Ji Yeon Yang

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.

TCPA: a resource for cancer functional proteomics data

To the Editor: Functional proteomics represents a powerful approach to understand the pathophysiology and therapy of cancer. However, comprehensive cancer proteomic data have been relatively limited. As a part of The Cancer Genome Atlas (TCGA) Project and other efforts, we have generated protein expression data over a large number of tumor and cell line samples using reverse-phase protein arrays (RPPAs). RPPA is a quantitative, antibody-based technology that can assess multiple protein markers in many samples in a cost-effective, sensitive and highthroughput manner1,2. This technology has been extensively validated for both cell line and patient samples3–5, and its applications range from building reproducible prognostic models6 to generating experimentally verified mechanistic insights7. Our RPPA profiling platform includes extensively validated antibodies to nearly 200 proteins and phosphoproteins (Supplementary Methods and Supplementary Table 1). We are in the process of extending it to 500 independent proteins, covering all major signaling pathways, including PI3K, MAPK, mTOR, TGF-b, WNT, cell cycle, apoptosis, DNA damage, Hippo and Notch pathways. The current data release covers 4,379 tumor samples and consists of three parts (Supplementary Table 2). These are (i) TCGA tumor tissue sample sets: 3,467 samples from 11 cancer types, to be extended to 25 cancer types; (ii) independent tumor tissue sample sets: one endometrial tumor set (244 samples)7 and two ovarian tumor sets (99 and 130 samples, respectively)6, with other independent sets to be added soon; and (iii) tumor cell lines: 439 samples in four cell line sets, including both baseline and drug-treated cell lines. To our knowledge, this represents the largest publicly available collection of cancer functional proteomics data with parallel DNA and RNA data. To facilitate broad access to these RPPA data sets, we developed a user-friendly data portal, The Cancer Proteome Atlas (TCPA; http://bioinformatics.mdanderson.org/main/ TCPA:Overview). TCPA provides six modules: Summary, My Protein, Download, Visualization, Analysis and Cell Line (Fig. 1, i). The Summary module provides an overview of the RPPA data with detailed descriptions of each set (Fig. 1, ii). The Download module allows users to obtain any RPPA data set for analysis through a tree-view interface (Fig. 1, iii). The My Protein module provides detailed information about each RPPA protein: protein name, corresponding gene symbol, antibody status and source for the antibody. Users can examine the expression pattern of a protein of interest across different tumor types (for example, HER2 expression shown in Fig. 1, iv). The Visualization module provides two ways to examine global protein expression patterns in a specific RPPA data set . One is through a “next-generation clustered heat map” (Fig. 1, v), which allows users to zoom, navigate and scrutinize clustering patterns of samples or proteins and link those patterns to relevant biological information sources. The other is through a network view (Fig. 1, vi), which overlays the correlation between any two interacting partners in the protein interaction network (curated in the Human Protein Reference Database8). The Analysis module provides three analysis methods. (i) For correlation analysis, given a user-specified data set, correlations between any pair of proteins are presented in a table (Fig. 1, vii). Users can search the results by protein name, rank correlations or visualize the scatter plot of a correlation of interest (for example, there is a strong correlation between PKC-a and its phosphorylated form PKC-a_pS657 in endometrial cancer, as shown in Fig. 1, vii). (ii) For differential analysis, differentially expressed protein markers between two tumor types or subtypes can be identified. Given user-defined comparison groups, the Krzywinski and Cairo reply: We are in full agreement with the core of Katz’s argument that “distortion,” “embellishment,” “concealment” and “unrepresentative displays” have no place in principled communication of scientific information1. There is no controversy here—Katz extrapolates our storytelling metaphor beyond the intended scope of our column and argues against a position we did not take. The Points of View series offers effective strategies for visual presentation of complex data. The scope of the Storytelling column2 was limited to the construction of multipanel figures, which summarize as much as they support detailed exposition of the text. The column did not address how this text should be composed or the broad subject of motivation and design of scientific experiments. We described an approach to structure the flow of concepts and data across panels in a figure as a way to achieve a narrative, not confabulation. The design of visual communication requires a distinct approach because we organize and interpret images very differently than words (Gestalt principles of perception3). Whereas text is a natural place for nuance and alternative interpretations, multiple lines of argument in a figure can easily interfere with our perception of all its parts. Our suggestion to “leave out detail that does not advance the plot” speaks to controlling the amount of information to avoid an incomprehensible image and deferring it to the text, where it can be more suitably framed. To interpret it as “inconvenient truths are [to be] swept away” is a misrepresentation. Readers often look to the abstract and then the figures to provide them with an initial impression and overview of the findings. These are not the only elements that are reported, merely the first elements to be read. At each step, from abstract to figure to text, the level of detail is expanded to accommodate the preparedness of the reader to assimilate new information. It is often impossible to “do justice to experimental complexities and their myriad of interpretations” with a figure. We support Katz’s position that authors should include all the details necessary to appreciate, understand and reproduce the science through the use of visual and written communication that is clear, concise and thoughtful.

Predicting time to ovarian carcinoma recurrence using protein markers

Patients with ovarian cancer are at high risk of tumor recurrence. Prediction of therapy outcome may provide therapeutic avenues to improve patient outcomes. Using reverse-phase protein arrays, we generated ovarian carcinoma protein expression profiles on 412 cases from TCGA and constructed a PRotein-driven index of OVARian cancer (PROVAR). PROVAR significantly discriminated an independent cohort of 226 high-grade serous ovarian carcinomas into groups of high risk and low risk of tumor recurrence as well as short-term and long-term survivors. Comparison with gene expression-based outcome classification models showed a significantly improved capacity of the protein-based PROVAR to predict tumor progression. Identification of protein markers linked to disease recurrence may yield insights into tumor biology. When combined with features known to be associated with outcome, such as BRCA mutation, PROVAR may provide clinically useful predictions of time to tumor recurrence.

Src Inhibition with Saracatinib Reverses Fulvestrant Resistance in ER-Positive Ovarian Cancer Models In Vitro and In Vivo

Purpose: More effective, less toxic treatments for recurrent ovarian cancer are needed. Although more than 60% of ovarian cancers express the estrogen receptor (ER), ER-targeted drugs have been disappointing due to drug resistance. In other estrogen-sensitive cancers, estrogen activates Src to phosphorylate p27 promoting its degradation and increasing cell-cycle progression. Because Src is activated in most ovarian cancers, we investigated whether combined Src and ER blockade by saracatinib and fulvestrant would circumvent antiestrogen resistance. Experimental Design: ER and Src were assayed in 338 primary ovarian cancers. Dual ER and Src blockade effects on cell cycle, ER target gene expression, and survival were assayed in ERα+ ovarian cancer lines, a primary human ovarian cancer culture in vitro, and on xenograft growth. Results: Most primary ovarian cancers express ER. Src activity was greater in ovarian cancer lines than normal epithelial lines. Estrogen activated Src, ER-Src binding, and ER translocation from cytoplasm to nucleus. Estrogen-mediated mitogenesis was via ERα, not ERβ. While each alone had little effect, combined saracatinib and fulvestrant increased p27 and inhibited cyclin E-Cdk2 and cell-cycle progression. Saracatinib also impaired induction of ER-target genes c-Myc and FOSL1; this was greatest with dual therapy. Combined therapy induced autophagy and more effectively inhibited ovarian cancer xenograft growth than monotherapy. Conclusions: Saracatinib augments effects of fulvestrant by opposing estrogen-mediated Src activation and target gene expression, increasing cell-cycle arrest, and impairing survival, all of which would oppose antiestrogen resistance in these ER+ ovarian cancer models. These data support further preclinical and clinical evaluation of combined fulvestrant and saracatinib in ovarian cancer. Clin Cancer Res; 18(21); 5911–23. ©2012 AACR.

YAP/TAZ-Mediated Upregulation of GAB2 Leads to Increased Sensitivity to Growth Factor–Induced Activation of the PI3K Pathway

The transcription regulators YAP and TAZ function as effectors of the HIPPO signaling cascade, critical for organismal development, cell growth, and cellular reprogramming, and YAP/TAZ is commonly misregulated in human cancers. The precise mechanism by which aberrant YAP/TAZ promotes tumor growth remains unclear. The HIPPO tumor suppressor pathway phosphorylates YAP and TAZ, resulting in cytosolic sequestration with subsequent degradation. Here, we report that the PI3K/AKT pathway, which is critically involved in the pathophysiology of endometrial cancer, interacts with the HIPPO pathway at multiple levels. Strikingly, coordinate knockdown of YAP and TAZ, mimicking activation of the HIPPO pathway, markedly decreased both constitutive and growth factor-induced PI3K pathway activation by decreasing levels of the GAB2 linker molecule in endometrial cancer lines. Furthermore, targeting YAP/TAZ decreased endometrial cancer tumor growth in vivo In addition, YAP and TAZ total and phosphoprotein levels correlated with clinical characteristics and outcomes in endometrial cancer. Thus, YAP and TAZ, which are inhibited by the HIPPO tumor suppressor pathway, modify PI3K/AKT pathway signaling in endometrial cancer. The cross-talk between these key pathways identifies potential new biomarkers and therapeutic targets in endometrial cancer. Cancer Res; 77(7); 1637-48. ©2017 AACR.

Integrative Protein-Based Prognostic Model for Early-Stage Endometrioid Endometrial Cancer

Purpose: Endometrioid endometrial carcinoma (EEC) is the major histologic type of endometrial cancer, the most prevalent gynecologic malignancy in the United States. EEC recurrence or metastasis is associated with a poor prognosis. Early-stage EEC is generally curable, but a subset has high risk of recurrence or metastasis. Prognosis estimation for early-stage EEC mainly relies on clinicopathologic characteristics, but is unreliable. We aimed to identify patients with high-risk early-stage EEC who are most likely to benefit from more extensive surgery and adjuvant therapy by building a prognostic model that integrates clinical variables and protein markers. Experimental Design: We used two large, independent early-stage EEC datasets as training (n = 183) and validation cohorts (n = 333), and generated the levels of 186 proteins and phosphoproteins using reverse-phase protein arrays. By applying an initial filtering and the elastic net to the training samples, we developed a prognostic model for overall survival containing two clinical variables and 18 protein markers and optimized the risk group classification. Results: The Kaplan–Meier survival analyses in the validation cohort confirmed an improved discriminating power of our prognostic model for patients with early-stage EEC over key clinical variables (log-rank test, P = 0.565 for disease stage, 0.567 for tumor grade, and 1.3 × 10−4 for the integrative model). Compared with clinical variables (stage, grade, and patient age), only the risk groups defined by the integrative model were consistently significant in both univariate and multivariate analyses across both cohorts. Conclusions: Our prognostic model is potentially of high clinical value for stratifying patients with early-stage EEC and improving their treatment strategies. Clin Cancer Res; 22(2); 513–23. ©2015 AACR.

Corrigendum: A pan-cancer proteomic perspective on The Cancer Genome Atlas

Nature Communications 5: Article number: 3887 (2014); Published 29 May 2014; Updated 28 Jan 2015 This Article contains an error in the Author contributions section that has resulted in incorrect credit for supervision of the network analysis. The correct Author contributions section is as follows: R.A.

Abstract 4262: A pan-cancer proteomic analysis of The Cancer Genome Atlas (TCGA) project

Protein levels and function are predicted poorly by genomic and transcriptomic analysis of patient tumors. Direct proteomic study can provide a wealth of information that complements those analysis in The Cancer Genome Atlas (TCGA) projects. We used 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 phospho-proteins. The resultant proteomic data were used for identifying commonalities, differences, emergent pathway properties, and novel network biology within and across tumor lineages. In general, tumor type and subtype were the dominant determinants of protein levels. However, we were able to identify several potentially targetable markers. E.g. luminal breast cancers demonstrated selective elevation of AR, BCL2, FASN, and pACC. SRC was activated in all but the hormone-responsive and bladder cancers, offering another potential therapeutic opportunity. Similarly, HER3 suggested itself as a potential target in renal cancer. EGFR activity, in general, paralleled SRC activity, but in GBM it was associated with NOTCH1 and HER3 activation, suggesting an opportunity for combination therapy. pSRC was highly expressed in a subset of head and neck tumors, suggesting that those may be more sensitive to EGFR targeting strategies. HER2 levels were elevated in a subset of endometrial, bladder, breast, and colorectal cancers. MYC was selectively amplified and expressed in high-grade serous ovarian cancers. We implemented a computational approach, MC, to decrease the effect of tissue-specific protein expression. MC allowed us to identify processes that drive cell behavior across tumor type and made it possible to find new therapeutic opportunities. We found 7 cross-tumor clusters, each driven by different markers and pathways. We found pan-cancer clusters with elevated HER2 and EGFR, elevated hormone signaling pathways, enriched MAPK and PI3K pathway activity, elevated EMT signatures, and cell cycle signatures. We also saw strong links between MYH11 and Rictor and between ETS1 and pPEA15 across tumor types. Those findings can provide useful clues for developing targeted, cross-tumor therapies. Pathway analysis of the data revealed several expected cross-tumor type associations, including pMEK with pERK, beta-catenin with E-cadherin and pPKCdelta with pPKCalpha and pPKCbeta. The findings support the ability of RPPA analysis to yield high-quality information from TCGA samples. A number of other links such as MYH11 with Rictor, cyclinB1 with FOXM1, and pACC with FASN were not expected and warrant further exploration, as does a negative link between p85 and claudin7 in lung squamous. Analysis of key nodes (e.g., CDK1) revealed other unexpected links to a wide range of protein pathways. Overall, those findings demonstrate the power of pan-cancer proteomic analysis, identifying several novel single-tumor and cross-tumor targets and pathways. Citation Format: Rehan Akbani, Kwok-Shing Ng, Henrica M. Werner, Fan Zhang, Zhenlin Ju, Wenbin Liu, Ji-Yeon Yang, Yiling Lu, John N. Weinstein, Gordon B. Mills. A pan-cancer proteomic analysis of The Cancer Genome Atlas (TCGA) project. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 4262. doi:10.1158/1538-7445.AM2014-4262