Characterization of fibrillar collagens and extracellular matrix of glandular benign prostatic hyperplasia nodules

Abstract

Objective: Recent studies have associated lower urinary tract symptoms (LUTS) in men with prostatic fibrosis, but a definitive link between collagen deposition and LUTS has yet to be demonstrated. The objective of this study was to evaluate ECM and collagen content within normal glandular prostate tissue and glandular BPH, and to evaluate the association of clinical parameters of LUTS with collagen content.

Methods: Fibrillar collagen and ECM content was assessed in normal prostate (48 patients) and glandular BPH nodules (24 patients) using Masson's trichrome stain and Picrosirius red stain. Second harmonic generation (SHG) imaging was used to evaluate collagen content. Additional BPH tissues (n = 47) were stained with Picrosirius red and the association between clinical parameters of BPH/LUTS and collagen content was assessed.

Results: ECM was similar in normal prostate and BPH (p = 0.44). Total collagen content between normal prostate and glandular BPH was similar (p = 0.27), but a significant increase in thicker collagen bundles was observed in BPH (p = 0.045). Using SHG imaging, collagen content in BPH (mean intensity = 62.52; SEM = 2.74) was significantly higher than in normal prostate (51.77±3.49; p = 0.02). Total collagen content was not associated with treatment with finasteride (p = 0.47) or α-blockers (p = 0.52), pre-TURP AUA symptom index (p = 0.90), prostate-specific antigen (p = 0.86), post-void residual (PVR; p = 0.32), prostate size (p = 0.21), or post-TURP PVR (p = 0.51). Collagen content was not associated with patient age in patients with BPH, however as men aged normal prostatic tissue had a decreased proportion of thick collagen bundles.

Conclusions: The proportion of larger bundles of collagen, but not total collagen, is increased in BPH nodules, suggesting that these large fibers may play a role in BPH/LUTS. Total collagen content is independent of clinical parameters of BPH and LUTS. If fibrosis and overall ECM deposition are associated with BPH/LUTS, this relationship likely exists in regions of the prostate other than glandular hyperplasia.

Figure 1. Extracellular matrix content is similar in normal prostate tissue and glandular BPH.

The tissue microarray (TMA) was stained with Masson's trichrome for detection of cellular and extracellular matrix (ECM) contents. Blue coloration, indicative of ECM, was separated for all cores by manual thresholding of hue (121–179), saturation (20–255), and brightness (10–255) values in ImageJ, and ECM content was quantified as mean blue intensity per tissue area. ECM composition was similar in normal prostate (mean = 6.308; SEM = 0.560) and glandular BPH (5.496±0.940; p = 0.44).

Figure 2. Analysis of fibrillar collagen content in normal prostate tissues and glandular BPH with Picrosirius red staining.

Total collagen content was defined as the sum of positive birefringent pixels divided by the total size of the region of interest. Total collagen content was not significantly different in normal prostate tissue (mean = 56.2%; SEM = 1.8) and glandular BPH (52.9±2.3; p = 0.27; A). The normalized proportion of green (p = 0.15), yellow (p = 0.39), and red (p = 0.09) birefringent collagen bundles were not significantly different between normal prostate tissues and BPH (B). The proportion of orange bundles was significantly higher in glandular BPH (normalized mean = 1.268; SEM = 0.138) than normal prostate (1.000±0.063; p = 0.045).

Figure 3. Evaluation of collagen content in normal prostate and glandular benign prostatic hyperplasia (BPH) with second harmonic generation (SHG) imaging.

Using SHG imaging, a total of 12–15 optical sections were acquired (40x objective) per core from a subset of normal prostate tissue (n = 10) and BPH (n = 11). Optical sections were stacked and flattened to create maximum intensity Z-projections using Fiji software. Collagen content was quantified as the mean gray intensity within the region of interest. Using SHG imaging, a significant increase in collagen content was observed in BPH (mean = 62.52, SEM = 2.738) compared to normal prostate tissue (51.77±3.492; p = 0.02).

Figure 4. Association of Picrosirius red fibrillar collagen content in glandular BPH specimens with pre-TURP treatment with α-blockers and finasteride.

A total of 47 patients undergoing transurethral resection of the prostate (TURP) for treatment of BPH at the University of Wisconsin Hospital were randomly selected for inclusion in this study. TURP samples were stained for Picrosirius red, and 3 representative acinar lobules were imaged under polarized light. Staining was quantified as total collagen content and proportion of colors within birefringent tissue, and triplicate images for each patient were averaged. Treatment with α-blockers (n = 37 patients) had no significant effect on total collagen content in glandular BPH nodules (p = 0.52; A). No changes were observed in the proportion of green (normalized mean = 1.157±0.010 vs. 1.000±0.180; p = 0.47), yellow (1.016±0.013 vs. 1.000±0.019; p = 0.55), orange (0.930±0.039 vs. 1.000±0.067; p = 0.40), or red (0.896±0.082 vs. 1.000±0.143; p = 0.55) birefringent tissue (B). Similarly, treatment with finasteride (n = 22 patients) had no effect on total collagen content within the tissues (p = 0.47; C). No significant changes in the proportion of green (0.828±0.082 vs. 1.000±0.112; p = 0.23), yellow (1.000±0.013 vs. 1.000±0.017; p = 1.00), orange (1.065±0.042 vs. 1.000±0.058; p = 0.38), or red (1.061±0.092 vs. 1.000±0.127; p = 0.71) collagen bundles were observed (D).

Figure 5. Association of patient age with Picrosirius red collagen content in prostate tissues.

The association of Picrosirius red collagen content and patient age at time of surgery was investigated in cohorts of BPH only tissues (TURP set), a combination of BPH and normal prostate tissue (TMA set [all patients]), and a subset of only normal prostate tissue (TMA set [normal only]). Total collagen content and the distribution of birefringent colors were not associated with age in the TURP set or combination TMA set (all patients) of tissues (p>0.05). Total collagen content was not associated with age in the TMA subset of normal only patients (p = 0.96), but a significant negative correlation between age and birefringent color proportions was observed for red (p = 0.009) and orange (p = 0.03) birefringence. No association was found between age and yellow (p = 0.41) or green (p = 0.15) birefringence in the normal only subset of prostate tissues.