Background: poor prognosis primary breast cancers are typically treated with cytotoxic chemotherapy. However, recurrences remain relatively common even after this aggressive therapy. Comparison of matched tumours pre- and post-chemotherapy can allow identification of molecular characteristics of therapy resistance and thereby potentially aid discovery of novel predictive markers or targets for chemosensitisation. Through this comparison, we aimed to identify microRNAs associated with chemoresistance, define microRNA target genes, and assess targets as predictors of chemotherapy response. Methods: cancer cells were laser microdissected from matched breast cancer tissues pre- and post-chemotherapy from estrogen receptor positive/HER2 negative breast cancers showing partial responses to epirubicin/cyclophosphamide chemotherapy (n = 5). MicroRNA expression was profiled using qPCR arrays. MicroRNA/mRNA expression was manipulated in estrogen receptor positive/HER2 negative breast cancer cell lines (MCF7 and MDA-MB-175 cells) with mimics, inhibitors or siRNAs, and chemoresponse was assessed using MTT and colony forming survival assays. MicroRNA targets were identified by RNA-sequencing of microRNA mimic pull-downs, and comparison of these with mRNAs containing predicted microRNA binding sites. Survival correlations were tested using the METABRIC expression dataset (n = 1979). Results: miR-195 and miR-26b were consistently up-regulated after therapy, and changes in their expression in cell lines caused significant differences in chemotherapy sensitivity, in accordance with up-regulation driving resistance. SEMA6D was defined and confirmed as a target of the microRNAs. Reduced SEMA6D expression was significantly associated with chemoresistance, in accordance with SEMA6D being a down-stream effector of the microRNAs. Finally, low SEMA6D expression in breast cancers was significantly associated with poor survival after chemotherapy, but not after other therapies. Conclusions: microRNAs and their targets influence chemoresponse, allowing the identification of SEMA6D as a predictive marker for chemotherapy response that could be used to direct therapy or as a target in chemosensitisation strategies.

In chronic lymphocytic leukemia (CLL), mutation and loss of p53 and ATM abrogate DNA damage signalling and predict poorer response and shorter survival. We hypothesised that poly (ADP-ribose) polymerase (PARP) activity, which is crucial for repair of DNA breaks induced by oxidative stress or chemotherapy, may be an additional predictive biomarker and a target for therapy with PARP inhibitors.We measured PARP activity in 109 patient-derived CLL samples, which varied widely (192 – 190052 pmol PAR/106 cells) compared to that seen in healthy volunteer lymphocytes (2451 – 7519 pmol PAR/106 cells). PARP activity was associated with PARP1 protein expression and endogenous PAR levels. PARP activity was not associated with p53 or ATM loss, Binet stage, IGHV mutational status or survival, but correlated with Bcl-2 and Rel A (an NF-kB subunit). Levels of 8-hydroxy-2′-deoxyguanosine in DNA (a marker of oxidative damage) were not associated with PAR levels or PARP activity. The potent PARP inhibitor, talazoparib (BMN 673), inhibited CD40L-stimulated proliferation of CLL cells at nM concentrations, independently of Binet stage or p53/ATM function.PARP activity is highly variable in CLL and correlates with stress-induced proteins. Proliferating CLL cells (including those with p53 or ATM loss) are highly sensitive to the PARP inhibitor talazoparib.

This is a repository copy of Genomic and expression analyses define MUC17 and PCNX1 as predictors of chemotherapy response in breast cancer.

ALK is the most commonly mutated oncogene in neuroblastoma with increased mutation frequency reported at relapse. Here we report the loss of an ALK mutation in two patients at relapse and a paired neuroblastoma cell line at relapse. ALK detection methods including Sanger sequencing, targeted next-generation sequencing and a new ALK Agena MassARRAY technique were used to detect common hotspot ALK variants in tumors at diagnosis and relapse from two high-risk neuroblastoma patients. Copy number analysis including single nucleotide polymorphism array and array comparative genomic hybridization confirmed adequate tumor cell content in DNA used for mutation testing. Case 1 presented with an ALK F1174L mutation at diagnosis with a variant allele frequency (VAF) ranging between 23.5% and 28.5%, but the mutation was undetectable at relapse. Case 2 presented with an ALK R1257Q mutation at diagnosis (VAF = 39%-47.4%) which decreased to <0.01% at relapse. Segmental chromosomal aberrations were maintained between diagnosis and relapse confirming sufficient tumor cell content for mutation detection. The diagnostic SKNBE1n cell line harbors an ALK F1174S mutation, which was lost in the relapsed SKNBE2c cell line. To our knowledge, these are the first reported cases of loss of ALK mutations at relapse in neuroblastoma in the absence of ALK inhibitor therapy, reflecting intra-tumoral spatial and temporal heterogeneity. As ALK inhibitors are increasingly used in the treatment of refractory/relapsed neuroblastoma, our

1788 Current treatment protocols for Chronic lymphocytic leukemia (CLL), including FCR (fludarabine, cyclophosphamide and rituximab) have improved response rates and newer approaches targeting the B-cell receptor show promise in early clinical trials. Despite these advances, CLL remains incurable, and since CLL patients are predominantly elderly, many cannot tolerate more aggressive therapies. Therefore, development of biomarkers to stratify subgroups of patients for personalised medicine is paramount to achieve the best therapeutic outcomes for these patients. Chemo-resistance in CLL frequently occurs due to loss of function of either p53 or ataxia telangiectasia mutated kinase (ATM) (del(17p) and del(11q) respectively). p53 and ATM mutation and/or deletion results in defective DNA double-strand break (DSB) response pathways, which confer greater genomic instability, poorer response and shorter overall survival (OS). ATM is a key DSB signaling kinase with roles in homologous recombination repair and cell cycle checkpoint activation, which facilitate DNA repair. We hypothesized that as well as defects in DSB-activated enzymes, genomic instability arising from unresolved single-strand DNA breaks may play a role in disease progression and resistance in CLL. Poly (ADP-ribose) polymerase (PARP) is activated by DNA damaging agents that induce single strand breaks (e.g. alkylating agents) and is pivotal in the signaling to direct repair of such DNA damage. Therefore, inhibitors of PARP sensitise cells to certain classes of chemotherapeutic agents by inhibiting DNA repair. PARP is also the first line of defence against oxidative stress, which is associated with poor prognosis in CLL. We developed the first clinically-used PARP inhibitor (PARPi) and subsequent studies identified the synthetic lethality of PARPi monotherapy in homologous recombination repair defective (HRD) cancer, which selectively kills HRD tumour cells, but is non-toxic to the HR-competent normal tissues. Importantly, PARP activity is reported to be higher in HRD cells, which may reflect increased genomic instability and oxidative stress in these cells. Currently, there are several PARPi undergoing clinical evaluation, including two studies in CLL patients. Here, we analysed the activity of PARP in CLL, to determine whether it could be used as a biomarker of response to conventional therapy and a means to stratify patients for novel therapies, including PARP inhibitor (PARPi) monotherapy. Using the clinically validated PARP activity assay, we found that PARP activity in PBMCs from healthy volunteers was in the same range as we had previously observed (Zaremba T, Biochem J 2011). However when we measured PARP activity in CLL cells (n=33 cases) it was up to 100-fold higher than the activity in PBMCs from healthy volunteers. Activity ranged from 100-50, 000 pmoL/106 cells and was higher in those patients with Binet stage C, or progressive disease. When patients were stratified according to ATM functional status (determined by measuring ATM activity) PARP activity was significantly higher in cases with confirmed ATM dysfunction (p=0.02), compared to those with functional ATM. We then designated cases as having PARP activity that was either higher than, or lower than the median (12,400 pmoL/106cells) and performed Kaplan Meier analysis. There was a trend towards high PARP activity being associated with shorter OS. These results could have significant ramifications: recent reports support the concept that ATM-defective CLL can call be selectively targeted by PARPi (Weston VJ, Blood 2010) and therefore current phase I clinical trials are investigating the use of PARPi in HRD (including ATM-defective) CLL.These trials are stratifying patients by del(11q) and ATM status (since ATM plays a key role in mediating successful HR) with the aim of selectively killing ATM-defective CLL cells with the PARPi. However, ATM status alone may not be sufficient for patient stratification. We propose that CLL cells with high PARP activity are “addicted” to PARP by virtue of its role protecting against increased DNA damage, and that PARP activity may be a key determinant of patient sensitivity to PARP inhibitors. Additionally, the underlying mechanism for high PARP activity in CLL requires further investigation, since this may reveal new therapeutic options, and strategies for patient stratification prior to clinical trial. Disclosures: No relevant conflicts of interest to declare.