Neuroendocrine prostate cancer (NEPC), a lethal subset of prostate cancer, is characterized by loss of AR signaling and resulting resistance to AR-targeted therapy during neuroendocrine transdifferentiation, for which the molecular mechanisms remain unclear. Here, we report that neuropilin 2 (NRP2) is upregulated in both de novo and therapy-induced NEPC, which induces neuroendocrine markers, neuroendocrine cell morphology, and NEPC cell aggressive behavior. NRP2 silencing restricted NEPC tumor xenograft growth. Mechanistically, NRP2 engages in reciprocal crosstalk with AR, where NRP2 is transcriptionally inhibited by AR, and in turn suppresses AR signaling by downregulating the AR transcriptional program and confers resistance to enzalutamide. Moreover, NRP2 physically interacts with VEGFR2 through the intracellular SEA domain to activate STAT3 phosphorylation and subsequently SOX2, thus driving NEPC differentiation and growth. Collectively, these results characterize NRP2 as a driver of NEPC and suggest NRP2 as a potential therapeutic target in NEPC.

Androgen receptor (AR) is the primary oncogenic driver of prostate cancer, including aggressive castration-resistant prostate cancer (CRPC). The molecular mechanisms controlling AR activation in general and AR reactivation in CRPC remain elusive. Here we report that monoamine oxidase A (MAOA), a mitochondrial enzyme that degrades monoamine neurotransmitters and dietary amines, reciprocally interacts with AR in prostate cancer. MAOA was induced by androgens through direct AR binding to a novel intronic androgen response element of the MAOA gene, which in turn promoted AR transcriptional activity via upregulation of Shh/Gli-YAP1 signaling to enhance nuclear YAP1-AR interactions. Silencing MAOA suppressed AR-mediated prostate cancer development and growth, including CRPC, in mice. MAOA expression was elevated and positively associated with AR and YAP1 in human CRPC. Finally, genetic or pharmacologic targeting of MAOA enhanced the growth-inhibition efficacy of enzalutamide, darolutamide, and apalutamide in both androgen-dependent and CRPC cells. Collectively, these findings identify and characterize an MAOA-AR reciprocal regulatory circuit with coamplified effects in prostate cancer. Moreover, they suggest that cotargeting this complex may be a viable therapeutic strategy to treat prostate cancer and CRPC. SIGNIFICANCE: MAOA and AR comprise a positive feedback loop in androgen-dependent and CRPC, providing a mechanistic rationale for combining MAOA inhibition with AR-targeted therapies for prostate cancer treatment.

Neuroendocrine prostate cancer (NEPC) is a lethal prostate cancer subtype arising as a consequence of more potent androgen receptor (AR) targeting in castration-resistant prostate cancer (CRPC). Its molecular pathogenesis remains elusive. Here, we report that the Wnt secretion mediator Wntless (WLS) is a major driver of NEPC and aggressive tumor growth in vitro and in vivo. Mechanistic studies showed that WLS is a transcriptional target suppressed by AR that activates the ROR2/PKCδ/ERK signaling pathway to support the neuroendocrine (NE) traits and proliferative capacity of NEPC cells. Analysis of clinical samples and datasets revealed that WLS was highly expressed in CRPC and NEPC tumors. Finally, treatment with the Wnt secretion inhibitor LGK974 restricted NE prostate tumor xenograft growth in mice. These findings collectively characterize the contribution of WLS to NEPC pathogenesis and suggest that WLS is a potential therapeutic target in NEPC.

Prostate cancer (PC) develops in a microenvironment where the stromal cells modulate adjacent tumor growth and progression. Here, we demonstrated elevated levels of monoamine oxidase B (MAOB), a mitochondrial enzyme that degrades biogenic and dietary monoamines, in human PC stroma, which was associated with poor clinical outcomes of PC patients. Knockdown or overexpression of MAOB in human prostate stromal fibroblasts indicated that MAOB promotes cocultured PC cell proliferation, migration, and invasion and co-inoculated prostate tumor growth in mice. Mechanistically, MAOB induces a reactive stroma with activated marker expression, increased extracellular matrix remodeling, and acquisition of a protumorigenic phenotype through enhanced production of reactive oxygen species. Moreover, MAOB transcriptionally activates CXCL12 through Twist1 synergizing with TGFβ1-dependent Smads in prostate stroma, which stimulates tumor-expressed CXCR4-Src/JNK signaling in a paracrine manner. Pharmacological inhibition of stromal MAOB restricted PC xenograft growth in mice. Collectively, these findings characterize the contribution of MAOB to PC and suggest MAOB as a potential stroma-based therapeutic target.

BACKGROUND: Artificial Intelligence(AI)-based solutions for Gleason grading hold promise for pathologists, while image quality inconsistency, continuous data integration needs, and limited generalizability hinder their adoption and scalability. METHODS: We present a comprehensive digital pathology workflow for AI-assisted Gleason grading. It incorporates A!MagQC (image quality control), A!HistoClouds (cloud-based annotation), Pathologist-AI Interaction (PAI) for continuous model improvement, Trained on Akoya-scanned images only, the model utilizes color augmentation and image appearance migration to address scanner variations. We evaluate it on Whole Slide Images (WSI) from another five scanners and conduct validations with pathologists to assess AI efficacy and PAI. RESULTS: Our model achieves an average F1 score of 0.80 on annotations and 0.71 Quadratic Weighted Kappa on WSIs for Akoya-scanned images. Applying our generalization solution increases the average F1 score for Gleason pattern detection from 0.73 to 0.88 on images from other scanners. The model accelerates Gleason scoring time by 43% while maintaining accuracy. Additionally, PAI improve annotation efficiency by 2.5 times and led to further improvements in model performance. CONCLUSIONS: This pipeline represents a notable advancement in AI-assisted Gleason grading for improved consistency, accuracy, and efficiency. Unlike previous methods limited by scanner specificity, our model achieves outstanding performance across diverse scanners. This improvement paves the way for its seamless integration into clinical workflows.