88 independent cohort 2 (C2)-CLL), 61 MCL, and 30 multiple myeloma (MM)] and compared with SSeq of the IGH V(D)J and/or LC and isotype expression. Results: IgCaller identified a complete IGH productive rearrangement [V(D)J] in 133 (88%) C1-CLL, 80 (91%) C2-CLL, 61 (100%) MCL, and 21 (70%) MM. A partial (VJ) rearrangement was detected in 8 (5%) C1-CLL and 1 (3%) MM. Available SSeq of the V(D)J or at least V gene for 131 C1-CLL, 10 C2-CLL, and 60 MCL successfully characterized by WGS highlighted only one discordant V(D)J rearrangement. Small discrepancies (only J or V disagreement) were found when the J (n=4) or V (n=1) gene identified by SSeq based on homology (IMGT/V-QUEST tool) did not correspond to the rearranged gene detected by WGS, which was the second scoring gene in IMGT/V-QUEST suggesting that our non-homology WGS-based approach might be more accurate. Of note, IgCaller identified the presence of two distinct IGH subclones in 1 case. Next, the comparison of the percentage of homology of the rearranged sequence to the germ line in 131 C1-CLL and 60 MCL with complete V gene both by SSeq and WGS showed a high correlation and concordance in both cohorts [R>0.95, p A productive IGK or IGL was found in 147 (97%) C1-CLL, 85 (97%) C2-CLL, 61 (100%) MCL, and 26 (87%) MM. These results fully agreed with the LC expression observed by flow cytometry (FC). Of clinical relevance, a total of 26/232 (11%) CLL carried IGLV3-21. The presence of IGLV3-21 was associated with a shorter time to first treatment (TTFT) independently of the IGHV status (p=0.045). Finally, IgCaller identified the CSR matching the isotype expressed by FC analysis in 26/30 (87%) MM. One of the 4 discordant cases with IgM by WGS expressed only LC by FC. Overall, IgCaller could not identify the isotype switch in 3 (10%) MM expressing IgG or IgA. In CLL, CSR was observed in 47/240 (20%) cases. Noteworthy, the presence of CSR identified 24% of M-CLL patients with a shorter TTFT than non-switched M-CLL (p=0.004), and similar to that of unmutated IGHV cases (p=0.15). Conclusions: IgCaller successfully characterized the entire IG gene of >90% B-cell neoplasms studied. The complete characterization of the rearranged IG gene based on WGS data, when available, could facilitate the analysis of LC rearrangements and CSR, and replace the traditional SSeq of the IG loci both in research and clinical settings. Disclosures No relevant conflicts of interest to declare.

While recurrent mutations in CLL have been extensively catalogued, how driver mutations affect disease phenotypes remains incompletely understood. To address this, we performed RNA sequencing on 184 CLL patient samples and linked gene expression changes to molecular subgroups, gene mutations and copy number variants. Library preparation was performed according to the Illumina TruSeq RNA sample preparation v2 protocol. Samples were paired-end sequenced and two to three samples were multiplexed per lane on Illumina HiSeq 2000, Illumina HiSeq3000/4000 or Illumina HiSeqX machines. Raw RNA-seq reads were demultiplexed and quality control was performed using FastQC version 0.11.5. Internal trimming with STAR version 2.5.2a was used to remove adapters before mapping. Mapping was performed using STAR version 2.5.2a against the Ensembl human reference genome release 75 (Homo sapiens GRCh37.75). STAR was run in default mode with internal adapter trimming using the clip3pAdapterSeq option. Mapped reads were summarized into counts using htseq-count version 0.9.0 with default parameters and union mode. Thus, only fragments unambiguously overlapping with one gene were counted. The count data were then imported into R (version 3.4) for subsequent analysis. We identified robust and previously unknown gene expression signatures associated with recurrent copy number variants (including trisomy 12, del11q22.3, del17p13, del18p12 and gain8q24), gene mutations (TP53, BRAF and SF3B1) and the mutation status of the immunoglobulin heavy-chain variable region (IGHV). The most profound gene expression changes were associated with IGHV, methylation groups and trisomy 12. We found evidence for a significant influence of CNVs beyond the gene dosage effect. In line with these observations, unsupervised clustering showed that these major biological subgroups form distinct clusters and are discernible by unsupervised clustering (IGHV, methylation groups and trisomy 12). We found 3275 genes significantly differentially expressed between M-CLL and U-CLL after adjustment for multiple testing using the method of Benjamini and Hochberg for FDR = 1% . In total 9.5 % of variance within gene expression was associated with the IGHV status. These data suggest a much larger impact on transcriptional changes than previously detected (Ferreira et al. 2014), a finding much more in line with the key impact of IGHV on clinical course and biology of disease. We found distinct expression pattern of up- and downregulated genes for trisomy 12 samples. Even though many upregulated genes are located on chromosome 12, the majority of differentially expressed genes are indeed distributed among the other chromosomes and cannot be therefore not be ascribed to a simple gene dosage effect. To investigate the role of genetic interactions, we tested the collaborative effect on gene expression phenotypes. We investigated epistatic gene expression changes for IGHV status and trisomy 12. Epistasis was defined as a non-linear effect on gene expression between sample with both variants co-occuring and the single variants alone. In total 893 genes showed specific expression pattern in a combined genotype (padj In summary, our study provides a comprehensive reference data set for gene expression in CLL. We show that IGHV mutation status, recurrent gene mutations and CNVs drive gene expression in a previously underappreciated fashion. This includes epistatic interaction between trisomy 12 and IGHV. Using a novel way to describe coordinated changes we can group genes into sets related to buffering, inversion and suppression. Disclosures Sellner: Takeda: Employment.

Introduction Ex-vivo drug response profiling (DRP) represents a promising strategy to better understand the tumor's pathway dependencies, genotype-phenotype associations and the underlying molecular networks. However, little is known about the predictive value of drug response profiling and its utility for individualized treatment recommendations. Methods We report the first analysis of our ongoing prospective non-interventional SMARTrial (Systematic and Mechanism-based Approach to Rational Treatment Trial of Blood Cancer; ClinicalTrials.gov: NCT03488641). This trial aims to demonstrate the feasibility of high-throughput DRP in the clinical routine setting and the ability to predict in-vivo outcome by DRP. Ex-vivo responses to 107 different drugs including chemotherapeutic agents and FDA-approved or clinically investigated small molecules were assessed by an ATP-based assay after 48 hours of incubation. Results were processed by our customized analysis pipeline, enabling immediate reporting of DRP results. Patients with hematological malignancies and scheduled treatment were eligible if sufficient tumor cells were available from either lymph node biopsies, the bone marrow or peripheral blood. Results 58 patients out of the target study population of 80 patients have been recruited (AML, n = 24; B-cell lymphomas, n = 28; T-cell-lymphoma, n = 5; ALL, n = 1) as of July 24th, 2019. The median follow-up was 6 months. The primary endpoint, defined as proportion of successfully completed DRP tests within 7 days, is used to evaluate the feasibility of DRP in the clinical routine. In our study cohort, 93% of all recruited patients met this primary endpoint. We will describe our optimized workflow for drug testing and data analysis, which enables us to provide DRP reports in the clinical routine. First, we analyzed if the DRP reflected expected genotype-drug response associations and systematically regressed mutational data on DRP phenotypes. This analysis exemplarily revealed that FLT3-ITD-mutated AML cells with a mutation ratio > 0.8 respond significantly stronger to FLT3 inhibitors than FLT3-ITD-mutated AML cells with a mutation ratio < 0.8 or wildtype AML cells (Quizartinib, t-test, p = 0.022; Gilteritinib, t-test, p < 0.001), confirming that the assay can uncover clinically meaningful vulnerabilities of cancer cells. 89.3% of all evaluable patients received treatments covered by our ex-vivo DRP assay, which facilitates the correlation of in-vivo and ex-vivo drug response. The majority of patients were scheduled for chemotherapy treatment (65%) or had received chemotherapy before (32%). In-vivo responses after scheduled treatment were categorized as response (Resp), incomplete response (IResp) or progressive disease (PD). We compared ex-vivo drug responses between these different in-vivo response categories. Cancer cells of patients with PD were significantly less sensitive to ex-vivo treatment with chemotherapy (mean of all ex-vivo tested chemotherapeutics) than patients with Resp (t-test, PD vs. Resp, p = 0.007; PD vs. IResp, p = 0.054). While the anthracycline daunorubicin (t-test, p = 0.034) and the purin analogon fludarabine (t-test, p = 0.004) were significantly more active in samples from patients with a clinical Resp than in samples from patients with PD, we did not observe this trend for vinca alkaloids (e.g. vincristine, vindesine) or antifolates (e.g. pralatrexate). Interestingly, individual patients exhibited an exceptional ex-vivo sensitivity to anti-folates, despite an almost complete insensitivity to all other chemotherapeutic drugs. For a patient with refractory Burkitt Lymphoma, we observed such a strong ex-vivo sensitivity to pralatrexate despite a general ex-vivo insensitivity to most drugs. This patient had progressed on treatment with DA-EPOCH and DHAP, but responded well to methotrexate, enabling him to undergo consolidating allogeneic stem cell transplantation. Conclusion Ex-vivo drug response profiling prior to treatment can be successfully implemented in the clinical routine. Our preliminary results suggest that DRP can predict in-vivo treatment resistance and support treatment guidance for individualized treatment. Disclosures Dreger: AbbVie, Gilead, Novartis, Riemser, Roche: Speakers Bureau; Neovii, Riemser: Research Funding; MSD: Membership on an entity's Board of Directors or advisory committees, Other: Sponsoring of Symposia; AbbVie, AstraZeneca, Gilead, Janssen, Novartis, Riemser, Roche: Consultancy.

As new generations of targeted therapies emerge and tumor genome sequencing discovers increasingly comprehensive mutation repertoires, the functional relationships of mutations to tumor phenotypes remain largely unknown. Here, we measured ex vivo sensitivity of 246 blood cancers to 63 drugs alongside genome, transcriptome, and DNA methylome analysis to understand determinants of drug response. We assembled a primary blood cancer cell encyclopedia data set that revealed disease-specific sensitivities for each cancer. Within chronic lymphocytic leukemia (CLL), responses to 62% of drugs were associated with 2 or more mutations, and linked the B cell receptor (BCR) pathway to trisomy 12, an important driver of CLL. Based on drug responses, the disease could be organized into phenotypic subgroups characterized by exploitable dependencies on BCR, mTOR, or MEK signaling and associated with mutations, gene expression, and DNA methylation. Fourteen percent of CLLs were driven by mTOR signaling in a non-BCR-dependent manner. Multivariate modeling revealed immunoglobulin heavy chain variable gene (IGHV) mutation status and trisomy 12 as the most important modulators of response to kinase inhibitors in CLL. Ex vivo drug responses were associated with outcome. This study overcomes the perception that most mutations do not influence drug response of cancer, and points to an updated approach to understanding tumor biology, with implications for biomarker discovery and cancer care.

Background: Signals provided by the microenvironment can modify and circumvent pathway activities that are therapeutically targeted by drugs. However, a systems-level understanding of how the microenvironment and the genetic and molecular alterations of the tumor interact with each other and contribute to drug resistance is lacking. Methods: To address this unmet need, we established an automated microscopy-based phenotyping platform that uses co-culture conditions mimicking the bone marrow environment. We cultured primary tumor cells from more than 100 leukemia patients (CLL, AML, MCL, T-PLL, HCL) with and without bone marrow stroma cell support in DMEM and 10% human serum and treated each condition with 57 drugs in 3 concentrations. After 72h of incubation, 22 000 images per patient were acquired and processed by our custom made image analysis pipeline. Our set-up allows us to increase sensitivity far beyond simple viability testing, as it reads out additional cell type specific features such as cell morphology, autophagy and cell-cell interactions. Results: Quality assessment revealed that in contrast to mono-culture conditions, assay plate edge effects can be avoided under stable stroma cell co-culture conditions. Correlation of replicated patient samples were comparable between mono- and co-cultures (R2>0.75). In the absence of their native microenvironment, primary leukemia cells undergo spontaneous apoptosis ex-vivo. Viability at culture start was always >90% and dropped to a median of 51% (viability range: 17%-90%) after 72h in mono-cultures. Among CLL samples spontaneous apoptosis was not dependent on either IGHV mutation status or any major cytogenetic risk group. Bone marrow stroma cell co-culture conditions protected tumor cells from spontaneous apoptosis (p=8.2e-6, paired t-test). Patient samples with a high degree of spontaneous apoptosis benefited most from co-culture conditions (p=7.2e-10, Pearson correlation). To model interactions of stroma cell conditions and drug-induced apoptosis we established the following linear model: Viability ~ drug-effect + culture-model + drug-effect:culture-model. While activity of some drugs was significantly altered under co-culture conditions, we could also identify drugs with similar activity in mono- and co-cultures. For instance, the activity of common chemotherapeutics (fludarabine: p=0.002 at 0.6µM, cytarabine: p=0.001 at 1.5µM, ANOVA) or bromodomain inhibitors (I-BET-762: p=5.9e-5 at 4.5µM, JQ1: p=1.5e-8 at 1.5µM, ANOVA) was significantly reduced under co-culture conditions. In contrast, PI3K inhibitors idelalisib and duvelisib had a similar activity in mono-culture and stroma co-culture conditions and might represent a starting point to overcome stroma cell mediated drug resistance. In CLL, we identified IGHV mutation status and trisomy 12 as important determinants of response to kinase inhibitors. We confirmed these findings in stroma cell co-cultures, e.g. a better activity of B-cell receptor inhibitors in trisomy 12 and IGHV unmutated CLL. A systematic comparison of ex-vivo drug response pattern in mono- and co-cultures across 171 drug conditions will be presented. Conclusion: Our results suggest that high throughput co-culture drug testing can be robustly performed and provide an unprecedented understanding of how the stroma cell microenvironment and the genetic make-up of tumor cells contribute to drug resistance and sensitivity. Figure: Over 2 million microscopy images were acquired and analysed to assess drug resistance and sensitivity in a co-culture model of primary leukemia and bone marrow stroma cells. blue= Hoechst33342, green=Calcein AM, red=lysosomal dye NIR Figure 1. Figure 1. Disclosures No relevant conflicts of interest to declare.

The M2 subtype Acute Myeloid Leukemia (AML-M2) with t(8;21) represents an unmet challenge because of poor clinical outcomes in a sizable portion of patients. In this study,we report that FTY720 (Fingolimod), a sphingosine analogue and an FDA approved drug for treating of multiple sclerosis, shows antitumorigenic activity against the Kasumi-1 cell line, xenograft mouse models and leukemic blasts isolated from AML-M2 patients with t(8;21) translocation. Primary investigation indicated that FTY720 caused cell apoptosis through caspases and protein phosphatase 2A (PP2A) activation. Transcriptomic profiling further revealed that FTY720 treatment could upregulate AML1 target genes and interfere with genes involved in ceramide synthesis. Treatment with FTY720 led to the elimination of AML1-ETO oncoprotein and caused cell cycle arrest. More importantly, FTY720 treatment resulted in rapid and significant increase of pro-apoptotic ceramide levels, determined by high-performance liquid chromatography-electrospray ionization tandem mass spectrometry based lipidomic approaches. Structural simulation model had also indicated that the direct binding of ceramide to inhibitor 2 of PP2A (I2PP2A) could reactivate PP2A and cause cell death. This study demonstrates, for the first time, that accumulation of ceramide plays a central role in FTY720 induced cell death of AML-M2 with t(8;21). Targeting sphingolipid metabolism by using FTY720 may provide novel insight for the drug development of treatment for AML-M2 leukemia.