Isogenic System Research Articles

FOXM1 drives HPV+ HNSCC sensitivity to WEE1 inhibition

Head and neck squamous cell carcinoma (HNSCC) associated with high-risk human papilloma virus (HPV) infection is a growing clinical problem. The WEE1 kinase inhibitor AZD1775 (WEE1i) overrides cell cycle checkpoints and is being studied in HNSCC regimens. We show that the HPV16 E6/E7 oncoproteins sensitize HNSCC cells to single-agent WEE1i treatment through activation of a FOXM1-CDK1 circuit that drives mitotic gene expression and DNA damage. An isogenic cell system indicated that E6 largely accounts for these phenotypes in ways that extend beyond p53 inactivation. A targeted genomic analysis implicated FOXM1 signaling downstream of E6/E7 expression and analyses of primary tumors and The Cancer Genome Atlas (TCGA) data revealed an activated FOXM1-directed promitotic transcriptional signature in HPV+ versus HPV- HNSCCs. Finally, we demonstrate the causality of FOXM1 in driving WEE1i sensitivity. These data suggest that elevated basal FOXM1 activity predisposes HPV+ HNSCC to WEE1i-induced toxicity and provide mechanistic insights into WEE1i and HPV+ HNSCC therapies.

Activity of Birinapant, a SMAC Mimetic Compound, Alone or in Combination in NSCLCs With Different Mutations.

Liver kinase B1 (LKB1/STK11) is the second tumor suppressor gene most frequently mutated in non-small-cell lung cancer (NSCLC) and its activity is impaired in about half KRAS-mutated NSCLCs. Nowadays, no effective therapies are available for patients having these mutations. To highlight new vulnerabilities of this subgroup of tumors exploitable to design specific therapies we screened an US FDA-approved drug library using an isogenic system of wild-type (WT) or deleted LKB1. Among eight hit compounds, Birinapant, an inhibitor of the Inhibitor of Apoptosis Proteins (IAPs), was the most active compound in LKB1-deleted clone only compared to its LKB1 WT counterpart. We validated the Birinapant cells response and its mechanism of action to be dependent on LKB1 deletion. Indeed, we demonstrated the ability of this compound to induce apoptosis, through activation of caspases in the LKB1-deleted clone only. Expanding our results, we found that the presence of KRAS mutations could mediate Birinapant resistance in a panel of NSCLC cell lines. The combination of Birinapant with Ralimetinib, inhibitor of p38α, restores the sensitivity of LKB1- and KRAS-mutated cell lines to the IAP inhibitor Birinapant. Our study shows how the use of Birinapant could be a viable therapeutic option for patients with LKB1-mutated NSCLCs. In addition, combination of Birinapant and a KRAS pathway inhibitor, as Ralimetinib, could be useful for patients with LKB1 and KRAS-mutated NSCLC.

Human Isogenic Cell Line Models for Neutrophils and Myeloid-Derived Suppressor Cells

Neutrophils with immunosuppressive activity are polymorphonuclear myeloid-derived suppressor cells (MDSCs) and may contribute to the resistance to cancer immunotherapy. A major gap for understanding and targeting these cells is the paucity of cell line models with cardinal features of human immunosuppressive neutrophils and their normal counterparts, especially in an isogenic manner. To address this issue, we employ the human promyelocytic cell line HL60 and use DMSO and cytokines (granulocyte macrophage-colony stimulating factor (GM-CSF) and interleukin 6 (IL6)) to induce the formation of either neutrophils or MDSCs. The induced MDSCs are CD11b+ CD33+ HLA-DR−/low and are heterogeneous for CD15 and CD14 expression. The induced MDSCs abrogate IL2 production and activation-induced cell death of the human T cell line Jurkat stimulated by CD3/CD28 antibodies, whereas the induced neutrophils enhance IL2 production from Jurkat cells. The induced MDSCs upregulate the expression of C/EBPβ, STAT3, VEGFR1, FATP2 and S100A8. Lastly, the immunosuppressive activity of the induced MDSCs is inhibited by all-trans retinoic acid and STAT3 inhibitor BP-1-102 through cellular differentiation and dedifferentiation mechanisms, respectively. Together, our study establishes a human isogenic cell line system for neutrophils and MDSCs and this system is expected to facilitate future studies on the biology and therapeutics of human immunosuppressive neutrophils.

Abstract 1434: Loss of PTEN confers resistance to neddylation inhibition through TOP2A in glioblastoma

Abstract As a disease that has not seen lasting advances in its standard of care for decades, glioblastoma (GBM) calls for novel therapeutic approaches, and given GBM's heterogeneous and aggressive nature, precision medicine stands out as an appealing treatment modality for this tragic disease. Neddylation is a post-translational modification process that regulates protein function and turnover. MLN4924, or Pevonedistat, inhibits the E1 ligase Nedd8 activating enzyme (NAE) and, in doing so, the E1, E2, and E3 neddylation cascade. In its most researched role, NEDD-8 is conjugated to the cullin subunit of cullin-RING ligases (CRLs), a family of E3 ligases that facilitate the ubiquitination of different substrates. Thus, inhibition of neddylation decreases ubiquitination and degradation of particular substrates: e.g., p21, p27, wee-1, and other cell cycle regulatory and pro-apoptotic proteins. We and others have shown that neddylation stands out a promising therapeutic approach to cancer and GBM, given the vulnerability to MLN4924 in subset of different in vitro and in vivo models. In this work, we establish the selective vulnerability to MLN4924 in GBM and explore the biological mechanisms that underlie this differential response. Using the Cancer Cell Line Encyclopedia (CCLE) and the Genomics of Drug Sensitivity in Cancer (GDSC) databases, we found that mutations and deletions in phosphatase and tensin homolog (PTEN) are associated with resistance to MLN4924 in both low-grade gliomas and GBM. Our established models validated these findings, as our sensitive models – GB1 (IC50= 0.28 μM) and LN18 (IC50 = 0.19 μM), – expressed PTEN, while our resistant models – M059K (IC50 = 5.5 μM) and SNU1105 (IC50 = 20.9 μM) – did not. Knocking down PTEN with siRNA in our sensitive models increased their IC50 values three- to four-fold. This relationship was further validated in multiple isogenic systems. We sought to uncover whether this PTEN-status dependent resistance was contingent on its well-studied lipid phosphatase activity and PI3K/AKT signaling. While the resistant models had higher baseline levels of phosphorylated AKT (p-AKT), we found that MLN4924 treatment did not consistently alter p-AKT levels in our sensitive and/or resistant models. We also used a PTEN null isogenic glioma stem cell system (GSC23), in which wild-type PTEN and lipid phosphatase dead (G129A) PTEN was re-expressed, and found that both PTEN models sensitized GSC23 to MLN4924 treatment equally. While independent of the PI3K/AKT axis, we did find that there were significant increases in topoisomerase II alpha (TOP2A) levels in the post-treatment lysate of our resistant (PTEN-null) models. Targeting this increase in TOP2A expression, we found strong synergy with two different TOP2 poisons – etoposide and doxorubicin – and MLN4924 in a combination DDR. For the first time, we show that PTEN status not only serves as a novel biomarker for MLN4924 response but also reveals a vulnerability to TOP2A inhibitors when used in combination with MLN4924. Citation Format: Brett Taylor, Shayesteh Ferdosi, Nanyun Tang, Rita Bybee, George Reid, Lauren Hartman, Darren Finlay, Kristiina Vuori, Matthew Lee, Sen Peng, Frank Furnari, Harshil Dhruv, Michael Berens. Loss of PTEN confers resistance to neddylation inhibition through TOP2A in glioblastoma [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1434.

Abstract 1058: Targeting chromatin remodeling-associated genetic vulnerabilities in cancer: PBRM1 defects are synthetic lethal with PARP and ATR inhibitors

Abstract Aim: Polybromo-1 (PBRM1), a specific subunit of the pBAF chromatin remodeling complex, is frequently inactivated in cancer. For example, 40% of clear cell Renal Cell Carcinoma (ccRCC) and 15% of cholangiocarcinoma present deleterious PBRM1 mutations. There is currently no precision medicine-based therapeutic approach that targets PBRM1 defects. To identify novel, targeted therapeutic strategies for PBRM1-defective cancers, we carried out high-throughput functional genomics and drug screenings followed by in vitro and in vivo validation studies. Methods: High-throughput siRNA-drug sensitization and drug sensitivity screens evaluating > 150 cancer-relevant small molecules in dose-response were performed in Pbrm1 siRNA-transfected mouse embryonic stem cells (mES) and isogenic PBRM1-KO or -WT HAP1 cells, respectively. After identification of PBRM1-selective small molecules, revalidation was carried out in a series of in-house-generated isogenic models of PBRM1 deficiency - including 786-O (ccRCC), A498 (ccRCC), U2OS (osteosarcoma) and H1299 (non-small cell lung cancer) human cancer cell lines - and non-isogenic ccRCC models, using multiple clinical compounds. Mechanistic dissection was performed using immunofluorescence, RT-qPCR, western blotting, DNA fiber assay, transcriptomics, proteomics and DRIP-sequencing to evaluate markers of DNA damage response (DDR), replication stress and cell-autonomous innate immune signaling. Preclinical data were integrated with TCGA tumor data. Results: Parallel high-throughput drug screens independently identified PARP inhibitors (PARPi) as being synthetic lethal with PBRM1 defects - a cell type-independent effect which was exacerbated by ATR inhibitors (ATRi) and which we revalidated in vitro in isogenic and non-isogenic systems and in vivo in a xenograft model. PBRM1 defects were associated with increased replication fork stress (higher γH2AX and RPA foci levels, decreased replication fork speed and increased ATM checkpoint activation), R-loop accumulation and enhanced genomic instability in vitro; these effects were exacerbated upon PARPi exposure. In patient tumor samples, we also found that PBRM1-mutant cancers possessed a higher mutational load. Finally, we found that ATRi selectively activated the cGAS/STING cytosolic DNA sensing pathway in PBRM1-deficient cells, resulting in increased expression of type I interferon genes. Conclusion: PBRM1-defective cancer cells present increased replication fork stress, R-loop formation, genome instability and are selectively sensitive to PARPi and ATRi through a synthetic lethal mechanism that is cell type-independent. Our data provide the pre-clinical rationale for assessing PARPi as a monotherapy or in combination with ATRi or immune-modulating agents in molecularly-selected patients with PBRM1-defective cancers. Citation Format: Roman Merial Chabanon, Daphné Morel, Léo Colmet-Daage, Thomas Eychenne, Nicolas Dorvault, Ilirjana Bajrami, Marlène Garrido, Suzanna Hopkins, Cornelia Meisenberg, Andrew Lamb, Theo Roumeliotis, Samuel Jouny, Clémence Astier, Asha Konde, Geneviève Almouzni, Jyoti Choudhary, Jean-Charles Soria, Jessica Downs, Christopher J. Lord, Sophie Postel-Vinay. Targeting chromatin remodeling-associated genetic vulnerabilities in cancer: PBRM1 defects are synthetic lethal with PARP and ATR inhibitors [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1058.

Abstract B10: Evaluating YAP and TAZ as therapeutic targets for treating malignant mesothelioma with Hippo pathway disruptions

Abstract Malignant mesotheliomas (MMs) are tumors originating from the mesothelial lining of internal organs. Hippo pathway disruptions are common in MMs; up to 50% of cases exhibit NF2 deficiency and other pathway genes such as LATS2 are also mutated (1). This suggests hyperactivity of YAP and TAZ as a likely oncogenic driver in MMs and a potential axis to exploit therapeutically (2). Before committing to such an approach, a dependence of MMs on YAP/TAZ for survival needs to be tested. We investigated this using different MM cell line models cultured in vitro. The impact of Hippo pathway mutations on YAP and TAZ activity in mesothelial cells was assessed using an NF2 mutant isogenic system in an immortalized mesothelial cell line (Met5A.) YAP activation levels were probed through immunoblotting for YAP S127 phosphorylation, immunofluorescence, and RNA sequencing (RNAseq). YAP and TAZ were inhibited through RNAi-mediated depletion or treatment with simvastatin or verteporfin in a panel of 7 MM cell lines, including 3 with mutations in known Hippo pathway genes and 4 lacking such mutations, and cell viability was measured using an Alamar Blue assay. We found that Hippo pathway-mediated phosphorylation of YAP at S127, and YAP cytoplasmic abundance, was reduced in NF2 mutant Met5A cells compared to isogenic nonmutant cells. Consistent with this, RNAseq analysis showed elevated YAP activity in NF2 mutant Met5A cells. However, RNAi-mediated depletion of YAP and TAZ in a panel of mesothelioma cell lines did not reveal a clear correlation between dependence on YAP/TAZ for survival and aberrant Hippo signaling. There was also a comparable level of sensitivity of NF2 wild-type and NF2 mutant Met5A cells to YAP/TAZ knockdown. Additionally, treatment with verteporfin and simvastatin did not show increased toxicity in cell lines with Hippo pathway mutations relative to nonmutant cells. Our findings confirm that Hippo pathway mutations drive YAP/TAZ hyperactivity in mesothelial cells. However, they show that pathway mutation status does not always correlate with a dependence on YAP and TAZ for mesothelioma cell survival, at least when cultured in vitro. Clearly, a more in-depth analysis of the Hippo pathway in MMs is required to determine the potential of this pathway as a therapeutic target for this disease.

Sequential role of RAD51 paralog complexes in replication fork remodeling and restart

Homologous recombination (HR) factors were recently implicated in DNA replication fork remodeling and protection. While maintaining genome stability, HR-mediated fork remodeling promotes cancer chemoresistance, by as-yet elusive mechanisms. Five HR cofactors – the RAD51 paralogs RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3 – recently emerged as crucial tumor suppressors. Albeit extensively characterized in DNA repair, their role in replication has not been addressed systematically. Here, we identify all RAD51 paralogs while screening for modulators of RAD51 recombinase upon replication stress. Single-molecule analysis of fork progression and architecture in isogenic cellular systems shows that the BCDX2 subcomplex restrains fork progression upon stress, promoting fork reversal. Accordingly, BCDX2 primes unscheduled degradation of reversed forks in BRCA2-defective cells, boosting genomic instability. Conversely, the CX3 subcomplex is dispensable for fork reversal, but mediates efficient restart of reversed forks. We propose that RAD51 paralogs sequentially orchestrate clinically relevant transactions at replication forks, cooperatively promoting fork remodeling and restart.

Influence of DNA damage repair defects on tumor immunogenicity

Abstract In recent years, immune checkpoint inhibitors (ICI) have dramatically altered the treatment landscape for patients with advanced stage and metastatic cancers. Durable benefit, however, is limited to subsets of histologies and patients, highlighting the importance of identifying predictive biomarkers to select for patients who may benefit. DNA damage repair and response (DDR) deficiency including mismatch repair deficiency and resulting increased tumor neoantigens have been associated with improved response. We hypothesize that a similar concept may extend to other defects in DDR pathways, the most common being homologous recombination (HR). In ongoing work, we have implicated mutations in the DDR pathways including the HR pathway with improved survival after ICI in a large retrospective cohort. However, divergent responses were noted depending on the individual HR gene mutated. Syngeneic murine models were used to study individual DDR genes using CRISPR-Cas9 mutagenesis and their role in altering the tumor microenvironment and response to immune checkpoint blockade in isogenic model systems. We report differential effects of mutations in genes of the homologous recombination (HR) pathway on response after ICI administration in mouse and human tumors, and show that truncating mutations in BRCA2 are associated with superior response to ICI compared to BRCA1-deficient tumors. Single cell RNA-sequencing of CD45+ cells sorted from tumors reveals distinct immune landscapes. These data have significant implications for elucidating the genetic and microenvironmental determinants of response to immunotherapy in the setting of DDR deficiency and ultimately expand the therapeutic window of immunotherapy.

Molecular underpinnings of RB status as a biomarker of poor outcome in advanced prostate cancer.

189 Background: There is emergent and compelling evidence to support RB status as a biomarker in advanced prostate cancer. RB loss is strongly associated with poor progression-free, disease-specific, and overall survival in prostate cancer (PCa). Preclinical studies in PCa have revealed RB positive tumors are more responsive to CDK4/6 inhibitors. An ongoing randomized Phase IB/II study of enzalutamide with and without ribociclib in patients with metastatic castration-resistant, chemotherapy naïve PCa has become a pioneer trial to include a positive RB status as inclusion criteria in a PCa study (NCT02555189). Beyond CDK4/6 inhibitors, therapeutic agents that target tumor metabolism have been introduced in the clinic. Current data suggests that RB status may be crucial to understand and predict therapeutic response to these agents within tumors. Methods: The biological significance of RB loss was studied utilizing isogenic model systems and human tumor xenografts of castration resistant prostate cancer (CRPC) with and without RB deletion. The mechanism that drives aggressive tumor phenotypes was identified through comprehensive transcriptomic, cistromic, and metabolomic analysis. Novel functions of RB were identified and the response to clinically-relevant therapeutics was examined. Results: Exclusively in CRPC, RB loss results in significant rewiring of cancer cell metabolism. Functional investigation revealed a causative link between RB loss and antioxidant production sufficient to alter responsiveness to genomic insult and selected chemotherapeutics. Observed changes in response to therapeutic intervention were attributed to RB-dependent modulation of intracellular reactive oxygen species. Conclusions: RB loss is strongly associated with poor outcome in advanced PCa. Molecular investigation identified RB-dependent rewiring of cancer cell metabolism as a significant consequence of RB loss, sufficient to alter response in model systems to therapeutic strategies of clinical relevance. These studies significantly advance understanding of the means by which RB loss enhances lethal tumor phenotypes, and are of relevance for development of RB status as a clinically actionable biomarker.

Disease Modeling of Severe Congenital Neutropenia Using CRISPR/Cas9 Gene Correction or Knockout of ELANE in Patients Derived Induced Pluripotent Stem Cells

Severe congenital neutropenia (CN) is a monogenic bone marrow failure syndrome with the frequency of 1:200,000 and is characterized by an absolute neutrophil count below 500 cells per microliter. Patients with CN suffer from severe life-threatening bacterial infections starting early after birth due to the absent or very low numbers of neutrophils in peripheral blood. While CN is a heterogeneous disease caused by many different gene mutations, autosomal-dominant ELANE mutations are the most common cause of CN. Although the majority of CN patients respond to daily treatment with granulocyte colony-stimulating factor (G-CSF), approximately 15 % do not respond at doses up to 20 μg/kg/day and approximately 20 % of G-CSF treated patients develop myelodysplasia (MDS) or acute myeloid leukemia (AML). In the present study, we first established an efficient gene-editing platform for induced pluripotent stem cells (iPSC) of CN patients using CRISPR/Cas9 technology. The platform uses ribonucleoprotein form of CRISPR/Cas9 making the editing approach safer as it is virus- or DNA free. Also, any further selection step or introducing extra modifications in the genome of edited cells such as silent mutation are not required. We generated and characterized iPSCs from ELANE-CN patients harboring p.A57V, p.C151Y, and p.G214R mutations, that are more severe hot-spot mutations associated with G-CSF non-response or MDS/AML. We corrected each mutation followed by EB-based hematopoietic differentiation, to evaluate and compare granulocytic differentiation of CN-patient specific iPSCs, with or without ELANE mutation, in an isogenic model. To study granulocytic differentiation, we performed live cell counts, flow cytometry analysis of myeloid-specific surface marker expression, CFU assay, cell morphology of cytospin preparations and neutrophil functional tests. Our isogenic model showed that correction of ELANE mutations led to fully normalized granulocytic differentiation. We have recently shown that CRISPR/Cas9 mediated ELANE knockout (KO) enables neutrophilic maturation of primary HSPCs and iPSCs of CN patients. We observed that granulocytic differentiation of ELANE KO iPSCs and primary HSPCs were comparable to healthy individuals. Phagocytic functions, ROS production, and chemotaxis of the ELANE KO neutrophils were also normal. To model CN in silico and to reveal the key driving pathomechanisms, we designed an isogenic patient-specific disease modeling system by comparing RNA-sequencing results of CN-ELANE corrected- or CN-ELANE KO hematopoietic stem and progenitor cells (HSPCs) to the original CN-ELANE patient cells. HSPCs were derived from iPSC lines. Our analysis showed a degree of similarity in enriched pathways upon ELANE correction or ELANE KO in a patient-specific manner. Thus, upon correction of p.C151Y mutation, TNF, IL4 and IL13 signaling pathways as well as MAPK signaling, PD-1 signaling and IL10 signaling were down-regulated. Interestingly, the same pathways were down-regulated upon ELANE KO in HSPCs of the same CN patient. Correction of p.A57V mutation led to down-regulation of IL12 expression which activates STAT family. Upon ELANE KO in the cells from the same patient, IL12, IL18, and IL1-beta expression were down-regulated. We also identified common pathways enriched in most of the isogenic samples upon ELANE correction or ELANE KO like down-regulation of MAPK or IFN α/β signaling as well as down-regulation of the Rap-1 signaling pathway leading to the Erk pathway activation. Analysis of putative transcription factor binding sites (TFBSs) that are enriched in the differentially expressed gene list upon ELANE mutation correction or ELANE KO showed that transcription factors GKLF (KLF4), MAZ, Kaiso (ZBTB33) and CHURCHILL are highly enriched in UP-regulated genes, for both, correction and KO samples. Taken together, we established a safe and efficient CRISPR/Cas9-RNP based ELANE gene-correction/knockout platform of iPSCs of ELANE-CN patients that may be used to establish an isogenic disease modeling system or provide novel stem cell-based therapy for CN patients with a high risk of leukemia development as well as for G-CSF-non-responsive patients. This platform could be also applied for other monogenic bone marrow failure syndromes. Disclosures No relevant conflicts of interest to declare.

Mutant TET2 Allele Dosage Affects Response to 5-Azacitidine in Acute Myeloid Leukemia

Personalised medicine is predicted to significantly improve outcomes for cancer patients, but implementation requires comprehensive genetic characterisation of malignant cells to identify therapeutically exploitable vulnerabilities. Using an isogenic cell model system with CRISPR-inactivated TET2 in HEL acute myeloid leukemia (AML) cells and an orthotopic mouse xenograft model we demonstrate that mutant TET2 allele dosage significantly affects sensitivity to 5-azacitidine hypomethylating therapy in AML, with biallelic mutation conferring hypersensitivity relative to monoallelic mutation. In the presence of 5-azacitidine, cell clones with biallelic TET2 mutation had significantly lower cloning efficiency (P = 3 x 10-3) and proliferation in liquid culture (P < 1 x 10-4) compared to isogenic clones with monoallelic TET2 mutation. Mixed populations of monoallelic and biallelicTET2 mutated HEL AML cells were transplanted via intrafemoral injection into Rag2−/−Il2rg−/−129×Balb/c mice, and treatment with 5-azacitidine resulted in significant negative in vivo selection against TET2 null cells relative to cells with monoallelic TET2 mutation (P = 4 x 10-4). Methylation analysis revealed the acquisition of an overall hypermethylation phenotype in TET2 null cells and RNA sequencing identified significant down-regulation of ABCB1 transcript, resulting in concomitant pronounced down-regulation of the MDR1 drug efflux transporter at the protein level. RNA sequencing pathway analysis also identified a global effect on ribosome pathway (KEGG pathway ko03010) transcript levels (Padjusted = 0.002), evidenced by down-regulation of numerous RNA polymerase II components in cells with bi-allelic TET2 mutation compared to cells with monoallelic TET2 mutation. Consistent with our isogenic model data, we characterise biallelic somatic TET2 mutation in a patient with AML that was chemoresistant to anthracycline/cytarabine-based chemotherapy but acutely sensitive to 5-azacitidine, resulting in durable cytomorphological remission. Integration of next generation sequencing, interphase FISH and SNP array analysis of bone marrow at AML presentation, relapse and during remission was used to infer tumour phylogeny which indicated that disease pathogenesis was initiated by a TET2 nonsense mutation (c.2815C>T, Q939*) with subsequent deletion of the second TET2 allele and a NPM1 mutation (c.863_864ins, TCTG) that arose after the acquisition of bi-allelic TET2 mutation. Furthermore, our data demonstrate that 5-azacitidine treatment almost completely eliminated the TET2/NPM1-mutated clone. 5-azacitidine also induced a modest reduction in ancestral pre-leukemic cells carrying bi-allelic TET2 mutation but negative for the NPM1 mutation, although the majority retained viability and re-acquired the ability to differentiate and recapitulate normal haematopoiesis rendering a cytomorphological remission. These observations suggest that bi-allelic TET2 mutation confers sensitivity to the cytotoxic effects of 5-azacitidine, but that the major effect of 5-azacitidine is the induction of phenotypic re-programming. The frequency of TET2 mutation in primary AML is estimated at 10-20%, with the majority of these being monoallelic. We determined the frequency of TET2 alterations in AML patients presenting with a chromosome 4 abnormality discernible cytogenetically. TET2 copy number and mutational status were determined using high density SNP arrays and gene sequencing, respectively. In a panel of 30 AML cases with a chromosome 4 abnormality, four patients were heterozygous for TET2 mutation (all deletions resulting in reduced copy number) and three patients were homozygous for TET2 mutation (deletion plus base substitution in two cases and homozygous base substitution resulting from uniparental disomy in one case). Furthermore, all seven cases with TET2 mutation were characterised by cytogenetics that included loss or gain of material on chromosome 4. In contrast, only 1 case with a TET2 mutation had a translocation affecting chromosome 4. In summary, our data argue in favour of using 5-azacitidine in patients with biallelic TET2-mutated AML and demonstrate the importance of considering mutant allele dosage in the implementation of personalised medicine for cancer patients. Disclosures Stoelzel: JAZZ Pharmaceuticals: Consultancy; Neovii: Other: Travel funding; Shire: Consultancy, Other: Travel funding. Jackson:Celgene, Amgen, Roche, Janssen, Sanofi: Honoraria. Meggendorfer:MLL Munich Leukemia Laboratory: Employment. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

LKB1 Deficiency Renders NSCLC Cells Sensitive to ERK Inhibitors.

Serine/threonine kinase 11 (LKB1/STK11) is one of the most mutated genes in NSCLC accounting for approximately one-third of cases and its activity is impaired in approximately half of KRAS-mutated NSCLC. At present, these patients cannot benefit from any specific therapy. Through CRISPR/Cas9 technology, we systematically deleted LKB1 in both wild-type (WT) and KRAS-mutated human NSCLC cells. By using these isogenic systems together with genetically engineered mouse models we investigated the cell response to ERK inhibitors both invitro and invivo. In all the systems used here, the loss of LKB1 creates vulnerability and renders these cells particularly sensitive to ERK inhibitors both invitro and invivo. The same cells expressing a WT LKB1 poorly respond to these drugs. At the molecular level, in the absence of LKB1, ERK inhibitors induced a marked inhibition of p90 ribosomal S6 kinase activation, which in turn abolished S6 protein activation, promoting the cytotoxic effect. This work shows that ERK inhibitors are effective in LKB1 and LKB1/KRAS-mutated tumors, thus offering a therapeutic strategy for this prognostically unfavorable subgroup of patients. Because ERK inhibitors are already in clinical development, our findings could be easily translatable to the clinic. Importantly, the lack of effect in cells expressing WT LKB1, predicts that treatment of LKB1-mutated tumors with ERK inhibitors should have a favorable toxicity profile.

Abstract 1509: Identification of immunotherapy resistance mechanisms within tumor microenvironment in a mouse model of oral squamous cell carcinoma

Abstract The response rate of 15-20% with anti-PD1 in head and neck squamous cell carcinoma (HNSCC) highlights the urgent need for strategies to overcome resistance. Our lab has previously developed a carcinogen-induced immunocompetent murine oral carcinoma (MOC) model to study HNSCC immunobiology. Specifically, MOC1 is an immunogenic cell line that, despite sensitivity to anti-PD-1, exhibits occasional development of escape tumors (MOC1esc). MOC1esc escape tumors display a resistance phenotype similar to those observed in HNSCC patients undergoing anti-PD1 therapy. When independent escape tumors are harvested and re-transplanted into naïve mice, they grow progressively and are resistant to anti-PD1 therapy. Intriguingly, whileMOC1esc is resistant to anti-PD1, it is completely rejected in tumor bearing mice treated with anti-CTLA4. Therefore, the anti-PD1 responsive MOC1 and resistant MOC1esc mouse model is an isogenic system that provides an excellent opportunity to study mechanism(s) in adaptive resistance to anti-PD1 therapy of HNSCC. To gain a comprehensive insight into the tumor microenvironment (TME) as a contributor to adaptive resistance, we analyzed tumor infiltrating lymphocytes (TIL) in naïve MOC1 and MOC1esc tumors using mass cytometry time-of-flight (CyTOF) with a 38-cell marker panel.MOC1esc tumors were highly infiltrated by CD103+ effector/memory regulatory T cells (Tregs) and M2-like tumor associated macrophages (TAMs), while MOC1 tumors have more M1-like TAMs and neutrophils. Furthermore, we observed that both anti-PD1 and anti-CTLA4 treatment dramatically expanded CD8+ T cells and decreased neutrophils in MOC1esc tumors. In responding MOC1esc tumors, anti-CTLA4 treatment resulted in depleted Tregs, decreased M2-like TAMs and neutrophils, as well as a striking increase in M1-like TAMs compared with isotype control treated tumors. In contrast, anti-PD1 treated resistant MOC1esc tumors showed decreased M1-like TAMs, while M2-like TAMs were increased compared with controls. Tregs were not affected by anti-PD1 treatment. Therefore, the comparison between the TME of anti-PD1 treated resistant tumors and anti-CTLA4 treated responding tumors suggeststhat Tregs, neutrophils, and TAMs may contribute to the sensitivity (or resistance) to checkpoint blockade therapy. Ongoing studies will test the functional contribution of these distinct TME components in antitumor immunity including cytokine production, proliferative capacity, and their roles in immunotherapy resistance. In summary, this study identified immune modulators within TME involved in adaptive immunotherapy resistance of HNSCC. Findings from this study have advanced our understanding of HNSCC immunotherapy resistance and will accelerate the discovery of new therapeutic targets and biomarkers for adaptive resistance. Citation Format: Liye Zhou, Yasutaka Nakahori, Fei Guo, Joshua Keegan, Rachel Riley, James A. Lederer, Ravindra Uppaluri. Identification of immunotherapy resistance mechanisms within tumor microenvironment in a mouse model of oral squamous cell carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1509.

Somatic loss of PIK3R1 may sensitize breast cancer to inhibitors of the MAPK pathway.

The PI3K pathway, which includes the PI3K catalytic subunits p110α (PIK3CA) and the PI3K regulatory subunit p85α (PIK3R1), is the most frequently altered pathway in cancer. We encountered a breast cancer patient whose tumor contained a somatic alteration in PIK3R1. Some commercial sequencing platforms suggest that somatic mutations in PIK3R1 may sensitize cancers to drugs that inhibit the mammalian target of rapamycin (mTOR). However, a review of the preclinical and clinical literature did not find evidence substantiating that hypothesis. The purpose of this study was to knock out PIK3R1 in order to determine the optimal therapeutic approach for breast cancers lacking p85α. We created an isogenic cellular system by knocking out both alleles of the PIK3R1 gene in the non-tumorigenic human breast cell line MCF-10A. Knockout cells were compared with wild-type cells by measuring growth, cellular signaling, and response to drugs. We observed hyperphosphorylation of MEK in these knockouts, which sensitized PIK3R1-null cells to a MEK inhibitor, trametinib. However, they were not sensitized to the mTOR inhibitor, everolimus. Our findings suggest that breast cancers with loss of p85α may not respond to mTOR inhibition, but may be sensitive to MEK inhibition.

G-quadruplex DNA drives genomic instability and represents a targetable molecular abnormality in ATRX-deficient malignant glioma

Mutational inactivation of ATRX (α-thalassemia mental retardation X-linked) represents a defining molecular alteration in large subsets of malignant glioma. Yet the pathogenic consequences of ATRX deficiency remain unclear, as do tractable mechanisms for its therapeutic targeting. Here we report that ATRX loss in isogenic glioma model systems induces replication stress and DNA damage by way of G-quadruplex (G4) DNA secondary structure. Moreover, these effects are associated with the acquisition of disease-relevant copy number alterations over time. We then demonstrate, both in vitro and in vivo, that ATRX deficiency selectively enhances DNA damage and cell death following chemical G4 stabilization. Finally, we show that G4 stabilization synergizes with other DNA-damaging therapies, including ionizing radiation, in the ATRX-deficient context. Our findings reveal novel pathogenic mechanisms driven by ATRX deficiency in glioma, while also pointing to tangible strategies for drug development.

Probing the contribution of individual polypeptide GalNAc-transferase isoforms to the O-glycoproteome by inducible expression in isogenic cell lines

The GalNAc-type O-glycoproteome is orchestrated by a large family of polypeptide GalNAc-transferase isoenzymes (GalNAc-Ts) with partially overlapping contributions to the O-glycoproteome besides distinct nonredundant functions. Increasing evidence indicates that individual GalNAc-Ts co-regulate and fine-tune specific protein functions in health and disease, and deficiencies in individual GALNT genes underlie congenital diseases with distinct phenotypes. Studies of GalNAc-T specificities have mainly been performed with in vitro enzyme assays using short peptide substrates, but recently quantitative differential O-glycoproteomics of isogenic cells with and without GALNT genes has enabled a more unbiased exploration of the nonredundant contributions of individual GalNAc-Ts. Both approaches suggest that fairly small subsets of O-glycosites are nonredundantly regulated by specific GalNAc-Ts, but how these isoenzymes orchestrate regulation among competing redundant substrates is unclear. To explore this, here we developed isogenic cell model systems with Tet-On inducible expression of two GalNAc-T genes, GALNT2 and GALNT11, in a knockout background in HEK293 cells. Using quantitative O-glycoproteomics with tandem-mass-tag (TMT) labeling, we found that isoform-specific glycosites are glycosylated in a dose-dependent manner and that induction of GalNAc-T2 or -T11 produces discrete glycosylation effects without affecting the major part of the O-glycoproteome. These results support previous findings indicating that individual GalNAc-T isoenzymes can serve in fine-tuned regulation of distinct protein functions.

Abstract B44: Identification of epigenomic changes induced by H3 K27M mutation in glioblastoma using patient-derived and CRISPR/Cas9 edited cell lines

Abstract Background: Glioblastoma is a grade IV malignant brain tumor with poor prognosis and rapid disease progression. Recurrent somatic mutations in histone H3 genes have been identified in the majority of pediatric glioblastoma cases. The K27M mutation in H3.1 and H3.3 histones globally inhibits lysine methylation at the K27 position, whereas H3.3 G34R/V possibly affects histone lysine methylation at the K36 position. H3 K27M mutation has been shown to dramatically decrease the total levels of H3K27me3 and H3K27me2 marks and increase H3K27ac levels. However, the effect of H3 K27M on global epigenomic changes is not fully characterized. Furthermore, standard profiling of histone marks by chromatin immunoprecipitation combined with next-generation sequencing (ChIP-seq) is not quantitative, a significant caveat when global levels of histone marks change so drastically. Methods: We assembled a collection of H3 K27M mutant and wild-type cell lines derived from the glioblastoma patients. The epigenomes of these cell lines were comprehensively characterized by profiling for six histone marks (H3K4me1, H3K4me3, H3K27ac, H3K27me3, H3K36me2, H3K36me3) using ChIP-seq. In addition, we derived isogenic cell lines overexpressing H3.3 K27M, as well as cell lines with knockin or knockout of the K27M mutation using the CRISPR/Cas9 genome editing system. These cell lines were profiled for H3K27me3 mark by ChIP-seq. We used a modified ChIP-seq protocol, chromatin immunoprecipitation with exogenous reference genome (ChIP-Rx), which allows quantitation of histone mark abundance by normalization to proportions of added Drosophila chromatin in the ChIP reaction. RNA sequencing was performed on both primary and isogenic cell lines. Results: The most striking difference we observed between H3 K27M and wild-type cells was in H3K27me3 mark. Using ChIP-Rx, we observe significantly lower levels of H3K27me3 mark in H3 K27M cell lines, both in primary cells and isogenic contexts. Despite very low total levels of H3K27me3 mark, K27M mutant cells display enrichment of the mark in certain regions, at comparable levels to wild-type cell lines. Using our isogenic cell line models, we show that K27M mutation is indeed responsible for those genome-wide changes in the epigenome. Correlating H3K27me3 distribution with transcriptome data, we show that expression changes mainly among the genes that are lowly expressed in these cells. Pathway analysis of differentially expressed genes shows enrichment for neural development and differentiation that suggests links to disease pathogenesis. Conclusions: Despite the fact that primary cell lines have different origins and a variety of additional driver mutations, their epigenomes appears to be remarkably similar, due to being shaped predominantly by the effects of histone mutations, as demonstrated in isogenic cell line systems. Global changes in H3K27me3 levels and distribution in H3 K27M mutant cells lead to specific changes in gene expression. The changes induced by K27M mutations also appear to be specific to the cell type and/or developmental context of origin. This may help better understand the effect they have in reshaping the epigenome to promote oncogenesis. Citation Format: Ashot S. Harutyunyan, Brian Krug, Simon Papillon-Cavanagh, Haifen Chen, Shriya Deshmukh, Warren A. Cheung, Rui Li, Jad Belle, Denise Bechet, Nicolas De Jay, Michele Zeinieh, Tenzin Gayden, Caterina Russo, Leonie Mikael, Damien Faury, Claudia Kleinman, Tomi Pastinen, Jacek Majewski, Nada Jabado. Identification of epigenomic changes induced by H3 K27M mutation in glioblastoma using patient-derived and CRISPR/Cas9 edited cell lines [abstract]. In: Proceedings of the AACR Special Conference: Pediatric Cancer Research: From Basic Science to the Clinic; 2017 Dec 3-6; Atlanta, Georgia. Philadelphia (PA): AACR; Cancer Res 2018;78(19 Suppl):Abstract nr B44.

PO-103 Regulation of prostate cancer cell growth and signalling by exercise

IntroductionMalignant mesothelioma (MM) is an aggressive cancer of the mesothelium. MM has a poor prognosis with a median post-diagnosis survival of 12 months. Currently, MM is treated with platinum-based chemotherapy, resulting in side-effects to the patient. Hence, efforts have been focused on genetically analysing MM tumours to improve treatment specificity. Genetically, MMs are characterised by the loss of key tumour suppressor genes such as NF2. NF2 functions as a member of the Hippo signalling pathway. The Hippo pathway consists of a tumour-suppressive kinase cascade which inactivates the proto-oncogenic transcription factors YAP and TAZ. NF2 up-regulates the Hippo pathway kinases to reduce the activity of YAP/TAZ. Hence, NF2 mutant MMs may exhibit hyperactivity of YAP/TAZ, driving oncogenic processes. So, YAP/TAZ may be ideal therapeutic targets for treating NF2 mutant MMs.Material and methodsA panel of 7 MM cell lines and one wildtype mesothelial cell line were used to assess the potential for selectively targeting YAP/TAZ to treat MMs with mutations in NF2 or in the Hippo pathway kinases LATS1/2. YAP and TAZ activity was abrogated in the cell lines using siRNA-mediated knockdown or using chemical inhibition with the drug compounds Verteporfin or Simvastatin. The viability of cell lines 96 hours post YAP/TAZ inhibition was measured using an Alamar blue viability assay. Statistical comparisons of cellular responses to YAP/TAZ inhibition were conducted using an unpaired student’s t-test.Results and discussionsThe effect of YAP/TAZ inhibition on the viability of a panel of 7 MM cell lines and a wildtype mesothelial cell line (Met5A) was assessed. Hippo pathway mutant cell lines showed the highest sensitivity to YAP/TAZ knockdown, however, two non-mutant MM cells also showed considerable sensitivity, indicating that there wasn’t a statistically significant dependence of mutant cell lines on YAP/TAZ for viability. Mutant cell lines did not show selective sensitivity to Simvastatin or Verteporfin, suggesting that these compounds may not be ideal treatment options. However, these observations were likely confounded by the heterogeneity of the panel of cell lines. As such, an isogenic system of NF2 mutant Met5A cells was developed to validate the findings.ConclusionBased on our findings, YAP and TAZ may not be ideal targets for treating Hippo pathway mutant MMs. However, the use of an NF2 mutant isogenic MM cell line system will provide a definitive outlook on their feasibility as therapeutic candidates.

PO-100 Targeting YAP and TAZ to treat hippo pathway mutant malignant mesotheliomas

IntroductionMalignant mesothelioma (MM) is an aggressive cancer of the mesothelium. MM has a poor prognosis with a median post-diagnosis survival of 12 months. Currently, MM is treated with platinum-based chemotherapy, resulting in side-effects to the patient. Hence, efforts have been focused on genetically analysing MM tumours to improve treatment specificity. Genetically, MMs are characterised by the loss of key tumour suppressor genes such as NF2. NF2 functions as a member of the Hippo signalling pathway. The Hippo pathway consists of a tumour-suppressive kinase cascade which inactivates the proto-oncogenic transcription factors YAP and TAZ. NF2 up-regulates the Hippo pathway kinases to reduce the activity of YAP/TAZ. Hence, NF2 mutant MMs may exhibit hyperactivity of YAP/TAZ, driving oncogenic processes. So, YAP/TAZ may be ideal therapeutic targets for treating NF2 mutant MMs.Material and methodsA panel of 7 MM cell lines and one wildtype mesothelial cell line were used to assess the potential for selectively targeting YAP/TAZ to treat MMs with mutations in NF2 or in the Hippo pathway kinases LATS1/2. YAP and TAZ activity was abrogated in the cell lines using siRNA-mediated knockdown or using chemical inhibition with the drug compounds Verteporfin or Simvastatin. The viability of cell lines 96 hours post YAP/TAZ inhibition was measured using an Alamar blue viability assay. Statistical comparisons of cellular responses to YAP/TAZ inhibition were conducted using an unpaired student’s t-test.Results and discussionsThe effect of YAP/TAZ inhibition on the viability of a panel of 7 MM cell lines and a wildtype mesothelial cell line (Met5A) was assessed. Hippo pathway mutant cell lines showed the highest sensitivity to YAP/TAZ knockdown, however, two non-mutant MM cells also showed considerable sensitivity, indicating that there wasn’t a statistically significant dependence of mutant cell lines on YAP/TAZ for viability. Mutant cell lines did not show selective sensitivity to Simvastatin or Verteporfin, suggesting that these compounds may not be ideal treatment options. However, these observations were likely confounded by the heterogeneity of the panel of cell lines. As such, an isogenic system of NF2 mutant Met5A cells was developed to validate the findings.ConclusionBased on our findings, YAP and TAZ may not be ideal targets for treating Hippo pathway mutant MMs. However, the use of an NF2 mutant isogenic MM cell line system will provide a definitive outlook on their feasibility as therapeutic candidates.

Abstract 4556: Phagocytosis response of macrophage to CD47 expression of tumor cells in anti-CD47 therapy

Abstract Macrophage as an important surveillance system is escaped by human cancer cells expressing CD47. To reinstate macrophage functionality in clearing tumor cells targeting CD47 and thereby blocking phagocytosis inhibitory CD47-SIRPalpha interaction has been an important strategy in immune-oncology therapy. We profiled a large panel of tumor cell lines, Crown's proprietary PDX models and clinical lung cancer tumors for CD47 expression. Although it has been reported as a ubiquitously expressed immune-escape marker in cancer cells, the expression levels of CD47 vary among different tumor cell lines and models when the cells were examined by flow cytometry. It is of interest that a portion of tumor cells in fact do not present CD47. To tackle the cause for the loss of CD47 expression, whole-exome sequencing was performed to examine the intactness of the gene in the CD47 loss models. The results will be presented. We next established a robust and reproducible in vitro assay platform to support anti-CD47 drug research. We applied a macrophage differentiation system using isolated CD14+ monocytes as starting material, and were able to achieve >80% of M1 and M2 populations, respectively, through the 6-day differentiation. Differentiated macrophage was co-cultured with target tumor cells labeled for flow cytometry analysis. In the presence of blocking anti-CD47 antibody BRICK126, steady and specific phagocytosis effect was observed when target tumor cells express high level of CD47 markers. BRIC126 antibody blocks the CD47-SIRPalpha interaction. In contrast, aCD274 antibody did not yield more than background phagocytosis effect in the same assay. To prove the CD47-specific effect, we constructed isogenic cell models from a couple of parental cell models in that CD47 gene is knocked out in one sub-cell line and overexpressed in another sub-cell line. The isogenic systems were then subject to the CD47 biology assay and more profound changes in phagocytosis responses were observed. Furthermore, the in vitro findings were validated in in vivo setting. Citation Format: Huajun Yang, Zhongliang Li, Phillip Wang, Beibei Tang, Qinyun Ma, Frank Xing, Qian Shi. Phagocytosis response of macrophage to CD47 expression of tumor cells in anti-CD47 therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4556.