Evaluating nuclear lamina-associated alterations as predictors of prostate cancer recurrence.

Abstract

e17119 Background: Prostate cancer exhibits marked disparities, and despite definitive treatment, some men with clinically localized disease develop biochemical recurrence (BCR). Nuclear architecture alterations are implicated in progression, but prognostic value of nuclear lamina-associated components remains unclear. We evaluated lamin A/C protein expression in a racially diverse cohort and leveraged a publicly available transcriptomic dataset to evaluate the prognostic relevance of nuclear envelope-associated biomarkers. Methods: Lamin A/C was measured by multiplex immunofluorescence on tissue microarrays representing multiple tumor cores per case from archival prostatectomy specimens (N=185, ~50% Black) at Boston Medical Center. Primary endpoint was BCR (2 consecutive post-operative PSA values ≥0.2 ng/mL). Lamin A/C was categorized into four groups by qualitative assessment and confirmed by quantitative digital image analysis in a subset to compute a continuous intensity ratio. Associations were tested using Fisher’s exact and chi-squared tests, with odds ratios estimated by logistic regression. Kaplan-Meier survival analyses and univariate and multivariate Cox proportional hazards models assessed associations between lamin A/C protein expression, clinicopathologic variables, and BCR. At the transcriptomic level, RNA-seq data with clinical outcomes from the TCGA prostate adenocarcinoma (PRAD) cohort (N=497) were analyzed, focusing on LMNA and other nuclear envelope-associated genes (e.g. LBR , EMD ). Results: Median follow up was 4.71 years; 53 BCR events (28.6%) occurred. Tumor lamin A/C protein expression was heterogeneous. Logistic regression revealed that Grade Group 2 had significantly higher odds of elevated lamin A/C protein expression vs Grade Group 1. Lamin A/C protein expression was not associated with pathologic features, race, or BCR, and lacked independent prognostic value in univariable and multivariable Cox analyses. Concordantly, LMNA mRNA expression in the TCGA-PRAD cohort showed no association with pathologic features or BCR. In contrast, transcriptomic analyses of additional nuclear lamina-associated genes demonstrated gene-specific prognostic signatures. Lamin B receptor ( LBR ) expression significantly increased with higher grade, while emerin ( EMD ) expression was independently associated with BCR in multivariate Cox models (HR=2.31, 95% CI 1.05-5.07; p=0.039). Conclusions: While lamin A/C reflects tumor differentiation, it does not independently predict BCR at either the protein or transcript level. In contrast, among nuclear lamina-associated components, only EMD demonstrates independent prognostic significance, supporting a functional distinction between relatively static nuclear lamin architecture and emerin-associated nuclear envelope dynamics, with potential utility for refining risk stratification.