Search

Acute Myeloid Leukaemia (AML) is a genetically heterogeneous disease of the blood and bone marrow primarily affecting older populations. Chemotherapy has been the standard of care for four decades, just 10% of elderly patients show a response and it will be curative in only a third of patients aged 18-60. Therefore, it's necessary to identify new therapeutic targets for improved treatment options. The search for novel mutations within the genomes of karyotypically normal AML patients identified Cohesin's STAG2 as a recurrently mutated gene (6%). Cohesin is a ring-shaped multimeric protein complex which has a multitude of roles including maintaining sister chromatid cohesion, DNA replication and repair, gene expression and chromatin architecture. STAG2 remains the most understudied Cohesin gene, how a mutation affects the ability of Cohesin to fulfil certain roles needs elucidation as well as how STAG2 functions as part of Cohesin to repair DNA breaks, promote faithful replication and control gene expression. Therefore, we sought to explore the consequences of loss-of-function STAG2 mutations on DNA repair and replication using a STAG2 knockout (ΔSTAG2) isogenic cell line model created using CRISPR/Cas9. Immunofluorescent staining of DNA damage repair proteins 53BP1 and RAD51 following 2Gy irradiation (IR) suggests ΔSTAG2 cells have a significant DNA repair defect. The average 53BP1 and RAD51 foci counts were significantly higher (p=0.001) in mutant cells which had three times as many foci at 24 hours after IR. In addition, 40% of ΔSTAG2 cells stain positive for 53BP1 and RAD51 respectively at 24 hours post-IR. This damage persists at 48 and 72 hours, whereas wild type (WT) damage returns to basal levels. Given that our data suggests a HR defect, we assessed the response of WT and ΔSTAG2 cells to PARP inhibitors (PARPi). Interestingly, ΔSTAG2 cells showed significantly greater sensitivity than WT cells to highly potent PARPi Talazoparib with an IC50 of 62nM. We next assayed single DNA fibres to evaluate the implications of a STAG2 mutation on the dynamics of individual replication forks. Interestingly, we observed a significantly reduced number of ongoing replication forks compared to WT cells (p=0.01) with a concomitant increase in the number of stalled forks (p=0.01) in the absence of exogeneous DNA damage. Following treatment with hydroxyurea (HU) this phenotype was exacerbated and ΔSTAG2 cells exhibit significantly more stalled replication forks than WTs (p=0.001). This indicates a role for STAG2 in maintaining replication fork stability and promoting fork restart following replication stress. Importantly, our data identifies a potential DNA damage repair deficiency that sensitises ΔSTAG2 cells to PARPi monotherapy as well as unearthing an essential role for STAG2 in replication fork progression and fork stability. This mutation could therefore be exploited therapeutically and represents a novel therapeutic target/approach for AML patients with mutated STAG2. Disclosures No relevant conflicts of interest to declare.

Acute Myeloid Leukemia (AML) is most commonly seen in people over the age of 65 and has a median age of 63. Globally there is an increasingly elderly population so the rate of incidence of AML is set to increase. The therapy landscape for AML has changed little over the past four decades. Cytarabine, first approved in 1969, is still the standard of care induction therapy for AML. There has been only modest improvements in survival rates during this time and there is currently no method of determining which patients will or will not respond to Cytarabine treatment. An assay, developed in 2014, used microarray data to determine which breast cancer patients had a DNA Damage Repair Deficiency (DDRD) and therefore would be more susceptible to DNA damaging agents. A negative DDRD (DDRD-) score predicts that patients do not to have a DNA Repair Deficiency whilst patients with a positive DDRD (DDRD+) score are predicted to have a DNA Repair Deficiency. This assay has been adapted to different solid cancer types such as ovarian and oesophageal cancer. This project has assessed the potential of using the DDRD assay for AML patients. The assay was applied to publically available microarray data of >600 AML patients (TCGA AML data &GSE6891), who were classed as DDRD- or DDRD+. Excluding patients not treated with Cytarabine, this left 639 patients, 405 DDRD+ and 234 DDRD-. Kaplan Meier analysis showed the DDRD+ patients survived significantly (p=0.00047) worse than the DDRD- cohort. Whole exome sequencing was available for 183 patients (131 DDRD+) and the mutations associated with each group were identified. As the DDRD+ patients had the worst outcome, we focused on group. The list of genes more commonly mutated in the DDRD+ patients (>2 instances and >50% occurring in this group) were subjected to pathway analysis. Deregulated pathways included "leukemogenisis" and "cell proliferation and regulation"; however, the most deregulated pathway was "metabolism of nucleobase containing compounds". As Cytarabine is a nucleobase-containing compound, this is potentially a contributing factor as to why these patients responded poorly to this treatment. The assay was applied to microarray data of a panel of myeloid cell lines, and DDRD-(NB4 & SKM1) and a DDRD+(HL-60) cell line were chosen as experimental models. Clonogenic assays, used to analyse the effect of Cytarabine on these cell lines, showed that the DDRD- cell lines were more sensitive with a lower colony growth rate than the DDRD+cell line. DNA damage induction and repair, following cytarabine treatment or 2gy radiation, were measured using RAD51 foci counts. Whilst foci counts were high in all cell lines 2hrs and 4hrs following radiation, the DDRD+ cell line continued to show high levels after 24hrs whereas the levels in the DDRD- cell lines returned to a basal level. RAD51 response to radiation treatment showed that a repair defect is present in DDRD+ cells as they fail to repair the damage induced by radiation. Following treatment with Cytarabine however, few foci were seen in the DDRD+ cell line 2hrs, 4hrs or 24hrs following treatment whereas the DDRD- cell lines responded in a similar fashion to radiation treatment. That RAD51 foci are not present following Cytarabine treatment indicates that Cytarabine fails to induce damage in these cells. The DDRD assay has shown to be an effective method for determining cellular response to Cytarabine in vivo. The non-response of the DDRD+ cell line to Cytarabine suggests that these cells do not elicit a DNA damage or an apoptotic response. This perhaps contributes to their poorer outcome and suggests that Cytarabine is not an effective treatment plan for patients deemed to be DDRD+. Although alternative induction treatment options are currently unavailable for DDRD+ AML patients, this DDRD assay could be used as a biomarker for Cytarabine response in the future. Disclosures No relevant conflicts of interest to declare.

Multiple large scale sequencing projects have identified a significant number of mutations in Acute Myeloid Leukaemia (AML) and high-risk Myelodysplastic syndromes (MDS) patients. Emerging from this are those affecting members of the cohesin complex, predominantly STAG2, and it has been highlighted that the poorer overall survival of this chromatin structure group is second only to patients carrying mutations affecting TP53. The Cohesin complex is an evolutionary conserved, multimeric protein complex consisting of SMC1, SMC3, RAD21 and STAG1/ STAG2. The complex plays a pivotal role during mitosis, ensuring sister chromatid cohesion however, substantial data exists highlighting roles in DNA damage repair, homologous recombination and most significantly, contribution to long-range interactions between cis-regulatory elements of the genome. Here, we have used the CRISPR/Cas9 genome editing system to create an isogenic model featuring a premature stop codon within STAG2 at a region which presents as highly mutated in patient data sets. This stop codon generates a loss of function phenotype due to the haplo-insufficient X-linked nature of the STAG2 gene. Following model validation, the impact on genome wide distribution of the cohesin complex was assessed using ChIP-Seq for STAG1, STAG2, SMC3 & the transcriptional insulator CTCF. A modest increase in overall CTCF binding from ~67,000 to ~72,000 high confidence peaks was observed, despite a ~7-fold decrease in enrichment, indicating decreased global insulator activity as a result of the inactivating STAG2 mutation. As expected, compensation between the STAG members was observed at a number of sites within transcriptionally active topologically associated domains, with an increase in STAG1 binding peaks (~17,000 to >25,000 sites) within the mutant model, despite no changes in average global enrichment. This data was integrated with RNA-seq of the isogeneic parental and mutant cells to determine the transcriptional impact of changes in cohesin complex binding on gene expression within active and repressive domains. Assessment of the active and repressive super-enhancer associated marks H3K27ac and H3K27me3 showed increased global levels of both marks without increase in the polycomb repressive complex 2 member, EZH2 total protein levels. These findings indicate that increased H3K27ac and H3K27me3 marks, in this model, are due to altered domain boundaries, associated with local activation of both H3K9ac and H3K9me3, leading to aberrant gene expression and consequently a neomorphic state within these mutant cells. With recent further stratification of de-novo AML, secondary AML and high-risk MDS emerging, this study on the aberrant and non-canonical nature of the cohesin complex provides avenues for an attractive therapeutic target and personalised treatment approach for patients within the chromatin subgroup. Disclosures No relevant conflicts of interest to declare.