Prognostic Biomarker-Based Identification of Drugs for Managing the Treatment of Endometrial Cancer

Dilraj Kaur,Chakit Arora,Gajendra Pal Singh Raghava

DOI: https://doi.org/10.1007/s40291-021-00539-1

Abstract:Introduction: Uterine corpus endometrial carcinoma (UCEC) causes thousands of deaths per year. To improve the overall survival of patients with UCEC, there is a need to identify prognostic biomarkers and potential drugs. Objectives: The aim of this study was twofold: the identification of prognostic gene signatures from expression profiles of pattern recognition receptor (PRR) genes and identification of the most effective existing drugs using the prognostic gene signature. Methods: This study was based on the expression profile of PRR genes of 541 patients with UCEC obtained from The Cancer Genome Atlas. Key prognostic signatures were identified using various approaches, including survival analysis, network, and clustering. Hub genes were identified by constructing a co-expression network. Representative genes were identified using k-means and k-medoids-based clustering. Univariate Cox proportional hazard (PH) analysis was used to identify survival-associated genes. 'cmap2' was used to identify potential drugs that can suppress/enhance the expression of prognostic genes. Results: Models were developed using hub genes and achieved a maximum hazard ratio (HR) of 1.37 (p = 0.294). Then, a clustering-based model was developed using seven genes (HR 9.14; p = 1.49 × 10-12). Finally, a nine gene-based risk stratification model was developed (CLEC1B, CLEC3A, IRF7, CTSB, FCN1, RIPK2, NLRP10, NLRP9, and SARM1) and achieved HR 10.70; p = 1.1 × 10-12. The performance of this model improved significantly in combination with the clinical stage and achieved HR 15.23; p = 2.21 × 10-7. We also developed a model for predicting high-risk patients (survival ≤ 4.3 years) and achieved an area under the receiver operating characteristic curve (AUROC) of 0.86. Conclusion: We identified potential immunotherapeutic agents based on prognostic gene signature: hexamethonium bromide and isoflupredone. Several novel candidate drugs were suggested, including human interferon-α-2b, paclitaxel, imiquimod, MESO-DAP1, and mifamurtide. These biomolecules and repurposed drugs may be utilised for prognosis and treatment for better survival.