Back to 2026 Abstracts
Spatially Defined SRD5A2-Independent Stromal Niches Predict 5ARI Response in Benign Prostatic Hyperplasia
Xingbo Long, M.D., Zahra Ansari, Ph.D., Aaron Fleishman, M.D., Ruslan Korets, M.D., Boris Gershman, M.D., Heidi J Rayala, M.D., Yue Sun, M.D., Aria F Olumi, M.D..
Beth Israel Deaconess Medical Center, Boston, MA, USA.
BACKGROUND:SRD5A2 is the therapeutic target of 5α-reductase inhibitors (5ARIs) for benign prostatic hyperplasia (BPH). However, SRD5A2 (Steroid 5 alpha Reductase 2; aka: 5-alpha reductase 2) expression varies substantially across patients, with approximately 30% exhibiting low or absent expression, leading to variable clinical responses to 5ARI therapy. The spatial organization underlying this heterogeneity and its impact on treatment response remain poorly understood, limiting effective patient stratification.
METHODS:We performed Xenium spatial transcriptomic profiling on 142 prostatic tissue regions from 63 patients with BPH to characterize spatial heterogeneity of SRD5A2 expression. Spatial proximity analysis was used to define functional niches, which were annotated as SRD5A2-dependent or -independent based on their association with SRD5A2 expression. Niche-specific gene signatures were applied to RNA-seq data from 108 patients in the MTOPS cohort to determine whether distinct niches predict response to 5ARI therapy.
RESULTS:SRD5A2 expression exhibited marked inter- and intra-patient heterogeneity, resulting in distinct stromal compositions and spatially organized niches. Spatially resolved transcriptomic profiling identified two SRD5A2-high fibroblast subtypes (Fib1/2) and one SRD5A2-low subtype (Fib3), which formed distinct spatial stromal niches. SRD5A2-high regions were enriched for Fib1/2-associated niches with activation of hormone and mTOR signaling, whereas SRD5A2-low regions were dominated by Fib3-associated niches enriched in metabolic and EGFR-KRAS signaling and transcriptionally independent of SRD5A2. Application of niche-specific signatures to the MTOPS cohort stratified patients into SRD5A2-dependent and -independent groups. While no differences were observed in untreated patients, the SRD5A2-independent group demonstrated significantly less improvement in AUASS with 5ARI therapy, indicating reduced therapeutic response and identifying patients unlikely to benefit from 5ARI therapy.
CONCLUSIONS:SRD5A2-high and -low regions define distinct stromal architectures. An SRD5A2-independent, EGFR-KRAS-driven Fib3-dominant niche underlies resistance to 5ARI therapy, identifies patients unlikely to benefit from treatment, and highlights EGFR-KRAS signaling as a therapeutic target in BPH.
Back to 2026 Abstracts
