AI-Powered Discovery of Natural Compounds for Benign Prostatic Hyperplasia Treatment

Background. Benign prostatic hyperplasia (BPH) is a common condition in aging men that causes urinary symptoms. Current treatments have limitations, necessitating new approaches. Single-cell RNA sequencing (scRNA-seq) provides detailed insights into cellular activity, while Artificial Intelligence (AI) techniques such as large language models (LLMs) can analyze complex queries to identify natural substances that may downregulate overexpressed genes, providing a novel approach for the treatment of BPH. Method. Single-cell RNA sequencing (scRNA-seq) data (GSE226237) were obtained from the National Institutes of Health (NIH) Gene Expression Omnibus (GEO) repository. Gene expression profiles from three large prostates were compared with those from three small prostates to identify differentially expressed genes associated with larger prostate tissue. These genes were evaluated as potential therapeutic targets. Candidate natural compounds that may modulate the activity of overexpressed genes were generated using a structured prompt submitted to the large language model ChatGPT. Results. Single-cell RNA sequencing analysis comparing three large prostates with three small prostates identified distinct gene expression differences associated with prostate enlargement. Several genes involved in immune regulation and cellular structure were downregulated in large BPH, while genes related to epithelial integrity, protein synthesis, and proliferation such as SLC14A1, KRT5, KRT14, KRT15, PRAC1, CSTA, RPL36A, GABARAP, RPS17, and FXYD3 were upregulated. To explore potential therapeutic interventions, these overexpressed genes were analyzed using ChatGPT 5.4 to identify candidate natural compounds capable of modulating their activity. The resulting gene/compound associations include natural products such as resveratrol, curcumin, epigallocatechin gallate (EGCG), genistein, quercetin, sulforaphane, berberine, ashwagandha, and lycopene, highlighting promising directions for natural product based modulation of BPH-associated gene expression. Conclusions. These findings highlight molecular changes in BPH progression and show how single-cell transcriptomics reveals gene expression linked to epithelial proliferation, immune modulation, and metabolism. Integrating AI with omics data offers a framework for identifying potential natural therapeutics, though these AI-generated recommendations remain hypothesis-generating rather than definitive.