Histology-resolved proteomics reveals distinct tumor and stromal profiles in low- and high-grade prostate cancer

Abstract

Background: Prostate cancer is one of the most frequently diagnosed cancers in men. Prostate tumor staging and disease aggressiveness are evaluated based on the Gleason scoring system, which is further used to direct clinical intervention. The Gleason scoring system provides an estimate of tumor aggressiveness through quantitation of the serum level of prostate specific antigen (PSA) and histologic assessment of Grade Group, determined by the Gleason Grade of the tumor specimen.

Methods: To improve our understanding of the proteomic characteristics differentiating low- versus high-grade prostate cancer tumors, we performed a deep proteomic characterization of laser microdissected epithelial and stromal subpopulations from surgically resected tissue specimens from patients with Gleason 6 (n = 23 specimens from n = 15 patients) and Gleason 9 (n = 15 specimens from n = 15 patients) prostate cancer via quantitative high-resolution liquid chromatography-tandem mass spectrometry analysis.

Results: In total, 789 and 295 grade-specific significantly altered proteins were quantified in the tumor epithelium and tumor-involved stroma, respectively. Benign epithelial and stromal populations were not inherently different between Gleason 6 versus Gleason 9 specimens. Notably, 598 proteins were exclusively significantly altered between Gleason 9 (but not Gleason 6) tumor-involved stroma and benign stroma, including several proteins involved in cholesterol biosynthesis and nucleotide metabolism.

Conclusions: Proteomic alterations between Gleason 6 versus Gleason 9 were exclusive to the disease microenvironment, observed in both the tumor epithelium and tumor-involved stroma. Further, the molecular alterations measured in the tumor-involved stroma from Gleason 9 cases relative to the benign stroma have unique significance in disease aggressiveness, development, and/or progression. Our data provide supportive evidence of a need for further investigations into targeting stromal reservoirs of cholesterol and/or deoxynucleoside triphosphates in PCa tumors and further highlight the necessity for independent examination of the TME epithelial and stromal compartments.

Keywords: Laser microdissection; Prostate cancer; Proteomics; Tumor microenvironment.

Fig. 1

Workflow diagram. Pre- and post-LMD images of representative Gleason 6 (J4256_461858; top) and Gleason 9 (J4255_461843; bottom) specimens

Fig. 2

Unsupervised heatmaps using the most variably abundant proteins (top 100 MAD) in the A epithelial (ET and BE) and B stromal (ES and BS) samples

Fig. 3

Differential analyses (limma adj. p < 0.05) of LMD enriched sample populations from PCa tissue specimens. A Pairwise limma analyses of proteomic abundances from LMD enriched samples. B The top five (by z-score) IPA canonical pathways and FDA-approved drug targets identified using differentially abundant proteins between Gleason 6 BE versus Gleason 6 ET, and from Gleason 9 BE versus Gleason 9 ET. C The top five IPA canonical pathways and FDA-approved drug targets identified using differentially abundant proteins between Gleason 6 ES versus Gleason 6 BS, and from Gleason 9 ES versus Gleason 9 BS. Pairwise differential analyses used for the (B) epithelial and (C) stromal comparative analyses are colored red in the upset plot (A)

Fig. 4

Boxplots depicting enrichment of IPA canonical pathways relating to A cholesterol biosynthesis and B nucleotide metabolism. The plots represent the differential analysis log₂-transformed FC values for each pathway-related protein in the notated subsets from comparative analysis. The highlighted number above each box-and-whisker plot represents the IPA z-score for pathway activation. Positive z-scores are enriched in the first LMD cell type comparator (i.e. ET or ES). Negative z-scores are enriched in the second LMD cell type comparator (i.e. BE or BS). Analysis was performed using proteins passing limma adj. p < 0.05 and selected IPA pathways passing p < 0.05

Fig. 5

Characterization of immune cell signatures in LMD enriched PCa samples. A xCell [18] immune cell types significantly differing (Welch’s two sample t-test p < 0.05) in TME cellular subpopulations from Gleason 6 versus Gleason 9 specimens. FC = fold change. % Coverage = the percentage of marker genes in the xCell cell type signature which were measured at the proteomic level in LMD enriched samples. B ProteoMixture immune scores of LMD enriched epithelial (top) and stromal (bottom) samples