Boosting Clear Cell Renal Carcinoma-Specific Drug Discovery Using a Deep Learning Algorithm and Single-Cell Analysis

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Clear cell renal carcinoma (ccRCC) is the most common kidney cancer subtype, characterized by a complex and heterogeneous tumor microenvironment (TME).
• Current treatments like targeted therapy and immunotherapy show limited efficacy against ccRCC due to TME complexity.
• Understanding the ccRCC TME is crucial for developing effective therapeutic strategies.

Purpose of the Study:

• To characterize the TME of ccRCC using single-cell transcriptome sequencing (scRNA-seq).
• To identify key transcription factors (TFs) regulating ccRCC tumor cells.
• To discover novel anti-ccRCC compounds through virtual drug screening and machine learning algorithms.

Main Methods:

• Analysis of scRNA-seq data from six ccRCC patients to profile TME components (T cells, TAMs, ECs, CAFs).
• Identification of tumor cell-specific regulatory programs and key TFs, including EPAS1/HIF-2α, via differential typing and virtual screening.
• Application of a deep graph neural network and machine learning approach to screen bioactive compound libraries for anti-ccRCC agents.

Main Results:

• Detailed characterization of the ccRCC TME, revealing specific regulatory programs.
• Identification of three key TFs driving tumor cell-specific programs and EPAS1/HIF-2α via virtual screening.
• Discovery of five potential anti-ccRCC compounds, including flufenamic acid, fludarabine, an endogenous metabolite, an immunology/inflammation compound, and a DNA methyltransferase inhibitor (N4-methylcytidine).

Conclusions:

• The study provides a comprehensive analysis of the ccRCC TME at the single-cell level.
• Key TFs and specific ccRCC-associated compounds have been identified.
• These findings offer promising directions for the clinical treatment of ccRCC patients.