Abstract
Quantitative proteomics/metabolomics investigation of laser-capture-microdissection (LCM) cell populations from clinical cohorts affords precise insights into disease/therapeutic mechanisms, nonetheless high-quality quantification remains a prominent challenge. Here, we devised an LC/MS-based approach allowing parallel, robust global-proteomics and targeted-metabolomics quantification from the same LCM samples, using biopsies from prostate cancer (PCa) patients as the model system. The strategy features: (i) an optimized molecular weight cutoff (MWCO) filter-based separation of proteins and small-molecule fractions with high and consistent recoveries; (ii) microscale derivatization and charge-based enrichment for ultrasensitive quantification of key androgens (LOQ = 5 fg/1k cells) with excellent accuracy/precision; (iii) reproducible/precise proteomics quantification with low-missing-data using a detergent-cocktail-based sample preparation and an IonStar pipeline for reproducible and precise protein quantification with excellent data quality. Key parameters enabling robust/reproducible quantification have been meticulously evaluated and optimized, and the results underscored the importance of surveying quantitative performances against key parameters to facilitate fit-for-purpose method development. As a proof-of-concept, high-quality quantification of the proteome and androgens in LCM samples of PCa patient-matched cancerous and benign epithelial/stromal cells was achieved (N = 16), which suggested distinct androgen distribution patterns across cell types and regions, as well as the dysregulated pathways involved in tumor-stroma crosstalk in PCa pathology. This strategy markedly leverages the scope of quantitative-omics investigations using LCM samples, and combining with IonStar, can be readily adapted to larger-cohort clinical analysis. Moreover, the capacity of parallel proteomics/metabolomics quantification permits precise corroboration of regulatory processes on both protein and small-molecule levels, with decreased batch effect and enhanced utilization of samples.