Platelet Transcriptome Yields Progressive Markers in Chronic Myeloproliferative Neoplasms and Identifies Putative Targets of Therapy

Abstract Predicting disease progression remains a particularly challenging endeavor in chronic degenerative disorders and cancer, thus limiting early detection, risk stratification, and preventive interventions. Here, profiling the spectrum of chronic myeloproliferative neoplasms (MPNs), as a model, we identify the blood platelet transcriptome as a proxy strategy for highly sensitive progression biomarkers that also enables prediction of advanced disease via machine learning algorithms. Using RNA sequencing (RNA-seq), we derive disease-relevant gene expression in purified platelets from 120 peripheral blood samples constituting two time-separated cohorts of patients diagnosed with one of three MPN subtypes at sample acquisition - essential thrombocythemia, ET (n=24), polycythemia vera, PV (n=33), and primary or post ET/PV secondary myelofibrosis, MF (n=42), and healthy donors (n=21). The MPN platelet transcriptome reveals an incremental molecular reprogramming that is independent of patient driver mutation status or therapy and discriminates each clinical phenotype. Differential markers in each of ET, PV and MF also highlight candidate genes as potential mediators of the pro-thrombotic and pro-fibrotic phenotypes in MPNs. In ET and PV, a strong thromboinflammatory profile is revealed by the upregulation of several interferon inducible transmembrane genes (IFITM2, IFITM3, IFITM10, IFIT3, IFI6, IFI27L1, IFI27L2), interleukin receptor accessory kinases/proteins (IRAK1, IL15, IL1RAP, IL17RC) and several solute carrier family genes (SLC16A1, SLC25A1, SLC26A8, SLC2A9) as glucose and other metabolic transport proteins, and coagulation factor V (F5). In MF, fibrosis-specific markers were identified by an additional focused comparison of MF patients versus ET and PV, showing increased expression of several pro-fibrotic growth factors (FGFR1, FGFR3, FGFRL1), matrix metalloproteinases (MMP8, MMP14), vascular endothelial growth factor A (VEGFA), insulin growth factor binding protein (IGFBP7), and cell cycle regulators (CCND1, CCNA2, CCNB2, CCNF). Also, focusing on the JAK-inhibitor ruxolitinib/RUX-specific signatures, we not only confirm previous observations on its anti-inflammatory and immunosuppressive effects (e.g. downregulation in our RUX-treated cohort of IL1RAP, CXCR5, CPNE3, ILF3) but also identify new gene clusters responsive to RUX - e.g. inhibition of type I interferon (e.g. IFIT1, IFIT2, IFI6), chromatin regulation (HIST2H3A/C, HIST1H2BK, H2AFY, SMARCA4, SMARCC2), epigenetic methylation in mitochondrial genes (ATP6, ATP8, ND1-6 and NDUFA5), and other proliferation, and proteostasis-associated markers as putative targets for MPN-directed therapy. Mechanistic insights from our data highlight impaired protein homeostasis as a prominent driver of MPN evolution, with a persistent integrated stress response. Preliminary ex vivo data on MPN patient bone-marrow-derived CD34+ cells and cultured megakaryocytes validate a proteostasis-focused subset of our peripheral platelet RNA-seq signatures. Further leveraging this substantive dataset, and in particular a progressive expression gradient across MPN, we develop a machine learning model (Lasso-penalized regression) predictive of the advanced subtype MF at high accuracy and under two conditions of validation: i) temporal Stanford internal (AUC-ROC of 0.96) and ii) geographic external cohorts (AUC-ROC of 0.97, using independently published data of an additional n=25 MF and n=46 healthy donors). Lasso-derived signatures offer a robust core set of < 5 MPN progression markers. Together, our platelet transcriptome snapshot of chronic MPNs demonstrates a methodological avenue for disease risk stratification and progression beyond genetic data alone, with potential utility in a wide range of age-related disorders. Part of the work contributing to this abstract has been posted as a preprint at this link: https://www.biorxiv.org/content/10.1101/2021.03.12.435190v2 Figure 1 Figure 1. Disclosures Gotlib: Blueprint Medicines: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Incyte: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Deciphera: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Abbvie: Membership on an entity's Board of Directors or advisory committees, Research Funding; BMS: Research Funding; Kartos: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; PharmaEssentia: Honoraria, Membership on an entity's Board of Directors or advisory committees; Cogent Biosciences: Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Chair for the Eligibility and Central Response Review Committee, Research Funding; Allakos: Consultancy.

Download Full-text

Transcriptome Profiling of Pediatric Myeloproliferative Neoplasms Demonstrates Dysregulation of Platelet-Relevant Genes and Enrichment of Inflammatory and JAK2 Mediated Complement Pathways

Blood ◽

10.1182/blood-2019-125103 ◽

2019 ◽

Vol 134 (Supplement_1) ◽

pp. 4203-4203

Author(s):

Nicole Kucine ◽

Amanda R. Leonti ◽

Aishwarya Krishnan ◽

Rhonda E. Ries ◽

Ross L. Levine ◽

...

Keyword(s):

Bone Marrow ◽

Board Of Directors ◽

Platelet Activation ◽

Research Funding ◽

Mononuclear Cells ◽

Myeloproliferative Neoplasms ◽

Differentially Expressed ◽

Rna Seq ◽

Advisory Committees ◽

Normal Controls

Introduction : Myeloproliferative neoplasms (MPNs) are rare clonal bone marrow disorders in children characterized by high blood counts, predisposition to clotting events, and the potential to transform to myelofibrosis or acute myeloid leukemia (AML). Children with MPNs have lower rates of the known driver mutations (in JAK2, MPL, and CALR) than adult patients, and the underlying pathways and molecular derangements in young patients remain unknown. Given the lack of knowledge about pediatric MPNs, it is critical that we gain a better understanding of the dysregulated pathways in these diseases, which is necessary for improving disease understanding and broadening treatment options in children. Therefore, the objective of this work was to identify differentially expressed genes and pathways between children with MPNs and healthy controls, as well as children with AML, to guide further study. Methods : Mononuclear cells were extracted from peripheral blood of pediatric MPN patients (n=20) and pediatric and young adult AML patients (n=1410), and bone marrow of normal controls (NC, n=68). AML patient samples were being evaluated as part of a Children's Oncology Group planned analysis. To identify an expression profile unique to MPNs, transcriptome data from MPN patients was contrasted against NC and AML patients. All samples were ribodepleted and underwent Illumina RNA-Seq to generate transcriptome expression data. All analyses were performed in R. Differentially expressed genes were identified using the voom function from the limma package (v. 3.38.3), and enriched pathways were identified using the pathfindR package (v. 1.3.1). Unsupervised hierarchical clustering and heatmap generation was performed using the ComplexHeatmap package (v. 1.20.0). Results : MPN patient samples showed a unique expression signature, distinct from both AML patients and normal controls. Unsupervised PCA plot (Figure 1A) and heatmaps (Figure 1B) show that MPN samples cluster together. There were 4,012 differentially expressed (DE) genes in MPNs compared to NC and 6,743 DE genes in MPNs compared to AML patients. There were 2,493 shared genes between the 2 groups (Figure 1C.) Significantly DE genes between MPNs and other groups included multiple platelet-relevant genes including PF4 (CXCL4), PF4V1, P2RY12, and PPBP (CXCL7). Interestingly, PF4V1 was the most DE gene in MPNs compared to AML, and third highest versus NC. Dysregulation of some of these genes has been seen in adult MPNs, as well as thrombosis. Further comparison of transcriptome profiles between children with (n=13) and without (n=7)JAK2 mutations showed upregulation of three genes, CFB, C2, and SERPING1, which are all known complement genes, implicating complement activation in JAK2-mutated MPN patients. Complement activation has previously been reported in adult MPNs. Pathway enrichment analysis shows a number of immune and inflammatory pathways as enriched in MPN patients compared to both AML and NC. There were 179 enriched pathways in MPNs compared to AML and 142 compared to NC, with 134 common pathways (Figure 1D.) The systemic lupus erythematosus pathway was the most heavily enriched pathway in MPNs compared to both AML and NC. Additional pathways with significant enrichment include hematopoietic cell lineage, cytokine-cytokine interactions, DNA replication, and various infection-relevant pathways. The JAK-STAT signaling pathway was also enriched in MPNs compared to both AML and NC, as was the platelet activation pathway. Conclusion: Transcriptome evaluation of childhood MPNs shows enrichment of numerous inflammatory and immune pathways, highlighting that, as in adult MPNs, inflammation is implicated in pediatric MPNs. Furthermore, specific complement genes were upregulated in JAK2-mutant MPN. Upregulation of platelet-specific genes implies potential insights into disease mechanisms and warrants more study. Variations in the cell populations may account for some of the differences seen, however all samples were largely mononuclear cells, making their comparisons reasonable. Further analysis of this early data is needed to better assess inflammatory changes and platelet activation in pediatric MPNs, as are larger sample sizes. Individual cells may have differential expression of various genes, and future experiments with single-cell RNA-seq would be helpful to further elucidate differences. Disclosures Levine: Novartis: Consultancy; Loxo: Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Research Funding; Gilead: Consultancy; Roche: Consultancy, Research Funding; Lilly: Honoraria; Amgen: Honoraria; Qiagen: Membership on an entity's Board of Directors or advisory committees; Imago Biosciences: Membership on an entity's Board of Directors or advisory committees; C4 Therapeutics: Membership on an entity's Board of Directors or advisory committees; Prelude Therapeutics: Research Funding; Isoplexis: Membership on an entity's Board of Directors or advisory committees.

Download Full-text

Safety and Efficacy of Combination Therapy of Interferon-Alpha2 + JAK1-2 Inhibitor in the Philadelphia-Negative Chronic Myeloproliferative Neoplasms. Preliminary Results from the Danish Combi-Trial - an Open Label, Single Arm, Non-Randomized Multicenter Phase II Study

Blood ◽

10.1182/blood.v126.23.824.824 ◽

2015 ◽

Vol 126 (23) ◽

pp. 824-824 ◽

Cited By ~ 12

Author(s):

Stine Ulrik Mikkelsen ◽

Lasse Kjær ◽

Vibe Skov ◽

Mads Emil Bjørn ◽

Christen Lykkegaard Andersen ◽

...

Keyword(s):

Combination Therapy ◽

Board Of Directors ◽

Research Funding ◽

Myeloproliferative Neoplasms ◽

Advisory Committees ◽

Off Label ◽

Chronic Myeloproliferative Neoplasms ◽

Major Response ◽

Travel Grant ◽

Grade 3

Abstract Background: The Philadelphia-negative, chronic myeloproliferative neoplasms (MPN) include essential thrombocythemia (ET), polycythemia vera (PV) and primary myelofibrosis (MF) (PMF). Chronic inflammation and a deregulated immune system are considered important for clonal evolution and disease progression. Ruxolitinib is a potent anti-inflammatory agent and has shown great benefit in MPN patients (pts), reducing spleen size and inflammation-mediated symptoms, thereby improving quality of life (QoL). Interferon-alpha2 (IFNa2) has been used successfully for decades in the treatment of MPN. However, 10-20% of pts are intolerant to IFNa2, yet others show limited response. Since concurrent inflammation might attenuate the efficacy of IFNa2 therapy, a combination therapy (CT) with the two agents may be more efficacious than monotherapy, likely reducing the inflammation-mediated adverse effects of IFNa2 as well. The purpose of this COMBI-trial is to evaluate the safety and efficacy of CT with IFNa2 and ruxolitinib. Patients and Methods: At the time of data cutoff, a total of 30 pts ≥18 years with prefibrotic or hyperproliferative PMF (n=7), PV (n=20) or post-PV MF (PPV-MF) (n=3) with or without prior treatment with IFNa2 and without serious comorbidity were enrolled. Evidence of active disease was required. Initial therapy was IFNa2 45 μg x 1 sc/week (Pegasys®) or 35 μg x 1 sc/week (PegIntron®) + ruxolitinib (Jakavi®) 20 mg x 2/day. Efficacy was evaluated by internationally accepted clinicohematological response criteria, with the modification that splenomegaly was assessed by palpation instead of imaging, by week 2 and 1, 3, 6 and 9 months. In addition, JAK2V617F-allele burden was monitored. Adverse events (AE) including serious AE (SAE) were recorded. Results: Median treatment duration was 24.4 weeks (range, 3.4 weeks-43.3 weeks). Twenty-seven pts were previously treated with IFNa2 (n=18 intolerant, n=5 unresponsive, n=4 both). Three pts were treatment-naïve. Twenty-seven pts (90%) remained on CT; 3 pts discontinued treatment due to an AE. One patient died from transformation to AML shortly after initiation of CT and was not included in this interim analysis. Marked improvements in pruritus, night sweats, and fatigue were recorded within the first 2-3 days in the large majority of pts and in all within 4 weeks. Palpable splenomegaly in 7 pts at baseline was significantly reduced by week 2. Hct control without phlebotomy was achieved by week 4 in 78 % of pts (7 of 9), who at baseline had an elevated hct. Only 3 pts required a total of 3 phlebotomies after initiation of CT. In Figure 1 median hct levels at 0, 1, 3, 6 months are shown. Overall, complete response (CR) was achieved as best response in 19 pts (63.3%) and partial response (PR) or major response in 8 pts (26.7%). Only 3 pts (10%) had no response (NR) to treatment. Among PV pts, 15 (75%) achieved CR (week 2, n=6; 1 month, n=6; 3 months, n=3). The other 5 PV pts achieved PR (week 2, n=3; 1 month, n=2). In PMF pts, CR (n=2) or major response (n=2) was achieved in 4 pts (57.1%) by week 2 or 1 month, and NR in 3 pts (42.8%). Among PPV-MF pts, 2 pts (66.7%) achieved CR and 1 patient (33,3%) PR by week 2. Furthermore, JAK2V617F% declined significantly as depicted in Figure 2 (JAK2V617F% over time for each patient) and 3 (median JAK2V617F% at 0, 3, 6 months). Anemia (n=15, 2 grade 3), granulocytopenia (n=13, 2 grade 3) or thrombocytopenia (n=6, 1 grade 3) were the most common AEs and were managed by dose reduction. One patient with PPV-MF (leuko- and thrombocytosis) developed pancytopenia within the first 2 weeks on CT, necessitating pausing medication for > 2 weeks. Eleven SAEs requiring hospitalization were recorded in 9 pts: pneumonia (n=3), fever (n=2), lipotymia, hematemesis, phlebitis, herpes zoster, angina pectoris and arterial hypertension, 1 patient each. Conclusion: CT with IFNa2 and ruxolitinib is highly efficacious and safein pts with PV or hyperproliferative MF,who were unresponsive or intolerant to monotherapy with IFNa2. Complete clinicohematological responses were achieved in the majority of pts in concert with a reduction in the JAK2V617F-allele burden. In general, the treatment was well tolerated. The preliminary results from this study are highly promising, encouraging a prospective study with CT in newly diagnosed pts. Additional follow-up data will be presented including QoL assessment and the impact of concurrent treatment with statins. Figure 1. Figure 1. Figure 2. Figure 2. Figure 3. Figure 3. Disclosures Off Label Use: The combination therapy with ruxolitinib (JAK1-2 inhibitor) and interferon-alpha is off-label in MPNs. The concept is dual myelosuppressive action and dual clonal suppression in addition to the anti-inflammatory properties of ruxolitinib.. Bjørn:Novartis Oncology: Research Funding. Bjerrum:Bristoll Myers Squibb, Novartis and Pfizer: Other: educational activities. El Fassi:Novartis Denmark: Honoraria, Other: Have conducted an educational session for Novartis Denmark, regarding MPNs and ruxolitinib, for this a honorarium was received.. Nielsen:Novartis: Research Funding. Pallisgaard:Roche: Other: travel grant; Amgen: Membership on an entity's Board of Directors or advisory committees, Other: travel grant, Speakers Bureau; Novartis: Other: travel grant, Research Funding, Speakers Bureau; Qiagen: Membership on an entity's Board of Directors or advisory committees; Bristol Meyer Squibb: Speakers Bureau. Hasselbalch:Novartis: Research Funding.

Download Full-text

Pre-Existing T Cell Subsets Determine Anti-PD1 Blockade Response in Richter's Transformation

Blood ◽

10.1182/blood-2020-138601 ◽

2020 ◽

Vol 136 (Supplement 1) ◽

pp. 42-43

Author(s):

Prajish Iyer ◽

Lu Yang ◽

Zhi-Zhang Yang ◽

Charla R. Secreto ◽

Sutapa Sinha ◽

...

Keyword(s):

T Cells ◽

T Cell ◽

Single Cell ◽

Board Of Directors ◽

Peripheral Blood ◽

Research Funding ◽

Gene Signature ◽

T Cell Subsets ◽

Rna Seq ◽

Advisory Committees

Despite recent developments in the therapy of chronic lymphocytic leukemia (CLL), Richter's transformation (RT), an aggressive lymphoma, remains a clinical challenge. Immune checkpoint inhibitor (ICI) therapy has shown promise in selective lymphoma types, however, only 30-40% RT patients respond to anti-PD1 pembrolizumab; while the underlying CLL failed to respond and 10% CLL patients progress rapidly within 2 months of treatment. Studies indicate pre-existing T cells in tumor biopsies are associated with a greater anti-PD1 response, hence we hypothesized that pre-existing T cell subset characteristics and regulation in anti-PD1 responders differed from those who progressed in CLL. We used mass cytometry (CyTOF) to analyze T cell subsets isolated from peripheral blood mononuclear cells (PBMCs) from 19 patients with who received pembrolizumab as a single agent. PBMCs were obtained baseline(pre-therapy) and within 3 months of therapy initiation. Among this cohort, 3 patients had complete or partial response (responders), 2 patients had rapid disease progression (progressors) (Fig. A), and 14 had stable disease (non-responders) within the first 3 months of therapy. CyTOF analysis revealed that Treg subsets in responders as compared with progressors or non-responders (MFI -55 vs.30, p=0.001) at both baseline and post-therapy were increased (Fig. B). This quantitative analysis indicated an existing difference in Tregs and distinct molecular dynamic changes in response to pembrolizumab between responders and progressors. To delineate the T cell characteristics in progressors and responders, we performed single-cell RNA-seq (SC-RNA-seq; 10X Genomics platform) using T (CD3+) cells enriched from PBMCs derived from three patients (1 responder: RS2; 2 progressors: CLL14, CLL17) before and after treatment. A total of ~10000 cells were captured and an average of 1215 genes was detected per cell. Using a clustering approach (Seurat V3.1.5), we identified 7 T cell clusters based on transcriptional signature (Fig.C). Responders had a larger fraction of Tregs (Cluster 5) as compared with progressors (p=0.03, Fig. D), and these Tregs showed an IFN-related gene signature (Fig. E). To determine any changes in the cellular circuitry in Tregs between responders and progressors, we used FOXP3, CD25, and CD127 as markers for Tregs in our SC-RNA-seq data. We saw a greater expression of FOXP3, CD25, CD127, in RS2 in comparison to CLL17 and CLL14. Gene set enrichment analysis (GSEA) revealed the upregulation of genes involved in lymphocyte activation and FOXP3-regulated Treg development-related pathways in the responder's Tregs (Fig.F). Together, the greater expression of genes involved in Treg activation may reduce the suppressive functions of Tregs, which led to the response to anti-PD1 treatment seen in RS2 consistent with Tregs in melanoma. To delineate any state changes in T cells between progressors and responder, we performed trajectory analysis using Monocle (R package tool) and identified enrichment of MYC/TNF/IFNG gene signature in state 1 and an effector T signature in state 3 For RS2 after treatment (p=0.003), indicating pembrolizumab induced proliferative and functional T cell signatures in the responder only. Further, our single-cell results were supported by the T cell receptor (TCR beta) repertoire analysis (Adaptive Biotechnology). As an inverse measure of TCR diversity, productive TCR clonality in CLL14 and CLL17 samples was 0.638 and 0.408 at baseline, respectively. Fifty percent of all peripheral blood T cells were represented by one large TCR clone in CLL14(progressor) suggesting tumor related T-cell clone expansion. In contrast, RS2(responder) contained a profile of diverse T cell clones with a clonality of 0.027 (Fig. H). Pembrolizumab therapy did not change the clonality of the three patients during the treatment course (data not shown). In summary, we identified enriched Treg signatures delineating responders from progressors on pembrolizumab treatment, paradoxical to the current understanding of T cell subsets in solid tumors. However, these data are consistent with the recent observation that the presence of Tregs suggests a better prognosis in Hodgkin lymphoma, Follicular lymphoma, and other hematological malignancies. Figure 1 Disclosures Kay: Pharmacyclics: Membership on an entity's Board of Directors or advisory committees, Research Funding; Oncotracker: Membership on an entity's Board of Directors or advisory committees; Rigel: Membership on an entity's Board of Directors or advisory committees; Juno Theraputics: Membership on an entity's Board of Directors or advisory committees; Agios Pharma: Membership on an entity's Board of Directors or advisory committees; Cytomx: Membership on an entity's Board of Directors or advisory committees; Astra Zeneca: Membership on an entity's Board of Directors or advisory committees; Morpho-sys: Membership on an entity's Board of Directors or advisory committees; Tolero Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees, Research Funding; Bristol Meyer Squib: Membership on an entity's Board of Directors or advisory committees, Research Funding; Acerta Pharma: Research Funding; Sunesis: Research Funding; Dava Oncology: Membership on an entity's Board of Directors or advisory committees; Abbvie: Research Funding; MEI Pharma: Research Funding. Ansell:AI Therapeutics: Research Funding; Takeda: Research Funding; Trillium: Research Funding; Affimed: Research Funding; Bristol Myers Squibb: Research Funding; Regeneron: Research Funding; Seattle Genetics: Research Funding; ADC Therapeutics: Research Funding. Ding:Astra Zeneca: Research Funding; Abbvie: Research Funding; Octapharma: Membership on an entity's Board of Directors or advisory committees; MEI Pharma: Membership on an entity's Board of Directors or advisory committees; alexion: Membership on an entity's Board of Directors or advisory committees; Beigene: Membership on an entity's Board of Directors or advisory committees; DTRM: Research Funding; Merck: Membership on an entity's Board of Directors or advisory committees, Research Funding. OffLabel Disclosure: pembrolizumab

Download Full-text

Metaphase Cytogenetics for Risk Stratification in Newly Diagnosed Multiple Myeloma

Blood ◽

10.1182/blood-2019-122291 ◽

2019 ◽

Vol 134 (Supplement_1) ◽

pp. 4396-4396

Author(s):

Patrick Mellors ◽

Moritz Binder ◽

Rhett P. Ketterling ◽

Patricia Griepp ◽

Linda B Baughn ◽

...

Keyword(s):

Multiple Myeloma ◽

Risk Stratification ◽

Board Of Directors ◽

Research Funding ◽

Multivariable Analysis ◽

Risk Modeling ◽

Newly Diagnosed ◽

Advisory Committees ◽

Cytogenetic Abnormalities ◽

Improve Risk Stratification

Introduction: Abnormal metaphase cytogenetics are associated with inferior survival in newly diagnosed multiple myeloma (MM). These abnormalities are only detected in one third of cases due to the low proliferative rate of plasma cells. It is unknown if metaphase cytogenetics improve risk stratification when using contemporary prognostic models such as the revised international staging system (R-ISS), which incorporates interphase fluorescence in situ hybridization (FISH). Aims: The aims of this study were to 1) characterize the association between abnormalities on metaphase cytogenetics and overall survival (OS) in newly diagnosed MM treated with novel agents and 2) evaluate whether the addition of metaphase cytogenetics to R-ISS, age, and plasma cell labeling index (PCLI) improves model discrimination with respect to OS. Methods: We analyzed a retrospective cohort of 483 newly diagnosed MM patients treated with proteasome inhibitors (PI) and/or immunomodulators (IMID) who had metaphase cytogenetics performed prior to initiation of therapy. Abnormal metaphase cytogenetics were defined as MM specific abnormalities, while normal metaphase cytogenetics included constitutional cytogenetic variants, age-related Y chromosome loss, and normal metaphase karyotypes. Multivariable adjusted proportional hazards regression models were fit for the association between known prognostic factors and OS. Covariates associated with inferior OS on multivariable analysis included R-ISS stage, age ≥ 70, PCLI ≥ 2, and abnormal metaphase cytogenetics. We devised a risk scoring system weighted by their respective hazard ratios (R-ISS II +1, R-ISS III + 2, age ≥ 70 +2, PCLI ≥ 2 +1, metaphase cytogenetic abnormalities + 1). Low (LR), intermediate (IR), and high risk (HR) groups were established based on risk scores of 0-1, 2-3, and 4-5 in modeling without metaphase cytogenetics, and scores of 0-1, 2-3, and 4-6 in modeling incorporating metaphase cytogenetics, respectively. Survival estimates were calculated using the Kaplan-Meier method. Survival analysis was stratified by LR, IR, and HR groups in models 1) excluding metaphase cytogenetics 2) including metaphase cytogenetics and 3) including metaphase cytogenetics, with IR stratified by presence and absence of metaphase cytogenetic abnormalities. Survival estimates were compared between groups using the log-rank test. Harrell's C was used to compare the predictive power of risk modeling with and without metaphase cytogenetics. Results: Median age at diagnosis was 66 (31-95), 281 patients (58%) were men, median follow up was 5.5 years (0.04-14.4), and median OS was 6.4 years (95% CI 5.7-6.8). Ninety-seven patients (20%) were R-ISS stage I, 318 (66%) stage II, and 68 (14%) stage III. One-hundred and fourteen patients (24%) had high-risk abnormalities by FISH, and 115 (24%) had abnormal metaphase cytogenetics. Three-hundred and thirteen patients (65%) received an IMID, 119 (25%) a PI, 51 (10%) received IMID and PI, and 137 (28%) underwent upfront autologous hematopoietic stem cell transplantation (ASCT). On multivariable analysis, R-ISS (HR 1.59, 95% CI 1.29-1.97, p < 0.001), age ≥ 70 (HR 2.32, 95% CI 1.83-2.93, p < 0.001), PCLI ≥ 2, (HR 1.52, 95% CI 1.16-2.00, p=0.002) and abnormalities on metaphase cytogenetics (HR 1.35, 95% CI 1.05-1.75, p=0.019) were associated with inferior OS. IR and HR groups experienced significantly worse survival compared to LR groups in models excluding (Figure 1A) and including (Figure 1B) the effect of metaphase cytogenetics (p < 0.001 for all comparisons). However, the inclusion of metaphase cytogenetics did not improve discrimination. Likewise, subgroup analysis of IR patients by the presence or absence of metaphase cytogenetic abnormalities did not improve risk stratification (Figure 1C) (p < 0.001). The addition of metaphase cytogenetics to risk modeling with R-ISS stage, age ≥ 70, and PCLI ≥ 2 did not improve prognostic performance when evaluated by Harrell's C (c=0.636 without cytogenetics, c=0.642 with cytogenetics, absolute difference 0.005, 95% CI 0.002-0.012, p=0.142). Conclusions: Abnormalities on metaphase cytogenetics at diagnosis are associated with inferior OS in MM when accounting for the effects of R-ISS, age, and PCLI. However, the addition of metaphase cytogenetics to prognostic modeling incorporating these covariates did not significantly improve risk stratification. Disclosures Lacy: Celgene: Research Funding. Dispenzieri:Akcea: Consultancy; Intellia: Consultancy; Alnylam: Research Funding; Celgene: Research Funding; Janssen: Consultancy; Pfizer: Research Funding; Takeda: Research Funding. Kapoor:Celgene: Honoraria; Sanofi: Consultancy, Research Funding; Janssen: Research Funding; Cellectar: Consultancy; Takeda: Honoraria, Research Funding; Amgen: Research Funding; Glaxo Smith Kline: Research Funding. Leung:Prothena: Membership on an entity's Board of Directors or advisory committees; Takeda: Research Funding; Omeros: Research Funding; Aduro: Membership on an entity's Board of Directors or advisory committees. Kumar:Celgene: Consultancy, Research Funding; Janssen: Consultancy, Research Funding; Takeda: Research Funding.

Download Full-text

Personalized Transcriptomic Analyses Identify Unique Signatures That Correlate with Genomic Subtypes in Acute Myeloid Leukemia (AML) Using Explainable Artificial Intelligence

Blood ◽

10.1182/blood-2020-139522 ◽

2020 ◽

Vol 136 (Supplement 1) ◽

pp. 33-34 ◽

Cited By ~ 1

Author(s):

Yazan Rouphail ◽

Nathan Radakovich ◽

Jacob Shreve ◽

Sudipto Mukherjee ◽

Babal K. Jha ◽

...

Keyword(s):

Machine Learning ◽

Board Of Directors ◽

Research Funding ◽

State Of The Art ◽

Data Monitoring Committee ◽

Biological Pathways ◽

Machine Learning Algorithms ◽

Advisory Committees ◽

Altered Expression ◽

Individual Level

Background Multi-omic analysis can identify unique signatures that correlate with cancer subtypes. While clinically meaningful molecular subtypes of AML have been defined based on the status of single genes such as NPM1 and FLT3, such categories remain heterogeneous and further work is needed to characterize their genetic and transcriptomic diversity on a truly individualized basis. Further, patients (pts) with NPM1+/FLT3-ITD- AML have a better overall survival compared to patients with NPM1-/FLT3-ITD+, suggesting that these pts could have different transcriptomic signature that impact phenotype, pathophysiology, and outcomes. Many current transcriptome analytic techniques use clustering analysis to aggregate samples and look at relationships on a cohort-wide basis to build transcriptomic signatures that correlate with phenotype or outcome. Such approaches can undermine the heterogeneity of the gene expression in pts with the same signatures. In this study, we took advantage of state of the art machine learning algorithms to identify unique transcriptomic signatures that correlate with AML genomic phenotype. Methods Genomic (whole exome sequencing and targeted deep sequencing) and transcriptomic data from 451 AML pts included in the Beat AML study (publicly available data) were used to build transcriptomic signatures that are specific for AML patients with NPM1+/FLT3-ITD+ compared to NPM1+/FLT3-ITD, and NPM1-/FLT3-ITD-. We chose these AML phenotypes as they have been described extensively and they correlate with clinical outcomes. Results A total of 242 patients (54%) had NPM1-/FLT3-, 35 (8%) were NPM1+/FLT3-, and 47 (10%) were NPM1+/FLT3+. Our algorithm identified 20 genes that are highly specific for NPM1/FLT3ITD phenotype: HOXB-AS3, SCRN1, LMX1B, PCBD1, DNAJC15, HOXA3, NPTXq, RP11-1055B8, ABDH128, HOXB8, SOCS2, HOXB3, HOXB9, MIR503HG, FAM221B, NRP1, NDUFAF3, MEG3, CCDC136, and HIST1H2BC. Interestingly, several of those genes were overexpressed or underexpressed in specific phenotypes. For example, SCRN1, LMX1B, RP11-1055B8, ABDH128, HOXB8, MIR503HG, NRP1 are only overexpressed or underexpressed in patients with NPM1-/FLT3-, while PCBD1, NDUFAF3, FAM221B are overexpressed or underexpressed in pts with NPM1+/FLT3+. These genes affect several important pathways that regulate cell differentiation, proliferation, mitochondrial oxidative phosphorylation, histone modification and lipid metabolism. All these genes had previously been reported as having altered expression in genomic studies of AML, confirming our approach's ability to identify biologically meaningful relationships. Further, our algorithm can provide a personalized explanation of overexpressed and underexpressed genes specific for a given patient, thus identifying targetable pathways for each pt. Figure 1 below shows three pts with the same genotype (NPM1+/FLT3-ITD+) but demonstrate different transcriptomic patterns of overexpression or underexpression that affect different biological pathways. Conclusions We describe the use of a state of the art explainable machine learning approach to define transcriptomic signatures that are specific for individual pts. In addition to correctly distinguishing AML subtype based on specific transcriptomic signatures, our model was able to accurately identify upregulated and downregulated genes that affecte several important biological pathways in AML and can summarize these pathways at an individual level. Such an approach can be used to provide personalized treatment options that can target the activated pathways at an individual level. Disclosures Mukherjee: Partnership for Health Analytic Research, LLC (PHAR, LLC): Honoraria; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees, Research Funding; EUSA Pharma: Consultancy; Celgene/Acceleron: Membership on an entity's Board of Directors or advisory committees; Bristol Myers Squib: Honoraria; Aplastic Anemia and MDS International Foundation: Honoraria; Celgene: Consultancy, Honoraria, Research Funding. Maciejewski:Alexion, BMS: Speakers Bureau; Novartis, Roche: Consultancy, Honoraria. Sekeres:BMS: Consultancy; Takeda/Millenium: Consultancy; Pfizer: Consultancy. Nazha:Jazz: Research Funding; Incyte: Speakers Bureau; Novartis: Speakers Bureau; MEI: Other: Data monitoring Committee.

Download Full-text

Progressive and predictive markers of disease evolution: platelet transcriptome in chronic myeloproliferative neoplasms

10.1101/2021.03.12.435190 ◽

2021 ◽

Author(s):

Zhu Shen ◽

Wenfei Du ◽

Cecelia Perkins ◽

Lenn Fechter ◽

Vanita Natu ◽

...

Keyword(s):

Machine Learning ◽

Risk Stratification ◽

Disease Risk ◽

Myeloproliferative Neoplasms ◽

Penalized Regression ◽

Machine Learning Algorithms ◽

Disease Evolution ◽

Chronic Myeloproliferative Neoplasms ◽

Age Related ◽

Wide Range

Predicting disease natural history remains a particularly challenging endeavor in chronic degenerative disorders and cancer, thus limiting early detection, risk stratification, and preventive interventions. Here, profiling the spectrum of chronic myeloproliferative neoplasms (MPNs), as a model, we identify the blood platelet transcriptome as a generalizable strategy for highly sensitive progression biomarkers that also enable prediction via machine learning algorithms. Using RNA sequencing (RNA seq), we derive disease relevant gene expression and alternative splicing in purified platelets from 120 peripheral blood samples constituting two independently collected and mutually validating patient cohorts of the three MPN subtypes: essential thrombocythemia, ET (n=24), polycythemia vera, PV (n=33), and primary or post ET/PV secondary myelofibrosis, MF (n=42), as well as healthy donors (n=21). The MPN platelet transcriptome discriminates each clinical phenotype and reveals an incremental molecular reprogramming that is independent of patient driver mutation status or therapy. Leveraging this dataset, in particular the progressive expression gradient noted across MPN, we develop a machine learning model (Lasso-penalized regression) predictive of the advanced subtype MF at high accuracy (AUC-ROC of 0.95-0.96) with validation under two conditions: i) temporal, with training on the first cohort (n=71) and independent testing on the second (n=49) and ii) 10 fold cross validation on the entire dataset. Lasso-derived signatures offer a robust core set of < 10 MPN progression markers. Mechanistic insights from our data highlight impaired protein homeostasis as a prominent driver of MPN evolution, with persistent integrated stress response. We also identify JAK inhibitor-specific signatures and other interferon, proliferation, and proteostasis associated markers as putative targets for MPN-directed therapy. Our platelet transcriptome snapshot of chronic MPNs establishes a methodological foundation for deciphering disease risk stratification and progression beyond genetic data alone, thus presenting a promising avenue toward potential utility in a wide range of age-related disorders.

Download Full-text

Nitric Oxide As a Mediator of Bone Marrow Fibrosis in Patients with Myelofibrosis

Blood ◽

10.1182/blood.v124.21.4587.4587 ◽

2014 ◽

Vol 124 (21) ◽

pp. 4587-4587

Author(s):

Ali Tabarroki ◽

Daniel Lindner ◽

Valeria Visconte ◽

Nikolaos Papandantonakis ◽

Jing Ai ◽

...

Keyword(s):

Nitric Oxide ◽

Bone Marrow ◽

Growth Factor ◽

Board Of Directors ◽

Transforming Growth Factor ◽

Myeloproliferative Neoplasms ◽

Inos Expression ◽

No Production ◽

Advisory Committees ◽

Extracellular Matrix Components

Abstract Bone marrow (BM) fibrosis is a key pathomorphologic feature of patients (pts) with primary myelofibrosis (PMF) and the fibrotic phases of essential thrombocythemia (post-ET MF) and polycythemia vera (post-PV MF). The degree of BM fibrosis appears to correlate with survival. Indeed worse survival has been associated with increased BM fibrosis. The BM stromal microenvironment is important in the pathogenesis of BM fibrosis. Cellular components (fibroblasts, macrophages, endothelial cells, adipocytes), structural fibrils (collagen, reticulin) and extracellular matrix components are all forming elements of the BM stroma. Increased stromal fibrosis has been linked to abnormalities in the number/ function of megakaryocytes and platelets in hematologic diseases. Several cytokines like Platelet Derived Growth Factor (PDGF) and Transforming Growth Factor-Beta (TGF-b) have been also linked to the pathophysiology of BM fibrosis. PDGF has been shown to increase fibroblast growth in megakaryocytes and platelets although increased PDGF did not correlate with increased production of either reticulin or collagenous fibrosis. Moreover, PMF pts have increased TGF-b levels in platelets, megakaryocytes, and monocytes. Nitric Oxide (NO) is a ubiquitous gas important in physiologic processes particularly vasodilatation. Dysregulation of NO levels has been implicated in pulmonary hypertension (PH), hemoglobinopathies, and cardiovascular diseases. In Peyronie’s disease, a localized fibrosis of the penile tunica albuginea, increased NO production by expression of iNOS decreases collagen deposition by neutralization of profibrotic reactive oxygen species and decreased myofibroblast formation. Aside from its role in maintaining normal vascular tone, NO also plays a role in fibroblast formation and collagen biosynthesis. We previously reported that ruxolitinib, a JAK1/2 inhibitor restores NO levels leading to improvement of PH in MF pts (Tabarroki et al., Leukemia 2014). We now hypothesize that plasma/serum NO level is a key regulator of BM fibrosis in MF and that ruxolitinib treatment (Tx) leads to improvement of BM fibrosis by NO modulation. Using a Sievers 280i NO analyzer we measured the plasma/serum NO level of a large cohort (n=75) of pts with myeloid and myeloproliferative neoplasms (MPN) [MDS, RARS/RCMD=8; MPN, ET=8, PV=8, MF=24, Mastocytosis=7; MDS/MPN, CMML=11, MDS/MPN-U, RARS-T=9]. Healthy subjects (n=10) were used as a control. MPN pts had low NO (nM) levels among the pts studied with the lowest level found in MF pts: MF=30.31±11.8, PV=39.0±16.1, ET=36±20.3, RARS=74.6±41.7 (P=.01), CMML=84.4±89.2 (P=.04), RCMD=163.4±103.8 (P<.001), RARS-T=131.1±99.8 (P<.001). In total, NO levels were lower in classic MPN (n=40, 35.3±16.6) compared to MDS (n=8, 119±62.8; P=.001) and MDS/MPN (n=20, 105±94.6; P=.008). When we looked at the correlation between NO levels and BM fibrosis grade we found that there is an inverse correlation between NO levels and worsening BM fibrosis grade from grade MF1 to MF3. NO levels in normal (n=10) vs MF1 (n=3) were 53.3 vs 39.1, P=.025; normal vs MF2 (n=7) were 53.3 vs 37, P=.021; normal vs MF3 (n=12) were 53.3 vs 34.4, P=.006. A total of 8 pts who were treated with ruxolitinib and had at least 1 pre and 1 post Tx (≥3 months from initiation of ruxolitinib) were tested for NO levels. Among the 8 pts, 4 pts who demonstrated improvement in BM scores had a trend towards improved NO levels after ruxolitinib Tx [NO pre vs post; pt #1: 6 vs 10.5; pt#2: 4.3 vs 6.4; pt#3 49.7 vs 52.1; pt#4 36 vs 41.3; P=.02] while 4 had worsening or had no change in BM fibrosis grade and had a minimal change or decline in the NO (pt#5: 18.4 vs 23, pt#6: 14.29 vs 12.1, pt#7: 32.7 vs 32.1, pt#8: 110.9 vs 40.4). One pt who had improvement in BM fibrosis grade after ruxolitinib Tx had increased iNOS expression by Western blotting (pt#1) while no iNOS expression (pt#5) was noted in the pt who did not have improvement in BM fibrosis. Of note, multi-analytic cytokines profile also showed an overall decrease in cytokines especially a 2.8 fold-decrease in IL8 levels post-Tx in the pt with improvement in BM fibrosis. In conclusion, NO is decreased in MPN particularly in MF and may be a key mediator of BM fibrosis in MF. Pharmacologic therapies such as JAK inhibitors may mediate improvement of BM fibrosis by modulation of NO levels in MF. Disclosures Tiu: Gilead: Membership on an entity's Board of Directors or advisory committees; Novartis: Speakers Bureau; Incyte : Membership on an entity's Board of Directors or advisory committees, Speakers Bureau.

Download Full-text

The Value of Risk Stratification Tools in Multiple Myeloma (MM) in the Real-World: Validation of the Revised- International Staging System (R-ISS) at Initiation of Treatment and Relevance of the ISS and the R-ISS for Risk Stratification in the Relapsed Setting Using Data from the Czech Registry of Monoclonal Gammopathies (RMG)

Blood ◽

10.1182/blood.v128.22.2418.2418 ◽

2016 ◽

Vol 128 (22) ◽

pp. 2418-2418 ◽

Cited By ~ 1

Author(s):

Roman Hájek ◽

Jiri Jarkovsky ◽

Walter Bouwmeester ◽

Maarten Treur ◽

DeCosta Lucy ◽

...

Keyword(s):

Risk Stratification ◽

Board Of Directors ◽

Real World ◽

Research Funding ◽

Risk Group ◽

Stage I ◽

Advisory Committees ◽

Standard Practice ◽

First Relapse ◽

Using Data

Abstract The ISS stratifies survival risk in patients with MM based on β2-microglobulin and albumin levels. The R-ISS is an improved stratification tool, which also uses chromosomal abnormalities (CA) and lactate dehydrogenase (LDH). It was developed based on clinical trial data in the first-line setting but, has not been validated outside clinical trials or for use in the relapsed setting. Using data from the RMG, we assessed the real-world validity of the R-ISS at diagnosis. Additionally, as it is standard practice to re-stage patients after first relapse, we explored the value of re-estimating ISS stage in the relapsed setting and exploring the carry on effect of R-ISS from diagnosis. Re-estimation of R-ISS at relapse is not possible as standard practice often does not include CA measurement at first relapse. Assessment of improvement in stratification was based on visual comparison of median OS, hazard ratios (HR) and confidence intervals. Eligible patients were diagnosed with symptomatic MM between May 2007 and April 2016. A Cox regression model and Kaplan-Meier analyses assessed the performance of the ISS and R-ISS for stratifying patients based on survival both at diagnosis and at first relapse. Overall, there were 3027 patients at diagnosis however only 493 were included in these analyses due to unavailable CA values (84% of patients). ISS and R-ISS stage distribution at diagnosis was ISS I 31.2%, II 29.1% and III 39.6%; and R-ISS I 12% II 57% and III 31% (Table 1). Median overall survival (OS) in months (95% confidence interval [CI]), from diagnosis was 73.5 (68.0-NE), 40.5 (31.0-50.0) and 29.0 (20.9-37.2) in patients with ISS stage I, II and III, respectively, and not reached (NR), 46.6 (39.2-54.1) and 26.0 (18.2-33.8) in patients with R-ISS stage I, II and III, respectively. Table 2 shows HR, which indicate OS assessed in alternative ways was significantly different among the three stages for both ISS and R-ISS. R-ISS provided refined stratification than ISS alone, since R-ISS stage III classified patients with higher risk than ISS III alone, (shorter median OS, narrower CI, and stronger HR vs. ISS I and II). From the original sample of 493 patients at diagnosis, only 250 went on to receive further treatment after first relapse. The median OS months (95% CI) after first relapse was 46.4 (32.0-60.8), 22.8 (13.4-32.1) and 14.9 months (9.0-20.8) in patients staged as ISS stage I, II and III at diagnosis, respectively. In patients staged as R-ISS stage I, II and III at diagnosis it was NR, 25.6 (20.8-30.3) and 10.4 (6.7-14.2), respectively. Data to enable re-estimation of ISS and R-ISS at first relapse were available for 187 patients (R-ISS re-stratification was using CA data at diagnosis only). Median OS months (95% CI) from first relapse was 32.2 (15.5-49.0), 25.6 (11.5-39.6) and 10.8 (8.6-13.0) in patients at ISS stage I, II and III, respectively, and 23.3 (NE-NE), 28.4 (20.8-36.0) and 9.7 (6.5-12.9) in patients at R-ISS stage I, II and III, respectively. The HRs comparing OS from first relapse stratified by ISS at diagnosis indicated that re-estimating ISS did not improve stratification (Table 2). For R-ISS, compared with staging patients at diagnosis, staging at first relapse resulted in refined stratification between stage II and III, however assessment of HRs comparing to stage I was difficult owing to small sample sizes. Re-estimation of R-ISS stage at first relapse resulted in 26% of patients having their stage reclassified; the main drivers of reclassification to a lower R-ISS risk group were β2-microglobulin and albumin levels, and to a higher risk group, LDH levels. Our real-world data show that, at diagnosis R-ISS provides refined risk stratification compared with ISS. Further refinement seemed to be added by restaging at first relapse using R-ISS, not ISS however, CA measurements are not currently routinely measured at first relapse, limiting the practical utility of the R-ISS at this stage. Therefore, re-estimating R-ISS stage after first relapse may enable physicians improve estimation of patient prognosis. Both ISS and R-ISS have been developed for use at diagnosis when there is less evidence to predict prognosis, therefore risk stratification after first relapse should also consider other historical patient, disease and treatment factors contributing to improved risk stratification and improved treatment selection and outcomes. Disclosures Hájek: Novartis: Consultancy, Research Funding; Celgene: Research Funding; Amgen: Honoraria, Research Funding; Takeda: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; BMS: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees. Bouwmeester:Amgen: Consultancy. Treur:Amgen: Consultancy. Lucy:Amgen: Employment, Other: Amgen Stock. Campioni:Amgen: Employment, Other: Holds Amgen Stock. Delforge:Celgene: Honoraria; Amgen: Honoraria; Janssen: Honoraria. Raab:BMS: Consultancy; Celgene: Membership on an entity's Board of Directors or advisory committees; Janssen: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Amgen: Consultancy, Research Funding; Novartis: Consultancy, Research Funding. Schoen:Amgen: Employment, Other: Holds Amgen Stock. Szabo:Amgen: Employment, Other: Holds Amgen Stock. Gonzalez-McQuire:Amgen: Employment, Other: Holds Amgen Stock.

Download Full-text

Machine Learning Prediction for Complete Response to Hypomethylating Agents with or without Additional Agents in Patients with Newly Diagnosed Myelodysplastic Syndrome

Blood ◽

10.1182/blood-2019-130155 ◽

2019 ◽

Vol 134 (Supplement_1) ◽

pp. 1720-1720

Author(s):

Koji Sasaki ◽

Guillermo Montalban Bravo ◽

Rashmi Kanagal-Shamanna ◽

Elias Jabbour ◽

Farhad Ravandi ◽

...

Keyword(s):

Machine Learning ◽

Board Of Directors ◽

Research Funding ◽

Area Under The Curve ◽

Complete Response ◽

Learning Model ◽

Hypomethylating Agents ◽

Newly Diagnosed ◽

Advisory Committees ◽

Machine Learning Model

Background: Myelodysplastic syndrome (MDS) is a heterogeneous malignant myeloid neoplasm of hematopoietic stem cells due to cytogenetic alterations and somatic mutations in genes (DNA methylation, DNA repair, chromatin regulation, RNA splicing, transcription regulation, and signal transduction). Hypomethylating agents (HMA) are the standard of care for MDS, and 40-60% of patients achieved response to HMA. However, the prediction for response is difficult due to the nature of heterogeneity and the context of clinical conditions such as the degree of cytopenias and the dependency on transfusion. Machine learning outperforms conventional statistical models for prediction in statistical competitions. Prediction with machine learning models may predict response in patients with MDS. The aim of this study is to develop a machine learning model for the prediction of complete response (CR) to HMA with or without additional therapeutic agents in patients with newly diagnosed MDS. Methods: From November 2012 to August 2017, we analyzed 435 patients with newly diagnosed MDS who received frontline therapy as follows; azacitidine (AZA) (3-day, 5-day, or 7-day) ± vorinostat ± ipilimumab ± nivolumab; decitabine (DAC) (3-day or 5-day) ± vorinostat; 5-day guadecitabine. Clinical variables, cytogenetic abnormalities, and the presence of genetic mutations by next generation sequencing (NGS) were included for variable selection. The whole cohort was randomly divided into training/validation and test cohorts at an 8:2 ratio. The training/validation cohort was used for 4-fold cross validation. Hyperparameter optimization was performed with Stampede2, which was ranked as the 15th fastest supercomputer at Texas Advanced Computing Center in June 2018. A gradient boosting decision tree-based framework with the LightGBM Python module was used after hyperparameter tuning for the development of the machine learning model with training/validation cohorts. The performance of prediction was assessed with an independent test dataset with the area under the curve. Results: We identified 435 patients with newly diagnosed MDS who enrolled on clinical trials as follows: 33 patients, 5-day AZA; 23, 5-day AZA + vorinostat; 43, 3-day AZA; 20, 5-day AZA + ipilimumab; 19 patients, AZA + nivolumab; 7, AZA + ipilumumab + nivolumab; 114, 5-day DAC; 74, 3-day DAC; 4, DAC + vorinostat; 97, 5-day guadecitabine. In the whole cohort, the median age at diagnosis was 68 years (range, 13.0-90.3); 117 (27%) patients had a history of prior radiation or cytotoxic chemotherapy; the median white blood cell count was 2.9 (×109/L) (range, 0.5-102); median absolute neutrophil count, 1.1 (×109/L) (range, 0.0-55.1); median hemoglobin count, 9.5 (g/dL) (range, 4.7-15.4); median platelet count, 63 (×109/L) (range, 2-881); and median blasts in bone marrow, 8% (range, 0-20). Among 411 evaluable patients for the revised international prognostic scoring system, 15 (4%) had very low risk disease; 42 (10%), low risk; 68 (17%), intermediate risk; 124 (30%), high risk; and 162 (39%), very high risk. Overall, 153 patients (53%) achieved CR. Hyperparameter tuning identified the optimal hyperparameters with colsample by tree of 0.175, learning rate of 0.262, the maximal depth of 2, minimal data in leaf of 29, number of leaves of 11, alpha regularization of 0.010, lambda regularization of 2.085, and subsample of 0.639. On the test cohort with 87 patients, the machine learning model accurately predicted response in 65 patients (75%); 53 non-CR among 56 non-CR (95% accuracy); and 12 CR among 31 CR (39% accuracy). The trend of accuracy improvement by iteration (i.e., the number of decision trees) is shown in Figure 1. The area under the curve was 0.761521 in the test cohort. Conclusion: Our machine learning model with clinical, cytogenetic, and NGS data can predict CR to HMA in patients with newly diagnosed MDS. This approach can identify patients who may benefit from HMA therapy with and without additional agents for response, and can optimize the timing of allogeneic stem cell transplant. Disclosures Sasaki: Otsuka: Honoraria; Pfizer: Consultancy. Jabbour:Takeda: Consultancy, Research Funding; BMS: Consultancy, Research Funding; Adaptive: Consultancy, Research Funding; Amgen: Consultancy, Research Funding; AbbVie: Consultancy, Research Funding; Pfizer: Consultancy, Research Funding; Cyclacel LTD: Research Funding. Ravandi:Cyclacel LTD: Research Funding; Selvita: Research Funding; Menarini Ricerche: Research Funding; Macrogenix: Consultancy, Research Funding; Amgen: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Xencor: Consultancy, Research Funding. Kadia:Pfizer: Membership on an entity's Board of Directors or advisory committees, Research Funding; Celgene: Research Funding; Bioline RX: Research Funding; Jazz: Membership on an entity's Board of Directors or advisory committees, Research Funding; AbbVie: Consultancy, Research Funding; BMS: Research Funding; Amgen: Membership on an entity's Board of Directors or advisory committees, Research Funding; Genentech: Membership on an entity's Board of Directors or advisory committees; Pharmacyclics: Membership on an entity's Board of Directors or advisory committees; Takeda: Membership on an entity's Board of Directors or advisory committees. Takahashi:Symbio Pharmaceuticals: Consultancy. DiNardo:syros: Honoraria; jazz: Honoraria; agios: Consultancy, Honoraria; celgene: Consultancy, Honoraria; notable labs: Membership on an entity's Board of Directors or advisory committees; medimmune: Honoraria; abbvie: Consultancy, Honoraria; daiichi sankyo: Honoraria. Cortes:Novartis: Consultancy, Honoraria, Research Funding; Bristol-Myers Squibb: Consultancy, Research Funding; Immunogen: Consultancy, Honoraria, Research Funding; Sun Pharma: Research Funding; Pfizer: Consultancy, Honoraria, Research Funding; Astellas Pharma: Consultancy, Honoraria, Research Funding; Jazz Pharmaceuticals: Consultancy, Research Funding; Merus: Consultancy, Honoraria, Research Funding; Forma Therapeutics: Consultancy, Honoraria, Research Funding; Daiichi Sankyo: Consultancy, Honoraria, Research Funding; BiolineRx: Consultancy; Biopath Holdings: Consultancy, Honoraria; Takeda: Consultancy, Research Funding. Kantarjian:AbbVie: Honoraria, Research Funding; Cyclacel: Research Funding; Pfizer: Honoraria, Research Funding; Astex: Research Funding; Agios: Honoraria, Research Funding; Jazz Pharma: Research Funding; Daiichi-Sankyo: Research Funding; Novartis: Research Funding; Actinium: Honoraria, Membership on an entity's Board of Directors or advisory committees; Immunogen: Research Funding; Takeda: Honoraria; BMS: Research Funding; Ariad: Research Funding; Amgen: Honoraria, Research Funding. Garcia-Manero:Amphivena: Consultancy, Research Funding; Helsinn: Research Funding; Novartis: Research Funding; AbbVie: Research Funding; Celgene: Consultancy, Research Funding; Astex: Consultancy, Research Funding; Onconova: Research Funding; H3 Biomedicine: Research Funding; Merck: Research Funding.

Download Full-text

Landscape of Recurrent Mutations in Non-Coding Genome with Functional Implications in Newly-Diagnosed Multiple Myeloma

Blood ◽

10.1182/blood-2018-190 ◽

2018 ◽

Vol 132 (Supplement 1) ◽

pp. 190-190 ◽

Cited By ~ 1

Author(s):

Mehmet Kemal Samur ◽

Chandraditya Chakraborty ◽

Raphael Szalat ◽

Anil Aktas Samur ◽

Mariateresa Fulciniti ◽

...

Keyword(s):

Board Of Directors ◽

Research Funding ◽

P Value ◽

Newly Diagnosed ◽

Rna Seq ◽

Coding Region ◽

Advisory Committees ◽

Coding Regions ◽

Somatic Alterations ◽

Equity Ownership

Abstract Multiple Myeloma (MM) is a complex disease with distinct molecular and clinical characteristics. Recent large collaborative efforts have identified number of driver genes. However over 95% of all somatic alterations occur in non-coding regions and very little is known about how they affect the disease. We performed a deep (average coverage > 80X) whole genome sequencing (WGS) on 260 MM samples (208 newly diagnosed and 52 first relapse after uniform treatment) to comprehensively analyze recurrent somatic alterations in non-coding regions. We detected median 11,852 (Range 4,802-87,396) mutations and indels per sample with overall more than 3.9M total somatic mutations. Introns (3.6 mutations/per Mb) and intergenic regions (4.06 mutations/per Mb) had significantly higher number of mutations per megabase compared to Exons (2.7 mutations/per Mb) (p < 1e-5). Mutations in coding regions in our data was similar to published whole exome sequencing studies. We observed 46 [range 7 - 219] structural variants (SVs) per sample with 98% involving non-coding regions. We found that number of SVs significantly correlated with overall survival (p value = 1.7e-5). We detected chromothripsis (>=7 oscillating copy number change and significant clustered SVs and/or clustered translocations) in 24% of newly diagnosed samples; and kataegis hotspots on chromosome 3q27-3q28 (24%), 11q13 (5.8%) and 12q24 (5.3%). By clustering SV breakpoints across the genome we have identified 3 SV hotspots on chromosome 17q21, 7q34, and 11q13. We next interrogated the non-coding regions to identify genomic loci with higher than expected mutation count compared to background mutation rate. We have identified 456 loci that are significantly enriched in non-coding regions (5' UTR, 3'UTR, promoter, intergenic, intronic, and distal regulatory regions) [adjusted p value < 1e-5 and observed in >=10% newly diagnosed MM]. These loci are then assigned to genes or gene neighborhoods to evaluate their potential impact. We have identified the most frequently involved genes affected by perturbation in neighboring non-coding region and integrate their expression using our matching deep RNA-seq data from the same patients. Of these the most prominent examples are 1.) 3'UTR mutations are enriched in CD93 gene, which plays critical role in B cell development with loss of expression in CD138+ MM cells compared to normal plasma cells (p value < 1e-5); 2.) Promoter region - we have identified 635 mutations in 2kb region in BCL6 coming from 76% of all newly diagnosed samples. BCL6 (p value < 1e-5) has significantly downregulated expression in MM. Interestingly, but not surprisingly this hypermutated region showed high intensity of H3K27Ac activity in normal cells; 3.) 5'UTR - BCL7A (27.9%) and LPP (11.7%) were top two 5' UTR mutated target genes and RNA-seq data confirmed significant downregulation of their expression (p values < 1e-5 and 0.0048 respectively) in the MM cells. Additionally, BCL7A (48%) also showed significant enrichment of intronic mutations. A similar mutational hotspots were observed within the vicinity of additional functionally important genes in myeloma including ROBO1/2, ILF2, IRF8 and BCL2A1. Our data also showed that these frequent mutations have higher cancer cell fraction (CCF) [median CCF > 0.75] suggesting their occurrence earlier in the disease development. To validate the function of these mutations, we have started to carry out gain/loss of function studies. Our analysis with BCL7A shows that BCL7A knockdown increases the cell viability while its overexpression decreased growth, colony formation and increased apoptosis. This tumor suppressor function of BCL7A is being further analyzed in light of our mutational data in the nearby non-coding region. In conclusion, this large deep whole genome sequencing data from newly-diagnosed MM patients identifies a vast majority of non-coding mutations with potentially significant functional and biological role in MM. Our integrative approach using both WGS and RNA-seq data from the patients now provides us important tools to further characterize the impact of these mutations and develop opportunities for targeted therapeutics. Disclosures Richardson: Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding; Oncopeptides: Membership on an entity's Board of Directors or advisory committees; BMS: Research Funding; Amgen: Membership on an entity's Board of Directors or advisory committees; Janssen: Membership on an entity's Board of Directors or advisory committees; Jazz Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees, Research Funding; Takeda: Membership on an entity's Board of Directors or advisory committees, Research Funding; Karyopharm: Membership on an entity's Board of Directors or advisory committees. Moreau:Janssen: Honoraria, Membership on an entity's Board of Directors or advisory committees; Takeda: Honoraria, Membership on an entity's Board of Directors or advisory committees; Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees; Amgen: Honoraria, Membership on an entity's Board of Directors or advisory committees; Abbvie: Honoraria, Membership on an entity's Board of Directors or advisory committees. Thakurta:Celgene Corporation: Employment, Equity Ownership. Anderson:Millennium Takeda: Consultancy; Gilead: Membership on an entity's Board of Directors or advisory committees; Bristol Myers Squibb: Consultancy; OncoPep: Equity Ownership, Other: Scientific founder; C4 Therapeutics: Equity Ownership, Other: Scientific founder; Celgene: Consultancy. Avet-Loiseau:Amgen: Consultancy, Membership on an entity's Board of Directors or advisory committees, Research Funding; Abbvie: Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees, Research Funding; Sanofi: Consultancy, Membership on an entity's Board of Directors or advisory committees, Research Funding; Janssen: Consultancy, Membership on an entity's Board of Directors or advisory committees; Takeda: Membership on an entity's Board of Directors or advisory committees, Research Funding. Munshi:OncoPep: Other: Board of director.

Download Full-text