Subclonal Mutations Research Articles

Selective advantage of epigenetically disrupted cancer cells via phenotypic inertia.

The evolution of established cancers is driven by selection of cells with enhanced fitness. Subclonal mutations in numerous epigenetic regulator genes are common across cancer types, yet their functional impact has been unclear. Here, we show that disruption of the epigenetic regulatory network increases the tolerance of cancer cells to unfavorable environments experienced within growing tumors by promoting the emergence of stress-resistant subpopulations. Disruption of epigenetic control does not promote selection of genetically defined subclones or favor a phenotypic switch in response to environmental changes. Instead, it prevents cells from mounting an efficient stress response via modulation of global transcriptional activity. This "transcriptional numbness" lowers the probability of cell death at early stages, increasing the chance of long-term adaptation at the population level. Our findings provide a mechanistic explanation for the widespread selection of subclonal epigenetic-related mutations in cancer and uncover phenotypic inertia as a cellular trait that drives subclone expansion.

The prognostic impact of subclonal IDH1 mutation in grade 2-4 astrocytomas.

Isocitrate dehydrogenase (IDH) mutations are thought to represent an early oncogenic event in glioma evolution, found with high penetrance across tumor cells; however, in rare cases, IDH mutation may exist only in a small subset of the total tumor cells (subclonal IDH mutation). We present 2 institutional cases with subclonal IDH1 R132H mutation. In addition, 2 large publicly available cohorts of IDH-mutant astrocytomas were mined for cases harboring subclonal IDH mutations (defined as tumor cell fraction with IDH mutation ≤0.67) and the clinical and molecular features of these subclonal cases were compared to clonal IDH-mutant astrocytomas. Immunohistochemistry (IHC) performed on 2 institutional World Health Organization grade 4 IDH-mutant astrocytomas revealed only a minority of tumor cells in each case with IDH1 R132H mutant protein, and next-generation sequencing (NGS) revealed remarkably low IDH1 variant allele frequencies compared to other pathogenic mutations, including TP53 and/or ATRX. DNA methylation classified the first tumor as high-grade IDH-mutant astrocytoma with high confidence (0.98 scores). In the publicly available datasets, subclonal IDH mutation was present in 3.9% of IDH-mutant astrocytomas (18/466 tumors). Compared to clonal IDH-mutant astrocytomas (n = 156), subclonal cases demonstrated worse overall survival in grades 3 (P = .0106) and 4 (P = .0184). While rare, subclonal IDH1 mutations are present in a subset of IDH-mutant astrocytomas of all grades, which may lead to a mismatch between IHC results and genetic/epigenetic classification. These findings suggest a possible prognostic role of IDH mutation subclonality, and highlight the potential clinical utility of quantitative IDH1 mutation evaluation by IHC and NGS.

Subclonal mutations in epigenetic regulators bring fitness in times of stress.

Mutations in epigenetic regulators are prevalent in cancer, but their functional impact is poorly delineated aside from disrupting differentiation. In this issue of Cancer Cell, Loukas etal. zoom in on the late occurring, subclonal subset of such mutations and define the molecular mechanism behind their enrichment in cancer.

GEN-10 WHOLE GENOME LANDSCAPE OF GLIOBLASTOMA, IDH-WILD TYPE

Abstract Introduction: Glioblastoma, IDH-wild type (GBM) is the most common malignant brain tumor with a dismal prognosis. Although its coding region is well-analyzed, little is yet known about the landscape of whole-genome in GBM. Here, we analyzed whole-genome sequencing data from large cohorts to completely uncover the genetic aberrations in GBM. Methods: We analyzed 281 whole-genome sequencing data of patients with GBM, of which 152 cases are from our cohort with deep coverage (×120) and 129 cases are from a public database. Results: The median numbers of somatic mutations and structural variants (SVs) per case are 3.0/Mb and 62.5, respectively. While a complex SV is rare in other malignant brain tumors such as IDH-mutant glioma (35% of samples), a large proportion of GBM cases (85%) have complex SV with ≥10 breakpoints. CDKN2A/B homozygous deletions (HDs) are usually comprised of a simple deletion in IDH-mutant glioma whereas about a quarter of CDKN2A/B HDs in GBM are induced by complex SVs. In addition, 30.5% of extrachromosomal DNA (ecDNA) involves multiple chromosomes. Taken together, complex SVs could play a key role in the initiation and progression of GBM. Our deep WGS enables us to delineate a fine view of clonal architecture, where mutational signature varies between clonal and subclonal mutations. The majority of clonal mutations consist of the clock-like signature, whereas subclonal mutations have a relatively low proportion of the clock-like signature. Instead, several other signatures including the APOBEC signature significantly increase in subclones, presuming that various mutational processes along with the clock-like signature contribute to the GBM pathogenesis in its progression phase. Conclusions: GBM evolves through exploiting complex structural variants involving multiple driver genes and the accumulation of genetic mutations caused by distinct mechanisms depending on its developmental stage.

Dynamic Mutational Landscape of Cerebrospinal Fluid Circulating Tumor DNA and Predictors of Survival after Proton Craniospinal Irradiation for Leptomeningeal Metastases.

Proton craniospinal irradiation (pCSI) is a promising treatment for patients with solid tumor leptomeningeal metastasis (LM). We hypothesize that genetic characteristics before and changes resulting after pCSI will reflect clinical response to pCSI. We analyzed the cerebrospinal fluid (CSF) circulating tumor DNA (ctDNA) from patients receiving pCSI for LM and explored genetic variations associated with response. We subjected CSF from 14 patients with LM before and after pCSI to cell-free DNA sequencing using a targeted-sequencing panel. In parallel, plasma ctDNA and primary tumors were subjected to targeted sequencing. Variant allele frequency (VAF) and cancer cell fraction (CCF) were calculated; clonality of observed mutations was determined. Kaplan-Meier analysis was used to associate genomic changes with survival. The median overall survival (OS) for the cohort was 9 months [interquartile range (IQR), 5-21 months]. We showed clonal evolution between tumor and ctDNA of the CSF and plasma with unique mutations identified by compartment. Higher CSF ctDNA mean VAF before pCSI (VAFpre) had worse OS (6 months for VAFpre ≥ 0.32 vs. 9 months for VAFpre < 0.32; P = 0.05). Similarly, increased VAF after pCSI portended worse survival (6 vs. 18 months; P = 0.008). Higher mean CCF of subclonal mutations appearing after pCSI was associated with worse OS (8 vs. 17 months; P = 0.05). In patients with solid tumor LM undergoing pCSI, we found unique genomic profiles associated with pCSI through CSF ctDNA analyses. Patients with reduced genomic diversity within the leptomeningeal compartment demonstrated improved OS after pCSI suggesting that CSF ctDNA analysis may have use in predicting pCSI response.

TLsub: A transfer learning based enhancement to accurately detect mutations with wide-spectrum sub-clonal proportion.

Mutation detecting is a routine work for sequencing data analysis and the trading of existing tools often involves the combinations of signals on a set of overlapped sequencing reads. However, the subclonal mutations, which are reported to contribute to tumor recurrence and metastasis, are sometimes eliminated by existing signals. When the clonal proportion decreases, signals often present ambiguous, while complicated interactions among signals break the IID assumption for most of the machine learning models. Although the mutation callers could lower the thresholds, false positives are significantly introduced. The main aim here was to detect the subclonal mutations with high specificity from the scenario of ambiguous sample purities or clonal proportions. We proposed a novel machine learning approach for filtering false positive calls to accurately detect mutations with wide spectrum subclonal proportion. We have carried out a series of experiments on both simulated and real datasets, and compared to several state-of-art approaches, including freebayes, MuTect2, Sentieon and SiNVICT. The results demonstrated that the proposed method adapts well to different diluted sequencing signals and can significantly reduce the false positive when detecting subclonal mutations. The codes have been uploaded at https://github.com/TrinaZ/TL-fpFilter for academic usage only.

Prognostic Impact of Clonal Hierarchy of Myelodysplasia-Related Gene Mutations in AML Patients

Introduction: The 2022 ICC classification identifies mutations in nine genes that classify AML with myelodysplasia-related gene mutations (AML-MRG). AML-MRG is associated with an unfavourable prognosis according to the 2022 ELN classification. However, it is unknown whether the clonal hierarchy of those mutations impacts prognosis. Methods: 485 adult newly-diagnosed AML patients (median age 54) with available genetic and follow-up data were included. Mutations present at diagnosis were identified by Illumina myeloid panel sequencing covering 56 AML-associated genes. Patients received standard induction and consolidation chemotherapy or underwent allogeneic hematopoietic cell transplantation (alloHCT). MRG mutations were classified as subclonal if their variant allele frequency (VAF) was ≥ 10% below the mutation with the highest VAF of other co-occuring mutations. Otherwise, MRG mutations were considered part of the dominant clone. Results: 177 (36.5%) patients carried at least one of the nine MRG mutations ASXL1, BCOR, EZH2, RUNX1, SF3B1, SRSF2, STAG2, U2AF1 and ZRSR2. The MRG mutation was found in the dominant clone in 134 (75.7%) and in a subclone in 43 patients (24.3%). BCOR, RUNX1, SRSF2, STAG2 and U2AF1 were more likely found in the dominant clone, whereas ASXL1, EZH2, SF3B1 and ZRSR2 were more often found in a subclone. Patients with MRG mutations in the dominant as well as a subclone were significantly older than patients without MRG mutations (p < 0.001 and p = 0.03, respectively). Patients with MRG mutations in the dominant or a subclone and patients without MRG mutations achieved CR/CRi in 80.6%, 83.7% and 87.3%, respectively (p = n.s.). In line with the ELN classification, patients with MRG mutations in the dominant clone had a worse OS than patients without MRG mutations (p = 0.004, HR = 1.505, 95%CI 1.136-1.995); however, patients with MRG mutations in a subclone had a similar OS compared to patients with MRG mutations in the dominant clone and to patients without MRG mutations (Figure 1A). TP53 mutations with ≥10% VAF are defined as AML with mutated TP53 by ICC 2022. 39 (8%) patients carried a TP53 mutation. One third of the cases had a subclonal TP53 mutation. Median TP53 VAF was 39.6% and 15.5% in patients with dominant or subclonal TP53 mutation (p < 0.001). OS was significantly worse in patients with dominant TP53 mutation compared to TP53 wildtype patients (p = 0.002, HR = 2.047, 95%CI 1.276-3.284), whereas OS was similar in patients with subclonal TP53 mutation compared to TP53 wildtype patients (Figure 1B). Conclusion: In our study, 76% of MRG mutations occur in the dominant clone in newly diagnosed AML patients. Dominant MRG mutations have a worse prognosis compared to patients without MRG mutations, while prognosis of subclonal MRG mutations should be evaluated in larger cohorts. AML patients with subclonal TP53 mutation have a similar outcome to patients without TP53 mutation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

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

Detecting the Unusual without Compromising Diagnostic Accuracy - a Prospective WGS/Wts Pilot Study in Acute Leukemias Provided Additional Information for Diagnosis, Prognosis and Treatment

Background: Following WHO and ELN guidelines, standard diagnostics and risk stratification for acute leukemias (AML and ALL) are based on genetic information gathered by a combination of cytogenetics and targeted molecular profiling. Today, comprehensive techniques, i.e WGS and WTS (whole genome and transcriptome sequencing) and the accompanying clinical bioinformatics workflows may even refine and individualize diagnostics. Aim: To evaluate the clinical utility and added value of WGTS in a real-world setting in comparison to standard of care genetic diagnostics. Patients and Methods: The study enrolled 113 patients, 97 with AML and 16 with ALL, diagnosed according to WHO criteria based on cytomorphology, immunophenotyping, cyto- and molecular genetics. WGS (100x) and WTS (50 Mio reads) were performed on NovaSeq instruments. For selected AML cases CD3+ MAC sorted cells were used as normal controls (30x). Variants were called with Strelka2, Manta and GATK using a tumor w/o normal pipeline, fusions with Arriba, STAR-Fusion and Manta. Results: All patients were analyzed for structural variants (SV), copy number variations (CNV) and copy-neutral loss of heterozygosity (cnLOH) based on WGS data and the presence of fusion transcripts and expression patterns based on WTS data. Samples without a normal control (n = 74) were analyzed for 121 genes recurrently mutated in myeloid/lymphoid neoplasms, whereas for 39 AML cases with CD3+ normal controls the whole exome was assessed. Overall, 308/333 (92%) gene mutations were consistently identified by targeted sequencing and WGS. Due to the higher read depth, 25/333 (8%) sub-clonal mutations were only detected by targeted sequencing, with 23/25 showing a variant allele fraction <10%. However, 40 gene mutations were identified exclusively by WGS due to the unrestricted analytical scope (e.g. PTPN11, DDX54). Every patient carried on average 3.3 gene mutations within the 121 genes of interest. The number increased to 63 variants (mean; range: 46-116) when analyzing the exome. 52/63 variants are most likely germline variants based on occurrence in CD3+ control sample. Hence, the CD3+ control samples reduced the number of somatic variants in the exome to 11 (mean; range: 1-23) mostly patient-specific mutations that can be explored for personalized treatment options. Addressing cytogenetics with relevance to risk stratification and assignment to AML/ALL genetic subgroups, WTS identified all entity defining fusions and WGS and chromosome banding analysis (CBA) showed a very high concordance. In 10 patients WGS missed subclonal aberrations due to small clone size (median clone size detected by IP-FISH analysis: 6%). However, WGS provided additional information in 40 patients, mostly SV, CNV and cnLOH of unknown significance, but also alterations of biological interest and prognostic and/or therapeutic importance: in one AML case WGTS revealed a KMT2A::MLLT10 rearrangement that was cytogenetically cryptic and not detectable by FISH (KMT2A break apart probe), assigning the patient to the adverse risk group. WGS also identified a cytogenetically cryptic SV leading to an ETV6::ABL1 rearrangement in an ALL patient with complex karyotype. Retrospective analysis showed that the aberration had already been present when the patient presented with an MPN and a normal karyotype (CBA) two years earlier. Hence, WGS led to the diagnosis of an ETV6::ABL1-positive neoplasm in blast phase (poor prognosis), which might be treated with tyrosine kinase inhibitors. Of note, in 3 AML patients a 7q deletion or complex SV leading to a 7q deletion resulted in a CDK6::NOM1 fusion transcript and high MNX1 expression. In addition, WGS revealed in one patient with ETP-ALL a cytogenetically cryptic translocation t(7;14)(q36;q32) affecting the IGH locus and the NOM1 gene, accompanied by high MNX1 expression. Elevated MNX1 expression as a result of t(7;12)(q36;p13) is a very rare event in pediatric AML, but our findings might suggest a more important role of MNX1 expression in adult acute leukemia than previously known. Conclusions: This prospective study in a real-world setting 1) demonstrated the diagnostic power of WGTS with an accuracy comparable to gold standard techniques, 2) added important prognostic information, 3) identified potential new therapeutic targets and 4) gave new insights into disease biology. WGTS significantly will improve and refine diagnostics in the future.

BCL6-Mediated Escape from Negative Selection Enables RAS-Driven B-Cell Transformation

Background: In B-cell acute lymphoblastic leukemia (B-ALL), RAS-pathway activating mutations (e.g. NRAS, KRAS, BRAF) are found in ~40% of cases (Irving et al., 2014; Jerchel et al., 2018; Zhang et al., 2011). Evidence suggests that RAS-pathway mutations in B-ALL subclones are selected for during therapy and are associated with relapse, which has a poor prognosis and remains one of the leading causes of death in children with malignancy (Irving et al., 2014; Jerchel et al., 2018). Beyond its role as a driver of B-ALL, ERK-activation can induce expression of PRDM1, which functions as a tumor suppressor and triggers negative B-cell selection and cell death (Yasuda et al., 2011; Hug et al., 2014; Setz et al., 2018). During B-cell development, PRDM1 and the proto-oncogene BCL6 are reciprocal antagonists (Shaffer et al., 2002). Here, we examined the mechanism by which oncogenic RAS-ERK signaling in B-ALL avoids negative selection and cell death to promote leukemogenesis. Results: Mimicking survival signals from pre-B cell receptor (pre-BCR), constitutive Erk activation upon oncogenic RAS (NRASG12D) expression in mouse B-cell precursors massively induced the mRNA (~390-fold) and protein (~50-fold) levels of the proto-oncogene Bcl6. Highlighting an essential role of Erk-signaling in upregulation of Bcl6 expression downstream of oncogenic RAS, pharmacological Erk-inhibition markedly reduced NRASG12D-mediated Bcl6 induction. Both the pre-BCR linker molecule Blnk and the surrogate light chains (SLCs) of the pre-BCR are required for Erk-mediated induction of Bcl6. Genetic ablation of Blnk or the SLC component λ5 abrogated RAS-mediated Bcl6-induction. Hence, oncogenic expression of Bcl6 in RAS-driven B-ALL cells depended on structural elements of the pre-BCR. Additionally, ERK-activation in patient-derived B-ALL cells had the same effect on BCL6 expression. Pharmacological reactivation of ERK (BCI-215) markedly induced BCL6 expression, whereas small molecule inhibition of ERK-activity (trametinib) reduced BCL6 levels in patient-derived B-ALL cells. Consistent with BCL6-transactivation downstream of ERK-activation, ChIP-seq analysis revealed binding of ERK-dependent transcription factors (CREB1, ELK1, EGR1, JUND and C-JUN) to the BCL6 promoter in human B cells. Altogether, our findings illuminated a critical role of ERK-signaling in positive regulation of BCL6 expression. Using a genetic model for Cre-mediated deletion of Bcl6 in NRASG12D B-ALL cells, we found that Bcl6 was required for leukemia-initiation in transplant recipient mice (P=0.007). Patient-derived RAS-driven B-ALL cells were highly sensitive to pharmacological inhibition of BCL6 using peptide (RI-BPI) and small molecule (FX1) inhibitors. Furthermore, pharmacological inhibition of BCL6 delayed onset of fatal disease and prolonged survival of transplant recipient mice bearing patient-derived RAS-driven B-ALL cells (P=0.009). Taken together, our findings demonstrated that suppressing BCL6 function is a promising option for targeted therapeutics in RAS-driven B-ALL. Mechanistically, genetic ablation or pharmacological inhibition of BCL6 increased PRDM1 expression. Gene expression analysis revealed that ectopic expression of Prdm1 in B-cell precursors promoted gene expression programs of apoptosis and cell cycle arrest (GSE111692). While Bcl6-deletion resulted in cell death of B-ALL cells, this was largely reversed by shRNA-mediated silencing of Prdm1. Altogether, our results showed that compromised leukemogenesis was a result of aberrant PRDM1 expression in BCL6-deficient RAS-driven B-ALL cells. Conclusions: We previously investigated targeted engagement of negative selection in B-ALL and showed that hyperactivation of SYK or AKT functionally mimics autoreactive pre-BCR signaling and results in negative selection and cell death (Chen et al., 2015; Shojaee et al., 2016). Here, we reinforced this concept by providing a mechanistic basis on how negative selection is downregulated in RAS-driven B-ALL and how it can be reactivated for therapeutic benefit - namely, BCL6 curbs ERK-mediated induction of PRDM1 expression to promote leukemogenesis and this mechanism can be exploited as synthetic lethality in the treatment of this disease.

DNMT3A/TET2 Mutant Clonal Hematopoiesis in Vexas Syndrome Results in DNA Hypomethylation and Transcriptional Activation of WT1 and MPL Oncogenic Pathways

Introduction: VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) syndrome is characterized by adult-onset inflammatory manifestations with predisposition to myelodysplastic syndromes (MDS). Given the higher median age at onset and prevalent inflammation, approximately 60% of patient have concurrent clonal hematopoiesis (CH) involving epigenetic regulators (epi), DNMT3A and TET2, with a negative impact on survival (see accompanying ASH abstract). DNMT inhibitors, considered standard of care for MDS, have anecdotally been used in VEXAS with mixed responses. However, to date, the epigenetic landscape in VEXAS has not been defined. We carried out this study to assess changes in DNA methylation in VEXAS syndrome. Methods: After IRB approval, Mayo Clinic and NIH patients with a confirmed diagnosis of VEXAS syndrome were included. DNA was extracted from blood mononuclear cells and error-corrected DNA sequencing was performed to define the CH landscape (see accompanying ASH abstract). DNA Methylation changes were identified using the Illumina Infinium MethylationEPIC BeadChip. Conventional bisulfite-treatment was completed using the TrueMethyl oxBS Module (Tecan Genomics, Männedorf, Switzerland). Briefly, DNA was purified using magnetic beads, denatured, and underwent bisulfite conversion followed by desulfonation and purification. The TrueMethyl converted DNA samples were then eluted and processed through the MethylationEPIC BeadChip array protocol. Subset-quantile Within Array Normalization (SWAN) was performed via the R package "minfi". The resultant β-values were used to identify changes in DNA methylation (Δβ) between groups. Whole transcriptome RNAseq is currently pending on a subset of patients. Results: Fifty male patients with VEXAS, without a diagnosis of MDS were included in this study; median age 70.5 years (range: 48 - 89). "Typical non-UBA1 CH mutations" were seen in 30 (60%) patients, predominantly involving DNMT3A and TET2 (DNMT3A n=10, TET2 n=8, DNMT3A/TET2 n=6; 48%). Subclonal CH mutations in PPM1D (N=2), KRAS and TP53 (n=1 each) were seen in 4 patients with variant allele fractions (VAF) <2%. The median VAF for DNMT3A was 28% and for TET2 was 2%, respectively. Thirty (60%) patients had UBA1Met41T substitutions, 14 (28%) had Met41V and 3 (8%) each had Met41L substitutions and splice site mutations, respectively. DNA methylation was first compared to 8 healthy males, with a median age of 59 years (Age range: 42 - 61). VEXAS patients had >2,800 differentially methylated CpGs, with 77% of the CpGs being hypermethylated compared to the healthy controls. Additional age matched controls are currently being assessed to eliminate the age bias in this analysis. We then compared methylation changes in VEXAS patients with DNMT3A and/or TET2 CH with those without (n=24 vs n=36, median ages not significantly different). VEXAS patients with CH-epi mutations had >900 differentially methylated CpGs when compared to VEXAS patients without CH-epi. Majority of these differentially methylated CpGs were hypomethylated (79%) in the VEXAS patients with CH-epi (Figure 1A). Ingenuity pathway analysis demonstrated that the hypomethylated loci resulted in transcriptional activation of the following pathways/genes; apoptosis of peripheral T lymphocytes (p=0.003, EZR and MAPK8) and multiple convergent myeloid neoplasm pathways (p<0.003, WT1 and MPL). Hypermethylated loci in CH-epi resulted in transcriptional repression of cell death of M0 macrophages (p=0.0048, IRF8), expansion of gamma/delta T cells (p=0.008, BTNL9), activation of peripheral T lymphocytes (p=0.025, MFN2) and aggregation of monocytes (p=0.002, BMP7) (Figure 1B). RNA sequencing confirming these changes is pending in 14 cases. We did not assess differences between DNMT3A and TET2 mutant groups given the smaller samples size and the fact that the median TET2 mutant VAF was 2%. Conclusions: We demonstrate DNA methylation changes in a large cohort of VEXAS patients, exploring the negative impact of CH-epi on survival. We show that VEXAS patients with CH-epi are enriched in hypomethylated CpG sites in comparison to those without CH-epi. Pathway analysis suggest roles for WT1, a critical transcription factor and oncogene and MPL, the platelet thrombopoietin receptor, in disease biology and survival. Integration with RNAseq and recruitment of additional patients and controls is ongoing. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Targeting Cell Adhesion Molecules in IDH Mutant Myeloid Malignances

Neomorphic mutations in the genes encoding isocitrate dehydrogenase (IDH) are common in human cancers including myeloid malignancies. The oncometabolite (R)-2-hydroxyglutarate (R-2-HG) produced by mutant IDH promotes leukemia development by competitively inhibiting α-ketoglutarate (α-KG)-dependent dioxygenases, causing epigenetic dysregulation of histone and DNA methylation in myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML). FDA-approved allosteric inhibitors (ivosidenib and enasidenib targeting mutant IDH1 and IDH2 respectively) are used for treating newly diagnosed and/or relapsed/refractory AML patients with IDH mutations. However, the low response rate and primary resistance to IDHmut inhibitors are major clinical challenges. Moreover, acquired resistance to IDH inhibition can occur through multiple mechanisms, including R-2-HG-restoring second-site IDH mutations and the selection of sub-clonal mutations in RUNX1/CEBPA or RAS-RTK pathway. These findings underscore the importance of developing new and more effective approaches for the treatment of IDH-mutant MDS and AML patients. To identify mutant IDH-mediated alterations and actionable dependencies, we generated isogenic TF-1 myeloid leukemia cell lines containing endogenous, hotspot IDH mutations (IDH1R132H, IDH2R140Q, IDH2R140W, and IDH2R172K) by CRISPR/Cas9-mediated base-editing. By comparative transcriptomic (RNA-seq) and epigenomic (ATAC-seq and ChIP-seq) profiling of base-edited versus parental isogenic leukemia cells, we identified cell adhesion molecules, especially CD44, as the top upregulated genes in IDH-mutant leukemia cells. CD44 expression is significantly upregulated in IDH-mutant AML patient samples relative to normal CD34+ hematopoietic stem/progenitor cells (HSPCs) or other AML samples without IDH mutations. Increased CD44 expression is also associated with overall poorer survival and therapy resistance in independent AML cohorts. Importantly, we found that ectopic CD44 expression promotes cytokine-independent cell growth of TF-1 myeloid leukemia cells, resembling the effect observed upon expression of mutant IDH. Moreover, CD44 depletion led to decreased R-2-HG levels in base-edited IDH-mutant TF-1 leukemia cells, and impaired the propagation of IDH-mutant leukemia cells in vitro and in xenotransplanted recipient mice. These studies establish direct evidence that IDH mutations cause aberrant activation of cell adhesion molecules such as CD44, and CD44 is required for the propagation of IDH-mutant myeloid leukemia cells. Mechanistically, we found that the aberrantly activated CD44 is first cleaved by matrix metalloproteinases MMP2 and MMP9 for proteolytic release of its ectodomain from membrane, followed by γ-secretase-mediated intramembranous cleavage to produce the CD44 intracellular domain (CD44ICD). CD44ICD subsequently activates the pentose phosphate pathway (PPP) for the production of NADPH, a critical cofactor for mutant IDH enzymes to convert α-KG to R-2-HG. Hence, mutant IDH-mediated activation of CD44 functions to rewire intracellular metabolism to facilitate IDHmut-catalyzed R-2-HG production, supporting a model for mutant IDH-induced feed-forward oncogenic programs. CD44 also activates AKT/ERK signaling pathway to promote the proliferation and cytokine-independent growth of IDH-mutant leukemia cells. More importantly, primary and resistant IDH-mutant leukemia cells are sensitive to combination regimens consisting of IDHmut inhibitors and humanized monoclonal antibody specific for CD44 (RG7536) or MMP2/9 inhibitors (SB-3CT) invitro and invivo. Taken together, our findings uncover the cell adhesion molecule-mediated metabolic signaling as a new targetable dependency of IDH-mutant leukemia cells, with therapeutic implications for more effective treatment of primary and resistant IDH-mutant myeloid malignancies.

Dynamic monitoring of cerebrospinal fluid circulating tumor DNA to identify unique genetic profiles of brain metastatic tumors and better predict intracranial tumor responses in non-small cell lung cancer patients with brain metastases: a prospective cohort study (GASTO 1028)

BackgroundDue to the blood-brain barrier, plasma is not an ideal source to evaluate the genetic characteristics of central nervous system tumors. Thus, cerebrospinal fluid (CSF) is becoming an alternative biopsy type to evaluate the genetic landscape of intracranial tumors. We aimed to explore the genetic profiles of CSF-derived circulating tumor DNA (ctDNA) to predict intracranial tumor responses and monitor mutational evolution during the treatment of non-small cell lung cancer (NSCLC) patients with brain metastases.MethodsWe conducted a prospective study of 92 newly diagnosed NSCLC patients with brain metastases. Paired CSF and plasma samples were collected at baseline, 8 weeks after treatment initiation, and disease progression. All samples underwent next-generation sequencing of 425 cancer-related genes.ResultsAt baseline, the positive detection rates of ctDNA in CSF, plasma, and extracranial tumors were 63.7% (58/91), 91.1% (82/90), and 100% (58/58), respectively. A high level of genetic heterogeneity was observed between paired CSF and plasma, while concordance in driver mutations was also observed. A higher number of unique copy number variations was detected in CSF-ctDNA than in plasma. ctDNA positivity of CSF samples at baseline was associated with poor outcomes (HR=2.565, P=0.003). Moreover, patients with ≥ 50% reductions in the concentrations of CSF ctDNA after 8 weeks of treatment had significantly longer intracranial progression-free survivals (PFS) than patients with < 50% reductions in CSF ctDNA concentrations (13.27 months vs 6.13 months, HR=0.308, P=0.017). A ≥ 50% reduction in CSF ctDNA concentrations had better concordance with radiographic intracranial tumor responses than plasma. A ≥ 50% reduction in plasma ctDNA concentrations was also associated with longer extracranial PFS (11.57 months vs 6.20 months, HR=0.406, P=0.033). Based on clonal evolution analyses, the accumulation of subclonal mutations in CSF ctDNA was observed after 8 weeks of treatment. The clonal mutations that remained in more than 80% in CSF after 8 weeks also predicted shorter intracranial PFS (HR=3.785, P=0.039).ConclusionsCSF ctDNA exhibited unique genetic profiles of brain metastases, and dynamic changes in CSF ctDNA could better predict intracranial tumor responses and track clonal evolution during treatment in NSCLC patients with brain metastases.Trial registrationClinicalTrials.gov identifier: NCT 03257735.

Resistance Mechanisms to Anti-Epidermal Growth Factor Receptor Therapy in RAS/RAF Wild-Type Colorectal Cancer Vary by Regimen and Line of Therapy.

Acquired resistance to anti-epidermal growth factor receptor (EGFR) inhibitor (EGFRi) therapy in colorectal cancer (CRC) has previously been explained by the model of acquiring new mutations in KRAS/NRAS/EGFR, among other MAPK-pathway members. However, this was primarily on the basis of single-agent EGFRi trials and little is known about the resistance mechanisms of EGFRi combined with effective cytotoxic chemotherapy in previously untreated patients. We analyzed paired plasma samples from patients with RAS/BRAF/EGFR wild-type metastatic CRC enrolled in three large randomized trials evaluating EGFRi in the first line in combination with chemotherapy and as a single agent in third line. The mutational signature of the alterations acquired with therapy was evaluated. CRC cell lines with resistance to cetuximab, infusional fluorouracil, leucovorin, and oxaliplatin, and SN38 were developed, and transcriptional changes profiled. Patients whose tumors were treated with and responded to EGFRi alone were more likely to develop acquired mutations (46%) compared with those treated in combination with cytotoxic chemotherapy (9%). Furthermore, contrary to the generally accepted hypothesis of the clonal evolution of acquired resistance, we demonstrate that baseline resistant subclonal mutations rarely expanded to become clonal at progression, and most remained subclonal or disappeared. Consistent with this clinical finding, preclinical models with acquired resistance to either cetuximab or chemotherapy were cross-resistant to the alternate agents, with transcriptomic profiles consistent with epithelial-to-mesenchymal transition. By contrast, commonly acquired resistance alterations in the MAPK pathway do not affect sensitivity to cytotoxic chemotherapy. These findings support a model of resistance whereby transcriptomic mechanisms of resistance predominate in the presence of active cytotoxic chemotherapy combined with EGFRi, with a greater predominance of acquired MAPK mutations after single-agent EGFRi. The proposed model has implications for prospective studies evaluating EGFRi rechallenge strategies guided by acquired MAPK mutations, and highlights the need to address transcriptional mechanisms of resistance.

The Dynamic Mutational Landscape of Cerebrospinal Fluid Circulating Tumor DNA can Predict Survival after Proton Craniospinal Irradiation for Leptomeningeal Metastasis

<h3>Purpose/Objective(s)</h3> Solid tumor leptomeningeal metastasis (LM) is a devastating disease with poor survival. Proton craniospinal irradiation (pCSI) is a promising treatment for LM, and there is a need for biomarkers for patient selection. Understanding the genetic variation of LM pre- and post-pCSI can elucidate differences in response to radiotherapy. We analyzed the observed mutations in cerebrospinal fluid (CSF) circulating tumor DNA (ctDNA) of LM patients treated with pCSI and explored whether the variant allele fraction (VAF) can predict survival. <h3>Materials/Methods</h3> We examined 14 patients from a phase I trial pCSI for solid tumor LM (NCT03281889) with pre- and post-pCSI ctDNA assessment. Mutations were identified using an institutional targeted-sequencing panel, MSK-IMPACT, and allele specific copy number variation was used to determine mutation clonality. Primary tumors or solid tumor metastases, and ctDNA in the plasma was also analyzed. VAF and cancer cell fraction (CCF) were calculated. An independent clinical sequencing cohort of 50 LM patients who did not receive pCSI was used for comparison. Kaplan Meier analysis was used to associate mutation measures and survival outcomes. <h3>Results</h3> The median overall survival for the cohort was 9 months (IQR: 5-21 months). The median number of mutations observed in the CSF ctDNA pre- pCSI was 10 (range:0-25) and post-pCSI was 9 (range:0-42). We observed the clonal evolution between solid tissue and ctDNA of the CSF and plasma, and we identified mutations that are unique by compartment, with more mutations seen in the CSF compared to plasma. A higher VAF pre-pCSI (VAF_pre) was associated with worse OS (8 months for VAF_pre≥0.32 [median] vs 21 months for VAF_pre<0.32, p=0.05). Patients with increased VAF after pCSI had significantly worse survival (6 months vs 18 months, p=0.008). The CCF of novel subclonal mutations after pCSI showed that higher values were associated with worse OS (8 months vs 17 months, p=0.05). In the comparison cohort, VAF≥0.37 [median] was associated with worse OS after CNS-directed RT (8 months vs 9 months, p=0.05). <h3>Conclusion</h3> In LM patients undergoing pCSI, there are unique baseline mutation profiles and distinct responses to pCSI that may be able to predict survival after radiotherapy. Analysis of the ctDNA of LM patients before and after pCSI shows that improved outcomes are associated with lower frequency of mutations. The quantification of these mutations has important implications for patient stratification in CNS-directed radiotherapy and escalation of therapy after pCSI.

Genomic characterization reveals distinct mutation landscapes and therapeutic implications in neuroendocrine carcinomas of the gastrointestinal tract.

Neuroendocrine carcinomas of the gastrointestinal tract (GI-NECs) remain a disease of grim prognosis with limited therapeutic options. Their molecular characteristics are still undefined. This study aimed to explore the underlying genetic basis and heterogeneity of GI-NECs. Comprehensive genomic analysis using whole-exome sequencing was performed on 143 formalin-fixed, paraffin-embedded samples of surgically resected GI-NEC with a thorough histological evaluation. Mutational signatures, somatic mutations, and copy number aberrations were analyzed and compared across anatomic locations and histological subtypes. Survival analysis was conducted to identify the independent factors. In total, 143 GI-NECs were examined: the stomach, 87 cases (60.8%); the esophagus, 29 cases (20.3%); the colorectum, 20 cases (14.0%); and the small intestine, 7 cases (4.9%). Eighty-three (58.0%) and 60 (42.0%) cases were subclassified into small cell and large cell subtypes, respectively. GI-NECs showed distinct genetic alterations from their lung counterparts and non-neuroendocrine carcinomas in the same locations. Obvious heterogeneity of mutational signatures, somatic mutations, and copy number variations was revealed across anatomic locations rather than histological subtypes. Except for tumor protein p53 (TP53) and retinoblastoma 1 (RB1), the most frequently mutated genes in the stomach, esophagus, colorectum, and small intestine were low-density lipoprotein receptor-related protein 1B (LRP1B), notch receptor 1 (NOTCH1), adenomatosis polyposis coli (APC), catenin beta 1 (CTNNB1), respectively. Mutations in the WNT-β-catenin, NOTCH and erythroblastic leukemia viral oncogene B (ERBB) pathways were prevalently identified in gastric, esophageal, and colorectal NECs, respectively. Importantly, 104 (72.7%) GI-NECs harbored putative clinically relevant alterations, and non-gastric location and RB1 bi-allelic inactivation with copy number alterations were identified as two independent poor prognostic factors. Furthermore, we found that tumor cells in GI-NECs first gain clonal mutations in TP53, RB1, NOTCH1 and APC, followed by subsequent whole-genome doubling (WGD) and post-WGD clonal mutations in LRP1B, CUB and Sushi multiple domains 3 (CSMD3), FAT tumor suppressor homolog 4 (FAT4) and erb-b2 receptor tyrosine kinase 4 (ERBB4), and finally develop subclonal mutations. GI-NECs harbor distinct genomic landscapes and demonstrate significant genetic heterogeneity across different anatomic locations. Moreover, potentially actionable alterations and prognostic factors were revealed for GI-NECs.

APOBEC mutagenesis, kataegis, chromothripsis in EGFR-mutant osimertinib-resistant lung adenocarcinomas

Studies of targeted therapy resistance in lung cancer have primarily focused on single-gene alterations. Based on prior work implicating apolipoprotein b mRNA-editing enzyme, catalytic polypeptide-like (APOBEC) mutagenesis in histological transformation of epidermal growth factor receptor (EGFR)-mutant lung cancers, we hypothesized that mutational signature analysis may help elucidate acquired resistance to targeted therapies. APOBEC mutational signatures derived from an Food and Drug Administration-cleared multigene panel [Memorial Sloan Kettering Cancer Center Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT)] using the Signature Multivariate Analysis (SigMA) algorithm were validated against the gold standard of mutational signatures derived from whole-exome sequencing. Mutational signatures were decomposed in 3276 unique lung adenocarcinomas (LUADs), including 93 paired osimertinib-naïve and -resistant EGFR-mutant tumors. Associations between APOBEC and mechanisms of resistance to osimertinib were investigated. Whole-genome sequencing was carried out on available EGFR-mutant lung cancer samples (10 paired, 17 unpaired) to investigate large-scale genomic alterations potentially contributing to osimertinib resistance. APOBEC mutational signatures were more frequent in receptor tyrosine kinase (RTK)-driven lung cancers (EGFR, ALK, RET, and ROS1; 25%) compared to LUADs at large (20%, P < 0.001); across all subtypes, APOBEC mutational signatures were enriched in subclonal mutations (P < 0.001). In EGFR-mutant lung cancers, osimertinib-resistant samples more frequently displayed an APOBEC-dominant mutational signature compared to osimertinib-naïve samples (28% versus 14%, P= 0.03). Specifically, mutations detected in osimertinib-resistant tumors but not in pre-treatment samples significantly more frequently displayed an APOBEC-dominant mutational signature (44% versus 23%, P < 0.001). EGFR-mutant samples with APOBEC-dominant signatures had enrichment of large-scale genomic rearrangements (P= 0.01) and kataegis (P= 0.03) in areas of APOBEC mutagenesis. APOBEC mutational signatures are frequent in RTK-driven LUADs and increase under the selective pressure of osimertinib in EGFR-mutant lung cancer. APOBEC mutational signature enrichment in subclonal mutations, private mutations acquired after osimertinib treatment, and areas of large-scale genomic rearrangements highlights a potentially fundamental role for APOBEC mutagenesis in the development of resistance to targeted therapies, which may be potentially exploited to overcome such resistance.

109P Improved accuracy of somatic variant detection from high throughput sequencing (HTS) data via site-specific noise estimation

109P Improved accuracy of somatic variant detection from high throughput sequencing (HTS) data via site-specific noise estimation

The spectrum of subclonal TP53 mutations in chronic lymphocytic leukemia: A next generation sequencing retrospective study.

Chronic lymphocytic leukemia (CLL) is a hematological disorder with complex clinical and biological behavior. TP53 mutational status and cytogenetic assessment of the deletion of the corresponding locus (17p13.1) are considered the most relevant biomarkers associated with pharmaco-predictive response, chemo-refractoriness, and worse prognosis in CLL patients. The implementation of Next Generation Sequencing (NGS) methodologies in the clinical laboratory allows for comprehensively analyzing the TP53 gene and detecting mutations with allele frequencies ≤10%, that is, "subclonal mutations". We retrospectively studied TP53 gene mutational status by NGS in 220 samples from 171 CLL patients. TP53 mutations were found in 60/220 (27.3%) samples and 47/171 (27.5%) patients. Interestingly, subclonal mutations could be detected in 31/60 samples (51.7%) corresponding to 25 patients (25/47, 53.2%). We identified 44 distinct subclonal TP53 mutations clustered in the central DNA-binding domain of p53 protein (exons 5-8, codons 133-286). Missense mutations were predominant (>80%), whereas indels, nonsense, and splice site variants were less represented. All subclonal TP53 variants but one [p.(Pro191fs)] were already described in NCI and/or Seshat databases as "damaging" and/or "probably damaging" mutations (38/44, 86% and 6/44, 14%, respectively). Longitudinal samples were available for 37 patients. Almost half of them displayed at least one TP53 mutant subclone, which could be alone (4/16, 25%) or concomitant with other TP53 mutant clonal ones (12/16, 75%); different patterns of mutational dynamics overtimes were documented. In conclusion, utilization of NGS in our "real-life" cohort of CLL patients demonstrated an elevated frequency of subclonal TP53 mutations. This finding indicates the need for precisely identifying these mutations during disease since the clones carrying them may become predominant and be responsible for therapy failures.

Triple MAPK inhibition salvaged a relapsed post-BCMA CAR-T cell therapy multiple myeloma patient with a BRAF V600E subclonal mutation

BackgroundMultiple Myeloma (MM) is a progressive plasma cell neoplasm characterized by heterogeneous clonal expansion. Despite promising response rates achieved with anti-BCMA CAR-T cell therapy, patients may still relapse and there are currently no clear therapeutic options in post-CAR-T settings. In this report, we present a case of a post-BCMA CAR-T relapsed/refractory (RR) MM patient with skin extramedullary disease (EMD) in which a novel MAPK inhibition combinatorial strategy was implemented based on next-generation sequencing and in vitro experiments.Case presentationA 61-year-old male with penta-refractory MM penta- (IgA lambda), ISS stage 3 with hyperdiploidy, gain of 1q21 and del13 was treated with anti-BCMA CAR-T cell therapy, achieving a best response of VGPR. He progressed after 6 months and was salvaged for a short period with autologous stem cell transplantation. Eventually, he progressed with extramedullary disease manifested as subcutaneous nodules. Based on whole-exome sequencing, we identified a BRAF (V600E) dominant subclone in both bone marrow and cutaneous plasmacytoma. Following in vitro experiments, and according to our previous studies, we implemented a triple MAPK inhibition strategy under which the patient achieved a very good partial response for 110 days, which allowed to bridge him to subsequent clinical trials and eventually achieve a stringent complete response (sCR).ConclusionHere, we show the applicability, effectiveness, and tolerability the triple MAPK inhibition strategy in the context of post-BCMA CAR-T failure in specific subset of patients. The triple therapy could bridge our hospice bound RRMM patient with BRAF (V600E) to further therapeutic options where sCR was achieved. We will further evaluate triple MAPK inhibition in patients with BRAF V600E in a precision medicine clinical trial launching soon.