Illuminating the conceptual structure of the space of moral violations with searchlight representational similarity analysis

Characterizing how representations of moral violations are organized, cognitively and neurally, is central to understanding how people conceive and judge them. Past work has identified brain regions that represent morally relevant features and distinguish moral domains, but has not yet advanced a broader account of where and on what basis neural representations of moral violations are organized. With searchlight representational similarity analysis, we investigate where category membership drives similarity in neural patterns during moral judgment of violations from two key moral domains: Harm and Purity. Representations converge across domains in a network of regions resembling the mentalizing network. However, Harm and Purity violation representations respectively converge in different regions: precuneus (PC) and left inferior frontal gyrus (LIFG). Examining substructure within moral domains, Harm violations converge in PC regardless of subdomain (physical harms, psychological harms), while Purity subdomains (pathogen-related violations, sex-related violations) converge in distinct sets of regions - mirroring a dissociation observed in principal-component analysis of behavioral data. Further, we find initial evidence for representation of morally relevant features within these two domain-encoding regions. The present analyses offer a case study for understanding how organization within the complex conceptual space of moral violations is reflected in the organization of neural patterns across the cortex.

In vitro anti-diabetic activity of flavonoids and pheophytins from Allophylus cominia Sw . on PTP1B, DPPIV, alpha-glucosidase and alpha-amylase enzymes

BACKGROUND

Ethno-botanical information from diabetic patients in Cuba led to the identification of Allophylus cominia as a possible source of new drugs for the treatment of type 2 diabetes mellitus (T2-DM).

EXPERIMENTAL

Chemical characterization of the extracts from A. cominia was carried out using chromatographic and spectroscopic methods. The extracts were tested for their activity on PTP1B, DPPIV, α-glucosidase enzymes and α-amylase.

RESULTS

The flavonoid rich fractions from A. cominia inhibited DPPIV enzyme (75.3±2.33%) at 30µg/ml and produced a concentration-dependent inhibition against DPPIV with a Ki value of 2.6µg/ml. At 30µg/ml, flavonoids and pheophytins extracts significantly inhibited PTP1B enzyme (100±2.6% and 68±1% respectively). The flavonoids, pheophytin A and pheophytin B fractions showed significant concentration-dependent inhibition against PTP1B with Ki values of 3µg/ml, 0.64µg/ml and 0.88µg/ml respectively. At 30µg/ml, the flavonoid fraction significantly inhibited α-glucosidase enzyme (86±0.3%) in a concentration-dependent pattern with a Ki value of 2µg/ml. None of the fractions showed significant effects on α-amylase. Fatty acids, tannins, pheophytins A and B, and a mixture of flavonoids were detected in the methanolic extract from A. cominia. The identified flavonoids were mearnsitrin, quercitrin, quercetin-3-alloside, and naringenin-7-glucoside.

CONCLUSION

The pharmacological effects of the extracts from A. cominia earlier observed in experimental diabetic models was confirmed in this study. Thus a new drug or formulation for the treatment of T2-DM could be developed from A. cominia.

Collaborating to Compete: Blood Profiling Atlas in Cancer (BloodPAC) Consortium

The cancer community understands the value of blood profiling measurements in assessing and monitoring cancer. We describe an effort among academic, government, biotechnology, diagnostic, and pharmaceutical companies called the Blood Profiling Atlas in Cancer (BloodPAC) Project. BloodPAC will aggregate, make freely available, and harmonize for further analyses, raw datasets, relevant associated clinical data (e.g., clinical diagnosis, treatment history, and outcomes), and sample preparation and handling protocols to accelerate the development of blood profiling assays.

Abstract P6-07-08: The complete spectrum of ESR1 mutations from 7590 breast cancer tumor samples

Background: Approximately 70% of newly diagnosed breast cancers express estrogen receptor alpha (ERα), and are treated with agents that block ER signaling. Acquired mutations in ESR1, the gene that encodes ERα, have been associated with resistance to aromatase inhibitor therapy in patients with ER positive metastatic breast cancer (ER+ mBC). The most frequently occurring ESR1 mutations are clustered between amino acids 536 to 538 within the ligand binding domain (LBD), although limited data exists characterizing the full mutation profile in a large number of breast cancer samples.

Methods: We surveyed the Foundation Medicine dataset of 7590 primary and metastatic breast cancer tumor samples for ESR1 short variants and copy number alterations. Hormone receptor status was unavailable, therefore two assumptions were made to provide an estimate of prevalence in the ER+ HER2- population: 70% of the tumor samples are from ER+ HER2- patients, and all ESR1 mutations from non-HER2 amplified metastatic sites are from ER+ HER2- patients. In a separate cohort of 48 ER+ mBC patients, circulating tumor DNA (ctDNA) was analyzed for ESR1 mutations using the BEAMing method by Sysmex and with Foundation Medicine's sequencing assay, FoundationACT (Assay for Circulating Tumor DNA).

Results: The prevalence of mutations in ER+ HER2- breast cancer was estimated to be 22% in samples from metastatic sites but less than 3% in samples from primary sites. ESR1 amplification was rare in samples from both primary and metastatic disease sites at 1.3% and 2.0% respectively. A total of 153 unique short variants of known and unknown status were identified. In addition to hotspot mutations at 537 and 538, previously undescribed rare mutations were identified throughout the entire length of the LBD, although 10 alterations at amino acids 380, 463, 536, 537, and 538 account for 86% of all ESR1 mutations in the ER+ HER2- metastatic sites. We also characterized the overlap of ESR1 alterations with commonly altered and clinically relevant genes in breast cancer, including PIK3CA mutations and HER2 amplification, and we report here a landscape of co-occurring alterations. In the cohort of patient samples where ctDNA was analyzed, BEAMing and FoundationAct assays both detected ESR1 mutations in 19 out of 48 samples, and overall concordance of mutation status (wild-type vs mutant) was 100%. A total of 51 individual mutations were detected with the BEAMing assay, 42 of which were detected with the FoundationACT assay. Seven mutations that were undetected by FoundationACT had mutant allele frequencies less than 0.1%. Ten ESR1 mutations were detected only by FoundationACT, 9 of which are not covered with the BEAMing assay. Alterations in PIK3CA, CDH1, TP53, ERBB2, and other breast cancer relevant genes were also detected with FoundationACT.

Conclusions: Understanding the mutational landscape of ESR1 and co-occurring alterations is important for diagnostic development in conjunction with the clinical development of novel anti-endocrine therapies. Our data demonstrate a large spectrum of mutations in the LBD in addition to known hotspot mutations. In addition, the FoundationACT assay offers a robust NGS-based method to screen for mutations in ctDNA that is highly concordant with digital PCR methods.

Citation Format: Spoerke JM, Schleifman E, Clark TA, Young G, Nahas M, Kennedy M, Young L, Chmielecki J, Otto GA, Lipson D, Wilson TR, Gendreau S, Lackner MR. The complete spectrum of ESR1 mutations from 7590 breast cancer tumor samples [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-07-08.

Abstract P3-04-18: The impact of circulating androgens on the androgen receptor in aromatase inhibitor resistant breast cancer

Aromatase Inhibitor (AI) therapy is the gold standard first line therapy for post-menopausal breast cancer. Following AI treatment the conversion of circulating androgens into estrone can be diminished by >99%, completely altering the tumour steroid microenvironment. Estrogens have been extensively studied in terms of the role of sex steroids as causative agents of breast cancer but interestingly, a case–control study nested within the European Prospective Investigation into Cancer and Nutrition reported that elevated levels of both serum androgens and estrogens are associated with increased breast cancer risk (E2 RR: 2.28 and androstenedione RR: 1.94). Also clinical studies have reported increases in the serum levels of androstenedione (4AD) in patients that recur on AI therapy.

In order to begin to address the role of circulating androgens and the androgen receptor (AR) in AI resistant breast cancer, RNA sequencing was performed on AI resistant (MCF7-aro-LetR) cells cultured with 4AD in the presence of letrozole. Gene expression analysis was performed using the cBioPortal software found that PI3K signalling pathways were upregulated in response to 4AD treatment as well as the upregulation of pathways involved in exocytosis and hormone secretion. We found elevated levels of PI3K and AR in endocrine resistant breast cancer compared to sensitive cell lines and that our AI resistant LetR cells showed significant reduction in cell growth in response to the combined pan class PI3K inhibitor BEZ235 and anti-AR treatment. Clinically, we evaluated the potential role of PI3K signalling in AI resistance a breast cancer patient TMA (n=488) was stained immunohistochemically for phospho-AKT and AR. The protective effect conferred by high AR expression in the total patient population (p=0.02) and tamoxifen treated population (p=0.05) was lost in the AI treated population (p=0.49). For p-AKT staining, patients that were positive for p-AKT significantly associate with the more aggressive luminal B classification yet exhibit an inverse association with luminal A subtype. Kaplan Meier analysis revealed that AR is associated with a more favourable outcome in p-AKT negative patients but that this is lost in the context of elevated p-AKT. Furthermore, interrogation of the breast invasive carcinoma TCGA dataset (n=963) revealed that patients with upregulated AR mRNA and wildtype PI3K exhibited more invasive disease compared with those harbouring a PI3K mutation (p=0.02) (n=56). Further investigations into the AR interacting partners will help elucidate potential mechanisms of action and the role of androgens and AR in facilitating breast cancer tumourigenesis.

Citation Format: Creevey L, Ali A, Bleach R, Hill ADK, Young L, McIlroy M. The impact of circulating androgens on the androgen receptor in aromatase inhibitor resistant breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-04-18.

Abstract PD1-05: Breast cancer brain metastases show limited intrinsic subtype switching, yet exhibit acquired ERBB2 amplifications and activating mutations

BACKGROUND: Metastasis is the major cause of mortality in breast cancer (BrCa) patients. Our understanding of brain metastasis (BrM) is limited, reflected by a lack of effective treatments. We aimed to (1) determine BrCa gene signature differences between primary tumors and matched BrM and (2) uncover BrM-specific alterations that may be clinically actionable.

MATERIALS and METHODS: NanoString expression profiling of 127 genes from 5 major prognostic tests (MammaPrint, EndoPredict, PAM50, OncotypeDX, MGI) was performed on 20 patient-matched primary (10 ER-neg, 10 ER-pos) and metastatic brain tumors. Subtype classification was performed using genefu. Protein changes in ER and HER2 (ERBB2) were confirmed by IHC. BrM-specific ERBB2 gains were corroborated in a publicly available dataset of 18 additional patient-matched cases (dbGAP phs000730.v1.p1). To test whether ERBB2 amplification and base pair mutation is metastasis-site specific, we further analyzed an expanded cohort of 7,884 breast tumors enriched for metastatic samples (52%) including liver (16.7%), lung (4.3%), bone (3.6%), and brain (2.0%) using comprehensive hybrid-capture sequencing of ERBB2.

RESULTS: 17/20 BrM retained the PAM50 subtype of the primary BrCa. Despite this concordance, 17/20 BrM harbored expression changes (< or > 2-fold) in clinically actionable genes including gains of FGFR4 (30%), FLT1 (20%), AURKA (10%) and loss of ESR1 expression (45%). The most recurrently upregulated gene was ERBB2, showing a >2-fold expression increase in 35% of BrM. 3 of 13 (23.3%) cases originally HER2-negative, and thus HER2-therapy naive, in the primary BrCa were IHC-positive (3+) in the paired BrM with an observed metastasis-specific amplification of the ERBB2 locus. In an independent dataset, 2 of 9 (22.2%) HER2-negative BrCa switched to HER2-positive with one BrM acquiring ERBB2 amplification and the other showing metastastic enrichment of the activating V777L ERBB2 mutation. Analysis of a large cohort of breast tumors (n=7,884) showed that across all organs ERBB2 amplification and/or base pair mutation was similar (p=0.18) between primary (13%) and metastatic disease (12%), however, a strong and significant enrichment was seen for BrM (primary 13% vs BrM 24%, p<0.0005).

CONCLUSIONS: Taken together, these results demonstrate that the majority (85%) of patient-matched BrM retain the intrinsic subtype of the primary cancer. However, despite this transcriptional similarity, alterations in clinically actionable genes are common, with BrM acquiring ERBB2 amplifications and/or base pair mutations at a frequency of ∼20%, even in HER2-therapy naive tumors. In a large cohort of primary and metastatic breast cancers, there is also a unique enrichment for ERBB2 alterations in BrM. This study provides a strong rationale to molecularly profile metastatic lesions to both better understand biological mechanisms of metastases and to perhaps refine therapeutic decision-making in advanced cancers.

Citation Format: Priedigkeit N, Hartmaier RJ, Chen Y, Vareslija D, Basudan A, Thomas R, Leone JP, Lucas PC, Bhargava R, Hamilton RL, Chmielecki J, Davidson NE, Oesterreich S, Brufsky AM, Young L, Lee AV. Breast cancer brain metastases show limited intrinsic subtype switching, yet exhibit acquired ERBB2 amplifications and activating mutations [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD1-05.

Abstract P3-04-17: Global transcriptional repression by the coactivator SRC-1 mediates disease progression in treatment-resistant breast cancer

Despite the effectiveness of endocrine therapy in treating estrogen receptor (ER) positive breast cancer, nearly 40% of breast cancer patients may develop resistance which can lead to metastatic disease progression. Steroid receptor co-activator-1 (SRC-1), a key regulator of ER signalling, is overexpressed in 35% of breast cancer patients and is strongly associated with the development of endocrine resistant metastasis(1). Recent studies highlighted the ability of SRC-1 to also act as transcriptional repressor. SRC-1 can therefore bi-directionally regulate gene expression to promote the resistant phenotype (2).

This study employed a genome-wide multi-omics sequencing approach to determine the SRC-1 repression signature in endocrine resistance and elucidate the mechanism by which SRC-1 mediates this repression. RNA-sequencing identified 736 genes significantly downregulated by SRC-1, with common functional pathways such as differentiation, cell morphogenesis and extracellular matrix enriched in the gene set. Parallel global methylation sequencing analysis revealed distinct differentially methylated regions specific to SRC-1 repressed target genes. Mechanistic studies in endocrine resistant cells revealed a role for methylation proteins, DNMTs and MBD, in SRC-1 directed repression. Through combined analysis of our global sequencing data we identified a network hub of five differentiation genes directly repressed by SRC-1. High expression of this signature predicts enhanced recurrence free survival in tamoxifen treated patients (n=335, p=0.032). A reduction in expression of these genes was shown to have functional output on proliferation, migration and mammosphere formation. Finally, we use tumour explant models to show that targeting DNA methylation can be used to reverse the SRC-1 driven suppression of this differentiation hub.

Here we report a novel mechanism by which SRC-1 may be driving endocrine resistant tumorigenesis. Our genome-wide discovery approach revealed a global epigenetic re-programming pathway whereby concerted differential DNA methylation is potentiated by the activation of SRC-1 in the presence of tamoxifen in the endocrine resistant setting. This study suggests that therapeutic strategies of combined targeted epigenetic therapy with estrogen deprivation could be a successful strategy to prevent acquired resistance to endocrine therapy.

Citation Format: Ward E, Varešlija D, Fagan A, Hill A, Young L. Global transcriptional repression by the coactivator SRC-1 mediates disease progression in treatment-resistant breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-04-17.

Transformation of galena to pyromorphite produces bioavailable sulfur for neutrophilic chemoautotrophy

The aqueous concentration of lead [Pb(II)] in geochemical environments is controlled by the solubility of Pb-bearing minerals and their weathering products. In contaminated soils, a common method for in situ stabilization of Pb(II) is the addition of phosphate to convert more redox sensitive sulfide minerals into sparingly soluble pyromorphite [Pb₅ (PO₄ )₃ X]. In this study, we conducted experimental studies to investigate the fate of reduced sulfur during the conversion of galena [PbS] to chloropyromorphite [Pb₅ (PO₄ )₃ Cl]. Powder X-ray diffraction analysis indicated that the reaction of phosphate with galena under oxic conditions resulted in the oxidation of sulfide and formation of elemental sulfur [S₈ ]. Under oxic abiotic conditions, the S₈ was retained in the solid phase, and negligible concentrations of sulfur as sulfide and thiosulfate were detected in the aqueous phase and only a small amount of sulfate. When PbS reacted in the presence of the chemoautotrophic organism Bosea sp. WAO, the S₈ in the secondary mineral was oxidized to sulfate. Strain WAO produced significantly more sulfate from the secondary S₈ than from the primary galena. Microscopic analysis of mineral-microbe aggregates on mineral-embedded slide cultures showed that the organism was colocalized and increased in biomass over time on the secondary mineral surface supporting a microbial role. The results of this study indicate that stimulation of sulfur-oxidizing activity may be a direct consequence of phosphate amendments to Pb(II)-contaminated soils.

Hetero-site-specific X-ray pump-probe spectroscopy for femtosecond intramolecular dynamics

New capabilities at X-ray free-electron laser facilities allow the generation of two-colour femtosecond X-ray pulses, opening the possibility of performing ultrafast studies of X-ray-induced phenomena. Particularly, the experimental realization of hetero-site-specific X-ray-pump/X-ray-probe spectroscopy is of special interest, in which an X-ray pump pulse is absorbed at one site within a molecule and an X-ray probe pulse follows the X-ray-induced dynamics at another site within the same molecule. Here we show experimental evidence of a hetero-site pump-probe signal. By using two-colour 10-fs X-ray pulses, we are able to observe the femtosecond time dependence for the formation of F ions during the fragmentation of XeF2 molecules following X-ray absorption at the Xe site.

Risk factors for low birth weight in a socio-economically disadvantaged population: parity, marital status, ethnicity and cigarette smoking

Low birth weight (LBW) is a public health problem, because it is associated with increased risk of morbidity and mortality. The principal aim of this study was to assess risk factors for LBW in a large multi-ethnic and socio-economically disadvantaged population. Data from 3242 mothers, who attended the Well Baby Clinic (Southwestern Sydney, Australia) for the first time, were analysed in relation to their demographic characteristics and socio-economic indices. The overall birthweight was 3377 +/- 577 g (mean +/- SD). In multiple linear regression analysis, smoking during pregnancy, marital status, parity, and country of birth were independently associated with birth weight. According to this analysis, lower birth weight was associated with mothers who had smoked during pregnancy (by 215.2 +/- 18.6 g), who were single (46.9 +/- 21 g), and of Asian background (108.5 +/- 38.2 g). However, higher parity was associated with significantly higher birth weight. The presence of each factor was coded as 1 and the absence, 0. A 'risk score' was then derived by summing up the individual scores. When birth weight was classified as 'low birth weight' (defined as those with birth weight being less than 2500 g) or normal birth weight, the overall prevalence of LBW was 1.9%. Each unit increase in the risk score was associated with a 1.9-fold (95% confidence interval: 1.5-2.6) increase in the risk of LBW. These data suggest that apart from marital status, ethnicity and parity, maternal smoking is the single most important preventable risk factors for LBW.

Validating FMEA output against incident learning data: A study in stereotactic body radiation therapy

PURPOSE

Though failure mode and effects analysis (FMEA) is becoming more widely adopted for risk assessment in radiation therapy, to our knowledge, its output has never been validated against data on errors that actually occur. The objective of this study was to perform FMEA of a stereotactic body radiation therapy (SBRT) treatment planning process and validate the results against data recorded within an incident learning system.

METHODS

FMEA on the SBRT treatment planning process was carried out by a multidisciplinary group including radiation oncologists, medical physicists, dosimetrists, and IT technologists. Potential failure modes were identified through a systematic review of the process map. Failure modes were rated for severity, occurrence, and detectability on a scale of one to ten and risk priority number (RPN) was computed. Failure modes were then compared with historical reports identified as relevant to SBRT planning within a departmental incident learning system that has been active for two and a half years. Differences between FMEA anticipated failure modes and existing incidents were identified.

RESULTS

FMEA identified 63 failure modes. RPN values for the top 25% of failure modes ranged from 60 to 336. Analysis of the incident learning database identified 33 reported near-miss events related to SBRT planning. Combining both methods yielded a total of 76 possible process failures, of which 13 (17%) were missed by FMEA while 43 (57%) identified by FMEA only. When scored for RPN, the 13 events missed by FMEA ranked within the lower half of all failure modes and exhibited significantly lower severity relative to those identified by FMEA (p = 0.02).

CONCLUSIONS

FMEA, though valuable, is subject to certain limitations. In this study, FMEA failed to identify 17% of actual failure modes, though these were of lower risk. Similarly, an incident learning system alone fails to identify a large number of potentially high-severity process errors. Using FMEA in combination with incident learning may render an improved overview of risks within a process.

Trends and costs of high concentration fluoride toothpaste prescribing in Scotland

Within Scotland there has been a significant increase in the prescription of 2,800 ppm and 5,000 ppm fluoride toothpaste. The objective of this paper was to analyse the trends in high concentration fluoride toothpaste (HCFT) in the five Scottish South East and Tayside (SEAT) health boards and consider the options for future national management of this prescribing. A retrospective analysis of routine prescribing data for the years 2006-2012 was carried out in primary care dental practices in Scotland. The cost of HCFT prescribing in the five Scottish health boards has increased from £15,243 (4,147 items) in 2006/07 to £206,529 (24,113 items), in 2011/12. Out of 2,430 dental list numbers, 100 list numbers (4.1%) accounted for 70% of the total prescribing costs (£144,367). The public dental service employs 153 (6%) of dentists working in Scotland, who in turn prescribe 11.6% HCFT. There is a need to ensure that the prescription of HCFT is both encouraged as best practice care but also managed appropriately to ensure that its delivery is targeted at those who are most in need.

Abstract P2-05-03: Whole genome transcriptome analysis of sequential breast to brain metastasis uncovers new signalling pathways and druggable targets

The occurrence of brain metastasis (BM) in breast cancer (BC) is currently on the rise across all molecular subtypes with 10-30% reported incidence. The need to uncover the mechanisms underlying this clinically devastating complication is apparent, and in the current study we sought to identify BC cell mediators of BM.

In our cohort of metastatic patients (n=196) we found that BM developed in 13% of the cases. Despite the previous reports of negative ER status being a risk factor for BM, the ER+ve patients accounted for 42% of all diagnosed BM. To elucidate the gene alterations required for successful colonisation of the brain we undertook RNA sequencing (RNA-seq) of sequential breast to brain metastasis of known receptor status (n=7).This study presents the first whole transcriptome next-generation RNA-seq analysis of resected BM and their matching primary breast tumours.

We identified 500 differentially expressed genes (DEGs) (< 0.05, fcThreshold >±1.5), accounting for those that were both upregulated and downregulated in BM compared to the primary. Analysis of protein-coding genes identified collective ER-specific metastatic pathways. Additionally, common functional pathways altered included ECM, cell adhesion and neuronal differentiation. Our analysis of the BM transcriptomic landscape and verification in cell line models that preferentially metastasise to the brain has unravelled a complex network of driver genes, cooperating with stromal derived factors, responsible for the organ-specific behaviour of the metastatic cells. Genes such as ANTRX1, THBS2, FAP, VCAN and TIMP2 were found to be part of the invasion and migration network that drives the extravasation of the BM cells. Furthermore, an EMT stemness signalling network driven by ANTRX1and WNT pathway driven RUNX was prominent in the cells acquiring the ability to migrate to the brain. Additional work is being carried out on uncovering the adaptations that re-activate the dormant brain metastatic cells and the contribution of the neuronal niche in the facilitating the colonisation by the MBC.

This study highlights the requirement of unique gene sets for the invasion, migration and colonisation to the brain and that functional characterisation of the DEGs will enable the identification of novel molecular targets for prevention and treatment of breast cancer BM.

Citation Format: Varešlija D, Fagan A, Buckley P, Farrell M, Hill A, Young L. Whole genome transcriptome analysis of sequential breast to brain metastasis uncovers new signalling pathways and druggable targets. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P2-05-03.