Pan Tong

A pan-cancer proteomic perspective on The Cancer Genome Atlas

Protein levels and function are poorly predicted by genomic and transcriptomic analysis of patient tumors. Therefore, direct study of the functional proteome has the potential to provide a wealth of information that complements and extends genomic, epigenomic and transcriptomic analysis in The Cancer Genome Atlas (TCGA) projects. Here we use reverse-phase protein arrays to analyze 3,467 patient samples from 11 TCGA “Pan-Cancer” diseases, using 181 high-quality antibodies that target 128 total proteins and 53 post-translationally modified proteins. The resultant proteomic data is integrated with genomic and transcriptomic analyses of the same samples to identify commonalities, differences, emergent pathways and network biology within and across tumor lineages. In addition, tissue-specific signals are reduced computationally to enhance biomarker and target discovery spanning multiple tumor lineages. This integrative analysis, with an emphasis on pathways and potentially actionable proteins, provides a framework for determining the prognostic, predictive and therapeutic relevance of the functional proteome.

Co-occurring genomic alterations define major subsets of KRAS - mutant lung adenocarcinoma with distinct biology, immune profiles, and therapeutic vulnerabilities

UNLABELLED The molecular underpinnings that drive the heterogeneity of KRAS-mutant lung adenocarcinoma are poorly characterized. We performed an integrative analysis of genomic, transcriptomic, and proteomic data from early-stage and chemorefractory lung adenocarcinoma and identified three robust subsets of KRAS-mutant lung adenocarcinoma dominated, respectively, by co-occurring genetic events in STK11/LKB1 (the KL subgroup), TP53 (KP), and CDKN2A/B inactivation coupled with low expression of the NKX2-1 (TTF1) transcription factor (KC). We further revealed biologically and therapeutically relevant differences between the subgroups. KC tumors frequently exhibited mucinous histology and suppressed mTORC1 signaling. KL tumors had high rates of KEAP1 mutational inactivation and expressed lower levels of immune markers, including PD-L1. KP tumors demonstrated higher levels of somatic mutations, inflammatory markers, immune checkpoint effector molecules, and improved relapse-free survival. Differences in drug sensitivity patterns were also observed; notably, KL cells showed increased vulnerability to HSP90-inhibitor therapy. This work provides evidence that co-occurring genomic alterations identify subgroups of KRAS-mutant lung adenocarcinoma with distinct biology and therapeutic vulnerabilities. SIGNIFICANCE Co-occurring genetic alterations in STK11/LKB1, TP53, and CDKN2A/B-the latter coupled with low TTF1 expression-define three major subgroups of KRAS-mutant lung adenocarcinoma with distinct biology, patterns of immune-system engagement, and therapeutic vulnerabilities.

MOABS: model based analysis of bisulfite sequencing data

Bisulfite sequencing (BS-seq) is the gold standard for studying genome-wide DNA methylation. We developed MOABS to increase the speed, accuracy, statistical power and biological relevance of BS-seq data analysis. MOABS detects differential methylation with 10-fold coverage at single-CpG resolution based on a Beta-Binomial hierarchical model and is capable of processing two billion reads in 24 CPU hours. Here, using simulated and real BS-seq data, we demonstrate that MOABS outperforms other leading algorithms, such as Fisher’s exact test and BSmooth. Furthermore, MOABS analysis can be easily extended to differential 5hmC analysis using RRBS and oxBS-seq. MOABS is available at http://code.google.com/p/moabs/.

A Patient-Derived, Pan-Cancer EMT Signature Identifies Global Molecular Alterations and Immune Target Enrichment Following Epithelial-to-Mesenchymal Transition

Purpose: We previously demonstrated the association between epithelial-to-mesenchymal transition (EMT) and drug response in lung cancer using an EMT signature derived in cancer cell lines. Given the contribution of tumor microenvironments to EMT, we extended our investigation of EMT to patient tumors from 11 cancer types to develop a pan-cancer EMT signature. Experimental Design: Using the pan-cancer EMT signature, we conducted an integrated, global analysis of genomic and proteomic profiles associated with EMT across 1,934 tumors including breast, lung, colon, ovarian, and bladder cancers. Differences in outcome and in vitro drug response corresponding to expression of the pan-cancer EMT signature were also investigated. Results: Compared with the lung cancer EMT signature, the patient-derived, pan-cancer EMT signature encompasses a set of core EMT genes that correlate even more strongly with known EMT markers across diverse tumor types and identifies differences in drug sensitivity and global molecular alterations at the DNA, RNA, and protein levels. Among those changes associated with EMT, pathway analysis revealed a strong correlation between EMT and immune activation. Further supervised analysis demonstrated high expression of immune checkpoints and other druggable immune targets, such as PD1, PD-L1, CTLA4, OX40L, and PD-L2, in tumors with the most mesenchymal EMT scores. Elevated PD-L1 protein expression in mesenchymal tumors was confirmed by IHC in an independent lung cancer cohort. Conclusions: This new signature provides a novel, patient-based, histology-independent tool for the investigation of EMT and offers insights into potential novel therapeutic targets for mesenchymal tumors, independent of cancer type, including immune checkpoints. Clin Cancer Res; 22(3); 609–20. ©2015 AACR.

The miR-200 family and the miR-183~96~182 cluster target Foxf2 to inhibit invasion and metastasis in lung cancers.

Metastatic lung cancer is one of the most lethal forms of cancer and molecular pathways driving metastasis are still not clearly elucidated. Metastatic cancer cells undergo an epithelial–mesenchymal transition (EMT) where they lose their epithelial properties and acquire a migratory and invasive phenotype. Here we identify that the expression of microRNAs from the miR-200 family and the miR-183~96~182 cluster are significantly co-repressed in non-small cell lung cancer cell lines and primary tumors from multiple TCGA dataset with high EMT scores. Ectopic expression of the miR-183~96~182 cluster inhibited cancer cell migration and invasion, whereas its expression was tightly modulated by miR-200. We identified Foxf2 as a common, novel and direct target of both these microRNA families. Foxf2 expression tightly correlates with the transcription factor Zeb1 and is elevated in mesenchymal-like metastatic lung cancer cells. Foxf2 expression induced robust EMT, migration, invasion and metastasis in lung cancer cells, whereas Foxf2 inhibition significantly repressed these phenotypes. We also demonstrated that Foxf2 transcriptionally represses E-cadherin and miR-200, independent of Zeb1, to form a double-negative feedback loop. We, therefore, identified a novel mechanism whereby the miR-200 family and the miR-183~96~182 cluster inhibit lung cancer invasion and metastasis by targeting Foxf2.

ZEB1 induces LOXL2-mediated collagen stabilization and deposition in the extracellular matrix to drive lung cancer invasion and metastasis

Lung cancer is the leading cause of cancer-related deaths, primarily due to distant metastatic disease. Metastatic lung cancer cells can undergo an epithelial-to-mesenchymal transition (EMT) regulated by various transcription factors, including a double-negative feedback loop between the microRNA-200 (miR-200) family and ZEB1, but the precise mechanisms by which ZEB1-dependent EMT promotes malignancy remain largely undefined. Although the cell-intrinsic effects of EMT are important for tumor progression, the reciprocal dynamic crosstalk between mesenchymal cancer cells and the extracellular matrix (ECM) is equally critical in regulating invasion and metastasis. Investigating the collaborative effect of EMT and ECM in the metastatic process reveals increased collagen deposition in metastatic tumor tissues as a direct consequence of amplified collagen gene expression in ZEB1-activated mesenchymal lung cancer cells. In addition, collagen fibers in metastatic lung tumors exhibit greater linearity and organization as a result of collagen crosslinking by the lysyl oxidase (LOX) family of enzymes. Expression of the LOX and LOXL2 isoforms is directly regulated by miR-200 and ZEB1, respectively, and their upregulation in metastatic tumors and mesenchymal cell lines is coordinated to that of collagen. Functionally, LOXL2, as opposed to LOX, is the principal isoform that crosslinks and stabilizes insoluble collagen deposition in tumor tissues. In turn, focal adhesion formation and FAK/SRC signaling is activated in mesenchymal tumor cells by crosslinked collagen in the ECM. Our study is the first to validate direct regulation of LOX and LOXL2 by the miR-200/ZEB1 axis, defines a novel mechanism driving tumor metastasis, delineates collagen as a prognostic marker, and identifies LOXL2 as a potential therapeutic target against tumor progression.

High-resolution profiling of histone h3 lysine 36 trimethylation in metastatic renal cell carcinoma

Mutations in SETD2, a histone H3 lysine trimethyltransferase, have been identified in clear cell renal cell carcinoma (ccRCC); however it is unclear if loss of SETD2 function alters the genomic distribution of histone 3 lysine 36 trimethylation (H3K36me3) in ccRCC. Furthermore, published epigenomic profiles are not specific to H3K36me3 or metastatic tumors. To determine if progressive SETD2 and H3K36me3 dysregulation occurs in metastatic tumors, H3K36me3, SETD2 copy number (CN) or SETD2 mRNA abundance was assessed in two independent cohorts: metastatic ccRCC (n=71) and the Cancer Genome Atlas Kidney Renal Clear Cell Carcinoma data set (n=413). Although SETD2 CN loss occurs with high frequency (>90%), H3K36me3 is not significantly impacted by monoallelic loss of SETD2. H3K36me3-positive nuclei were reduced an average of ~20% in primary ccRCC (90% positive nuclei in uninvolved vs 70% positive nuclei in ccRCC) and reduced by ~60% in metastases (90% positive in uninvolved kidney vs 30% positive in metastases) (P<0.001). To define a kidney-specific H3K36me3 profile, we generated genome-wide H3K36me3 profiles from four cytoreductive nephrectomies and SETD2 isogenic renal cell carcinoma (RCC) cell lines using chromatin immunoprecipitation coupled with high-throughput DNA sequencing and RNA sequencing. SETD2 loss of methyltransferase activity leads to regional alterations of H3K36me3 associated with aberrant RNA splicing in a SETD2 mutant RCC and SETD2 knockout cell line. These data suggest that during progression of ccRCC, a decline in H3K36me3 is observed in distant metastases, and regional H3K36me3 alterations influence alternative splicing in ccRCC.

Activation of the PI3K/mTOR Pathway following PARP Inhibition in Small Cell Lung Cancer

Small cell lung cancer (SCLC) is an aggressive malignancy with limited treatment options. We previously found that PARP is overexpressed in SCLC and that targeting PARP reduces cell line and tumor growth in preclinical models. However, SCLC cell lines with PI3K/mTOR pathway activation were relatively less sensitive to PARP inhibition. In this study, we investigated the proteomic changes in PI3K/mTOR and other pathways that occur following PAPR inhibition and/or knockdown in vitro and in vivo. Using reverse-phase protein array, we found the proteins most significantly upregulated following treatment with the PARP inhibitors olaparib and rucaparib were in the PI3K/mTOR pathway (p-mTOR, p-AKT, and pS6) (p≤0.02). Furthermore, amongst the most significantly down-regulated proteins were LKB1 and its targets AMPK and TSC, which negatively regulate the PI3K pathway (p≤0.042). Following PARP knockdown in cell lines, phosphorylated mTOR, AKT and S6 were elevated and LKB1 signaling was diminished. Global ATP concentrations increased following PARP inhibition (p≤0.02) leading us to hypothesize that the observed increased PI3K/mTOR pathway activation following PARP inhibition results from decreased ATP usage and a subsequent decrease in stress response signaling via LKB1. Based on these results, we then investigated whether co-targeting with a PARP and PI3K inhibitor (BKM-120) would work better than either single agent alone. A majority of SCLC cell lines were sensitive to BKM-120 at clinically achievable doses, and cMYC expression was the strongest biomarker of response. At clinically achievable doses of talazoparib (the most potent PARP inhibitor in SCLC clinical testing) and BKM-120, an additive effect was observed in vitro. When tested in two SCLC animal models, a greater than additive interaction was seen (p≤0.008). The data presented here suggest that combining PARP and PI3K inhibitors enhances the effect of either agent alone in preclinical models of SCLC, warranting further investigation of such combinations in SCLC patients.

An integrated molecular analysis of lung adenocarcinomas identifies potential therapeutic targets among TTF1-negative tumors, including DNA Repair Proteins and Nrf2

Purpose: Thyroid transcription factor-1 (TTF1) immunohistochemistry (IHC) is used clinically to differentiate primary lung adenocarcinomas (LUAD) from squamous lung cancers and metastatic adenocarcinomas from other primary sites. However, a subset of LUAD (15%–20%) does not express TTF1, and TTF1-negative patients have worse clinical outcomes. As there are no established targeted agents with activity in TTF1-negative LUAD, we performed an integrated molecular analysis to identify potential therapeutic targets. Experimental Design: Using two clinical LUAD cohorts (274 tumors), one from our institution (PROSPECT) and The Cancer Genome Atlas, we interrogated proteomic profiles (by reverse phase protein array, RPPA), gene expression, and mutational data. Drug response data from 74 cell lines were used to validate potential therapeutic agents. Results: Strong correlations were observed between TTF1 IHC and TTF1 measurements by RPPA (Rho = 0.57, P < 0.001) and gene expression (NKX2-1, Rho = 0.61, P < 0.001). Established driver mutations (e.g., BRAF and EGFR) were associated with high TTF1 expression. In contrast, TTF1-negative LUAD had a higher frequency of inactivating KEAP1 mutations (P = 0.001). Proteomic profiling identified increased expression of DNA repair proteins (e.g., Chk1 and the DNA repair score) and suppressed PI3k/mTOR signaling among TTF1-negative tumors, with differences in total proteins confirmed at the mRNA level. Cell line analysis showed drugs targeting DNA repair to be more active in TTF1-low cell lines. Conclusions: Combined genomic and proteomic analyses demonstrated infrequent alteration of validated lung cancer targets (including the absence of BRAF mutations in TTF1-negative LUAD), but identified novel potential targets for TTF1-negative LUAD, including KEAP1/Nrf2 and DNA repair pathways. Clin Cancer Res; 21(15); 3480–91. ©2015 AACR.

Epithelial–Mesenchymal Transition Predicts Polo-Like Kinase 1 Inhibitor–Mediated Apoptosis in Non–Small Cell Lung Cancer

Purpose: To identify new therapeutic targets for non–small cell lung cancer (NSCLC), we systematically searched two cancer cell line databases for sensitivity data on a broad range of drugs. We identified polo-like kinase 1 (PLK1) as the most promising target for further investigation based on a subset of sensitive NSCLC cell lines and inhibitors that were in advanced clinical development. Experimental Design: To identify potential biomarkers of response of NSCLC to PLK1 inhibition and mechanisms of PLK1 inhibitor–induced apoptosis, integrated analysis of gene and protein expression, gene mutations, and drug sensitivity was performed using three PLK1 inhibitors (volasertib, BI2536, and GSK461364) with a large panel of NSCLC cell lines. Results: The NSCLC cell lines had different sensitivities to PLK1 inhibition, with a minority demonstrating sensitivity to all three inhibitors. PLK1 inhibition led to G2–M arrest, but only treatment-sensitive cell lines underwent substantial apoptosis following PLK1 inhibition. NSCLC lines with high epithelial–mesenchymal transition (EMT) gene signature scores (mesenchymal cell lines) were more sensitive to PLK1 inhibition than epithelial lines (P < 0.02). Likewise, proteomic profiling demonstrated that E-cadherin expression was higher in the resistant cell lines than in the sensitive ones (P < 0.01). Induction of an epithelial phenotype by expression of the miRNA miR-200 increased cellular resistance to PLK1 inhibition. Also, KRAS mutation and alterations in the tight-junction, ErbB, and Rho signaling pathways correlated with drug response of NSCLC. Conclusions: In this first reported large-scale integrated analysis of PLK1 inhibitor sensitivity, we demonstrated that EMT leads to PLK1 inhibition sensitivity of NSCLC cells. Our findings have important clinical implications for mesenchymal NSCLC, a significant subtype of the disease that is associated with resistance to currently approved targeted therapies. Clin Cancer Res; 22(7); 1674–86. ©2015 AACR.