Conzen, Suzanne D.

Abstract Background: Epithelial ovarian cancer (EOC), commonly known as ovarian cancer (OvCa), is the deadliest gynecological malignancy. High tumor expression of either NR3C1 mRNA (encoding the glucocorticoid receptor or GR) or GR protein (IHC) are associated with poor overall survival (OS) in patients (pts) with advanced stage OvCa. We hypothesized that high GR expression and associated GR transcriptional activity favor OvCa cell survival and chemotherapy resistance, thereby reducing OS. Here we investigated whether a GR transcriptome-based gene signature (GR-sig) would be a better prognosticator than tumor NR3C1 mRNA expression alone. Methods: mRNAs up- or downregulated >1.5 fold after glucocorticoid treatment (Dex) in one or both cultured OvCa cell lines (OVSAHO and HEYA8) were defined as differentially expressed genes (DEGs). DEGs were then identified in n=1247 unique OvCa microarrays from 15 publicly available Gene Expression Omnibus (GEO) datasets with matched pt OS (training set). We then used the OS data to train a Cox proportional hazards model. The “best subset selection” (BeSS) algorithm and bootstrap sampling identified those GR-regulated DEGs whose tumor mRNA expression correlated most significantly with pt OS, creating a GR-sig “risk score.” Subsequently, all n=522 microarray and n=374 RNA-seq publicly available samples from the OvCa TCGA were used as two independent validation sets. Microarray and RNA-seq data were normalized using the MAS5 and FPKM-UQ algorithms respectively. The two validation sets were independently stratified relative to the median training set risk score into low- and high-risk score groups. To examine the prognostic ability of NR3C1 mRNA expression alone, the median training set NR3C1 expression value was used to stratify validation set pts into low- and high-expression groups. In all cases, OS was estimated by the Kaplan-Meier method. Results: We identified n=2288 protein-coding DEGs in one or both cell lines. Of these genes, n=1251 were expressed in each GEO microarray sample (training set). A 10-gene GR-signature (GR-Sig) was then identified using the BeSS algorithm and bootstrap sampling, yielding a median training set risk score of 1.02. In both the microarray and RNA-seq validation datasets, the 10-gene GR-sig risk score was associated with significant OS differences between low- and high-risk score groups (microarray p = 0.018; RNA-seq p = 0.00049). In contrast, the OS difference between low- and high- NR3C1 mRNA expression groups did not reach significance (microarray p = 0.12; RNA-seq p = 0.28). Conclusion: In this retrospective analysis, a 10-gene GR-sig derived from OvCa GR-regulated DEGs performed better as a prognosticator than NR3C1 mRNA expression alone. Prospective application of the GR-sig may be useful in stratifying patients for treatment response with novel agents in OvCa clinical trials. Citation Format: Ananth V. Panchamukhi, Lei Huang, Eleanore P. Gray, Yuanhang Liu, Manisha Taya, Jennifer T. Veneris, Xiaonan Hou, S John Weroha, Balázs Győrffy, Gini F. Fleming, Chen Wang, Suzanne D. Conzen. Development of a prognostic ovarian cancer gene signature based on glucocorticoid receptor transcriptional activity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4739.

e17039 Background: Treatment strategies are needed for platinum-resistant recurrent ovarian cancer. Glucocorticoid receptor (GR) is associated with poor prognosis in ovarian cancer. GR signaling inhibits chemotherapy-induced cell death via transcriptional regulation of anti-apoptotic proteins. Non-selective (mifepristone) and selective GR antagonists (Relacorilant, Corcept Therapeutics) are under clinical development (eg. NCT02046421, NCT03776812). We hypothesized that GR antagonism would improve chemotherapy sensitivity in ovarian cancer patient-derived xenograft (PDX) models. Methods: Five PDX models were selected based on GR expression above median RNA seq levels. Tissues were transplanted into 300 female SCID-Bg mice at the Mayo Clinic. Once tumor diameter reached 0.2 cm, mice were randomized into treatment arms (2-25 mice per arm) and treated intraperitoneal (i.p): control (9% saline, CTL) on day 0 (D0), 1, 7, 8, 14, 15, 21, 22; mifepristone (MIF), 15 mg/kg D0, 1, 7, 8, 14, 15, 21, 22; relacorilant (134), 20 mg/kg D0, 1, 7, 8, 14, 15, 21, 22; carboplatin (CBDCA) at the MTD, 51 mg/kg D1, 8, 15, 22; CBDCA + MIF; CBDCA + 134. Tumor area was measured by ultrasound weekly (D0-D28), and the ratio calculated relative to D0. Tumor growth curves were compared between arms for each PDX via a linear mixed effects regression test of coincident curves. Time to progression was also evaluated and will be presented. Results: In 1/5 models (PH160), CBDCA + MIF significantly suppressed tumor growth compared to CBDCA alone (p < 0.05, Table 1). Addition of 134 to CBDCA did not suppress tumor growth compared to CBDCA alone in any model. In the 2 models with no difference of CBDCA versus CTL (PH160, PH454), addition of a GR antagonist to CBDCA suppressed tumor growth versus CTL. Conclusions: Addition of GR antagonists can improve response to chemotherapy. Work is ongoing to identify the GR transcriptional network, tumor, and host characteristics which predict responses to GR antagonists. [Table: see text]

e17038 Background: Carcinosarcomas (CS) are aggressive gynecologic malignancies defined by a biphasic tumor histology of carcinomatous and sarcomatous components. High expression of glucocorticoid receptor (GR) is associated with aggressive tumor biology in triple-negative breast, ovarian, and endometrial cancers. GR activation results in transcriptional regulation of epithelial-mesenchymal transition (EMT) genes. Therefore, we hypothesized that the sarcomatous component of Müllerian CS may be associated with relatively higher GR expression compared to the carcinomatous portion. Methods: Müllerian CSs from the University of Chicago tissue bank diagnosed from 5/19/2009 – 8/25/2014 were identified and pathology confirmed (RL). In addition, we evaluated GR expression in a mouse PDX CS model. Immunohistochemical expression of nuclear estrogen (ER), progesterone (PR), androgen (AR), and GR was quantified using the H-score. Scores between the sarcomatous and carcinomatous components were compared using a paired t-test. Results: Mean GR H-score was relatively higher in the sarcomatous vs carcinomatous component, overall (144.44 vs 38.89, P = 0.0014). Mean H-score also differed between sarcomatous and carcinomatous components for ER (3.06 vs 63.06, P = 0.016), PR (1.67 vs 47.22, P = 0.004), and AR (0 vs 2.78, P = 0.056). The significantly higher sarcomatous GR expression was replicated in IHC evaluation of the mouse PDX model. Conclusions: GR expression is significantly higher in the sarcomatous than the carcinomatous component of Müllerian CS, while ER, PR, and AR expression is significantly lower. Recent data suggest that an EMT gene signature is more prominently expressed in the sarcomatous versus carcinomatous component. We hypothesize that GR expression and activity may mediate EMT gene expression and represent a therapeutic target for CS. [Table: see text]

Abstract Purpose: Although high glucocorticoid receptor (GR) expression in early-stage estrogen receptor (ER)-negative breast cancer is associated with shortened relapse-free survival (RFS), how associated GR transcriptional activity contributes to aggressive breast cancer behavior is not well understood. Using potent GR antagonists and primary tumor gene expression data, we sought to identify a tumor-relevant gene signature based on GR activity that would be more predictive than GR expression alone. Experimental Design: Global gene expression and GR ChIP-sequencing were performed to identify GR-regulated genes inhibited by two chemically distinct GR antagonists, mifepristone and CORT108297. Differentially expressed genes from MDA-MB-231 cells were cross-evaluated with significantly expressed genes in GR-high versus GR-low ER-negative primary breast cancers. The resulting subset of GR-targeted genes was analyzed in two independent ER-negative breast cancer cohorts to derive and then validate the GR activity signature (GRsig). Results: Gene expression pathway analysis of glucocorticoid-regulated genes (inhibited by GR antagonism) revealed cell survival and invasion functions. GR ChIP-seq analysis demonstrated that GR antagonists decreased GR chromatin association for a subset of genes. A GRsig that comprised n = 74 GR activation-associated genes (also reversed by GR antagonists) was derived from an adjuvant chemotherapy-treated Discovery cohort and found to predict probability of relapse in a separate Validation cohort (HR = 1.9; P = 0.012). Conclusions: The GRsig discovered herein identifies high-risk ER-negative/GR-positive breast cancers most likely to relapse despite administration of adjuvant chemotherapy. Because GR antagonism can reverse expression of these genes, we propose that addition of a GR antagonist to chemotherapy may improve outcome for these high-risk patients. Clin Cancer Res; 24(14); 3433–46. ©2018 AACR.

BACKGROUNDThe objective of this study was to demonstrate that computer‐extracted image phenotypes (CEIPs) of biopsy‐proven breast cancer on magnetic resonance imaging (MRI) can accurately predict pathologic stage.METHODSThe authors used a data set of deidentified breast MRIs organized by the National Cancer Institute in The Cancer Imaging Archive. In total, 91 biopsy‐proven breast cancers were analyzed from patients who had information available on pathologic stage (stage I, n = 22; stage II, n = 58; stage III, n = 11) and surgically verified lymph node status (negative lymph nodes, n = 46; ≥ 1 positive lymph node, n = 44; no lymph nodes examined, n = 1). Tumors were characterized according to 1) radiologist‐measured size and 2) CEIP. Then, models were built that combined 2 CEIPs to predict tumor pathologic stage and lymph node involvement, and the models were evaluated in a leave‐1‐out, cross‐validation analysis with the area under the receiver operating characteristic curve (AUC) as the value of interest.RESULTSTumor size was the most powerful predictor of pathologic stage, but CEIPs that captured biologic behavior also emerged as predictive (eg, stage I and II vs stage III demonstrated an AUC of 0.83). No size measure was successful in the prediction of positive lymph nodes, but adding a CEIP that described tumor “homogeneity” significantly improved discrimination (AUC = 0.62; P = .003) compared with chance.CONCLUSIONSThe current results indicate that MRI phenotypes have promise for predicting breast cancer pathologic stage and lymph node status. Cancer 2016;122:748–757. © 2015 American Cancer Society.