Global expression analysis of endometrial cancer cells in response to progesterone identifies new therapeutic targets.

Progesterone prevents development of endometrial cancers through its receptor (PR) although the molecular mechanisms have yet to be fully characterized. In this study, we performed a global analysis of gene regulation by progesterone using human endometrial cancer cells that expressed PR endogenously or exogenously. We found progesterone strongly inhibits multiple components of the platelet derived growth factor receptor (PDGFR), Janus kinase (JAK), signal transducer and activator of transcription (STAT) pathway through PR. The PDGFR/JAK/STAT pathway signals to control numerous downstream targets including AP-1 transcription factors Fos and Jun. Treatment with inhibitors of the PDGFR/JAK/STAT pathway significantly blocked proliferation in multiple novel patient-derived organoid models of endometrial cancer, and activation of this pathway was found to be a poor prognostic signal for the survival of patients with endometrial cancer from The Cancer Genome Atlas. Our study identifies this pathway as central to the growth-limiting effects of progesterone in endometrial cancer and suggests that inhibitors of PDGFR/JAK/STAT should be considered for future therapeutic interventions.

Blocking autophagy overcomes resistance to dual histone deacetylase and proteasome inhibition in gynecologic cancer.

Histone deacetylase (HDAC) inhibitors and proteasome inhibitors have been approved by the FDA for the treatment of multiple myeloma and lymphoma, respectively, but have not achieved similar activity as single agents in solid tumors. Preclinical studies have demonstrated the activity of the combination of an HDAC inhibitor and a proteasome inhibitor in a variety of tumor models. However, the mechanisms underlying sensitivity and resistance to this combination are not well-understood. This study explores the role of autophagy in adaptive resistance to dual HDAC and proteasome inhibition. Studies focus on ovarian and endometrial gynecologic cancers, two diseases with high mortality and a need for novel treatment approaches. We found that nanomolar concentrations of the proteasome inhibitor ixazomib and HDAC inhibitor romidepsin synergistically induce cell death in the majority of gynecologic cancer cells and patient-derived organoid (PDO) models created using endometrial and ovarian patient tumor tissue. However, some models were not sensitive to this combination, and mechanistic studies implicated autophagy as the main mediator of cell survival in the context of dual HDAC and proteasome inhibition. Whereas the combination of ixazomib and romidepsin reduces autophagy in sensitive gynecologic cancer models, autophagy is induced following drug treatment of resistant cells. Pharmacologic or genetic inhibition of autophagy in resistant cells reverses drug resistance as evidenced by an enhanced anti-tumor response both in vitro and in vivo. Taken together, our findings demonstrate a role for autophagic-mediated cell survival in proteasome inhibitor and HDAC inhibitor-resistant gynecologic cancer cells. These data reveal a new approach to overcome drug resistance by inhibiting the autophagy pathway.

Advantages of Tyrosine Kinase Anti-Angiogenic Cediranib over Bevacizumab: Cell Cycle Abrogation and Synergy with Chemotherapy.

Angiogenesis plays a crucial role in tumor development and metastasis. Both bevacizumab and cediranib have demonstrated activity as single anti-angiogenic agents in endometrial cancer, though subsequent studies of bevacizumab combined with chemotherapy failed to improve outcomes compared to chemotherapy alone. Our objective was to compare the efficacy of cediranib and bevacizumab in endometrial cancer models. The cellular effects of bevacizumab and cediranib were examined in endometrial cancer cell lines using extracellular signal-related kinase (ERK) phosphorylation, ligand shedding, cell viability, and cell cycle progression as readouts. Cellular viability was also tested in eight patient-derived organoid models of endometrial cancer. Finally, we performed a phosphoproteomic array of 875 phosphoproteins to define the signaling changes related to bevacizumab versus cediranib. Cediranib but not bevacizumab blocked ligand-mediated ERK activation in endometrial cancer cells. In both cell lines and patient-derived organoids, neither bevacizumab nor cediranib alone had a notable effect on cell viability. Cediranib but not bevacizumab promoted marked cell death when combined with chemotherapy. Cell cycle analysis demonstrated an accumulation in mitosis after treatment with cediranib + chemotherapy, consistent with the abrogation of the G2/M checkpoint and subsequent mitotic catastrophe. Molecular analysis of key controllers of the G2/M cell cycle checkpoint confirmed its abrogation. Phosphoproteomic analysis revealed that bevacizumab and cediranib had both similar and unique effects on cell signaling that underlie their shared versus individual actions as anti-angiogenic agents. An anti-angiogenic tyrosine kinase inhibitor such as cediranib has the potential to be superior to bevacizumab in combination with chemotherapy.

Successful Patient-Derived Organoid Culture of Gynecologic Cancers for Disease Modeling and Drug Sensitivity Testing.

Developing reliable experimental models that can predict clinical response before treating the patient is a high priority in gynecologic cancer research, especially in advanced or recurrent endometrial and ovarian cancers. Patient-derived organoids (PDOs) represent such an opportunity. Herein, we describe our successful creation of 43 tumor organoid cultures and nine adjacent normal tissue organoid cultures derived from patients with endometrial or ovarian cancer. From an initial set of 45 tumor tissues and seven ascites fluid samples harvested at surgery, 83% grew as organoids. Drug sensitivity testing and organoid cell viability assays were performed in 19 PDOs, a process that was accomplished within seven days of obtaining the initial surgical tumor sample. Sufficient numbers of cells were obtained to facilitate testing of the most commonly used agents for ovarian and endometrial cancer. The models reflected a range of sensitivity to platinum-containing chemotherapy as well as other relevant agents. One PDO from a patient treated prior to surgery with neoadjuvant trastuzumab successfully predicted the patient's postoperative chemotherapy and trastuzumab resistance. In addition, the PDO drug sensitivity assay identified alternative treatment options that are currently used in the second-line setting. Our findings suggest that PDOs could be used as a preclinical platform for personalized cancer therapy for gynecologic cancer patients.

Creation and validation of models to predict response to primary treatment in serous ovarian cancer.

Nearly a third of patients with high-grade serous ovarian cancer (HGSC) do not respond to initial therapy and have an overall poor prognosis. However, there are no validated tools that accurately predict which patients will not respond. Our objective is to create and validate accurate models of prediction for treatment response in HGSC. This is a retrospective case-control study that integrates comprehensive clinical and genomic data from 88 patients with HGSC from a single institution. Responders were those patients with a progression-free survival of at least 6 months after treatment. Only patients with complete clinical information and frozen specimen at surgery were included. Gene, miRNA, exon, and long non-coding RNA (lncRNA) expression, gene copy number, genomic variation, and fusion-gene determination were extracted from RNA-sequencing data. DNA methylation analysis was performed. Initial selection of informative variables was performed with univariate ANOVA with cross-validation. Significant variables (p < 0.05) were included in multivariate lasso regression prediction models. Initial models included only one variable. Variables were then combined to create complex models. Model performance was measured with area under the curve (AUC). Validation of all models was performed using TCGA HGSC database. By integrating clinical and genomic variables, we achieved prediction performances of over 95% in AUC. Most performances in the validation set did not differ from the training set. Models with DNA methylation or lncRNA underperformed in the validation set. Integrating comprehensive clinical and genomic data from patients with HGSC results in accurate and robust prediction models of treatment response.

Characterization of a TP53 Somatic Variant of Unknown Function From an Ovarian Cancer Patient Using Organoid Culture and Computational Modeling.

In our proof-of-concept study of 1 patient with stage IIIC carcinosarcoma of the ovary, we discovered a rare mutation in the tumor suppressor, TP53, that results in the deletion of N131. Immunofluorescence imaging of the organoid culture revealed hyperstaining of p53 protein. Computational modeling suggests this residue is important for maintaining protein conformation. Drug screening identified the combination of a proteasome inhibitor with a histone deacetylase inhibitor as the most effective treatment. These data provide evidence for the successful culture of a patient tumor and analysis of drug response ex vivo.

Prediction of Epithelial Ovarian Cancer Outcomes With Integration of Genomic Data.

Some of the patients with epithelial ovarian cancer will not respond to initial therapy. These patients have a poor prognosis. Our aim was to identify patients with a worse prognosis by integrating clinical, pathologic, and genomic data. Using publicly available genomic data and integrating it with clinical data, we significantly improved the prediction of patients with worse surgical outcomes and those who do not respond to initial chemotherapy. We further improved these models with more precise data collection and better understanding of the genetic background of the studied population. Better prediction will lead to better patient classification and opportunities for individualized treatment.

Differential DNA methylation in high-grade serous ovarian cancer (HGSOC) is associated with tumor behavior.

The epigenome offers an additional facet of cancer that can help categorize patients into those at risk of disease, recurrence, or treatment failure. We conducted a retrospective, nested, case-control study of advanced and recurrent high-grade serous ovarian cancer (HGSOC) patients in which we assessed epigenome-wide association using Illumina methylationEPIC arrays to characterize DNA methylation status and RNAseq to evaluate gene expression. Comparing HGSOC tumors with normal fallopian tube tissues we observe global hypomethylation but with skewing towards hypermethylation when interrogating gene promoters. In total, 5,852 gene interrogating probes revealed significantly different methylation. Within HGSOC, 57 probes highlighting 17 genes displayed significant differential DNA methylation between primary and recurrent disease. Between optimal vs suboptimal surgical outcomes 99 probes displayed significantly different methylation but only 29 genes showed an inverse correlation between methylation status and gene expression. Overall, differentially methylated genes point to several pathways including RAS as well as hippo signaling in normal vs primary HGSOC; valine, leucine, and isoleucine degradation and endocytosis in primary vs recurrent HGSOC; and pathways containing immune driver genes in optimal vs suboptimal surgical outcomes. Thus, differential DNA methylation identified numerous genes that could serve as potential biomarkers and/or therapeutic targets in HGSOC.

An integrated prediction model of recurrence in endometrial endometrioid cancers.

Objectives: Endometrial cancer incidence and mortality are rising in the US. Disease recurrence has been shown to have a significant impact on mortality. However, to date, there are no accurate and validated prediction models that would discriminate which individual patients are likely to recur. Reliably predicting recurrence would be of benefit for treatment decisions following surgery. We present an integrated model constructed with comprehensive clinical, pathological and molecular features designed to discriminate risk of recurrence for patients with endometrioid endometrial adenocarcinoma. Subjects and methods: A cohort of endometrioid endometrial cancer patients treated at our institution was assembled. Clinical characteristics were extracted from patient charts. Primary tumors from these patients were obtained and total tissue RNA extracted for RNA sequencing. A prediction model was designed containing both clinical characteristics and molecular profiling of the tumors. The same analysis was carried out with data derived from The Cancer Genome Atlas for replication and external validation. Results: Prediction models derived from our institutional data predicted recurrence with high accuracy as evidenced by areas under the curve approaching 1. Similar trends were observed in the analysis of TCGA data. Further, a scoring system for risk of recurrence was devised that showed specificities as high as 81% and negative predictive value as high as 90%. Lastly, we identify specific molecular characteristics of patient tumors that may contribute to the process of disease recurrence. Conclusion: By constructing a comprehensive model, we are able to reliably predict recurrence in endometrioid endometrial cancer. We devised a clinically useful scoring system and thresholds to discriminate risk of recurrence. Finally, the data presented here open a window to understanding the mechanisms of recurrence in endometrial cancer.

Molecular Characterization of Non-responders to Chemotherapy in Serous Ovarian Cancer.

Nearly one-third of patients with high-grade serous ovarian cancer (HGSC) do not respond to initial treatment with platinum-based therapy. Genomic and clinical characterization of these patients may lead to potential alternative therapies. Here, the objective is to classify non-responders into subsets using clinical and molecular features. Using patients from The Cancer Genome Atlas (TCGA) dataset with platinum-resistant or platinum-refractory HGSC, we performed a genome-wide unsupervised cluster analysis that integrated clinical data, gene copy number variations, gene somatic mutations, and DNA promoter methylation. Pathway enrichment analysis was performed for each cluster to identify the targetable processes. Following the unsupervised cluster analysis, three distinct clusters of non-responders emerged. Cluster 1 had overrepresentation of the stage IV disease and suboptimal debulking, under-expression of miRNAs and mRNAs, hypomethylated DNA, "loss of function" TP53 mutations, and the overexpression of genes in the PDGFR pathway. Cluster 2 had low miRNA expression, generalized hypermethylation, MUC17 mutations, and significant activation of the HIF-1 signaling pathway. Cluster 3 had more optimally cytoreduced stage III patients, overexpression of miRNAs, mixed methylation patterns, and "gain of function" TP53 mutations. However, the survival for all clusters was similar. Integration of genomic and clinical data from patients that do not respond to chemotherapy has identified different subgroups or clusters. Pathway analysis further identified the potential alternative therapeutic targets for each cluster.

Population Substructure Has Implications in Validating Next-Generation Cancer Genomics Studies with TCGA.

In the era of large genetic and genomic datasets, it has become crucially important to validate results of individual studies using data from publicly available sources, such as The Cancer Genome Atlas (TCGA). However, how generalizable are results from either an independent or a large public dataset to the remainder of the population? The study presented here aims to answer that question. Utilizing next generation sequencing data from endometrial and ovarian cancer patients from both the University of Iowa and TCGA, genomic admixture of each population was analyzed using STRUCTURE and ADMIXTURE software. In our independent data set, one subpopulation was identified, whereas in TCGA 4⁻6 subpopulations were identified. Data presented here demonstrate how different the genetic substructures of the TCGA and University of Iowa populations are. Validation of genomic studies between two different population samples must be aware of, account for and be corrected for background genetic substructure.

A Prediction Model for Preoperative Risk Assessment in Endometrial Cancer Utilizing Clinical and Molecular Variables.

The utility of comprehensive surgical staging in patients with low risk disease has been questioned. Thus, a reliable means of determining risk would be quite useful. The aim of our study was to create the best performing prediction model to classify endometrioid endometrial cancer (EEC) patients into low or high risk using a combination of molecular and clinical-pathological variables. We then validated these models with publicly available datasets. Analyses between low and high risk EEC were performed using clinical and pathological data, gene and miRNA expression data, gene copy number variation and somatic mutation data. Variables were selected to be included in the prediction model of risk using cross-validation analysis; prediction models were then constructed using these variables. Model performance was assessed by area under the curve (AUC). Prediction models were validated using appropriate datasets in The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. A prediction model with only clinical variables performed at 88%. Integrating clinical and molecular data improved prediction performance up to 97%. The best prediction models included clinical, miRNA expression and/or somatic mutation data, and stratified pre-operative risk in EEC patients. Integrating molecular and clinical data improved the performance of prediction models to over 95%, resulting in potentially useful clinical tests.

Prediction of chemo-response in serous ovarian cancer.

BACKGROUND: Nearly one-third of serous ovarian cancer (OVCA) patients will not respond to initial treatment with surgery and chemotherapy and die within one year of diagnosis. If patients who are unlikely to respond to current standard therapy can be identified up front, enhanced tumor analyses and treatment regimens could potentially be offered. Using the Cancer Genome Atlas (TCGA) serous OVCA database, we previously identified a robust molecular signature of 422-genes associated with chemo-response. Our objective was to test whether this signature is an accurate and sensitive predictor of chemo-response in serous OVCA. METHODS: We first constructed prediction models to predict chemo-response using our previously described 422-gene signature that was associated with response to treatment in serous OVCA. Performance of all prediction models were measured with area under the curves (AUCs, a measure of the model's accuracy) and their respective confidence intervals (CIs). To optimize the prediction process, we determined which elements of the signature most contributed to chemo-response prediction. All prediction models were replicated and validated using six publicly available independent gene expression datasets. RESULTS: The 422-gene signature prediction models predicted chemo-response with AUCs of ~70 %. Optimization of prediction models identified the 34 most important genes in chemo-response prediction. These 34-gene models had improved performance, with AUCs approaching 80 %. Both 422-gene and 34-gene prediction models were replicated and validated in six independent datasets. CONCLUSIONS: These prediction models serve as the foundation for the future development and implementation of a diagnostic tool to predict response to chemotherapy for serous OVCA patients.

Association analysis of a chemo-response signature identified within The Cancer Genome Atlas aimed at predicting genetic risk for chemo-response in ovarian cancer.

BACKGROUND: A gene signature associated with chemo-response in ovarian cancer was created through integration of biological data in The Cancer Genome Atlas (TCGA) and validated in five independent microarray experiments. Our study aimed to determine if single nucleotide polymorphisms (SNPs) within the 422-gene signature were associated with a genetic predisposition to platinum-based chemotherapy response in serous ovarian cancer. METHODS: An association analysis between SNPs within the 422-gene signature and chemo-response in serous ovarian cancer was performed under the log-additive genetic model using the 'SNPassoc' package within the R environment (p<0.0001). Subsequent validation of statistically significant SNPs was done in the Ovarian Cancer Association Consortium (OCAC) database. RESULTS: 19 SNPs were found to be associated with chemo-response with statistical significance. None of the SNPs found significant in TCGA were validated within OCAC for the outcome of interest, chemo-response. CONCLUSIONS: SNPs associated with chemo-response in ovarian cancer within TGCA database were not validated in a larger database of patients and controls from OCAC. New strategies integrating somatic and germline information may help to characterize genetic predictors for treatment response in ovarian cancer.