Search

XLS file 38K, Single-biomarker tumor growth inhibition projection

PDF file 155K, Principal component analysis (PCA) of the ten biomarkers provides a biomarker profile for each cell line

PDF file 285K, Biomarker validation by Western blot

PDF file 252K, In vivo pharmacodynamic IHC-based biomarker dose-response quantifications

PDF file 64K, Unsupervised clustering of the core biomarker modulation in cell lines, xenograft tumors, and in tumors from three patients in the 100mg cohort

PDF file 168K, Pearson and Spearman correlation with survival fraction for all cell lines together or individual cell lines

XLS file 37K, RPPA quantification of laser captured microscopy needle biopsies in patients from the 100mg cohort

Purpose: The oncogenic PI3K/Akt/mTOR pathway is an attractive therapeutic target in cancer. However, it is unknown whether the pathway blockade required for tumor growth inhibition is clinically achievable. Therefore, we conducted pharmacodynamic studies with GDC-0068, an ATP competitive, selective Akt1/2/3 inhibitor, in preclinical models and in patients treated with this compound.Experimental Design: We used a reverse phase protein array (RPPA) platform to identify a biomarker set indicative of Akt inhibition in cell lines and human-tumor xenografts, and correlated the degree of pathway inhibition with antitumor activity. Akt pathway activity was measured using this biomarker set in pre- and post-dose tumor biopsies from patients treated with GDC-0068 in the dose escalation clinical trial.Results: The set of biomarkers of Akt inhibition is composed of 10 phosphoproteins, including Akt and PRAS40, and is modulated in a dose-dependent fashion, both in vitro and in vivo. In human-tumor xenografts, this dose dependency significantly correlated with tumor growth inhibition. Tumor biopsies from patients treated with GDC-0068 at clinically achievable doses attained a degree of biomarker inhibition that correlated with tumor growth inhibition in preclinical models. In these clinical samples, compensatory feedback activation of ERK and HER3 was observed, consistent with preclinical observations.Conclusion: This study identified a set of biomarkers of Akt inhibition that can be used in the clinical setting to assess target engagement. Here, it was used to show that robust Akt inhibition in tumors from patients treated with GDC-0068 is achievable, supporting the clinical development of this compound in defined patient populations. Clin Cancer Res; 19(24); 6976–86. ©2013 AACR.

PDF file 161K, Tumor PIK3CA and PTEN status

XLS file 41K, RPPA biomarkers

PDF file 117K, Supplementary Materials and Methods -Supplementary Figure Legends -Supplementary Table Legends

KRAS- and BRAF-mutant tumors are often dependent on MAPK signaling for proliferation and survival and thus sensitive to MAPK pathway inhibitors. However, clinical studies have shown that MEK inhibitors are not uniformly effective in these cancers indicating that mutational status of these oncogenes does not accurately capture MAPK pathway activity. A number of transcripts are regulated by this pathway and are recurrently identified in genome-based MAPK transcriptional signatures. To test whether the transcriptional output of only 10 of these targets could quantify MAPK pathway activity with potential predictive or prognostic clinical utility, we created a MAPK Pathway Activity Score (MPAS) derived from aggregated gene expression. In vitro, MPAS predicted sensitivity to MAPK inhibitors in multiple cell lines, comparable to or better than larger genome-based statistical models. Bridging in vitro studies and clinical samples, median MPAS from a given tumor type correlated with cobimetinib (MEK inhibitor) sensitivity of cancer cell lines originating from the same tissue type. Retrospective analyses of clinical datasets showed that MPAS was associated with the sensitivity of melanomas to vemurafenib (HR: 0.596) and negatively prognostic of overall or progression-free survival in both adjuvant and metastatic CRC (HR: 1.5 and 1.4), adrenal cancer (HR: 1.7), and HER2+ breast cancer (HR: 1.6). MPAS thus demonstrates potential clinical utility that warrants further exploration., Biomarker: Gene signature predicts drug responses and patient outcomes A clinical score based on the activity of genes that regulate cell signaling can predict drug sensitivity and patient outcomes across a range of cancer types. Marie-Claire Wagle, Daniel Kirouac and their colleagues at Genentech in South San Francisco, California, USA developed an index that aggregates expression levels of 10 genes involved in modulating the mitogen-activated protein kinase (MAPK) pathway. This “MAPK Pathway Activity Score”, or MPAS, performed as well or better than other more complicated, genome-based tools at predicting whether drugs that inhibit MAPK-related enzymes were active against tumor cell lines. Retrospective analyses of clinical datasets also showed that MPAS correlated with survival outcomes in patients with melanoma, colon cancer, and breast cancer. The authors suggest that MPAS should be evaluated more broadly and perhaps implemented as a clinically informative biomarker.

Key molecular drivers that underlie transformation of colonic epithelium into colorectal adenocarcinoma (CRC) are well described. However, the mechanisms through which clinically targeted pathways are activated during CRC progression have yet to be elucidated. Here, we used an integrative genomics approach to examine CRC progression. We used laser capture microdissection to isolate colonic crypt cells, differentiated surface epithelium, adenomas, carcinomas and metastases, and used gene expression profiling to identify pathways that were differentially expressed between the different cell types. We identified a number of potentially important transcriptional changes in developmental and oncogenic pathways, and noted a marked upregulation of EREG in primary and metastatic cancer cells. We confirmed this pattern of gene expression by in situ hybridization and observed staining consistent with autocrine expression in the tumor cells. Upregulation of EREG during the adenoma-carcinoma transition was associated with demethylation of two key sites within its promoter, and this was accompanied by an increase in the levels of epidermal growth factor receptor (EGFR) phosphorylation, as assessed by reverse-phase protein analysis. In CRC cell lines, we demonstrated that EREG demethylation led to its transcriptional upregulation, higher levels of EGFR phosphorylation, and sensitization to EGFR inhibitors. Low levels of EREG methylation in patients who received cetuximab as part of a phase II study were associated with high expression of the ligand and a favorable response to therapy. Conversely, high levels of promoter methylation and low levels of EREG expression were observed in tumors that progressed after treatment. We also noted an inverse correlation between EREG methylation and expression levels in several other cancers, including those of the head and neck, lung and bladder. Therefore, we propose that upregulation of EREG expression through promoter demethylation might be an important means of activating the EGFR pathway during the genesis of CRC and potentially other cancers.

Significance Programmed death-ligand 1 (PD-L1) expression on tumor cells and tumor-infiltrating immune cells is regulated by distinct mechanisms and has nonredundant roles in regulating anticancer immunity, and PD-L1 on both cell types is important for predicting best response to atezolizumab in non-small cell lung cancer., Programmed death-ligand 1 (PD-L1) expression on tumor cells (TCs) by immunohistochemistry is rapidly gaining importance as a diagnostic for the selection or stratification of patients with non-small cell lung cancer (NSCLC) most likely to respond to single-agent checkpoint inhibitors. However, at least two distinct patterns of PD-L1 expression have been observed with potential biological and clinical relevance in NSCLC: expression on TC or on tumor-infiltrating immune cells (ICs). We investigated the molecular and cellular characteristics associated with PD-L1 expression in these distinct cell compartments in 4,549 cases of NSCLC. PD-L1 expression on IC was more prevalent and likely reflected IFN-γ–induced adaptive regulation accompanied by increased tumor-infiltrating lymphocytes and effector T cells. High PD-L1 expression on TC, however, reflected an epigenetic dysregulation of the PD-L1 gene and was associated with a distinct histology described by poor immune infiltration, sclerotic/desmoplastic stroma, and mesenchymal molecular features. Importantly, durable clinical responses to atezolizumab (anti–PD-L1) were observed in patients with tumors expressing high PD-L1 levels on either TC alone [40% objective response rate (ORR)] or IC alone (22% ORR). Thus, PD-L1 expression on TC or IC can independently attenuate anticancer immunity and emphasizes the functional importance of IC in regulating the antitumor T cell response.

Breast cancers are categorized into three subtypes based on protein expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2/ERBB2). Patients enroll onto experimental clinical trials based on ER, PR, and HER2 status and, as receptor status is prognostic and defines treatment regimens, central receptor confirmation is critical for interpreting results from these trials. Patients enrolling onto experimental clinical trials in the metastatic setting often have limited available archival tissue that might better be used for comprehensive molecular profiling rather than slide-intensive reconfirmation of receptor status. We developed a Random Forests-based algorithm using a training set of 158 samples with centrally confirmed IHC status, and subsequently validated this algorithm on multiple test sets with known, locally determined IHC status. We observed a strong correlation between target mRNA expression and IHC assays for HER2 and ER, achieving an overall accuracy of 97 and 96 %, respectively. For determining PR status, which had the highest discordance between central and local IHC, incorporation of expression of co-regulated genes in a multivariate approach added predictive value, outperforming the single, target gene approach by a 10 % margin in overall accuracy. Our results suggest that multiplexed qRT-PCR profiling of ESR1, PGR, and ERBB2 mRNA, along with several other subtype associated genes, can effectively confirm breast cancer subtype, thereby conserving tumor sections and enabling additional biomarker data to be obtained from patients enrolled onto experimental clinical trials. Electronic supplementary material The online version of this article (doi:10.1007/s10549-014-3163-8) contains supplementary material, which is available to authorized users.

Purpose: Tailoring cancer treatment to tumor molecular characteristics promises to make personalized medicine a reality. However, reliable genetic profiling of archived clinical specimens has been hindered by limited sensitivity and high false-positive rates. Here, we describe a novel methodology, MMP-seq, which enables sensitive and specific high-throughput, high-content genetic profiling in archived clinical samples. Experimental Design: We first validated the technical performance of MMP-seq in 66 cancer cell lines and a Latin square cross-dilution of known somatic mutations. We next characterized the performance of MMP-seq in 17 formalin-fixed paraffin-embedded (FFPE) clinical samples using matched fresh-frozen tissue from the same tumors as benchmarks. To demonstrate the potential clinical utility of our methodology, we profiled FFPE tumor samples from 73 patients with endometrial cancer. Results: We demonstrated that MMP-seq enabled rapid and simultaneous profiling of a panel of 88 cancer genes in 48 samples, and detected variants at frequencies as low as 0.4%. We identified DNA degradation and deamination as the main error sources and developed practical and robust strategies for mitigating these issues, and dramatically reduced the false-positive rate. Applying MMP-seq to a cohort of endometrial tumor samples identified extensive, potentially actionable alterations in the PI3K (phosphoinositide 3-kinase) and RAS pathways, including novel PIK3R1 hotspot mutations that may disrupt negative regulation of PIK3CA. Conclusions: MMP-seq provides a robust solution for comprehensive, reliable, and high-throughput genetic profiling of clinical tumor samples, paving the way for the incorporation of genomic-based testing into clinical investigation and practice. Clin Cancer Res; 20(8); 2080–91. ©2014 AACR.

Breast cancer is a heterogeneous disease and patients are managed clinically based on ER, PR, HER2 expression, and key risk factors. We sought to characterize the molecular landscape of high-risk breast cancer patients enrolled onto an adjuvant chemotherapy study to understand how disease subsets and tumor immune status impact survival. DNA and RNA were extracted from 861 breast cancer samples from patients enrolled onto the United States Oncology trial 01062. Samples were characterized using multiplex gene expression, copy number, and qPCR mutation assays. HR+ patients with a PIK3CA mutant tumor had a favorable disease-free survival (DFS; HR 0.66, P=0.05), however, the prognostic effect was specific to luminal A patients (Luminal A: HR 0.67, P=0.1; Luminal B: HR 1.01, P=0.98). Molecular subtyping of triple-negative breast cancers (TNBCs) suggested that the mesenchymal subtype had the worst DFS, whereas the immunomodulatory subtype had the best DFS. Profiling of immunologic genes revealed that TNBC tumors (n=280) displaying an activated T-cell signature had a longer DFS following adjuvant chemotherapy (HR 0.59, P=0.04), while a distinct set of immune genes was associated with DFS in HR+ cancers. Utilizing a discovery approach, we identified genes associated with a high risk of recurrence in HR+ patients, which were validated in an independent data set. Molecular classification based on PAM50 and TNBC subtyping stratified clinical high-risk patients into distinct prognostic subsets. Patients with high expression of immune-related genes showed superior DFS in both HR+ and TNBC. These results may inform patient management and drug development in early breast cancer.

Mutations in ESR1 have been associated with resistance to aromatase inhibitor (AI) therapy in patients with ER+ metastatic breast cancer. Little is known of the impact of these mutations in patients receiving selective oestrogen receptor degrader (SERD) therapy. In this study, hotspot mutations in ESR1 and PIK3CA from ctDNA were assayed in clinical trial samples from ER+ metastatic breast cancer patients randomized either to the SERD fulvestrant or fulvestrant plus a pan-PI3K inhibitor. ESR1 mutations are present in 37% of baseline samples and are enriched in patients with luminal A and PIK3CA-mutated tumours. ESR1 mutations are often polyclonal and longitudinal analysis shows distinct clones exhibiting divergent behaviour over time. ESR1 mutation allele frequency does not show a consistent pattern of increases during fulvestrant treatment, and progression-free survival is not different in patients with ESR1 mutations compared with wild-type patients. ESR1 mutations are not associated with clinical resistance to fulvestrant in this study., Fulvestrant degrades the oestrogen receptor. Here, the authors report on a clinical trial using fulvestrant and show that mutations in the oestrogen receptor alpha gene are prevalent in circulating tumour DNA and do not influence the clinical outcome of patients to fulvestrant.

Cancer of the ovary confers the worst prognosis among women with gynecological malignancies, primarily because most ovarian cancers are diagnosed at late stage. Hence, there is a substantial need to develop new diagnostic biomarkers to enable detection of ovarian cancer at earlier stages, which would confer better prognosis. In addition, the identification of druggable targets is of substantial interest to find new therapeutic strategies for ovarian cancer.The expression of 22,500 genes in a series of 67 serous papillary carcinomas was compared with 9 crudely enriched normal ovarian tissue samples by RNA hybridization on oligonucleotide microarrays. Multiple genes with near-uniformly expression were elevated in carcinomas of varying grade and malignant potential, including several previously described genes (e.g., MUC-1, CD9, CD24, claudin 3, and mesothelin). We performed immunohistochemical staining with antibodies against several of the proteins encoded by differentially expressed genes in an independent cohort of 71 cases of paraffin-embedded ovarian cancer samples.We found striking differences in EpCAM (p0.005), CD9 (p0.001), MUC-1 (p0.001), and claudin 3 proteins (p0.001) but not for mesothelin (p0.05) using the Mann-Whitney U test.Protein expression of a majority of the differentially expressed genes tested was found to be elevated in ovarian carcinomas and, as such, define potential new biomarkers or targets.

Activation of the PI3K pathway occurs commonly in a wide variety of cancers. Experience with other successful targeted agents suggests that clinical resistance is likely to arise and may reduce the durability of clinical benefit. Here, we sought to understand mechanisms underlying resistance to PI3K inhibition in PTEN-deficient cancers. We generated cell lines resistant to the pan-PI3K inhibitor GDC-0941 from parental PTEN-null breast cancer cell lines and identified a novel PIK3CB D1067Y mutation in both cell lines that was recurrent in cancer patients. Stable expression of mutant PIK3CB variants conferred resistance to PI3K inhibition that could be overcome by downstream AKT or mTORC1/2 inhibitors. Furthermore, we show that the p110β D1067Y mutant was highly activated and induced PIP3 levels at the cell membrane, subsequently promoting the localization and activation of AKT and PDK1 at the membrane and driving PI3K signaling to a level that could withstand treatment with proximal inhibitors. Finally, we demonstrate that the PIK3CB D1067Y mutant behaved as an oncogene and transformed normal cells, an activity that was enhanced by PTEN depletion. Collectively, these novel preclinical and clinical findings implicate the acquisition of activating PIK3CB D1067 mutations as an important event underlying the resistance of cancer cells to selective PI3K inhibitors. Cancer Res; 76(5); 1193–203. ©2016 AACR.

Stratification of Colorectal Cancer (CRC) into actionable molecular subtypes has tremendous clinical value. Recently, a consolidated classifier identified four molecularly distinct CRC subtypes (CMS1-4) that were associated with unique biology and clinical outcomes based on global transcriptional analysis of frozen tissues. Here, we developed and applied a novel CRC panel that is ideally suited for transcriptional classification of archival clinical samples. Findings from in silico analysis demonstrated that the 800 genes on our panel could accurately classify CRC samples from external public datasets into the correct CMS subtypes. We applied our panel in the analysis of a novel cohort of 312 formalin-fixed paraffin-embedded (FFPE) tissues from 205 patients, and were able to detect all 4 CMS subtypes in primary CRCs and in metastases. When we examined the CMS subtypes of primary tumors and matched metastases from 50 patients we found 70% of cases to be concordant, as were key biologies, such as WNT/MYC pathway activation in CMS2 and EMT features in tumors of the CMS4 subtypes. This was confirmed by in situ hybridization (ISH) using the markers ASCL2 for CMS2 and SPARC for CMS4, respectively. Discordance in the CMS subtypes between primary tumors and matched metastases were observed in 30% of cases and may reflect tumor heterogeneity. Our findings suggest that our CRC-focused panel many have clinical utility for CMS classification of FFPE samples, and point to potential risks of using CMS subtypes of primary tumors to inform clinical decision-making at the metastatic stage in a subset of patients. Citation Format: Ling-Yuh Huw, Robert Piskol, Felipe de Sosa e Melo, Doris Kim, Xueping Qu, Hartmut Koeppen, Mark Lackner, Garret Hampton, Omar Kabbarah, Rachel Tam. A custom gene expression panel for consensus molecular subtype classification of archival primary and metastatic colorectal cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2708. doi:10.1158/1538-7445.AM2017-2708

Lung cancer remains the primary cause of cancer-related deaths worldwide. Improved tools for early detection and therapeutic stratification would be expected to increase the survival rate for this disease. Alterations in the molecular pathways that drive lung cancer, which include epigenetic modifications, may provide biomarkers to help address this major unmet clinical need. Epigenetic changes, which are defined as heritable changes in gene expression that do not alter the primary DNA sequence, are one of the hallmarks of cancer, and prevalent in all types of cancer. These modifications represent a rich source of biomarkers that have the potential to be implemented in clinical practice. This perspective describes recent advances in the discovery of epigenetic biomarkers in lung cancer, specifically those that result in the methylation of DNA at CpG sites. We discuss one approach for methylation-based biomarker assay development that describes the discovery at a genome-scale level, which addresses some of the practical considerations for design of assays that can be implemented in the clinic. We emphasize that an integrated technological approach will enable the development of clinically useful DNA methylation-based biomarker assays. While this article focuses on current literature and primary research findings in lung cancer, the principles we describe here apply to the discovery and development of epigenetic biomarkers for other types of cancer.

Patients with newly diagnosed, early stage estrogen receptor positive (ER+) breast cancer often show disease free survival in excess of five years following surgery and systemic adjuvant therapy. An important question is whether diagnostic tumor tissue from the primary lesion offers an accurate molecular portrait of the cancer post recurrence and thus may be used for predictive diagnostic purposes for patients with relapsed, metastatic disease. As the class I phosphatidylinositol 3' kinase (PI3K) pathway is frequently activated in ER+ breast cancer and has been linked to acquired resistance to hormonal therapy, we hypothesized pathway status could evolve over time and treatment. Biomarker analyses were conducted on matched, asynchronous primary and metastatic tumors from 77 patients with ER+ breast cancer. We examined whether PIK3CA and AKT1 alterations or PTEN and Ki67 levels showed differences between primary and metastatic samples. We also sought to look more broadly at gene expression markers reflective of proliferation, molecular subtype, and key receptors and signaling pathways using an mRNA analysis platform developed on the Fluidigm BioMark™ microfluidics system to measure the relative expression of 90 breast cancer related genes in formalin-fixed paraffin-embedded (FFPE) tissue. Application of this panel of biomarker assays to matched tumor pairs showed a high concordance between primary and metastatic tissue, with generally few changes in mutation status, proliferative markers, or gene expression between matched samples. The collection of assays described here has been optimized for FFPE tissue and may have utility in exploratory analyses to identify patient subsets responsive to targeted therapies.

Purpose: The oncogenic PI3K/Akt/mTOR pathway is an attractive therapeutic target in cancer. However, it is unknown whether the pathway blockade required for tumor growth inhibition is clinically achievable. Therefore, we conducted pharmacodynamic studies with GDC-0068, an ATP competitive, selective Akt1/2/3 inhibitor, in preclinical models and in patients treated with this compound. Experimental Design: We used a reverse phase protein array (RPPA) platform to identify a biomarker set indicative of Akt inhibition in cell lines and human-tumor xenografts, and correlated the degree of pathway inhibition with antitumor activity. Akt pathway activity was measured using this biomarker set in pre- and post-dose tumor biopsies from patients treated with GDC-0068 in the dose escalation clinical trial. Results: The set of biomarkers of Akt inhibition is composed of 10 phosphoproteins, including Akt and PRAS40, and is modulated in a dose-dependent fashion, both in vitro and in vivo. In human-tumor xenografts, this dose dependency significantly correlated with tumor growth inhibition. Tumor biopsies from patients treated with GDC-0068 at clinically achievable doses attained a degree of biomarker inhibition that correlated with tumor growth inhibition in preclinical models. In these clinical samples, compensatory feedback activation of ERK and HER3 was observed, consistent with preclinical observations. Conclusion: This study identified a set of biomarkers of Akt inhibition that can be used in the clinical setting to assess target engagement. Here, it was used to show that robust Akt inhibition in tumors from patients treated with GDC-0068 is achievable, supporting the clinical development of this compound in defined patient populations. Clin Cancer Res; 19(24); 6976–86. ©2013 AACR.

The objective of this study was to evaluate circulating hepatocyte growth factor (cHGF) as a pharmacodynamic biomarker of Met inhibition for onartuzumab (MetMAb, OA5D5v2) in a phase I trial in patients with advanced cancers and a phase II trial in non–small cell lung cancer (NSCLC). The phase I study was a dose escalation trial with onartuzumab administered i.v. once every three weeks. The phase II study was a randomized two-arm trial in which onartuzumab or placebo was administered in combination with erlotinib in 137 patients with second and third line (2/3L) NSCLC. cHGF levels were evaluated by ELISA at multiple time points over the treatment period. Onartuzumab administration resulted in an acute and sustained rise in cHGF in both the phase I and phase II studies. Elevation in cHGF was independent of dose or drug exposure and was restricted to onartuzumab treatment. Neither higher baseline nor elevated change in cHGF levels upon treatment could simply be attributed to tumor burden or number of liver metastasis. We have shown that elevated cHGF can consistently and reproducibly be measured as a pharmacodynamic biomarker of onartuzumab activity. The elevation in cHGF is independent of tumor type, dose administered, or dose duration. Although these studies were not powered to directly address the contribution of cHGF as a predictive, on-treatment, circulating biomarker, these data suggest that measurement of cHGF in future expanded studies is warranted. Mol Cancer Ther; 12(6); 1122–30. ©2013 AACR.

Purpose Tumors with oncogenic dependencies on the HER family of receptor tyrosine kinases (RTKs) often respond well to targeted inhibition. Our previous work suggested that many cell lines derived from squamous cell carcinomas of the head and neck (SCCHNs) depend on autocrine signaling driven by HER2/3 dimerization and high-level co-expression of HRG. Additionally, results from a Phase I trial of MEHD7495A, a dual-action antibody that blocks ligand binding to EGFR and HER3, suggest that high-level HRG expression was associated with clinical response in SCCHN patients. Here we explore the hypothesis that high-level HRG expression defines a subpopulation of SCCHNs with activated HER3. Experimental Design qRT-PCR expression profiling was performed on >750 tumors of diverse origin, including >150 therapy-naive, primary, and recurrent SCCHNs. Activated HER3, defined by immunoprecipitation of phospho-HER3, was compared to HRG expression in SCCHN samples. Paracrine versus autocrine expression was evaluated using RNA-in situ hybridization. Results SCCHN tumors express the highest levels of HRG compared to a diverse collection of other tumor types. We show that high HRG expression is associated with activated HER3, whereas low HRG expression is associated with low HER3 activation in SCCHN tumors. Furthermore, HRG expression is higher in recurrent SCCHN compared to patient-matched therapy naive specimens. Conclusions HRG expression levels define a biologically distinct subset of SCCHN patients. We propose that high-level expression of HRG is associated with constitutive activation of HER3 in SCCHN and thus defines an actionable biomarker for interventions targeting HER3.

In the age of personalized medicine stratifying tumors into molecularly defined subtypes associated with distinctive clinical behaviors and predictable responses to therapies holds tremendous value. Towards this end, we developed a custom microfluidics-based bladder cancer gene expression panel for characterization of archival clinical samples. In silico analysis indicated that the content of our panel was capable of accurately segregating bladder cancers from several public datasets into the clinically relevant basal and luminal subtypes. On a technical level, our bladder cancer panel yielded robust and reproducible results when analyzing formalin-fixed, paraffin-embedded (FFPE) tissues. We applied our panel in the analysis of a novel set of 204 FFPE samples that included non-muscle invasive bladder cancers (NMIBCs), muscle invasive disease (MIBCs), and bladder cancer metastases (METs). We found NMIBCs to be mostly luminal-like, MIBCs to include both luminal- and basal-like types, and METs to be predominantly of a basal-like transcriptional profile. Mutational analysis confirmed the expected enrichment of FGFR3 mutations in luminal samples, and, consistently, FGFR3 IHC showed high protein expression levels of the receptor in these tumors. Our bladder cancer panel enables basal/luminal characterization of FFPE tissues and with further development could be used for stratification of bladder cancer samples in the clinic.

First two authors contributed equally Last two authors contributed equally Recent studies have shown that ex vivo propagation of normal tissues or patient-derived tumor cells in presence of irradiated fibroblast feeder cells and ROCK inhibitor can rapidly establish conditionally reprogramed cells (CRCs). In case of normal tissues, the induction of CRCs was reversible when the ROCK inhibitor and the feeder cells were removed, resulting in CRCs differentiating to its tissue origin (Liu et al.2012). Previous publications suggested that the establishment of such cell models provides new strategies to understand acquired resistance during treatment (Crystal et al 2015) and to predict treatment response (Liu et al. 2014). However, gene expression modulations and genomic drifting during the establishment of CRC propagation have not been thoroughly studied. The primary goal of this study is to molecularly characterize alterations between the original tumor tissues and the derived models growing with or without ROCK inhibitor. Understanding in-depth molecular fluctuations in this patient-derived ex vivo system will facilitate its appropriate use for tumor biology experimentation. Herein, tumors from prostate cancer and breast cancer patients were surgical removed and cryopreserved at the clinical site then processed and cultured as previously described (Liu et al. 2012). Gene expression profiling and next-generation sequencing were carried out on the original tumor tissues and cellular models passaged during the CRC propagation in the presence or absence of ROCK inhibitor. Gene expression analysis of the prostate cancer cells and the breast cancer cells were carried out using a 93-gene prostate cancer-focused Fluidigm panel and a 800 gene NanoString breast cancer-focused panel, respectively. Cancer hotspot mutations were analyzed using the Ion Torrent Cancer Hotspot v2 NGS assay. Through aforementioned genomic and transcriptomic interrogations, we demonstrated the extent of indication-relevant gene expression modulation during establishment and propagation of these cells. We also characterize cancer hotspot mutations in the primary tumor cells and the stability of those mutations during ex vivo propagation. These results should begin to inform the appropriate use of the CRC model for tumor biology experimentation. Citation Format: Jessica Li, Bonnie Liu, Rin Nakamura, Heidi Savage, Shoji Ikeda, Timothy Wilson, Teiko Sumiyoshi, Garret Hampton, Lukas Amler, Mark Lackner, Shih-Min A. Huang, Walter C. Darbonne. Gene expression and genomic drift comparative analysis between patient-derived conditionally reprogrammed cells and original tumors. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4257.

KRAS mutations occur in approximately 25% of NSCLC (1). Tumors with these mutations are predicted to be sensitive to MEK inhibition due to activation of MAPK signaling. However, MEK inhibitors in multiple clinical trials, either as a monotherapy or in combination with chemotherapies, have not shown superior efficacy in the KRAS mutant subgroup when compared to the KRAS wild-type subgroup, indicating a limitation of utilizing KRAS mutation status as a predictive biomarker of efficacy to MEK inhibition (2, 3). Furthermore, stratification based on KRAS mutation status may inadvertently miss wild-type KRAS tumors that could be addicted to MAPK signaling regardless of KRAS mutation status, thus denying patients potential benefit from MEK inhibitors. Here we describe a novel predictive model that more accurately forecasts the sensitivity of the KRAS wild-type NSCLC subpopulation to MEK inhibitors such as cobimetinib and trametinib. Cell viability data from cobimetinib or trametinib-treated cells, with concomitant gene expression data (RNAseq), from 46 colon, 106 lung, and 37 pancreatic cell lines were used to create an elastic net regression model trained on gene expression features (alpha = 0.5 and optimal lambda chosen by 5-fold cross validation) (4). From the model, we established two distinct predictive gene lists: (1) a longer low cross-validation list, and (2) a shorter low error list. Initial analysis of the model demonstrated that predicted mean viabilities of the cell lines used to create the model correlated well with their actual mean viabilities (R: 0.65-0.7 for trametinib and cobimetinib respectively). Predicted mean viabilities of 40 previously unscreened NSCLC cell lines were then generated by the model based on their expression features (RNAseq). The 40 cell lines were categorized either as sensitive or resistant by the median of predicted mean viabilities derived from the model. Subsequently, the actual experimental GI50 values of cobimetinib were obtained for each of these cell lines. We found that KRAS mutational status predicted that only 8 of the 40 cell lines screened would be sensitive to MEK inhibition. In contrast, our model predicted that 15 of the 40 cell lines screened would be sensitive. Experimentally, we demonstrated that 24 of the cell lines were sensitive to MEK inhibition with a measured GI50 References (1) Blumenschein GR. et al., (2015) Annal Oncol. 26(5):894-901. (2) Gandara DL et al., (2013) J Clin Oncol 31, (suppl; abstr 8028). (3) Laethem JLV et al, (2014) J Clin Oncol 32:5s, (suppl; abstr 4025). (4) Barretina J. et al. (2012) Nature 483, 603-607. Citation Format: Marie Wagle, Christiaan Klijn, Bonnie Liu, Shilpi Mahajan, Peter Haverty, John Moffat, Mark Merchant, Bob Yauch, Garret Hampton, Lukas Amler, Mark Lackner, Shih-Min A. Huang. A novel predictive biomarker model for MEK sensitivity. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 417.

The molecular drivers that underlie transformation of colonic epithelium into colorectal adenocarcinoma (CRC) have been well described. However, the mechanisms through which some of the clinically targeted pathways are activated during CRC progression have yet to be elucidated. Here, we employed an integrative genomics approach to examine CRC progression. Transcriptional profiling of laser capture microdissected colonic crypt cells, differentiated surface epithelium, adenomas, carcinomas, and metastases, showed distinctive patterns in the activation of developmental and oncogenic pathways, including the clinically important EGFR axis. We observed a dramatic up-regulation of the EGFR ligand EREG in primary and metastatic cancer cells as compared to normal and adenomatous tissues that was indicative of autocrine tumor production, and confirmed this pattern of gene expression by in situ hybridization. Global methylation analysis indicated that up-regulation of EREG during the adenoma-carcinoma transition was associated with de-methylation of two key sites within the EREG promoter and this was accompanied by an increase in the levels of EGFR phosphorylation, as assessed by reverse phase protein analysis. In a clinical trial setting, we observed that low levels of EREG methylation in patients who received cetuximab as part of a Phase II study were associated with high expression of the ligand and a favorable response to therapy. Conversely, high levels of promoter methylation and low levels of EREG expression were observed in tumors that progressed after treatment. We also noted an inverse correlation between EREG methylation and expression levels in several other cancers from the TCGA datasets, including those of the head and neck, lung, and bladder. We propose that up-regulation of EREG expression through promoter de-methylation might be an important means of activating the EGFR pathway during the genesis of CRC and, potentially, other types of cancer. Citation Format: Xueping Qu, Thomas Sandmann, Henry Frierson, Ling Fu, Eloisa Fuentes, Kimberly Walter, Kwame Okrah, Craig Rumpel, Christopher Moskaluk, Shan Lu, Yulei Wang, Richard Bourgon, Elicia Penuel, Andrea Pirzkall, Lukas Amler, Mark Lackner, Josep Tabenero, Garret Hampton, Omar Kabbarah. Integrated genomic analysis of colorectal cancer progression reveals activation of EGFR through demethylation of the EREG promoter. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 5155.

Activation of the phosphoinositide 3-kinase (PI3K) pathway occurs commonly in breast cancers via mechanisms that include loss of function of the tumor suppressor phosphatase and tensin homolog (PTEN). Experience with other successful targeted agents suggests that clinical resistance to PI3K inhibitors is likely to arise and may reduce the durability of clinical benefit. Here, we sought to understand mechanisms underlying resistance to PI3K inhibition in PTEN deficient breast cancers. PTEN null breast cancer cell lines were selected for resistance to a pan-PI3K inhibitor, GDC-0941. Comprehensive molecular and cellular profiling was conducted to identify the mechanism of resistance. We generated GDC-0941 resistant derivatives of PTEN deficient EVSA-T and ZR-75-1 cell lines and identified a novel PIK3CB D1067Y mutation in both cell lines. We found that the PIK3CB mutation at D1067 position was recurrent in cancer patients. Stable expression of mutant PIK3CB variants in PTEN null breast cancer cells conferred resistance to PI3K inhibition that could be overcome by downstream AKT or mTORC1/2 inhibitors. We showed further that the p110β D1067Y mutant is highly activated and elevates PIP3 level at the cell membrane, promoting localization and activation of AKT and PDK1 at the cell membrane and driving PI3K signaling to a level that can overcome treatment with proximal inhibitors. Finally, we show that the PIK3CB D1067Y mutant can behave as an oncogene and transform normal cells, and that PTEN knock down enhances this activity. These novel preclinical and clinical findings implicate PIK3CB D1067 alterations as a novel oncogene that may cause resistance to selective PI3K inhibitor treatment. Taken together with previous findings that PTEN deficient cancers tend to signal through p110β rather than p110α, this work also suggests PTEN deficient breast cancers may depend on p110β and are thereby susceptible to PIK3CB mutation as an escape mechanism from PI3K inhibition. Citation Format: Yoshito Nakanishi, Kimberly M. Walter, Jill M. Spoerke, Carol O'Brien, Ling Y. Huw, Garret M. Hampton, Mark R. Lackner. Activating mutations in PIK3CB confer resistance to PI3K inhibition in PTEN-deficient breast cancer and define a novel oncogenic role for p110β. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Breast Cancer Research; Oct 17-20, 2015; Bellevue, WA. Philadelphia (PA): AACR; Mol Cancer Res 2016;14(2_Suppl):Abstract nr B28.

Purpose: Class 1 phosphatidylinositol 3-kinase (PI3K) plays a major role in cell proliferation and survival in a wide variety of human cancers. Here, we investigated biomarker strategies for PI3K pathway inhibitors in non–small-cell lung cancer (NSCLC). Experimental Design: Molecular profiling for candidate PI3K predictive biomarkers was conducted on a collection of NSCLC tumor samples. Assays included comparative genomic hybridization, reverse-transcription polymerase chain reaction gene expression, mutation detection for PIK3CA and other oncogenes, PTEN immunohistochemistry, and FISH for PIK3CA copy number. In addition, a panel of NSCLC cell lines characterized for alterations in the PI3K pathway was screened with PI3K and dual PI3K/mTOR inhibitors to assess the preclinical predictive value of candidate biomarkers. Results: PIK3CA amplification was detected in 37% of squamous tumors and 5% of adenocarcinomas, whereas PIK3CA mutations were found in 9% of squamous and 0% of adenocarcinomas. Total loss of PTEN immunostaining was found in 21% of squamous tumors and 4% of adenocarcinomas. Cell lines harboring pathway alterations (receptor tyrosine kinase activation, PI3K mutation or amplification, and PTEN loss) were exquisitely sensitive to the PI3K inhibitor GDC-0941. A dual PI3K/mTOR inhibitor had broader activity across the cell line panel and in tumor xenografts. The combination of GDC-0941 with paclitaxel, erlotinib, or a mitogen-activated protein–extracellular signal-regulated kinase inhibitor had greater effects on cell viability than PI3K inhibition alone. Conclusions: Candidate biomarkers for PI3K inhibitors have predictive value in preclinical models and show histology-specific alterations in primary tumors, suggesting that distinct biomarker strategies may be required in squamous compared with nonsquamous NSCLC patient populations. Clin Cancer Res; 18(24); 6771–83. ©2012 AACR.

Background: The PI3K/mTOR pathway is frequently activated in cancer by multiple mechanisms. GDC-0980 is a dual pan-PI3K and mTOR (TORC1/2) inhibitor that has been evaluated in two Phase II studies (J Clin Oncol 32:5s, 2014 [suppl; abstr 4525 and 5513]). The randomized ROVER study showed that the PI3K/mTOR inhibitor GDC-0980 did not improve efficacy over the TORC1 inhibitor everolimus in metastatic renal cell carcinoma (mRCC). The single arm MAGGIE study evaluated the activity of GDC-0080 in patients with advanced endometrial cancer (EC). Although some single agent anti-tumor activity was observed, overall evaluation of anti-tumor activity of GDC-0980 was limited by tolerability, especially in diabetic patients. Comprehensive biomarker analysis, including targeted next generation sequencing (NGS) and a panel of biomarkers tailored to each tumor type, was conducted in both Phase II studies to determine the prevalence of PI3K/mTOR pathway alterations, and to assess the association between anti-tumor activity and candidate predictive biomarkers. Methods: The primary and secondary endpoints included progression-free survival (PFS) and objective response rate (ORR). Archival tissue samples were collected for biomarker analysis, and correlations with efficacy were retrospectively explored. Samples were subjected to targeted NGS (Illumina) covering 88 oncogenes and tumor suppressors, copy number analysis using quantitative-PCR, PTEN immunohistochemistry (IHC), HIF1A IHC (ROVER), and gene expression analysis (NanoString nCounter® System, ROVER). Results: In ROVER, the median PFS was significantly shorter for GDC-0980 than everolimus. Retrospective biomarker analyses revealed a relationship between VHL mutation status (by NGS) and improved PFS with everolimus but not GDC-0980. High HIF1A protein expression was associated with longer PFS in both arms, whereas low expression of STK11/LKB1 mRNA was associated with benefit with everolimus only. Additional gene expression analysis of PI3K pathway, apoptosis/cell cycle, and tumor immunity related genes will be presented. In MAGGIE, PFS at 6 months was estimated to be 20%, and the ORR was 9% (unconfirmed). Evaluable archival tumor samples were obtained from 88% of the patients and 52% of patients had at least one alteration in PIK3CA, PTEN or AKT1. PTEN loss by IHC was generally well correlated with mutation status determined by NGS. All 5 patients with either a confirmed or investigator assessed partial response had at least one PI3K pathway alteration. Conclusions: Clinical data to date have suggested that identification of predictive biomarkers for agents targeting PI3K/mTOR signaling is challenging and will require tailoring to specific tumor types. Here we provide comprehensive assessment from two phase II clinical studies of GDC-0980. Our data, although retrospective in nature and requiring confirmation, suggest that pathway activation along the VHL-HIF1A axis may be predictive of anti-tumor activity for mTOR-targeting agents in mRCC. Our results in EC suggest that at least in this study population, frequency of pathway alterations was somewhat lower than observed in prior published data, but the presence of PIK3CA or AKT1 mutations or PTEN loss enriched for anti-tumor activity. Clinical trial information: NCT01442090 (mRCC), NCT01455493 (EC). Citation Format: Jill M. Spoerke, Vicky Makker, Carol Aghajanian, Powles Thomas, Robert J. Motzer, Jennifer O. Lauchle, Hema Parmar, Houston Gilbert, Wei Lin, Bridget O'Keeffe, Michelle Byrtek, Hartmut Koeppen, Yulei Wang, Shan Lu, Ling-Yuh Huw, Garret M. Hampton, Mark R. Lackner. Comprehensive predictive biomarker evaluation in two phase II clinical trials of the PI3K/mTOR inhibitor GDC-0980 in metastatic renal cell carcinoma and advanced endometrial cancer. [abstract]. In: Proceedings of the AACR Special Conference: Targeting the PI3K-mTOR Network in Cancer; Sep 14-17, 2014; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(7 Suppl):Abstract nr A03.

Background: The PI3K/Akt pathway is frequently altered in cancer by multiple mechanisms including PI3K activating mutations, PTEN loss, RTK activation, and other oncogenic mutations. GDC-0941 and GDC-0980 are selective PI3K and dual PI3K/mTOR inhibitors, respectively, which are currently in clinical development. Pharmacodynamic changes in biomarkers in response to dose and exposures were analyzed from sequential biopsies from phase I studies of GDC-0941 and GDC-0980. The purpose was to evaluate pathway inhibition at tolerable doses, as well as look for associations between modulation of phosphorylation and gene expression of downstream PI3K factors and interactions with cancer immune cell infiltration. Methods: Pre- and post-treatment biopsies were collected from a subset of patients who were treated with escalating doses of GDC-0941 (NCT00876122, NCT00876109) or GDC-0980 (NCT00854126). In addition to previously described phospho-S6, phospho-AKT, and phospho-PRAS40 analysis by immunohistochemistry (IHC), CyclinD1, phospho-ERK, Ki-67, and markers of T-cell infiltration (CD8, PD-L1) were also assessed by IHC. Gene expression analysis was also performed with the nCounter® Analysis System (NanoString Technologies) to determine if pathway modulation can be assessed more quantitatively across a broader set of markers, and to determine whether feedback upregulation of pathway components was observed in treated patient samples. Genes analyzed included PI3K pathway, apoptosis/cell cycle, and tumor immunity related genes. Results: Pharmacodynamic biomarker assays were conducted on 23 paired samples from the GDC-0941 study and 22 paired samples from the GDC-0980 study. Post-dose samples were obtained within hours of anticipated Cmax of both drugs. Reduction of phospho-S6, phospho-AKT, and phospho-PRAS40 were observed in a dose and exposure dependent manner. Upregulation of immune-related proteins was not observed after two weeks dosing with GDC-0941 or GDC-0980, which could be impacted by inhibition of T-cell signaling through PI3K. We report here the pharmacodynamic gene expression analysis, as measured by the NanoString nCounter® system, in patients from whom tissue was available, and analysis of the extent to which these the pharmacodynamic biomarkers are associated with each other. Based on PK modeling and PD, the doses achieved in Phase I studies enable future studies to be conducted at doses associated with tumor xenograft shrinkage (J Clin Oncol 29: 2011 [suppl; abstr 3052, 3021], Mol Cancer Ther 2009;8[12 Suppl]:B137). Conclusions: Pharmacodynamic assays confirmed effective broad pathway knockdown of multiple signaling components at safe and tolerated clinical doses of GDC-0941 and GDC-0980. What remains unclear is duration and magnitude of pathway inhibition required will translate to clinical efficacy and translatability across tumor types that may have different PI3K pathway dependencies and alterations. These data supports further clinical testing to evaluate efficacy in these different patient subsets. Citation Format: Yoshito Nakanishi, Jill M. Spoerke, Mika Derynck, Jennifer O. Lauchle, Hartmut Koeppen, Jill Fredrickson, Joseph Ware, Garret Hampton, Yibing Yan, Mark R. Lackner. Pharmacodynamic biomarker evaluation in phase I clinical trials of selective PI3K and PI3K/mTOR inhibitors. [abstract]. In: Proceedings of the AACR Special Conference: Targeting the PI3K-mTOR Network in Cancer; Sep 14-17, 2014; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(7 Suppl):Abstract nr B03.

Patients with newly diagnosed, early stage estrogen receptor positive (ER+) breast cancer often show disease free survival in excess of five years following surgery and systemic adjuvant therapy. As such, an important question is whether diagnostic tumor tissue from the primary lesion offers an accurate molecular portrait of the cancer post recurrence and thus may be used for predictive diagnostic purposes for patients with relapsed, metastatic disease. To address this question, we performed detailed biomarker analyses on matched, asynchronous primary and metastatic tumors from 77 patients with ER+ breast cancer. The class I phosphatidylinositol 3' kinase (PI3K) is thought to be an important driver in ER+ breast cancer and has been linked to acquired resistance to hormonal therapy. We thus examined whether mutations in PIK3CA and loss of PTEN showed differences in primary and metastatic samples. We also sought to look more broadly at markers reflective of proliferation, molecular subtype, and key receptors and signaling pathways. To accomplish this, we developed an analysis platform using the Fluidigm BioMark™ microfluidics system to measure the relative expression of 90 breast cancer related genes in formalin-fixed paraffin-embedded (FFPE) tissue. Application of this panel of assays to matched tumor pairs showed a very high concordance between primary and metastatic tissue, with generally few changes in mutation status, proliferative markers, or gene expression between matched samples. Thus, archival primary tumor tissue may still provide an accurate portrait of biomarker status in patients with disease recurrence. Citation Format: Jill M. Spoerke, Erica B. Schleifman, Rupal Desai, Yuanyuan Xiao, Cheryl Wong, Ilma Abbas, Carol O'Brien, Rajesh Patel, Teiko Sumiyoshi, Ling Fu, Rachel Tam, Hartmut Koeppen, Timothy Wilson, Rajiv Raja, Garret M. Hampton, Mark R. Lackner. Comprehensive biomarker profiling of matched primary and metastatic estrogen receptor positive breast cancers. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Breast Cancer Research: Genetics, Biology, and Clinical Applications; Oct 3-6, 2013; San Diego, CA. Philadelphia (PA): AACR; Mol Cancer Res 2013;11(10 Suppl):Abstract nr B024.

Background: The PI3K/Akt pathway is frequently activated in cancer by multiple mechanisms including PI3K activating mutations, PTEN loss, RTK activation, and other means. GDC-0941 and GDC-0980 are selective PI3K and dual PI3K/mTOR inhibitors, respectively, which are currently in phase II clinical development. Assays for candidate predictive and pharmacodynamic biomarkers were conducted on patient samples collected from phase I studies of GDC-0941 and GDC-0980. The purpose was to confirm pathway inhibition at tolerable doses, as well as look for association between anti-tumor activity and candidate predictive biomarkers. Methods: Archival tumor tissue samples were analyzed using a 6 gene qPCR mutation assay (PIK3CA, EGFR, KRAS, BRAF, NRAS, AKT1), an immunohistochemistry assay for PTEN, and a fluorescence in situ hybridization (FISH) assay for PIK3CA. Select samples were analyzed for an expanded qPCR mutation panel and subjected to targeted next-generation sequencing (Illumina). For pharmacodynamic biomarker assays, pre- and post-treatment biopsies were collected from a subset of patients. In addition to previously described pS6 analysis, samples were analyzed for phospho-AKT, phospho-PRAS40, and CyclinD1 by immunohistochemistry. Results: Predictive biomarker assays were conducted on over 200 samples from the phase I studies. Overall we found a prevalence of 7% PIK3CA mutations and 12% loss of PTEN in these samples. PIK3CA amplification was observed in several samples from ovarian cancer patients. Based on several means of evaluating tumor response (FDG-PET, RECIST, time on study), activity was seen at or below clinically relevant doses in several different tumor types, including breast, ovarian, and mesothelioma. We report here the predictive biomarker status in all patients from whom tissue was available, and analysis of the extent to which these alterations are associated clinical outcome, to the extent such associations can be determined from a phase I dose escalation study designed to look a safety and tolerability. Conclusions: Pharmacodynamic assays confirmed effective pathway knockdown at safe and tolerated clinical doses of GDC-0941 and GDC-0980. Anti-tumor activity was observed in patients with PIK3CA mutations, as well as some patients whose tumors did not harbor pathway alterations. These data support patient stratification in phase II clinical studies to determine whether predictive biomarkers will be useful in identifying responsive patients. Citation Format: Jill Spoerke, Rupal Desai, Rajesh Patel, Jill Fredrickson, Yulei Wang, Gallia Levy, Steve Gendreau, Jennifer Lauchle, Mika Derynck, Rajiv Raja, Hartmut Koeppen, Garret Hampton, Yibing Yan, Mark R. Lackner. Biomarker evaluation in phase I clinical trials of selective PI3K and PI3K/mTOR inhibitors. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 4567. doi:10.1158/1538-7445.AM2013-4567

Bladder cancer is a common malignancy, and in the US approximately 15,000 patients a year succumb to metastatic disease. The transition from the non-invasive to the lethal invasive variety is poorly understood, and this is reflected by the lack of efficacious treatment options for patients presenting with advanced disease. Notably, a significant fraction of superficial cancers can recur, requiring additional surgeries and conferring a higher risk for disease progression. Stratifying bladder cancers into molecular subtypes with defined clinical attributes could highlight non-invasive tumors with high risk of recurrence and reveal opportunities for therapeutic intervention in advanced disease. To this end, we molecularly characterized a collection of ∼200 clinically-annotated, formalin-fixed, paraffin-embedded tissues that represent non-invasive and invasive/advanced stage histopathologies. Tumors were assessed for mutation status at ∼100 mutation hotspots in key oncogenes as well as for the expression levels of ∼100 genes on a custom Fluidigm™ platform to interrogate key bladder cancer pathways, such as the FGFR3, PI3K and MAPK signaling axes. Integrative analysis of gene expression, mutation, and clinical data identified non-invasive subtypes that were FGFR3 mutation positive and exhibited a pathway up regulation gene expression signature. In contrast, invasive tumors were FGFR3 wild type and displayed less prominent pathway up regulation. As expected, invasive tumors had significantly worse disease-free survival (DFS) than their non-invasive counterparts (HR = 0.54; P = 0.03). On the molecular level, advanced tumors exhibited dysregulation of key pathways, including p53 and PI3K. Although similar histologically, and of common FGFR3 mutation status, tumors of the non-invasive type could be further classified into two distinct transcriptional subtypes associated with remarkably different DFS profiles (HR = 0.29; P = 0.004). Our molecular stratification of bladder cancer identified distinct subtypes associated with their respective clinical behaviors. In advanced disease, we defined molecular alterations that highlight opportunities for therapeutic intervention. We also identified a novel subtype of non-invasive malignancies associated with a surprisingly high risk of recurrence, highlighting the value of molecular stratification for identifying bladder cancer patients who might benefit from more aggressive treatment than the current standard of care. Citation Format: Doris Kim, YJ Choi, Dorothy French, Rajesh Patel, Ling Fu, Cheryl Wong, Ilma Abbas, Rachel Tam, Erica Schleifman, Teiko Sumiyoshi, Anna Faarborg, Bob Yauch, Garret Hampton, Lukas Amler, Rajiv Raja, Omar Kabbarah. Molecular stratification of bladder cancer reveals distinct subtypes associated with unique clinical behaviors. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 1203. doi:10.1158/1538-7445.AM2013-1203

Activation of the PI3K pathway has been reported in endometrial cancers, and there are a number of inhibitors targeting different nodes of the PI3K pathway currently in clinical trials, including the PI3K/mTOR dual inhibitor GDC-0980 that is in a phase II trial for endometrial cancer (NCT01455493). In this study, we set to examine the prevalence of overlapping mutational activation of the PI3K pathway with other pathways, such as MAPK. We also tested whether targeting different nodes of PI3K pathway, such as PI3K, AKT, or PI3K/mTOR could exert differential effects on endometrial cancer cells, and whether combination with a MEK inhibitor could provide an additional benefit in inhibiting endometrial cancer cell growth. Alteration of oncogenes and tumor suppressor genes were profiled in 80 endometrial cancer samples by qRT-PCR, targeted deep sequencing, and IHC staining. Multiple PI3K inhibitors, including GDC-0941 (PI3K), GDC-0068 (AKT), GDC-0980 (PI3K/mTOR), and the MEK inhibitor GDC-0973 were tested in a panel of 27 endometrial cancer cell lines for their effects on cell growth. Cell signaling status at baseline and on-treatment was profiled by reverse phase protein array. Mutations and copy number variations of relevant oncogenes and tumor suppressor genes were examined in these cells by targeted deep sequencing. In the 80 endometrial tumor tissues we profiled, there were multiple instances of PI3K pathway alteration including 31% PIK3CA mutations, 9% AKT1 mutations, 16% MET mutations, and 38% PTEN null. In addition, KRAS mutations significantly overlapped with PTEN null or PIK3CA mutation in the same specimen. Endometrial cancer cell lines had distinct patterns of sensitivity to inhibitors targeting different nodes of the PI3K pathway. Preliminary results showed that GDC-0068 was more effective in cell lines with PTEN mutations, while GDC-0941 had a greater effect on those with PIK3CA mutations, and the cell lines were broadly sensitive to GDC-0980. In spite of the widespread activation of the PI3K pathway in endometrial cancer cells, synergistic inhibition of cell growth was observed when GDC-0973 was combined with either GDC-0068 or GDC-0941 in most of the cell lines tested. We will seek to identify genomic and/or proteomic features that will allow better selection as to which node of the PI3K pathway to target, and when to combine with a MEK inhibitor. Citation Format: Matthew J. Wongchenko, Yinghui Guan, Marie-Claire Wagle, Lisa Ryner, Shan Lu, Hartmut Koeppen, Garret Hampton, Mark Lackner, Yulei Wang, Yibing Yan. Sensitivity of endometrial cancer cells to inhibitors targeting different nodes of the PI3K pathway and their combination with a MEK inhibitor. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3479. doi:10.1158/1538-7445.AM2013-3479

The class I phosphatidylinositol 3’ kinases (PI3K) play a major role in proliferation and survival in a wide variety of human cancers, and activation of the PI3K pathway is thought to be an important driver in estrogen receptor positive (ER+) breast cancer. A key factor in successful development of drugs targeting this pathway will be development in appropriate molecular subsets. Important questions relevant to PI3K inhibitor development in ER+ breast cancers are whether these inhibitors will work equally well in luminal A compared to luminal B tumors, and whether gene expression signatures of pathway activation may have additional utility in patient stratification beyond PIK3CA mutation status alone. The goal of this study was to develop a methodology for high throughput profiling of ER+ breast cancers, in order to enable molecular subtyping of patients enrolled in clinical studies. To accomplish this, we developed an analysis platform to measure the relative expression of 90 breast cancer and PI3K pathway specific genes in formalin-fixed paraffin-embedded (FFPE) tissue. The content for this panel consists of genes known to be important for epithelial-mesenchymal biology, proliferation rate, and transcriptional output of the PI3K pathway. The 96 assay panel (including 6 housekeeping genes) was developed on the Fluidigm Biomark microfluidics platform and was extensively validated using well-characterized breast cancer cell lines and FFPE breast cancer samples of known subtypes based on immunohistochemistry for HER2, ER, and PR. All assays showed high levels of inter-and intra-chip reproducibility and were sensitive on standard curves down to 3ng RNA input. Using this method we were able to separate breast cancers into distinct molecular subtypes, as well as identify more proliferative luminal B type tumors. In addition, PIK3CA mutation status, a potential biomarker, was determined using a highly specific and sensitive qRT-PCR mutation assay, in order to allow comparison with the PI3K pathway activation signature. We extended these analyses to a small cohort of patient samples consisting of matched primary and metastatic tumor tissues, and report here the correlation of primary and matched metastatic ER+ breast cancer FFPE tumor samples at both the gene expression and mutational levels. We found that the majority of matched pairs were concordant for both mutation status and gene expression, though a subset did show differences. Future studies will examine the prognostic significance and clinical relevance of this gene signature. Citation Format: Erica B. Schleifman, Rupal M. Desai, Jill Spoerke, Cheryl Victoria Wong, Ilma Abbas, Carol O'Brien, Garret Hampton, Timothy Wilson, Hartmut Koeppen, Rajesh Patel, Teiko Sumiyoshi, Ling Fu, Rachel Tam, Rajiv Raja, Mark Lackner. High-throughput qRT-PCR expression profiling of estrogen receptor positive breast tumors. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3462. doi:10.1158/1538-7445.AM2013-3462

Somatic copy number variations (CNVs) are frequently observed in cancer, likely a result high genomic and structural instability characteristic of cancer. CNVs have been shown to be important drivers for many oncogenic pathways and are associated with drug sensitivity and resistance. Therefore, these genetic alterations are valuable candidate biomarkers for patient stratification and predictors for therapeutic responses to targeted therapies. A fundamental challenge for copy number determination in tumor FFPE DNA is that the material is often degraded and modified by fixation, processing and storage. Here we present a PCR-based copy number assay utilizing target specific pre-amplification and a microfluidics detection system. 45 genes were selected based on literature and their potential relevance to therapeutic targets. This high-throughput platform showed excellent specificity and dynamic range from 1 to 100ng input genomic DNA. A panel of matched frozen and FFPE DNAs from cell lines were used for cross -validation with copy number identified by array CGH. A medium-center method is employed to calculate copy number and assess the extent of degradation in sample DNAs. To evaluate clinical utility of the technology, copy number alteration from a collection of matched primary and metastatic breast tumors as well as correlation between copy number changes and gene expression will be reported. We will also report results from a survey of other tumor types, including gastric cancer, mesothelioma, and HER2+ breast cancer. Citation Format: Ling-Yuh Huw, Rajesh Patel, Carol O'Brien, Ling Fu, Rajiv Raja, Lukas Amler, Garret Hampton, Mark Lackner. Development of robust copy number assays for tumor FFPE tissues. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3505. doi:10.1158/1538-7445.AM2013-3505

Introduction: The PI3K/AKT pathway is activated in a wide variety of human cancers, and it has previously been shown that its dysregulation can lead to EGFR inhibitor resistance. In colorectal cancer, activating mutations of the PIK3CA gene have been identified in 13–17% of patients, and typically occur in either the helical domain (E542K or E545K) or the kinase domain (H1047R). Here, we used the colorectal cancer cell SW48, which has an EGFR G719S activating mutation, as well as the isogenic lines with knock-in of either PIK3CA E545K or H1047R mutation to study the differential effects of these two mutations on cell signaling, and the impact that each mutation has on the activity of the EGFR inhibitors erlotinib (Tarceva®), and GDC-0068, an AKT inhibitor. Results: The effect on cell growth of the erlotinib was measured in parental, PIK3CA E545K, and PIK3CA H1047R SW48 cells. Compared to the parental SW48 cells, knock-in of E545K led to erlotinib resistance, while knock-in of H1047R significantly sensitized cells to erlotinib. The GI50 of cells exposed to erlotinib was 2,400, 21,000, and 20 nM, respectively. The EGFR/ErB2 inhibitor lapatinib (Tykerb®) had GI50 values of >30,000, >30,000, and 18 nM, respectively. To explore the mechanism of resistance and sensitivity rendered by the two PIK3CA mutations, we measured the phosphorylation levels of 80 proteins that represent several key signaling pathways in these isogenic cells. Compared to parental SW48 cells, H1047R cells were found to have increased activation of EGFR and ErbB2, as well as PDGFRβ and Ron, receptor tyrosine kinases that have been shown to heterodimerize with members of the ErbB family. Furthermore, these cells showed increased activity of the MEK/ERK pathway. Treatment with erlotinib led to attenuation of EGFR, ErbB2, PDGFRβ, Ron, and MEK/ERK pathway signaling, providing a mechanistic basis for the increased effect of erlotinib in H1047R cells. While cells with both the E545K and H1047R mutations had increased activity of the PI3K/AKT pathway, a greater increase was seen in E545K cells. To determine the impact of each mutation on the cell's dependence on the PI3K/AKT pathway, we measured the effect of GDC-0068 on cell proliferation. SW48 parental, E545K, and H1047R cells had GI50 values of 2,500, 520, and 1,800 nM, respectively, highlighting the differential dependence on the PI3K/AKT pathway in cells with the E545K. While GDC-0068 inhibited downstream targets of PI3K/AKT in all three lines, increased suppression was seen in the E545K mutant. In addition, more substantial inhibition of targets that converge downstream of both PI3K/AKT and MEK/ERK was seen in E545K cells. Conclusion: Previous studies have shown that activating mutations of the PIK3CA gene lead to independence from receptor tyrosine kinase signaling, and insensitivity to EGFR inhibition. However, we have demonstrated that two distinct types of activating mutation of PIK3CA lead to differential sensitivity to EGFR inhibitors in CRC cells. PI3K with the H1047R mutation led to increased receptor tyrosine kinase activities and increased sensitivity to EGFR inhibitors, whereas PI3K with the E545K mutation caused hyperactivity of the PI3K/AKT pathway, which led to resistance to EGFR inhibitors and increased sensitivity to an Akt inhibitor. Our results, which added to potential explanations for the variability of clinical outcomes of EGFR inhibitors in patients with colorectal cancer, demonstrate the importance of informed patient selection.

Purpose: Chronic myeloid leukemia (CML) can be effectively treated with BCR-ABL inhibitors such as imatinib, dasatinib, and nilotinib; however, resistance to these inhibitors develops over time causing patients to relapse. Recently, MEK inhibition was shown to synergize with BCR-ABL inhibitors to inhibit resistant CML cells harboring the T315I mutation (1). In this study we characterized mechanisms of resistance to BCR-ABL inhibitor in acquired resistant cells without BCR-ABL mutation by genomic and phosphoprotein profiling. We further evaluated the sensitivity of these resistant clones to MEK inhibitor, GDC0973, PI3K inhibitor, GDC0941, alone or in combination. Methods: The CML cell K562 was grown in an increasing concentration of imatinib or/and dasatinib to induce acquired resistance. Clones were isolated from the resistant pool and the GI50 of each clone was measured using the Cell-Titer Glo® viability assay. Genomic DNA from selected clones was tested using Oncoscan® (Affymetrix), which surveys whole genome for copy number variation (CNV) and 400 oncogenic mutations. Cell lysates were subjected to reverse phase protein array (RPPA) analysis of phosphorylation levels of ∼100 proteins. Genomic mutation and CNV profiles, as well as protein phosphorylation patterns of the clones were compared to the parental cells to identify potential pathways involved in resistance. Results: Four imatinib-resistant clones with varying GI50s (4–30 times higher than parental cells) were fully characterized with both Oncoscan and RPPA. None contained the T315I mutation. A genomic mutation scan showed that all the clones had multiple mutations affecting various signaling pathways including TGFβ (Smad-2), Wnt (APC), PI3K (PIK3CA) and EGFR. Likewise, whole genome copy number scan showed that CNV occurred in multiple genes including copy number gains in ADAM29, BMP2, Rspondin4 (Wnt) and PKA in multiple clones, and copy number losses in ADAMTS14, LIM-kinase and FzD9 (Wnt) in multiple clones. The protein phosphorylation patterns also showed multiple changes in the resistant clones. Three of the resistant clones (K14, 15 and 25) had 6-10-fold increased pMEK and pERK, relative to the parental cells. MEk inhibition resulted in a cytostatic response in the parental cells (GI50: 1900 nM), whereas resistant clones 14, 15, and 25 were highly sensitive to GDC0973-mediated apoptosis, with GI50s ranging from 60–75 nM. Clone 6 had lower pMEK and pERK levels than the others, but was still sensitive to GDC0973 (GI50: 600 nM). Phosphorylation of proteins in the PI3K pathway such as pS6, p70S6K and p-mTOR were elevated in all of the clones as compared to the parental cells (7–11 fold); however, none of these clones were sensitive to GDC0941 alone. Our data suggest that either GDC0973 or GDC0973 alone or in combination with GDC0941 could be effective at inducing cell death in resistant clones despite their various genetic backgrounds. Conclusion: The MEK inhibitor GDC0973 caused robust cellular apoptosis in all of the resistant clones in contrast to the cytostatic effect in parental cells. All of the clones had elevated PI3K signaling as shown by enhanced PI3K substrate phosphorylation. Despite the many genetic aberrations that occurred as cells became resistant; most noticeably changes in Wnt, PKA, ADAMs and BMP2 in all of the clones, resistant cells appear to channel escape/survival signals through the MAPK and/or PI3K pathway rendering them sensitive to GDC0973 alone or in combination with GDC0941.

Lung cancer is the leading cause of cancer deaths worldwide. Several oncogenes and tumor suppressor genes have been implicated in lung carcinogenesis. Among them, PIK3CA pathway activation is one of the key pathways that link cell surface receptor kinases to effectors for tumor growth, metabolism and survival. To support predictive diagnostic biomarker efforts for the selective PI3K inhibitor GDC-0941, we conducted detailed molecular characterization of NSCLC tumor samples. We evaluated the key oncogenic mutations (KRAS, EGFR, PIK3CA, BRAF and NRAS), gene expression, protein expression (PTEN) and gene copy number alterations in genes in the PI3K pathway in NSCLC tumors. We found that oncogenic mutations and copy number alternations were associated with histological subtype. Of note, mutations in KRAS and EGFR were found with high frequency (30%) in adenocarcinoma, as was loss of the tumor suppressor LKB1. On the other hand, amplification of PI3K was found in the squamous subtype, as were rare mutations in PI3K. The tumor suppressor PTEN showed a high frequency of reduced expression in both squamous and adenocarcinomas, though total loss of PTEN was more common in squamous carcinomas. Our studies suggest that the PI3K pathway is an attractive therapeutic target in NSCLC, and that distinct biomarker strategies may be required for development in squamous and non-squamous patient populations. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1147. doi:1538-7445.AM2012-1147

GDC-0941 is a class I PI3 kinase inhibitor of all 4 isoforms currently in phase II clinical trials in multiple indications, including ER+ breast cancer. The PI3 kinase pathway plays an important role in proliferation and survival and is known to be one of the most altered pathways in cancer. These alterations include PI3 kinase amplification, PI3 kinase mutations and PTEN loss, as well as numerous others. We have previously shown in-vitro and in-vivo data that alterations in this pathway can potentially help identify patients that are more likely to respond to a PI3 kinase inhibitor. A major outstanding question is mechanisms of intrinsic and acquired resistance to PI3K inhibitors. To address these questions we have selected multiple breast cancer cell lines harboring different pathway alterations (HER2, PIK3K mutations, PTEN loss) for resistance to PI3 kinase inhibition. Some of the resistant lines show cross resistance to inhibitors of mTOR and AKT (ie general pathway resistance), while others show specific resistance to PI3K or PI3K/mTOR inhibitors. We have characterized these resistant cell lines in detail at the molecular level and in terms of signaling pathway activation. Analyses performed include deep sequencing, reverse phase protein arrays, as well as focused Western blots and RTK arrays. We will discuss these results in detail and comment on candidate resistance mechanisms. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 5609. doi:1538-7445.AM2012-5609

Background: The PI3K/Akt pathway is frequently activated in cancer by multiple mechanisms including PI3K activating mutations, PTEN loss, RTK activation, and other means. GDC-0941 and GDC-0980 are selective PI3K and dual PI3K/mTOR inhibitors, respectively, which are currently in phase II clinical development. Assays for candidate predictive and pharmacodynamic biomarkers were conducted on patient samples collected from phase I studies of GDC-0941 and GDC-0980. The purpose was to confirm pathway inhibition at tolerable doses, as well as look for association between anti-tumor activity and candidate predictive biomarkers. Methods: Archival tumor tissue samples were analyzed using a 6 gene mutation assay (PIK3CA, EGFR, KRAS, BRAF, NRAS, AKT1), an immunohistochemistry assay for PTEN, and a fluorescence in situ hybridization (FISH) assay for PIK3CA. Select samples were analyzed for an expanded mutation panel and genome-wide copy number alterations using the Oncoscan platform (Affymetrix). For pharmacodynamic biomarker assays, analyses were initially limited to pre and post treatment phospho-AKT levels in platelet rich plasma. Once substantial downmodulation of pAKT in surrogate tissue was observed, pre- and post treatment biopsies were collected from a subset of patients and analyzed for phospho-S6 levels by immunohistochemistry. Results: Predictive biomarker assays were conducted on over 150 samples from the phase I study. Overall we found a prevalence of 7% PIK3CA mutations and 15% loss of PTEN in these samples. PIK3CA amplification was observed in several samples from ovarian cancer patients. Overall there were five confirmed partial responses in the studies, and two of these occurred in patients with activating mutations in PIK3CA. Robust PD knockdown of pAKT in surrogate tissue was observed in the early stage of dose-escalation. Moreover, substantial knockdown of pS6 was seen in tumor tissues in patients treated at or below the recommended phase II dose. Conclusions: Pharmacodynamic assays confirmed effective pathway knockdown at safe and tolerated clinical doses of GDC-0941 and GDC-0980. Anti-tumor activity was observed in patients with PIK3CA mutations, as well as some patients whose tumors did not harbor pathway alterations. These data support patient stratification in phase II clinical studies to determine whether a predictive biomarker will be useful in identifying responsive patients. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4821. doi:1538-7445.AM2012-4821

Purpose: Non-small cell lung cancers (NSCLC) comprise multiple distinct biological groups with different prognoses. For example, patients with epithelial-like (EL) tumors have a better prognosis and exhibit greater sensitivity to inhibitors of the epidermal growth factor receptor (EGFR) pathway than patients with mesenchymal-like (ML) tumors. Here we test the hypothesis that EL NSCLCs can be distinguished from ML NSCLCs on the basis of global DNA methylation patterns. Experimental Design: To determine whether phenotypic subsets of NSCLC can be defined based on their DNA methylation patterns, we combined microfluidics-based gene expression analysis and genome-wide methylation profiling. We derived robust classifiers for both gene expression and methylation in cell lines and tested these classifiers in surgically resected NSCLC tumors. We validate our approach using quantitative RT-PCR and methylation specific PCR in formalin-fixed biopsies from NSCLC patients who went on to fail front-line chemotherapy. Results: We show that patterns of methylation divide NSCLCs into EL and ML subsets as defined by gene expression and that these signatures are similarly correlated in NSCLC cell lines and tumors. We show that DNA methylation differences predict EL or ML phenotypes NSCLC cell lines, surgically resected tumors, and formalin-fixed biopsies from NSCLC patients who went on to fail front-line chemotherapy. Conclusions: Our data demonstrate that patterns of DNA methylation can divide non-small lung cancers into two phenotypically distinct subtypes of tumors and provide proof of principle that differences in DNA methylation can be used as a platform for predictive biomarker discovery and development. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1732. doi:1538-7445.AM2012-1732