SF3B1 Mutations Research Articles

Molecular Characterization of Clinical Response in Relapsed/Refractory Acute Myeloid Leukemia and High-Risk Myelodysplastic Syndrome Patients Treated with Single Agent Emavusertib

Introduction Acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) are myeloid malignancies that exhibit a dynamic mutational landscape as the disease progresses. Genetic mutations in the splicing factors SF3B1 and U2AF1 drive overexpression of a highly active long isoform of interleukin-1 receptor-associated kinase 4 (IRAK4), which is critical in triggering inflammation, oncogenesis, and survival of cancer cells. Splicing factor mutations are frequently reported in AML and MDS and have been associated with disease progression and poor prognosis in relapsed/refractory (R/R) diseases. Emavusertib is a potent oral inhibitor of IRAK4 as well as Fms-like tyrosine kinase 3 (FLT3). The safety, clinical activity, and potential biomarkers of emavusertib in AML or high-risk MDS (HR-MDS) are investigated in the ongoing open-label, phase 1/2a TakeAim Leukemia trial (NCT04278768). Methods Eligible patients with R/R AML or HR-MDS were enrolled and treated with emavusertib. The exploratory objectives of the study are to assess the association between clinical activity and target-related biomarkers, gene expression signatures, and molecular subtypes. Bone marrow or peripheral blood were collected at the baseline and on treatment. To characterize the mutational profiles of enrolled patients and further define the targeted population, targeted next generation sequencing (NGS) of 68 genes was performed on genomic DNA from bone marrow or peripheral blood mononuclear cells. A minimum coverage of 100x, variant allele frequency (VAF) of >1%, and a VARSCAN-assigned variant p<0.001 were used as thresholds for single nucleotide variants (SNV) calling and presumed somatic driver mutations were identified as previously reported (Metzeler et al. Blood 2016). Mutations from patients’ molecular pathology reports were also documented. Results As of December 16, 2021, 49 patients have been treated with emavusertib monotherapy at one of dose levels of 200, 300, 400, or 500 mg BID, including 30 R/R AML and 19 R/R HR-MDS. The median number of prior anti-cancer therapies was 2 (range 1-5). This abstract is intended to add additional molecular findings to the patient subset presented in 2022 ASCO, of which 6 AML and 7 HR-MDS had splicing factor SF3B1 or U2AF1 mutations and 3 AML had FLT3 mutations, including 2 AML with both splicing factor and FLT3 mutations (hereafter referred to as "targeted mutations"). In the patients with targeted mutations, with available mutational profile data or molecular reports from sites, there were 6 responders: 1 CR, 1 CRh, and 4 mCR. In this responder population, the most common co-mutation was ASXL1. During emavusertib treatment, FLT3-mutated clones became undetectable in 2 of 3 (67%) FLT3-mutated AML patients. One of the 3 patients (33%) with a FLT3 mutation achieved CR and the remaining 2 patients achieved SD (1 with > 90% blast reduction). Additional gene mutations, such as ASXL1, RUNX1, CEBPA, ETNK1, ETV6, and SRSF2, became undetectable during emavusertib treatment. New mutations that emerged after ≥ 1 cycle of emavusertib treatment include NRAS, SETBP1, RUNX1, WT1, STAG2, KRAS, CCND2, DNMT3A, ASXL1, and NOTCH1. Analysis of genomic and transcriptomic profiles will be updated. Conclusion Emavusertib demonstrated anti-cancer activity in heavily pretreated AML and HR-MDS patients, especially in those with U2AF1, SF3B1, or FLT3 mutations. Our preliminary NGS data is suggestive of molecular responses and disease-modifying activity of emavusertib. Patients with targeted mutations responded to emavusertib, even in the presence of co-mutations of ASXL1 that are typically associated with poor prognosis. We will continue to explore potential genomic biomarkers of emavusertib sensitivity and resistance

Activation of Distinct Inflammatory Pathways in LR-MDS Is Determined By Genetics

Background Dysregulated innate immune signaling has been identified as a driver of the complex pathophysiology of myelodysplastic syndromes (MDS), predominantly in low risk (LR)-MDS (Basiorka et al., Blood 2016). The pharmacological targeting of inflammation may therefore help to retard disease development. However, since MDS is a heterogeneous disease, a personalized approach is likely to be necessary, requiring a more detailed knowledge of distinct inflammatory patterns and their variation between genetically defined subgroups of LR-MDS patients (Winter et al., J Clin Oncol 2020). Aims The aim of this work was to describe inflammation states in the bone marrow of LR-MDS patients (IPSS-R ≤ 3.5) compared to controls with and without clonal hematopoiesis of indeterminate potential (CHIP) and to high-risk (HR)-MDS patients (IPSS-R ≥ 4). Methods Cryopreserved bone marrow mononuclear cells (BM-MNCs) and corresponding bone marrow plasma samples from non-CHIP controls (n = 15), CHIP individuals (n = 12) and patients with LR-MDS (n = 47) or HR-MDS (n = 14) were obtained from the MDS registry and the BoHemE study (NCT02867085) at the University Hospitals in Dresden and Leipzig. Targeted expression analysis of 11 inflammasome-related genes (S100A9, NLRP3, PYCARD, CASP1, IL1B, IL18, NLRP1, NLRC4, AIM2, CASP3, CASP4) and multiplex cytokine measurements were performed. To investigate heterogeneity in the LR-MDS cohort, a principle component analysis (PCA) was performed and a cumulative gene expression score was calculated as summed log2 fold changes (mean non-CHIP = 0) of all 11 mRNA expression values for each sample. Furthermore, BM-MNCs of selected non-CHIP (n = 3) and CHIP (n = 3) individuals as well as LR-MDS (n = 14) and HR-MDS (n = 5) patients were FACS-sorted for hematopoietic stem and progenitor cells (HSPCs), monocytes, monocytic and polymorphonuclear myeloid-derived suppressor cells (MDSCs) and lymphocytes. The expression of IL1B, IL18, S100A9, NLRP3, PYCARD, CASP1 and NLRC4 was assessed in each population by qRT-PCR. Finally, colony forming assays were used to assess the effect of the IL-1β-neutralizing antibody canakinumab (100 µg/ml) on the hematopoietic potential of healthy CD34+ HSPCs pre-cultured with CD14+ monocytes from patients with SF3B1-mutated (n = 3) or del(5q) (n = 3) LR-MDS. Results The expression of inflammasome-related genes in the bone marrow increased from non-CHIP through CHIP to LR-MDS individuals, with overall levels of inflammasome-related gene expression being 40% higher in LR-MDS than in non-CHIP controls (p = 0.002). The precise signature of inflammation gene expression varied widely within LR-MDS patients. A PCA identified two separable inflammation states with low (cluster 1, n = 24) and high (cluster 2, n = 23) inflammation patterns (Figure 1), with CASP1, PYCARD and NLRC4 contributing most to the PCA dimensions. The clusters were associated with distinct clinical features and disease genotypes: more than 80% of the SF3B1 mutations were found in cluster 1, while all del(5q) cases were in cluster 2 (Figure 1). Compared to del(5q) samples, those with SF3B1-mutation showed lower expression of IL-1β at both mRNA (3.9-fold decrease, p = 0.0005) and protein level (8.3-fold decrease, p = 0.03). Targeted gene expression analysis of sorted cell populations revealed distinct contributions to the differential patterns of inflammation. IL1B, NLRP3, NLRC4 and S100A9 were expressed in all disease states predominantly by monocytes and monocytic MDSCs, consistent with a dominant role for monocytic cells in determining the inflammatory bone marrow environment. However, IL18 mRNA levels were highest in HSPCs. Compared to untreated controls, canakinumab augmented the colony forming activity of healthy HSPCs pre-cultured with monocytes from SF3B1-mutated MDS (+70%, p = 0.002). The effect in co-cultures with del(5q) MDS-derived monocytes was weaker (+56%, p = 0.188), showing that canakinumab efficacy in this assay does not depend on IL-1β expression level alone. Conclusion We conclude that inflammation patterns in LR-MDS are heterogeneous, highlighting the need for personalization of targeted anti-inflammatory therapies currently entering clinical trials. Our analysis suggests that high resolution studies of gene expression in specific populations or single cells will further resolve a spectrum of inflammation states relevant to prognosis and personalized therapy. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Assessment of the Frequency of Variants of Undetermined Significance (VUS) in Adult Patients with Chronic Idiopathic Neutropenia Studied for Myeloid Gene Mutations By Next Generation Sequencing

Background: Chronic Idiopathic Neutropenia (CIN) is a neutrophil disorder characterized by the persistent and unexplained reduction in the number of peripheral blood (PB) neutrophils. The diagnosis of CIN is based on exclusion criteria, namely the absence of clinical/laboratory evidence of any underlying disease or use of drugs that might be associated with neutropenia, absence of history of exposure to irradiation or use of chemical compounds, negative antineutrophil antibodies, and normal bone marrow (BM) morphology and karyotype. We have previously performed next generation sequencing (NGS) analysis of genes recurrently mutated in myeloid malignancies in a cohort of CIN patients and we found that 11.35% of the patients display clonal hematopoiesis in term of presence of pathogenic/likely pathogenic variants with VAF higher than 2%. Variants of undetermined significance (VUS) were ruled out from the analysis and their potential significance was not assessed. Aims: In the current study we have assessed the presence, frequency and potential clinical significance of VUS detected in our cohort of CIN patients previously studied for clonal hematopoiesis. Methods: Genomic DNA was extracted from CIN patients' BM or PB samples, sequencing libraries were prepared and subjected to targeted NGS on an Ion S5 Prime Sequencer (Thermo Fisher Scientific) using a panel of 38 genes recurrently mutated in myeloid malignancies (Tsaknakis G et al. Blood. 2021;138:1249-1257). Results: NGS analysis of myeloid genes has been performed in 139 CIN patients (106 female subjects and 33 male subjects, aged 21 to 89 years with a median age of 59 years). Re-analysis of patients' NGS data and assessment of VUS revealed that 35 out of 139 patients (25.18%) carried one or more VUS (a total of 39 VUS were detected, all with an allelic frequency of 50% and all missense variants). Genetic variants were evaluated according to the ACMG guidelines and classified as VUS. The majority of VUSs were observed in ATM (n=7) and SETBP1 (n=6) genes followed by TET2, NOTCH1 (4 occurrences respectively) and PLCG2 (n=3). Two different patients shared the same variant in TET2 gene whereas two siblings shared the same variant in SETBP1 gene. Pathogenicity score calculation based on the combined evidence from multiple other in-silico predictors identified 3 variants in 3 different genes as having the highest probability to be pathogenic. Namely these were PLCG2K456E,CXCR4R235C and SF3B1T56I. PLCG2 variants have been implicated in familial cold autoinflammatory syndrome 3 (PLCG2 associated antibody deficiency and immune dysregulation, PLAID) and in autoinflammation, antibody deficiency, and immune dysregulation syndrome (APLAID) whereas CXCR4 variants are known to be associated with Warts, hypogammaglobulinemia, infections, and myelokathexis (WHIM) syndrome. SF3B1 mutations have not yet been shown to be involved in hereditary disease. Conclusions: Our study of VUS assessment in a cohort of 139 well characterized adult CIN patients identified 39 VUS, all heterozygous, in a total of 35 patients. Further steps include paired analysis with constitutional reference tissue to determine the germline nature of the mutations. Moreover, these variants should be functionally tested and integrated with functional results and clinical phenotypes to more accurately interpret their potential contribution to disease onset. This study of VUS analysis in CIN patients is anticipated to contribute to the better understanding of the pathogenetic causes of this rare disease.

Molecular Landscape of Chronic Lymphocytic Leukemia Using Targeted Gene Panel Sequencing

Introduction Chronic lymphocytic leukemia (CLL) is a clinically heterogeneous disease with various genetic abnormalities that can affect the treatment and survival of patients. Recently, by the advent of next-generation sequencing (NGS) techniques recurrently mutated genes have been described in CLL. Driver gene mutations such as NOTCH1, SF3B1, TP53, and MYD88 have been considered to have prognostic impact in CLL. The aim of this study was to detect pathogenic variants of commonly mutated genes in CLL patients and to explore their associations with IGHV mutational status and cytogenetic abnormalities. Patients and Methods The samples were collected from bone marrow or peripheral blood of 124 CLL patients. Genomic DNA was isolated using QIAamp Blood Mini kit. NGS analysis was performed using Twist Custom Panel and bidirectional sequencing with a minimum coverage of 1000x on Illumina NextSeq according to the manufacturers’ instructions. Data were analyzed by NextGene software. The following 27 recurrently mutated CLL genes were studied: PIK3CD, PIK3CA, SF3B1, MYD88, FBXW7, BRAF, NOTCH1, PTEN, BIRC3, PTPN6, MAP2K1, TP53, STAT5B, STAT3, CD79B, TCF3, MAPK1, DDX3X,ATM, BCL2, BTK, KRAS, NRAS, CARD11, CXCR4, PLCG2, POT1. Variants below 5% variant allele frequency (VAF) were filtered out from the analysis. Variants were considered subclonal with 5-10% VAF and clonal when VAF was >10%. Classification of variants were performed according to the American College of Medical Genetics and Genomics (ACMG) criteria. Results Our NGS study revealed a total of 96 variants in 124 CLL patients. Overall, 63/124 patients (50.8%) harbored at least one mutation, the majority of them (60/96, 62.5%) was clonal. In 36 patients only one pathogenic variant was detected, while 88 patients carried two or more (2-4) genetic alterations. The most frequently mutated gene was the NOTCH1, followed by TP53, SF3B1, FBXW7, and BRIC3. Subclonal variants were also detected in 36/124 patients (29%), particularly in NOTCH1,FBXW7, BIRC3, and PIK3CA genes. TP53 and NOTCH1 variants occured mostly in unmutated IGHV CLL cases. Patients with mutated IGHV harbored mainly FBXW7 and BIRC3 mutationss. FBXW7 variants were detected in 9/125 (7%) patients, that is a higher frequency compared to other studies. Fluorescence in situ hybridization (FISH) was performed in 113 cases. The majority of these patients (84 cases, 74%) had one or more cytogenetic abnormality. In this cohort, 13q deletion was the most common aberration (65%, 55/84 patients), followed by 11q deletion (26%, 22/84 patients) and 17q deletion (14%, 12/84 patients). Deletion of 17p co-occured with TP53 single nucleotide variants in 10 of 12 patients. SF3B1, NOTCH1, BIRC3 and TP53 variants occured with 11q deletion in 10 cases. Trisomy 12 was detected with NOTCH1, BIRC3 and FBXW7 variants, and showed negative association with unfavorable SF3B1 mutations. Conclusions Our results confirmed the presence of high number of clonal and subclonal variants in the most frequently mutated genes (NOTCH1, TP53, SF3B1, FBXW7, BIRC3). TP53, NOTCH1 and SF3B1 mutations were enriched with unfavorable biologic factors, as they were found to be associated with unmutated IGHV and cytogenetic aberrations, such as del17p, del11q and trisomy 12. Cases with NOTCH1 variants harbored all types of the investigated cytogenetic alterations (del13q, del11q, de17p, trisomy 12) and were associated exclusively with unmutated IGHV. TP53 variants occured mostly in unmutated IGHV CLL cases together with del17p. Patients with SF3B1 variants showed del11q, del13q and mostly unmutated IGHV. FBXW7 variants enriched in both of unmutated and mutated IGHV cases and harbored del13q, de17p and trisomy 12. FBXW7 and NOTCH1 pathogenic variants have the same biological consequences in CLL, therefore presence of FBXW7 mutations may have clinical relevance regarding anti-CD20 therapy. Taken together, these NGS results complemented with the study of IGHV mutational status and cytogenetic data can contribute to better prognostic workup and management of our CLL patients.

Validation of the 5th Edition of the World Health Organization Classification of Myelodysplastic Neoplasms on 852 Consecutive De Novo Patients from a Single Institute

Background: A summary of the World Health Organization (WHO) 5th edition (2022) classification of myelodysplastic neoplasms (MDS) has been published recently. As compared to the 2016 (revised 4th) WHO criteria, the newest proposal places greater emphasis on genomic and pathology features. In this study, we assess the validity of the WHO 2022 classification for MDS patients. Methods: Data from 852 consecutive subjects ≥ 18 years with newly-diagnosed MDS according to the WHO 2016 criteria in our center from August 30, 2016 to September 22, 2021 were interrogated. Subjects were re-diagnosed and re-classified according to WHO 2022 guidelines. The last follow-up was on June 4, 2022, with a median follow-up of 24 months. Results: Applying the WHO 2022 guidelines, 30 subjects with an NPM1 mutation were re-classified as acute myeloid leukemia (AML) who were previously classified as MDS with excess blast type2 (MDS-EB2; n = 13), MDS with multi-lineage dysplasia (MDS-MLD; n = 9), MDS with excess blast type1 (MDS-EB1; n = 6) and MDS unclassifiable (MDS-U; n = 2) according to WHO 2016 criteria. Nine subjects previously classified as (MDS-U) were re-classified to clonal cytopenia of undetermined significance (CCUS). Re-classification of remaining 813 subjects between the WHO 2016 to WHO 2022 classification are displayed in Figure1. Classification of the 11 subjects with MDS with low blasts and isolated del(5q) (MDS-5q) remained unchanged. In addition to prior MDS with ring sideroblasts (MDS-RS) patients (n=45), 25 subjects were reclassified as MDS with low blasts and SF3B1 mutation (MDS-SF3B1) because the new criteria have no limitation on numbers of ring sideroblasts. We re-classified 53 subjects as MDS-biTP53, most commonly those with excess blasts (34/53; 64%). Amongst subjects without the genetic abnormalities defined above, 80 previously classified as MDS with singe/multi-lineage dysplasia (MDS-SLD/MLD) or MDS-U were re-classified as hypoplastic MDS (MDS-h) and the remaining 293 as MDS with low blasts (MDS-LB). Subjects previously classified as MDS-EB1 or EB2 were re-classified as MDS with increased blasts type1 (MDS-IB1; n = 161) or MDS with increased blasts type2 (MDS-IB2; n = 103) after those with MDS with fibrosis (MDS-f; n = 42) were excluded. Survival according to WHO 2022 classification for MDS are shown in Figure2. Median survival of MDS subjects was 45 months (95% Confidence Interval [CI], [34, 56] months). There were significantly briefer median survivals of subjects with MDS-biTP53 (10 months [8, 12] months) and MDS-f (15 months [8, 23 months]) compared with other subtypes. Median survivals of MDS-h and MDS-LB cohorts were unreached but those with MDS-h had longer survival (P = 0.07). Median survivals for MDS-IB1 and MDS-IB patients were 24 months ([18, 30] months) and 26 months ([17, 35] months), respectively. Survival rate between MDS-IB1 and MDS-IB2 patients were comparable (P=0.707). We next explored the prognostic value of myelofibrosis in subjects without increased blasts. 40 subjects in our cohort had grade-2/-3 myelofibrosis. 37 were re-classified from the MDS-LB cohort and 3 from the MDS-h cohort. Median survival of subjects with low blasts and myelofibrosis (MDS-LBf) was 23 months (8, 38 months), significantly briefer compared with those with MDS-LB or MDS-h (unreached; P < 0.001) comparable to subjects with MDS-IB1 (24 months; P = 0.80) or MDS-IB2 (26 months; P = 0.79) and longer compared with subjects with MDS-f (15 months; P = 0.04). We also tested the prognostic value of dysplastic lineages in subjects with MDS-LB according to WHO 2022 classification. Based on bone marrow histology 28 subjects were considered MDS-LB with single lineage dysplasia (MDS-LB-SLD) and 228, MDS-LB with multiple lineage dysplasia (MDS-LB-MLD). Median survivals of both cohorts were unreached however MDS-LB-SLD had longer survival (P = 0.01). Based on these findings, we suggest the revised WHO 2022 classification divide subjects with low blasts in to MDS-LBf, MDS-h, MDS-LB-SLD and MDS-LB-MLD. The first supersedes other categories and MDS-h overrides the diagnosis of MDS-LB-SLD or MDS-LB-MLD. Conclusion: We validate the WHO 2022 classification of MDS as thoughtful and propose some refinements to provide more accurate prognostic information. Because our study is from 1 center, external validation is needed. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Downregulation of UBA1 in Myelodysplastic Syndrome

Somatic loss-of-function mutations of major E1 ubiquitin activating enzyme gene UBA1 were identified in VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) and myelodysplastic syndromes (MDS) recently, suggesting that impairment of UBA1 activity contributes to pathogenesis of these diseases that have systemic inflammatory disorders. We sought to determine whether abnormal expression of UBA1 also occurs in MDS and contributes to development of disease. We analyzed RNA sequencing results obtained from the CD34+ bone marrow hematopoietic stem and progenitor cells (BM HSPC) in a cohort containing treatment naive MDS (N=55), chronic myelomonocytic leukemia (CMML, N=22), and heathy donors (N=12). Since UBA1 is an X-chromosome linked gene that escapes X-chromosomal silencing in females, all donors in this cohort were males. Results indicated that the level of UBA1 RNA was significantly downregulated in MDS but not in CMML. In the subtypes of MDS-low blasts (N=23, including MDS with ring sideroblasts, multilineage dysplasia, and single lineage dysplasia), forty-eight percent (N=11) had an over 1.5-fold downregulation of UBA1 RNA expression. Based on gene set enrichment assay, the UBA1 downregulation in MDS-low blasts was associated with altered autophagy and innate immune signaling such as complement, IL6, and TNF signals. Moreover, we also identified that the UBA1 downregulation in BM CD34+ cells of MDS-low blasts co-occurred with SF3B1 mutations detected by targeted next generation sequencing in BM mononuclear cells: Among the 11 MDS-low blasts patients with downregulated UBA1 expression, six patients (55%) carried mutant SF3B1, whereas none of the 12 patients with normal UBA1 expression had SF3B1 mutation. In another word, all SF3B1 mutant patients with MDS-low blasts (N=6) presented decreased UBA1 RNA. To characterize whether impairment of UBA1 affects hematopoietic functions of BM HSPC, we applied TAK-243, an UBA1 specific inhibitor, to BM CD34+ cells isolated from healthy donors (N=5, all males) and cultured in vitro. Colony formation unit (CFU) assays were performed to evaluate hematopoietic repopulating activity. Results indicated the CFU in CD34+ cells treated by TAK-243 was significantly reduced compared with cells treated by vehicle control (P<0.05). Moreover, when hematopoietic colonies formed by CD34+ cells were collected from methylcellulose plates and characterized by flow cytometry analysis, the frequency of CD71-/ CD235+ erythroid cells in TAK-243 treated plates was significantly lower than that in control plates (P<0.05). These results implicated that impairment of UBA1 activity perturbed hematopoietic functions of BM HSPC, particularly the erythropoietic function. We next evaluated in vivo hematopoietic impact of UBA1 impairment by applying TAK-243 to wildtype male B6 mice. Compared to vehicle control treated mice (N=4), mice treated by TAK-243 (N=4) exhibited significant decreases in hemoglobin and white cell counts (P<0.05) in peripheral blood (PB) and splenomegaly. In BM compartment, TAK-243 treated mice had decreased CD71+ and Ter119+ erythroid populations and increased Gr1+/ CD11b+ myeloid population (P<0.05). TAK-243 treatment also altered BM HSPCs, including reduced megakaryocyte-erythroid progenitor (MEP) and increased granulocyte-monocyte progenitor (GMP) in BM, as well as reduced CD150+/ CD48- stem cells in the BM Lin-/ Sca1+/ cKit+ (LSK) population. Moreover, BM LSK isolated from TAK-243 treated mice (N=4) formed significantly reduced number of colonies compared to vehicle control treated mice (N=4) in CFU assay. Consistently, in competitive transplantation assay, repopulating activity of BM cells of TAK-243 treated mice was significantly impaired, as measured by the donor-derived CD45.2 chimerism in PB of the CD45.1 recipients. Moreover, we performed bulk RNA-Seq in BM LSK of drug treated mice, which revealed that the ribosomal function and cell cycle regulation signals in TAK-243 treated mice were significantly affected. In summary, our study indicated that, in addition to somatic mutations, there was also a significant downregulation of UBA1 expression in BM HSPC of MDS. Compromised UBA1 function impairs hematopoietic function of BM HSPC, which could be an important mechanism for pathogenesis of MDS.

Luspatercept in Combination with Recombinant-Erythropoietin in MDS RS Patients: Stimulating Early and Late Erythropoiesis

Myelodysplastic syndrome with ring sideroblasts (MDS RS) and SF3B1 mutation is generally marked by severe anemia, limited response to recombinant erythropoietin (rEPO) and consequent transfusion burden. Recently, luspatercept, a TGF-beta-pathway inhibitor, was shown beneficial in approximately 1/3 of patients by reactivating late-stage erythropoiesis. Since rEPO mainly acts on early-stage erythropoiesis, we hypothesized a possible role of combination therapy in this subset of patients. Seven patients with MDS RS treated with luspatercept in combination with rEPO at two tertiary hematologic centers in Milan, Italy were retrospectively evaluated. Luspatercept was administered according to current data sheet (1 to 1.75mg/kg every 3 weeks). Endogenous (e)EPO was evaluated before rEPO start, and patients with values < 200 IU/L received subcutaneous epoetin alpha 40,000 U/week, while those with higher values received 80,000 U/week. Hematologic data and number of RBC transfusions were collected at each cycle of luspatercept, and hematologic improvement (HI) was assessed according to the revised International Working Group (IWG) 2018. All 7 patients were males, with a median age of 83 years old (70-90). According to IPSS-R, three patients were classified as intermediate risk, while the others were lower risk. At the time of luspatercept start all patients were anemic with median Hb values of 7.3 g/dL (6.1-9.9), while median neutrophil (3.1x103/mm3, 0.8-4.9) and platelet (200x103/mm3,43-282) counts were normal. The median number of RBC units transfused in the last 8 weeks before luspatercept was 8 (3-11). Figure 1 details the timing and doses of luspatercept and rEPO treatment. In total, 5 patients received rEPO as add-on treatment due to non- or loss of response to luspatercept. Median eEPO values in these subjects were 250 (99-445) IU/L. Overall, three patients (Case 1, 2 and 3) achieved HI, and one of them became transfusion independent. Specifically, Case 1 achieved minimal HI with luspatercept after cycles 2 and 6, and lost response; rEPO was added at cycle 12 with HI after cycle 14. Case 2 initially reached HI after cycle 4 but showed several oscillations of pre-transfusion Hb (7 to 9 g/dL). At cycle 13 rEPO was added obtaining HI after cycle 20. Case 3 achieved HI after 2 cycles of luspatercept but lost response after cycle 10; rEPO was added at cycle 14 and he became transfusion independent. Patient 6 had rEPO as a pre-existing therapy and experienced a minimal HI after luspatercept start; rEPO had not been stopped to avoid a drop of pre-transfusion Hb. The other patients did not respond to luspatercept, nor to combination with rEPO, which was added at cycles 14 and 7 (Case 4 and Case 5), or already ongoing at luspatercept start in Case 7. Notably, no treatment emerging adverse events were registered with the combination. In conclusion, despite the small number of cases, these data show for the first time that combination therapy of luspatercept plus rEPO is feasible and beneficial in about one third of subjects. Concomitant stimulation of early and late-stage erythropoiesis may be an interesting strategy that requires further investigation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Abstract A055: Altered RNA splicing causes pancreatic cancer and exposes a therapeutic vulnerability

Abstract With the aim to identify new second-hit mutations that cause pancreatic ductal adenocarcinoma (PDAC), other than mutations in p53, we performed a genetic interaction analysis based on the concept of mutually exclusive mutations, which suggests that mutations that are exclusive to one another function within the same pathway. Using our unbiased quantitative methods to evaluate the effects of somatic mutations on cancer we analyzed &amp;gt;3000 PDAC cases to prioritize mutations among hundreds that appear at low frequencies (5-10%) and that are mutually exclusive with mutations in p53. We found that mutations in RNA splicing factors SF3B1 and RBM10 (~15% of cases), were among the most significant, and mutually exclusive, to mutant p53. Thus, we hypothesized that aberrant RNA splicing promoted by SF3B1K700E and loss of RBM10 cause an oncogenic program that leads to tumorigenesis and therapy resistance. To test this hypothesis, first, we engineered mouse models to co-express KrasG12D with either Sf3b1K700E or loss of Rbm10 in Mist1 and Pdx1 driven expressing pancreatic cells. Interestingly, mice harboring KrasG12D and Sf3b1 K700E or loss of RBM10—like mutant p53—caused PDACs in mice. To identify the RNA splicing mechanism that leads to PDAC formation, we performed deep RNA sequencing and unbiased transcriptome-wide splicing analyses in isolated PDAC cells bearing wild-type, SF3B1K700E and loss of RBM10, derived from murine and patient PDACs. We found that like the splicing defects induced by p53mut—SF3B1K700E or loss of RBM10—induce splicing defects in sequence-specific coding exons in mRNAs derived from chromosomes 19, 7, 1 and 2 and impacting mainly GTPase activation and lipid metabolism pathways across species models. Consistent with this common mechanism, we found that PDAC cells do not tolerate the forced expression of mutant p53 with either SF3B1K700E or loss of Rbm10, supporting the mutual exclusivity findings from patient samples. To determine a personalized therapy pathway for patients bearing mutant splicing-factor tumors, we found that PDAC cells with SF3B1K700E or RBM10 loss cells are more sensitive to Gemcitabine rather than 5FU, which are the two main chemotherapeutic components of the two standard-of-care therapies. In addition, these mutations sensitize cells to small-molecule splicing inhibitor H3B-8800, which has advanced to phase 2 trial for hematologic malignancies. Lastly, we found that combination of Gemcitabine and H3B-8800 synergize to selective kill mutant splicing cells. We are currently testing these combination therapies in our genetically engineered mouse models and patient derived organoids. We are starting a phase 2 trial for Gemcitabine/Abraxane with escalating doses of H3B-8800 for tumors that have either SF3B1K700E, RBM10 loss or neomorphic p53, all which have aberrant RNA splicing. This work has determined novel drivers and mechanisms of PDAC development and a precision therapeutic strategy for the treatment of 45% pancreatic cancer patients (SF3B1K700E [5%], RBM10 loss [10%] and neomorphic p53 [30%]). Citation Format: Natasha Pinto Medici, Diana Martinez Saucedo, Danny Lee, Li-Ting Ku, Robert Tseng, Vincent Cannataro, Deanne Yugawa, Sumedha Chowdhury, Jeffrey Townsend, Christine A. Iacobuzio-Donahue, Omar Abdel-Wahab, Steven D. Leach, Luisa Escobar-Hoyos. Altered RNA splicing causes pancreatic cancer and exposes a therapeutic vulnerability [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer; 2022 Sep 13-16; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2022;82(22 Suppl):Abstract nr A055.

Development of a Novel Chromatin Activation Score As a Powerful Independent Prognostic Factor in Chronic Lymphocytic Leukemia

Background & Aim: In a previous study, we identified a signature of 498 regulatory regions with de novo increased H3K27ac levels (a mark of chromatin activation) in patients with chronic lymphocytic leukemia (CLL), which seemed to be present in all cases studied (Beekman et al., Nat Med 2018). In this abstract, we show that such de novo chromatin activation is not stable among CLL patients but heterogeneous, and therefore, we aim at exploring whether such heterogeneity has a biological and clinical impact. Results: The signature of 498 de novo active regulatory regions was related to 434 specific H3K27ac peaks based on chromatin immunoprecipitation sequencing (ChIPseq) data from 104 CLLs. Analyzing the signal variability among patients, we observed a high degree of heterogeneity. An unsupervised clustering revealed two different sets of regions whose mean H3K27ac signal was associated with time to first treatment (TTFT) using a univariate Cox model. In detail, a higher mean H3K27ac intensity was associated with a shorter TTFT in one of the sets (P=0.001 and hazard ratio (HR)=3.5) while, intriguingly, showed a trend towards longer TTFT (P=0.09 HR=0.6) in the other set. These findings suggest that not only disease progression but also disease stability seem to be active processes in CLL involving de novo chromatin changes. Then, we argued that the clinical behavior of CLL patients could be better captured by the balance between the activation levels of these two sets of regions. Indeed, the ratio of both means was statistically more significant (P=1.4x10-9 HR=15.4). Next, we further refined these groups by selecting individual regions associated with shorter (HR>1, 77 peaks) or longer (HR<1, 144 peaks) TTFT (FDR 0.2). We subsequently defined the score SPP as the mean signal in the pro-progressive (PP) group (HR>1), the score SPS as the mean signal in the pro-stable (PS) disease group (HR<1), and SB as their ratio. Ultimately, this led to an improvement of the statistical significance of the SB with TTFT (P=1.4x10-14 HR=8.6). From a biological perspective, we explored the validity of the SB performing ChIPseq for H3K27ac in 10 primary CLLs before and after a co-culture system inducing strong proliferation (Mangolini et al., Nat Comms in press). In these experiments, we observed an increased SB (P<10-3) that was mostly associated with a decreased SPS while maintaining the SPP levels. To further validate the biological implications of our findings, we mined RNAseq data in 64 cases with concurrent ChIPseq experimentation. We identified target genes whose expression levels correlated with H3K27ac levels (43 genes associated with the PP regions and 62 target genes associated with the PS regions). Both lists included genes previously described in CLL (e.g. CTLA4, FMOD or KSR2). Analogously to H3K27ac, we defined three new scores based on expression levels: SPP-exp, SPS-exp and SB-exp. These scores highly correlated with the H3K27ac-based scores (Pearson r between 0.82 to 0.88). Interestingly, cases with higher SB-exp were enriched in TNF-α and NF-κB pathways and in RELB, FOS, JUN, JUNB and EGR1 transcription factors regulons. We also identified 21 genetic driver alterations related to a higher SB-exp including trisomy 12, 11q loss and mutations in IGLV3-21R110, ATM, SF3B1, NOTCH1 and U1. Finally, we evaluated the potential of SB and SB-exp as independent prognostic factors in multivariate models. In the ChIPseq series (n=104), SB was significant independently of the IGHV mutational status, Binet stage and number of genetic driver alterations. Similarly, using three independent CLL series of gene expression data (n=110, 174 and 140), we observed that the SB-exp was a highly significant prognostic factor independent from IGHV status, Binet stage and number of drivers. To favour potential clinical translation, we generated a reduced SB-exp with only 12 genes which maintained the clinical impact. Conclusions: This study revealed an unexpected heterogeneity within de novo active chromatin regions in CLL, and that such heterogeneity has a dual clinical impact on CLL. A balance score between regions related to the two tendencies (i.e. more progressive vs more stable disease) identified groups of patients with different biological features. Finally, using both chromatin and gene expression data as a surrogate, the balance score was a highly significant independent prognostic factor in independent CLL series.

The Growth Kinetics of Clonal Hematopoiesis Inform Risk of Blood Cancer Progression

Introduction: Clonal hematopoiesis of indeterminate potential (CHIP) is a frequent consequence of aging associated with increased risk of blood cancers, although overall individual risk is < 1% per annum. Proposed risk factors for blood cancer progression include a variant allele frequency (VAF) > 10%, specific gene mutations or the presence of more than one CHIP clone. However, the relative value of monitoring CHIP clone growth with respect to blood cancer risk is unknown. To address this evidence gap, we measured CHIP at baseline and 3 years later in 11,298 healthy individuals aged ≥70y from the ASPREE randomized clinical trial (www.aspree.org). Methods: A custom, high sensitivity (0.5%) and reproducible amplicon-based targeted sequencing assay was designed to assess CHIP in 19 genes, except for SRSF2, which had poor coverage. CHIP was defined as a VAF ≥ 2% with a minimum of 5 variant reads for previously reported CHIP or 20 for variants not previously reported. The analysis involved measuring CHIP variants in two serial blood samples collected at baseline and year 3 of follow-up for longitudinal analysis. Associations between CHIP and blood cancers were assessed using Cox proportional hazards models. Results: CHIP at baseline was present in 34% of participants who had a corresponding year 3 sample with adequate sequencing depth. Unexpectedly, 1,245/3,110 (40%) CHIP clones present at baseline fell below 2%, while 1,520 new CHIP clones appeared at 3 years. This equated to an incident CHIP rate of 7% per annum. CHIP genes more likely to expand and contract were TP53 and ASXL1 (odds ratio 1.25 and 1.22 respectively). To measure growth characteristics of CHIP clones, we calculated a Fitness Score, FS (Log2(VAF3-year/VAFBaseline) for the 3,110 clones present at baseline. We observed no significant mean fitness advantage for any recurrent CHIP variant other than PPM1D R552* (Figure 1). CHIP variants of DNMT3A (including DNMT3A R882), TET2 and ASXL1 had a statistically significant negative FS. Even variants generally considered most pathogenic (SF3B1 K700E, JAK2 V617F, U2AF1 Q157, TP53 R273) had no significant fitness advantage. Some CHIP clones had a significant growth disadvantage (FS < -1). For example, only 1/28 DNMT3A S393N clones present at baseline remained above 2% and almost half of TP53 variants reduced in size by more than 2-fold at 3 years. We observed 299 blood cancers over a median time of 6.9 years. Individuals with CHIP clones >10% VAF (n= 652) had increased incidence of blood cancers (adjusted hazard ratio, aHR 2.14, confidence interval, CI 1.34-3.42, p=0.002). Genes associated with increased risk for blood cancers were SF3B1/U2AF1 (aHR 3.02, CI 1.63-5.61, p<0.001) and TET2 (n=aHR 1.89, CI 1.18-3.02, p=0.008). Participants with small clones (VAF 2-10%) at baseline or year 3 had no increased risk for blood cancers compared to those without CHIP. To investigate if the growth kinetics of CHIP predicts blood cancer risk, we subdivided CHIP clones into expanding (FS > +1, n= 104), stable (FS -1 to +1, n= 2043) and contracting (FS < -1, n= 963). Baseline VAF of expanding clones were smaller than stable clones: geometric mean 3.9% and 5.0%, respectively (p<0.001). Expanding clones were enriched 2-fold for PPM1D, SF3B1 and JAK2 V617F. Participants with expanding clones (n=99) were more likely to have more than one CHIP variant (50% compared with 39% for stable and 37% for contracting), although 90% of these additional variants were stable or contracting, indicating that the expanding clone was dominant. For similar VAF levels at 3 years, participants with expanding CHIP clones had a higher risk of blood cancer than a stable or contracting one (Figure 2). Conclusions: This is the largest longitudinal study of CHIP variants and associated blood cancer outcomes to date. In contrast to other longitudinal studies, there was little evidence for selective growth advantage of CHIP clones with a VAF ≥ 2%, including those with pathogenic mutations of DNMT3A, SF3B1, U2AF1, TP53 and JAK2. While clone size (VAF > 10%) and gene mutation (SF3B1, U2AF1 and TET2) were associated with increased risk of blood cancer progression, our study cautions clinical intervention of CHIP based upon a single timepoint analysis irrespective of the gene mutation. Instead, we propose monitoring of clone size may be a clinically more useful strategy to inform counselling and patient selection for early intervention studies. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Proteogenomic Analysis of SF3B1mut CLL Reveals Altered Protein Signaling Cascades and Metabolomic Pathways

Although mutation in the U2 snRNP spliceosome component SF3B1 has been appreciated as a common driver of chronic lymphocytic leukemia (CLL) for over a decade, we still know little of its mechanism of action. Prior transcriptome-based characterizations have demonstrated widespread changes in alternative splicing in SF3B1mut CLL. We hypothesized that protein level characterization could potentially yield fresh insights on alternative expression of proteins and on the impact of altered signaling pathways in patients bearing this mutation. To this end, we assembled CLL specimens from a cohort of 105 patients enrolled on the German CLL Study Group (GCLLSG) trials CLL8 and CLL11, collected prior to therapy. Of these, 52 patients expressed SF3B1wt and 53 patients harbored clonal SF3B1mut (CCF>0.8); 25 of these were the K700E mutation. Additionally, we characterized these samples by WES (available for 97 patients) and RNA-Seq (available for 96 patients). Proteomes and phosphoproteomes from all 105 samples were analyzed using a multiplexed isobaric-tagged (TMT) high resolution mass spectrometry (MS) workflow. Briefly, cells were lysed and proteins were digested into peptides with LysC and trypsin. Protein yield per sample ranged from 20 μg to 300 μg. Peptides were labeled with TMT-11 reagent and combined into 11 multiplexes while accounting for input material availability and randomizing for SF3B1 mutation, IGHV status and ATM/TP53 mutation. To enable cross TMT-plex comparison, we constructed a common reference consisting of 40 samples with highest protein amount and an equal representation between SF3B1wt and SF3B1mut samples. Each plex was fractionated into 24 fractions by high pH reversed phase separation and 500 ng protein per fraction were analyzed by MS (Orbitrap Exploris, Thermo Fisher Scientific). For phosphoproteome analysis, peptides were pooled into 12 fractions and enriched by IMAC prior to MS analysis. Spectra were analyzed by Spectrum Mill v7.1 (Broad Institute) and searched against a human reference proteome database supplemented with non-canonical ORFs supported by Riboseq evidence of translation (nuORFdb v1.0). Patient specific proteins containing single amino variants, indels, and splice isoforms, derived from WES and RNA-Seq, were included in the search database. We obtained a total of 10,419 proteins, with a median of 8,882 proteins per TMT-plex and 6,969 proteins that were quantifiable across all samples. The phosphoproteome analysis identified 32,410 sites; while each TMT-plex yielded a median of 19,380 phosphosites per plex, 5,810 were quantifiable across all patients. We integrated proteome data with patient specific mutation calls, RNA-Seq and copy number alteration (CNA) data using Panoply, a cloud-based platform for proteogenomic data analysis (Broad Institute). We confirmed the impact of chr11 and chr13 deletion as well as chr12 amplification on protein expression. We readily identified peptides spanning the SF3B1mut amino acid regions in corresponding samples. By unsupervised clustering, samples were grouped by SF3B1 status on the CNA and proteome level, a trend not observed based on RNA or phosphoproteome analysis. Gene set enrichment analysis revealed enrichment of TNFA and PI3K signaling in the SF3B1wt clusters, while the SF3B1mut associated clusters showed a strong association with metabolism changes. This analysis was highly concordant with a parallel interrogation comparing proteins expressed in SF3B1wt CLL versus SF3B1mut CLL. Here 1,955 proteins and 1,737 phosphosites were significantly altered (5% FDR). Proteins involved in B cell receptor and MAP kinase signaling were highly expressed in SF3B1wt samples, while again amino acid metabolism pathways as well as oxidative phosphorylation were upregulated in SF3B1mut samples. To more deeply investigate the role of metabolism pathways in SF3B1mut CLL, we subjected 6 independent CLL samples to MS-based metabolomics analysis. Inosine (P=0.029) and guanosine (P=0.047) were lower in the SF3B1mut extracts (P=0.001), while acylcarnitines (P=0.0004) were increased, indicating fatty acid oxidation as a preferred energy source in SF3B1mut CLL. Altogether, our study reveals fresh evidence of altered energy metabolism in SF3B1mut CLL. Ongoing studies focus on defining the mechanism by which SF3B1mut exerts these metabolic changes and novel therapeutic opportunities afforded by these findings.

The Adverse Effect of Myelodysplasia-Related Mutations in De Novo Acute Myeloid Leukemia Is Associated with Higher Variant Allelic Frequency (VAF): A Proposal for a Numeric Cutoff for Variant Allelic Frequency

Mutations in myelodysplasia-related genes ASXL1, BCOR, EHZH2, RUNX1, SF3B1, SRSF2, STAG2, U2AF1 and ZRSR2 (MRmut) are class-defining alterations in acute myeloid leukemia (AML) in the latest World Health Organization (WHO, with the exception of RUNX1) and International Consensus Classification (ICC) schemes and indicate adverse-risk per European Leukemia Net (ELN) 2022 guidelines, unless occurring in favorable-risk AML. The lower limit of detection (LOD) for most clinical-grade next generation sequencing (NGS) is ~1-2% variant allelic frequency (VAF); however, data on the prognostic significance of MRmut relative to VAF is lacking. No guidelines that provide a numeric cutoff for VAF are available for utilizing these mutations to classify AML. Our aim was to define a clinically meaningful VAF cutoff for MR mutations in de novo AML. We included patients (pts) with de novo, untreated AML with MRmut treated at our institution between 2017 to 2022. We excluded pts with acute promyelocytic leukemia, core binding factor AML, and other ELN adverse-risk categories (adverse karyotype and/or TP53 mutation). NGS at baseline was performed using a panel of 81 genes recurrently mutated in myeloid neoplasms (LOD ~1-2%). We performed a best fit classifier on individual MRmut to determine optimal VAF cutoffs using event-free survival (EFS) and overall survival (OS) as endpoints. VAF for MRmut involving X-linked genes was adjusted for men. Pts (n=145) included 92 (63%) men and 53 (37%) women with a median age of 68 years (range, 58-75). 85 (59%) pts had diploid karyotype AML, 51 (35%) had other intermediate-risk non-diploid karyotype AML, and 9(6%) had insufficient metaphases for analysis. 34 (23%) pts had ELN favorable-risk and 111(77%) had ENL adverse-risk AML. Pts were treated with low-intensity+venetoclax (VEN) (72/50%), low-intensity (26/18%), high-intensity (22/15%), and high-intensity+VEN (12/17%) regimens. Median follow-up was 26.9 months (range, 13.9 - 40.8). 72 (49%) pts had 1 MRmut; the remaining had 2-4 concurrent MRmut. MRmut had the following frequencies and VAFs: ASXL1 (28%; median VAF: 26%, [range 2-48%]); BCOR (18%; median VAF 37% [3-100%]); EHZH2 (9%; median VAF 7% [1-55%]), RUNX1 (32%; median VAF 47% [1-96%]), SF3B1 (9%; median VAF: 38% [1-55%]), SRSF2 (46%; median VAF: 37% [1-76%]), STAG2 (20%; median VAF: 44% [15-98%]), U2AF1 (12%; median VAF: 33% [2-43%]) and ZRSR2 (2%; median VAF: 94% [93-95%]). 34 (23%) pts had ELN favorable-risk AML including 28 (23%) with mutated NPM1 without FLT3-ITD and 6 (4%) with CEBPA bZIP mutations, respectively; in 12/34 pts NPM1 or CEBPA bZIP mutations were subclonal compared with clonal MR mutation(s). A summary of baseline genetic and treatment data is provided in Fig 1A. The overall composite remission rate (CR+CRi) was 88% (64% and 24%, respectively) Median OS was 30.3 months and estimated 12/24 month OS were 77%/60%. The median EFS was 24.6 months and estimated 12/24 month EFS were 65%/51%. Pts with ELN favorable-risk AML with MRmut had longer OS and EFS compared with the adverse-risk group (median OS 34.9 vs 27.5 months, median EFS 34.9 vs 22.9 months) but the difference was not significant (p=0.60 and 0.30, respectively). The VAF cutoff for MRmut associated with EFS and OS, respectively, was: ASXL1: 21%/21%, BCOR 25%/25% , EHZH2 15%/15%, RUNX1 25%/27%, SF3B1 31%/31%, SRSF2 32%/35%, STAG2 40%/21%, U2AF1 44%/44% and ZRSR2 47/47%.There was no significant difference in OS or EFS between pts with 1 MRmut vs those involving ≥2 MR genes. Pts treated with high-intensity, high intensity+ VEN and low intensity+VEN regimens had longer OS (Fig 1B) and EFS compared with pts treated with low intensity regimens without VEN (OS: not reached, 31.7 months, 31.4 months vs 8.9 months, respectively, p<0.001; EFS: not reached, 31.7 months, 29.6 months vs 3.7 months, respectively, p<0.001). Our data suggest that the poor prognostic effect of MRmut in de novo AML is associated with higher VAF and not a binary presence or absence of MRmut. The minimum VAFs associated with worse outcome for each MRmut ranged from 15-47%. Additional analysis to create predictive models is in progress to determine the optimal VAF cutoff when using MRmut to assign AML to the WHO/ICC MR categories and to ELN adverse-risk group. Importantly, our results suggest that pts with MRmut and non-adverse karyotype have significantly better outcomes when treated with high-intensity and/or venetoclax-based treatment regimens. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Establishment and Characterization of Mouse Models with Primary and Acquired BTK Inhibitor Resistance

Background: The second-generation Bruton's tyrosine kinase (BTK) inhibitor acalabrutinib is effective for treatment of patients with chronic lymphocytic leukemia (CLL), including high-risk and relapsed patients. Resistance to BTK inhibitors can develop not only due to on-target mutations, but also by bypassing the BTK signaling pathway. Mouse models represent useful tools in studying mechanisms of resistance as well as testing the efficacy of alternative therapy options to overcome resistance. Eµ-TCL1-transgenic (TCL1-tg) mice have the advantage of immunocompetence and resemble the human CD49d-high CLL cohort (Szenes et al., 2020). Transplantation of tumors from these mice to wild-type recipients allows acceleration of pathophysiology, selection of specific tumor clones, and more efficient testing of drug combinations. Here, we developed acalabrutinib resistance models by serial transplantation and long-term administration of the drug in drinking water, characterized these models, and compared molecular resistance signatures with those of human patient samples. Methods: Models were developed starting from four primary TCL1-tg mice at overt disease with tumor burden in spleen ranging up to 95% of lymphocytes. Serial retransplantation to wild-type C57BL/6J recipients for two to four rounds and acalabrutinib therapy pressure resulted in several drug resistant pedigrees. Mice were sacrificed at end-stage disease and tumor burden in spleen, bone marrow and blood was analyzed by flow cytometry using anti-CD5/CD19 antibodies. Purified leukemic cells were subjected to RNA-seq, real-time quantitative PCR and mutational analyses. Results: Three response patterns were identified in recipient mice: a) sustained sensitivity (models were stopped after 23 weeks of treatment), b) acquired resistance (with initial response followed by development of resistance under continuous therapy pressure and c) retransplantation-induced resistance in a case of an aggressive tumor clone that was responding to acalabrutinib in the parental mouse. The sensitive models displayed a 63% reduction in spleen weight upon treatment and a concomitant 65% reduction in blood tumor burden. In acquired or primary resistant mice, spleen weight and tumor burden was not significantly different from that of untreated mice, with blood tumor burden in the magnitude of 100,000 leukemic cells/µl and spleen tumor burden of 500-1000 million leukemic cells. Resistance could not be attributed to the mutational landscape of BTK and PLCγ2. Mutations in p53, Myc, Notch1, Atm, and Sf3b1 were detected but found to be intrinsic in individual ancestors and further present in the follow-up transplantations. RNA-seq analysis revealed 863 differentially expressed genes in mice with acquired resistance to acalabrutinib (compared to untreated, Figure 1) and 114 shared differentially expressed genes in mice with acquired or transplantation-based resistance. Gene set enrichment analysis demonstrated the presence of MAPK signaling and NF-kB signaling gene signatures within the resistant tumors. In particular, acquired resistance was accompanied by changes in the B cell receptor downstream components. We also compared the murine 863 gene signature to human samples from acalabrutinib resistant patients (progression on drug versus baseline) and found 13 genes are differently expressed in both, human and mouse. Validation for this gene expression data is ongoing and updated results will be presented. Conclusion: We have developed novel TCL1-tg based murine models of acalabrutinib resistance. Our models provide well-characterized tools to study primary as well as acquired BTK inhibitor resistance and allow the current testing of potential new therapeutic approaches to overcome resistance. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Correlation of Mutational Profiles and Cytogenetics with Morphologic Dysplasia in Myelodysplastic Syndromes

Introduction: Myelodysplastic syndromes (MDS) are a spectrum of clonal bone marrow failure disorders demonstrating cytopenias, dysplasia, and risk of progression. Although ~50% of MDS patients have normal karyotype (NK), a subset of MDS exhibits genetic alterations that result in an abnormal karyotype. However, the relationship between the molecular landscape and histologic features in MDS have not been well explored. Using next generation sequencing (NGS) and cytogenetics, we sought to correlate genetic abnormalities in MDS with morphologic dysplasia. Methods: After IRB approval, we retrospectively identified NGS studies performed with a 164 gene panel in patients with MDS at Stanford (2018-2021). Molecular and cytogenetic data were collected for 225 MDS patients. Peripheral blood and bone marrow (BM) slides were collected from 61 patients for histological review. As controls, 32 staging BMs were obtained from untreated lymphoma patients. A hematopathologist performed a blinded intense morphologic review. A subset of cases was reviewed by 2 other hematopathologists and the intraclass correlation coefficient (ICC) was calculated to evaluate concordance. The degree of dysplastic features and overall dysplasia was scored. The standardized mean difference (SMD) was calculated to quantify the degree of dysplasia between comparison groups. Statistical analysis was performed using R (3.6.3); p value < 0.05 was considered statistically significant. Results: MDS cases had higher degrees of dysplasia than control cases in megakaryocytic (median 37.5% vs 0%), granulocytic (median 20% vs 0%), and erythroid (median 10% vs 0%) lineages. To assess specificity, using a cutoff of 10%, 1/31 control cases (3.1%) showed megakaryocytic dysplasia, 2/32 (6.3%) had erythroid dysplasia, while none showed granulocytic dysplasia. Dysplasia in > 2 lineages in >10% of cells was specific for MDS, detected in 49/61 (80.3%) of MDS cases vs 0/32 control cases (Table 1). Within the MDS group, lack of multilineage dysplasia was more common in NK (8/25; 32%) vs non-NK (3/36; 8.3%), suggesting MDS-NK is morphologically subtle. Using a cutoff of 10%, 13/25 (52%) NK lacked granulocytic dysplasia vs 10/36 (27.8%) in non-NK, 8/25 (32%) of NK lacked erythroid dysplasia vs 10/36 (27.8%) in non-NK, and 5/25 (20%) NK lacked megakaryocytic dysplasia vs 3/35 (8.3%) in non-NK. The most strongly associated dysplastic features for MDS included erythroid megaloblastic change (SMD=1.25), erythroid nuclear features (SMD=1.041), atypical megakaryocyte nuclei (SMD=1.189), separated megakaryocyte nuclei (SMD=1.131), and hypogranulated granulocytes (SMD=1.096). Based on 10 MDS and 5 control cases, rough concordance on overall dysplasia was demonstrated among 3 hematopathologists. Granulocytic dysplasia showed the highest reproducibility (ICC for granulocytic = 0.948, erythroid = 0.547, megakaryocytic = 0.531). The percent of cells with dysplasia in the MDS cohort did not correlate with number of mutations. The overall number of mutations between MDS with NK and non-NK were relatively similar (Figure 1). The non-NK group exhibited a greater frequency of pathogenic TP53 variants than NK (31.5% vs 2.6%, p < 0.0001). Conversely, SF3B1 and TET2variants occurred more frequently in the NK cohort than non-NK (SF3B1: 32.5% vs 12.6%; TET2: 29.8% vs 12.6%, both p values < 0.003). In MDS cases with a TET2 variant, NK was associated with a significantly higher total number of mutations than non-NK cases (median [IQR]: 5[5-7] vs 4[3-6], p = 0.031). Using Random Forest models, exploratory multivariate analysis identified the presence of mutations in TP53, SF3B1, TET2, and overall megakaryocytic, granulocytic, and erythroid dysplasia as the prominent parameters that distinguished NK from non-NK MDS cases. There was no significant difference in age at diagnosis between NK and non-NK MDS in all NGS cases (median age 74 vs 70, p = 0.10) and morphologically scored cases (median age 70 in both groups). Conclusion: Our study demonstrates that granulocytic dysplasia was the most specific morphologic finding for MDS with high reproducibility among 3 hematopathologists. Detection of dysplasia in > 2 lineages in >10% cells was also specific for MDS, while MDS-NK appear morphologically subtle. Mutations in TET2 were more frequent in MDS-NK. Our findings highlight the importance of NGS and histopathology in helping establish a diagnosis of MDS-NK. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Genomic Characterization of Patients in a Phase 2 Study of Zanubrutinib in BTK Inhibitor-Intolerant Patients with Relapsed/Refractory B-Cell Malignancies

Background/Introduction: Targeting Bruton tyrosine kinase (BTK) to inhibit B-cell receptor signaling is an effective way to treat B-cell malignancies. However, some patients (pts) have experienced toxicities to BTK inhibitors ibrutinib (ibr) and acalabrutinib (acala), which lead to dose reduction or treatment discontinuation. Zanubrutinib (zanu) is a potent and selective next-generation BTK inhibitor (Blood. 2021;138[suppl 1]:1410). BGB-3111-215 (NCT04116437) is an ongoing, phase 2 study of the safety and efficacy of zanu monotherapy in pts with chronic lymphocytic leukemia (CLL), small lymphocytic lymphoma (SLL), Waldenström macroglobulinemia (WM), mantle cell lymphoma (MCL), or marginal zone lymphoma (MZL) who discontinued ibr, acala, or acala and ibr because of intolerance. Here, we present the gene mutation profile of pts in this study to demonstrate the mutational landscape of pts intolerant to ibr or acala and to explore the association between gene mutations and response to zanu in ibr- and acala- intolerant pts. Methods: Peripheral blood from pts was collected before treatment and at or after the time of disease progression. Genomic DNA was isolated from peripheral blood mononuclear cells or plasma, and the mutational status of targeted genes was assessed using a targeted 106-gene next-generation sequencing (NGS) panel (PredicineHEMETM, Predicine). Samples were sequenced to a median depth of >20,000 reads, with a validated sensitivity of 0.25% mutant allele frequency for all genomic regions. Results: Samples from 63 pts were analyzed, including 63 at baseline and 5 at or after progressive disease (PD). There were 41 pts with CLL, 6 with SLL, 10 with WM, 3 with MCL, and 3 with MZL. The most common mutations at baseline were TP53 (31.7%), SF3B1 (22.2%), ATM (17.5%), NOTCH1 (17.5%), CHEK2 (14.3%), and KRAS (12.7%). Copy number variants were detected, including deletions at ATM locus (9.5%) and RB1 locus (9.5%) and amplifications at CCND2 locus (6.3%). For CLL/SLL, the frequencies of TP53 (27.7%), NOTCH1 (21.3%), SF3B1 (25.5%), ATM (21.3%), and BIRC3 (10.4%) mutations were comparable to those in other studies with high-risk relapsed/refractory CLL pts (Leukemia. 2018;32:3-91; Blood. 2014;123:3247-3254; Leuk Lymphoma. 2018;59[10]:2318-26). For WM, 5 of 10 (50%) samples harbored TP53 mutations, and 4 of 10 samples had MYD88 mutations; all harbored the L265P mutation. Among the 4 pts with MYD88 mutation, CXCR4 gene mutation was detected in 1 case. All 6 pts with MYD88 wild type also had CXCR4 wild type gene, consistent with published data (N Engl J Med. 2015;373:584-6). Baseline mutations were available for the 10 pts with PD, including 8 CLL, 1 SLL, and 1 MCL. Compared to pts without PD, pts with PD had increased frequency of certain mutations: TP53 (60%, P=0.071), ATM (50%, P=0.017), SF3B1 (50%, P=0.035), RB1 (40%, P=0.009), SETD2 (40%, P=0.009), and CDKN2A (30%, P=0.011). This finding suggests that these particular baseline mutations are associated with resistance to zanu. Five pts also had mutations detected at or after PD. Of these, 3 pts had BTK or PLCG2 mutations (submitted). One patient with CLL and without BTK or PLCG2 mutations had ATM and FBXW7 mutations (variant allele frequencies were 86% and 50%, respectively), suggesting DNA damage pathway mutations and NOTCH1 dysregulations may have contributed to disease progression in this patient. Conclusion: Exploratory analysis results confirmed that cell cycle, DNA damage, and NOTCH1 pathway genes were frequently mutated in pts with B-cell malignancies on study BGB-3111-215 (pts intolerant to ibr and/or acala). Pts with mutations associated with poor prognosis at baseline were more likely to develop PD.

Subclonal KIT D816V Mutations Are Prevalent in Chronic Myelomonocytic Leukemia and Correlate with Distinct Phenotypic Features

Background: Activating mutations in KIT are infrequently observed in myeloid neoplasms, but are a pathognomonic finding in systemic mastocytosis (SM) where the hotspot D816V variant is documented in 90-95% of patients. In SM, KIT variants may occur at a low variant allele frequency (VAF) below the limits of detection of most commercial next generation sequencing (NGS) assays necessitating digital droplet PCR (ddPCR) for accurate detection. Aggressive SM often presents concurrently with an associated hematologic neoplasm (SM-AHN). Chronic myelomonocytic leukemia (CMML) is the most frequently co-occurring hematologic malignancy, but KIT mutations are infrequently detected by conventional NGS. The limitations in detecting KIT variants by NGS and increased frequency of CMML in SM-AHN suggest that subclonal KIT mutations may be underrecognized in CMML. Further, identification may have prognostic and therapeutic implications given the emergence of KIT-targeted therapies in SM. Methods: A multicenter, retrospective study was performed to assess the prevalence of subclonal KIT D816V mutations in CMML patients and examine clinical implications. Eligible patients had pathologically confirmed CMML by WHO 2017 criteria, with available bone marrow (n=59) or peripheral blood (n=14) sample obtained prior to therapy. Samples underwent DNA sequencing using a 285-gene NGS panel (QIAseq), with a VAF threshold of 5% utilized for variant calling. KIT D816V mutations were assessed by ddPCR, with limit of detection of 0.03%. Clinical data was collected from baseline and follow up for phenotypic and prognostic comparisons. Fishers exact and Chi-square tests were used to compare categorical variables and the Mann-Whitney test for continuous variables. The Kaplan-Meier method was used to estimate overall survival (OS), with the log rank test used to compare survival between groups. The study was approved by institutional review boards at each institution. Results: A total of 73 patients were included with a median age of 71, and 66% were male. By WHO criteria, 50% were classified as CMML-0, 28% CMML-1, and 22% CMML-2; 59% of patients had proliferative CMML by FAB criteria. KIT D816V mutations were detected in 15 patients (21%) by ddPCR, compared to two (3%) by NGS. The median VAF of mutations detected by ddPCR was 0.06%, whereas the two variants detected by NGS had VAFs of 44.1% and 47.0%. Baseline characteristics of KIT wild-type (WT) and mutated (MUT) patients are detailed in Table 1. KIT MUT patients had lower white blood cell (p=0.007) and absolute monocyte counts (p=0.04), less frequent splenomegaly (p=0.05), more frequent karyotype abnormalities (p=0.14), and lower frequency of proliferative disease by FAB criteria (p=0.01). There was a median of 4 variants (range 0-7) detected per patient by NGS, and the most frequently mutated genes were TET2 (59%), SRSF2 (44%), ASXL1 (33%),RUNX1 (25%), and NRAS (25%). KIT MUT patients had a median of 3 total variants (range 0-6) detected by NGS compared to 4 (range 0-7) in KIT WT (p=0.63). There was a trend toward higher frequency of SF3B1 (23% vs. 5%, p=0.07) mutations in KIT MUT patients. Conversely, no mutations were observed in KRAS or CBL in KIT MUT patients, compared to 14% and 16% in KIT WT, respectively (p=0.34, p=0.20). When evaluating mutations in signaling pathways, there was a trend toward lower frequency in KIT MUT patients compared to WT (31% vs. 59%, p=0.12). With a median follow up of 48 months, median OS of the cohort was 30 months. KIT MUT patients demonstrated an OS of 27.5 months compared to 30 months in WT patients (HR 1.06, p=0.87). Transformation to AML was observed in 42% of KIT MUT patients compared to 29% of WT patients (p=0.49). A similar proportion of patients were treated with hypomethylating agents (HMAs), with a response rate of 14% in KIT MUT patients compared to 32% in WT (p=0.63), and complete remission rates of 0% vs. 27%, respectively. Conclusions: This study is the first to document the prevalence of subclonal KIT D816V mutations by ddPCR in CMML, revealing a high prevalence in comparison to detection by standard NGS assays. While sample size limited statistical comparisons, KIT mutations were associated with distinct phenotypic features, including less proliferative disease and poor response to HMAs. Given the efficacy of KIT inhibitors in SM-CMML, these findings could represent an avenue for targeted therapy in CMML. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Molecular Patterns Identify Distinct Subclasses of Myeloid Neoplasia

The diverse molecular pathogenesis of MDS underpins its pathomorphologic and clinical heterogeneity, ultimately converging in a stepwise progression to leukemia. The identification of hallmarks morphological features, hampered by inter-observer variability, limited a precise resolution of pathogenic differences, partially mitigated by the inclusion of cytogenetic abnormalities, the two pillars of MDS classification. The advent of NGS led to the discovery of a multitude of mutations, and their combinations unveiled a tremendous molecular diversity in MDS. Intuitively, molecular patterning has been used in supervised fashion to fit morphologic classification, but this approach ensured only a limited success. Machine learning (ML) analytic methods, as demonstrated in acute myeloid leukemia (AML), provide new opportunities to integrate molecular data and identify functionally convergent patterns manifesting with similar clinical outcomes. Particularly, ML approaches can identify cryptic molecular interactions and pathogenic mechanisms irrespective of their often comparable impact on outcomes. Here, we took advantage of a large, well-annotated multicenter cohort of MDS and secondary AML (sAML) cases (N=3588) to test the hypothesis that related molecular patterns can be analyzed in an unbiased/unsupervised fashion to characterize molecularly-defined configurations. Our aim was not meant to compete with or replace current prognostication tools (including the new molecular prognostic score)2 but rather to illuminate the genetic sub-classification of MDS and related conditions in an operator-independent fashion, according to molecular correlations and mutual functional proximity, irrespective of their impact on clinical outcomes. Molecular signatures driving each cluster were identified and used for genomic subclassification that separated functionally related subtypes in an outcome independent fashion. We then validated the model on 412 patients from two external institutions. Among the 3588 patients included in this cohort, 735 (20%) had sAML, 774 (22%) had higher-risk MDS (HR-MDS), and 2079 (58%) had lower-risk MDS (LR-MDS). Abnormal karyotype was found in 43% of the cases, and 77% had at least one somatic mutation, with 8% harboring >4 mutations. ML analysis identified 14 distinctive molecular clusters (MCs) with unique pathobiological associations, treatment responses, and prognosis (Fig.1A). Normal karyotype (NK) was enriched in MC2, MC4, MC6, MC9, MC10, and MC12 with different distributions of ASXL1, DNMT3A,JAK2, RAS, SF3B1, and TET2 mutations. Complex karyotype and trisomy 8 were enriched in MC13 and MC1, respectively. Remarkably, our ML approach identified functionally relevant genomic clusters (Fig.1A). For instance, MC2, MC6, and MC8 demonstrated distinct genomic signatures: MC2 included cases with NK only (100%) and some DNMT3A (11%), JAK2 (11%), and RAS pathway (10%) mutations; MC6 cases had similar features to MC2 but were also enriched in SRSF2 (49%) and RAS mutations (23%); MC8 was instead characterized by the presence of del5q/-5 (100%), DNMT3A (17%), and TP53 (17%) mutations. MC13 contained cases with complex karyotype (100%) and TP53 (44%) mutations (Fig1.B). We then identified five risk groups to reflect the biological differences between clusters. Our clustering model highlighted the significant survival differences among patients assigned within similar IPSS-R risk groups but displaying heterogeneous molecular configurations. Different response rates to hypomethylating agents (e.g., MC9 and MC13 [OR: 2.2 and 0.6, respectively]) paralleled the biological differences across the clusters. Interestingly, our MCs retained significant survival differences regardless of the bone marrow blast percentage. Our model is available via web-based access resource (https://drmz.shinyapps.io/mds_latent). Despite the complexity and the diversity of molecular alterations in MDS/sAML, by deploying ML we were able to identify and distinguish functional and pathologically related MCs irrespective of the anamnestic clinico-morphological features. In recognition that similar outcomes do not warrant similar molecular features, our model provides genomic signatures for a better understanding of the pathobiological mechanisms of disease, progression to higher stages, and identification of future therapeutic vulnerabilities. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Molecular Characterization of Adult Acute Lymphoblastic Leukemia Identifies a Subgroup with Myeloid Mutations and Pre-Existing Clonal Hematopoiesis

Acute lymphoblastic leukemia (ALL) represents 20% of adult leukemias. Genomic risk stratification of ALL relies on the data generated from large cohorts of pediatric and young adult patients, where older patients with ALL have been underrepresented. It is imperative to understand the unique leukemogenic pathways in this group in order to develop novel therapeutic approaches. Clonal hematopoiesis (CH) is a precursor lesion for development of myeloid leukemias and mature lymphoid neoplasms (Niroula, Nat Med 2021), but its impact on the biology and outcomes of ALL is unknown. Based on our observations on high incidence of myeloid mutations in older adults with ALL, we hypothesized that CH constitutes a fertile genomic background for lymphoblastic leukemogenesis, and ALL with myeloid mutations is a unique entity with distinct clinical and molecular characteristics. We studied 400 adult patients with ALL treated at the University of Chicago, Moffitt Cancer Center, and City of Hope between 2014 and 2022 (Fig 1A). Genetic profiling of pre-treatment bone marrow samples was performed with a comprehensive panel covering 147 myeloid and lymphoid genes. Overall, 33% of ALL cases had at least one myeloid CH-associated mutation with the following frequencies: TP53 (18%), DNMT3A (5%), TET2 (4%), ASXL1 (3%), RUNX1 (3%), IDH2 (2%), BCORL1 (2%), U2AF1 (1%), CUX1 (1%), BCOR (1%), CBL (1%), and SF3B1 (1%). We further investigated the association between mutations and clinical variables by calculating the odds-ratios (OR) with adjustments for multiple testing by using the Benjamini-Hochberg method. Myeloid CH mutations in TP53, DNMT3A, TET2, and ASXL1 genes were more common in older adults with ALL. Mutations in TP53 and DNMT3A were enriched in therapy-related B-ALL group (OR 2.6, 95% CI= 1.3-5.5, q= .007 and OR 4.7, CI= 1.2-18.1, q= .01, respectively), and were more commonly associated with hypodiploid (OR 7.1, 95% CI= 1.5-32.7, q=.004 and OR 5.2, 95% CI= 0.6-48, q=.1, respectively) and hyperdiploid (OR 5.9, 95% CI= 2.7-13.1, q=<0.001 and OR 3.9, 95% CI= 1.1-16.5, q=.04, respectively) cytogenetics. We stratified the ALL cohort into three groups: TP53-mutated ALL, non-TP53 CH-associated ALL, and ALL without myeloid CH mutations. Patients in both groups of CH-associated B-ALL had worse overall survival (OS) than patients without myeloid CH mutations (median OS, 23 vs 25 vs 48 months, respectively, p=.001). In a multivariable analysis of OS for B-ALL, adjusting for age and therapy-related ALL, both TP53 (HR 1.95, 95% CI=1.2-3.17, p=.006) and non-TP53 myeloid CH (HR 1.95, 95% CI=1.14-3.35, p=.01) were associated with worse OS. Based on the data indicating high allelic frequencies of myeloid CH mutations in ALL, we hypothesized that these mutations are ancestral events arising in HSCs. To gain insights into the clonal architecture, we performed single cell DNA+protein sequencing using the Mission Bio, lnc platform. Data from 4,180 cells obtained from a CH-associated B-ALL patient are shown in Fig 1B. Cells were clustered by immunophenotype, and mutations were overlayed into 8 clusters. IDH2 R140Q mutation was shared between ALL and myeloid compartments. PTPN11 mutation was more prevalent in B-lymphoblasts. We observed a similar pattern in 7 additional patients, indicating that both lymphoblasts and non-malignant myeloid cells share CH mutations in ALL patients. Next, we investigated whether CH can be detected as a precursor lesion prior to ALL. In a patient with therapy-related B-ALL, a TET2 CH clone was detectable 5 years prior to ALL, and additional SF3B1 and CDKN2A mutations were detectable at the time of B-ALL diagnosis. We confirmed the presence of these mutations in B-lymphoblasts by single cell sequencing. This patient had persistent TET2/SF3B1 mutated CH despite achieving remission from B-ALL. We observed a similar pattern in four additional patients, for whom the CH mutations were detectable several years prior to their ALL diagnosis and were present in the dominant ALL clone. Collectively, these data support a model in which ALL can arise from pre-existing clones harboring myeloid CH mutations in a significant fraction of older adults. ALL with myeloid mutations is a distinct clinicopathologic entity with inferior survival outcomes. Future studies may decipher mechanisms driving lymphoid bias of CH-mediated leukemogenesis that can be targeted to prevent and treat ALL. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Clonal Hematopoiesis in Vexas Syndrome

Background VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) syndrome is a recently described systemic inflammatory disease of older men, due to somatic mutations in UBA1 (UBA1mut). VEXAS predisposes to myelodysplastic syndrome (MDS) and plasma cell dyscrasia (PCD). A few studies have reported concomitant somatic mutations in genes associated with myeloid neoplasia but their true prevalence and the clonal landscape is unknown in UBA1mut cases. We hypothesized that clonal hematopoiesis in myeloid-related genes (typical CH) may be linked to heterogenous clinical phenotypes and progression to MDS seen in VEXAS. Methods We performed error-corrected DNA sequencing (ECS) in 80 VEXAS patients and correlated findings with clinical outcomes in 77 patients. Targeted single-cell proteogenomic sequencing (scDNA) was performed on three patients to define clonal architectures. DNA methylation was also investigated in 38 cases. ResultsUBA1mut p.M41T c.122T>C (n=52; 65%) was most frequent followed by p.M41V c.121A>G (n=15;18%), p.M41L c.121A>C (n=10,12%), and three (4%) splice site mutations (c.118-1G>C); one patient had two UBA1mut in independent clones: a p.M41T and c.118-2A>C. Median variant allele frequency (VAF) of UBA1 mut variant was 75%. Forty-eight patients (60%) had one (n=28; 35%) or ≥2 typical CH mutations (n=20; 23%) concomitant with UBA1mut, a frequency much higher than in healthy age-matched controls (>50% in VEXAS patients ³40 years vs. 18% in controls). Mutations in DNMT3A and TET2 were most frequent, DNMT3A mutations at high VAF (median 27%), in contrast to TET2 (median 1.5%; Figure 1). Patients with typical CH were divided into 4 subgroups: with DNMT3A or TET2 mutations alone (D/T, n=19), concomitant DNMT3A and TET2 mutations (D&T; n=7), mutations in other genes regardless of DNMT3A/TET2 mutations (Others; n=20), and no typical CH mutations (n=31). UBA1mut genotype and co-occurring CH mutations did not predict inflammatory or hematologic manifestations (MDS, PCD, thrombosis, cytopenia). An MDS diagnosis was only associated with transfusion-dependent anemia and thrombocytopenia but not with typical CH mutation subgroups, UBA1 VAF or abnormal karyotype. Overall survival was 60% at 10 years. MDS diagnosis did not confer poorer prognosis but transfusion-dependent anemia, thrombocytopenia, and the presence of typical CH mutation at VAF>2% (particularly DNMT3A/TET2) increased the risk of mortality (Figure 2). VEXAS patients had marked epigenetic age acceleration with a median of 11 years (range 7-23), especially in those with DNMT3A or TET2 mutations. By scDNA, >90% of myeloid but not lymphoid cells were UBA1mut. The majority (60-90%) of HSPC in the marrow were also UBA1mut. There was a shift towards increased neutrophils (with typically high CD16/CD62L/CD10 expression), plasmacytoid DC and classical monocytes in PB, compared to a healthy control. In UPN1, a TET2 and DNMT3A mutations preceded UBA1mut in the same clone, resulting in striking clonal expansion after acquisition of UBA1mut; all variants were at VAF >45% by ECS. In UPN2, UBA1mut was a driver of a clone that dominated the clonal landscape; subclones with SF3B1 and TET2 mutations in the same cell remained stable at lower VAF by ECS (<5%). In UPN3, UBA1mut itself resulted in clonal expansion of mutated cells. Conclusions In VEXAS, UBA1mut cells are the primary cause of inflammatory and hematological manifestations, defining a new molecularly defined somatic entity that is associated with MDS. scDNA data demonstrate that UBA1mut clones have competitive proliferative advantage, therefore, the expansion of clones with typical CH mutations is dependent on co-occurrence with UBA1mut. In VEXAS, typical CH mutations may serve as a biomarker of shortened overall survival, without an increased rate of myeloid neoplastic transformation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Nonsense-Mediated RNA Decay Disruption Induces Altered Splicing in Spliceosome Mutant Cells

Introduction: Expression of mutant splicing factors U2AF1(S34F), SF3B1(K700E), or SRSF2(P95H) alter RNA splicing in myeloid malignancies, including increased production of nonsense transcripts that rely on nonsense-mediated RNA decay (NMD) for clearance. Cell lines expressing spliceosome mutants are more sensitive to NMD inhibition than wild-type cells and the differential sensitivity is partially dependent on R-loop formation. The impact of NMD disruption on the viability of primary splicing factor mutant hematopoietic cells and the consequence on downstream RNA splicing are not known. Methods: Cell viability of primary mouse hematopoietic cells (untransformed and AML cells) and a human AML cell line that expresses mutant or wild-type U2AF1 after treatment with a highly-specific inhibitor of SMG1 (SMG1i), the only known protein kinase in the NMD pathway, was measured using a CellTiter-Glo assay. Gene expression and RNA splicing changes induced by mutant U2AF1 and/or SMG1 inhibition were identified by short and long-read RNA sequencing (RNA-seq). Results: We treated primary untransformed mouse U2AF1S34F or U2AF1WT expressing cKit+ hematopoietic cells ex vivo with SMG1 inhibitor and monitored cell viability. We found that SMG1i reduced the viability of primary mouse cells expressing mutant U2AF1S34F compared to U2AF1WT (IC50 52.6 nM vs 173.5 nM, respectively, p=<0.0001). Next, we tested the sensitivity of primary mouse AML cells expressing a doxycycline-inducible U2AF1S34F or U2AF1WT transgene to SMG1i that were generated using retroviral expression of the fusion oncogene MLL-AF9 to induce AML. We found that NMD disruption using SMG1i also reduced the ex vivo viability of primary U2AF1S34F compared to U2AF1WT expressing mouse AML cells (IC50 40.9 nM vs 102.2 nM, respectively, p=<0.0001), highlighting the potential utility of SMG1i treatment for spliceosome mutant myeloid malignancies. To explore the underlying mechanism for the sensitivity of spliceosome mutant cells to NMD inhibition, we generated, K562 cell lines expressing doxycycline-inducible U2AF1S34F or U2AF1WT and confirmed that NMD disruption using SMG1i reduced cell viability of mutant expressing cells in vitro (IC50 128.1nM vs 578.5nM at 72 hours, respectively, p≤0.0001). Next, we performed RNA-seq on K562 cells expressing U2AF1WT or U2AF1S34F that were treated with vehicle or 1uM SMG1i for 24 hours (n=3 replicates). We identified reduced expression of genes involved in the cell cycle, DNA replication/repair, telomere maintenance and mRNA splicing in SMG1i treated U2AF1S34F compared to U2AF1WT cells (GSEA; FDR<0.008). Next, we identified 3,393 RNA splicing alterations induced by SMG1i treatment of U2AF1S34F compared to U2AF1WT cells (rMATS; FDR<0.1, percent spliced-in >10%). Genes with altered splicing were enriched in similar pathways, including the cell cycle, DNA replication, and RNA splicing and processing pathways (Cluster Profiler; FDR≤0.02). Finally, we performed long-read RNA-seq (Nanopore) on RNA isolated from K562 cells (not expressing spliceosome mutants) that were treated with vehicle or SMG1i (0.5uM, 24hours, n=2) to directly identify full length transcripts containing premature termination codons (PTC) that are normal targets of NMD. Our preliminary data indicate that SMG1i significantly increases the number of detectable PTC containing transcripts (known and novel) compared to vehicle treatment (mean 19,291 vs 5,621 PTCs, respectively; p=0.008). In addition, PTC-containing transcripts that were overexpressed in SMG1i vs vehicle treated cells also occur in genes involved in the cell cycle, DNA repair and mRNA splicing (FDR£0.001), suggesting they are regulated by NMD under basal conditions. These gene pathways may be particularly vulnerable to further dysregulation in mutant treated cells because they are also altered by mutant U2AF1S34F expression. Conclusion: A vulnerability of primary hematopoietic cancer cells with spliceosome mutations to NMD inhibition suggests the possibility that NMD can be targeted for treating MDS with aberrant splicing. Excessive R-loop accumulation in SMG1i treated U2AF1S34F cells may contribute to the downstream gene expression and splicing alterations and reduced viability of mutant cells. Ongoing pre-clinical in vivo efficacy studies will test the potential for SMG1i to treat hematologic malignancies that have splicing factor mutations.