Pathway Enrichment Analysis of Microarray Data Fom Human Penis of Diabetic and Peyronie's Patients, in Comparison With Diabetic Rat Erectile Dysfunction Models

Abstract

Background: Erectile dysfunction (ED) is a debilitating medical condition in which current treatments are minimally effective in diabetic patients due to neuropathy of the cavernous nerve, a peripheral nerve that innervates the penis. Loss of innervation causes apoptosis of penile smooth muscle, remodeling of corpora cavernosa (penile erectile tissue) morphology, and ED.

Aim: In this study, microarray and pathway analysis were used to obtain a global understanding of how signaling mechanisms are altered in diabetic patients and animal models as ED develops, in order to identify novel targets for disease management, and points of intervention for clinical therapy development.

Methods and outcomes: Human corpora cavernosal tissue was obtained from diabetic (n = 4) and Peyronie's (control, n = 3) patients that were undergoing prosthesis implant to treat ED, and BB/WOR diabetic (n = 5) and resistant (n = 5) rats. RNA was extracted using TRIzol, DNase treated, and purified by Qiagen mini kit. Microarray was performed using the Human Gene 2.0 ST Array. (i) Alterations in patient and diabetic rat pathway signaling were examined using several analytical tools (ShinyGO, Metascape, WebGestalt, STRING) and databases, (ii) Strengths/weaknesses of the different pathway analysis tools were compared, and (iii) Comparison of human and rat (BB/WOR and Streptozotocin) pathway analysis was performed. Two technical replicates were performed. P value (FDR) < .15 was used as threshold for differential expression. FDR < 0.05 was considered significant.

Results: Microarray identified 182 differentially expressed protein-coding genes. Pathway analysis revealed similar enrichments with different analytical tools. Down regulated pathways include development, tubular structure, sprouting, cell death, ischemia, angiogenesis, transcription, second messengers, and stem cell differentiation. ED patients, who have diabetes, incur significant loss of normal regulatory processes required for repair and replacement of injured corpora cavernosal tissue. Combined with loss of apoptotic regulatory mechanisms, this results in significant architectural remodeling of the corpora cavernosa, and loss of regenerative capacity in the penis.

Clinical translation: This first report of microarray and pathway analysis in human corpora cavernosa, is critical for identification of novel pathways pertinent to ED and for validating animal models.

Strengths and limitations: The analysis of tissue specific gene expression profiles provides a means of understanding drivers of disease and identifying novel pathways for clinical intervention.

Conclusion: Penis from diabetic ED patients lacks capacity for maintenance of corpora cavernosal architecture and regeneration, which are critical points for intervention for therapy development. Searl T, Ohlander S, McVary KT, et al., Pathway Enrichment Analysis of Microarray Data Fom Human Penis of Diabetic and Peyronie's Patients, in Comparison With Diabetic Rat Erectile Dysfunction Models. J Sex Med 2022;19:37-53.

Keywords: Corpora Cavernosa; Diabetic; Erectile Dysfunction; Microarray; Pathway Analysis; Penis; Peyronie's.

Figure 1:

A diagrammatic comparison of different web-based analysis tools and databases. The analytical sites utilized provide analysis from a combination of partially overlapping databases, that have individualized approaches to purpose, organization and categorization.

Figure 2:

Heatmap of differentially expressed genes comparing corpora cavernosa from Peyronie’s and diabetic ED patients. The majority of genes were downregulated (green) in the diabetic corpora cavernosa versus Peyronie’s controls. Upregulated genes are presented in red.

Figure 3:

Pathway analysis of human diabetic and Peyronie’s (control) corpora cavernosal tissue. (A) ShinyGO analysis presented as hierarchical clustering trees for GO Biological Process, GO Cellular Component, GO Molecular Function, KEGG, Wikipathway and Reactome. Enrichment terms with many shared genes are clustered together. Significant changes are high lighted with black dots. FDR values shown with larger dots indicate more significant p values. (B) WebGestalt analysis of GO Biological Process, GO Molecular Function, GO Cellular Component, KEGG, and DisGeNET. (C) Metascape and STRING analysis. (1) Metascape analysis. (2) The main Metascape enrichments are shown as a bar graph (-log₁₀ P scale). (3) The Metascape derived enrichment analysis of DEG associated pathways is depicted as a bar graph (-log₁₀ P scale). (4) STRING protein-protein network analysis presented with the nodes highlighted with selected GO Biological Processes. (Red: Tube Development; Blue: Response to organonitrogen compound; Green: Response to cAMP; Purple: Regulation of immune system process; Yellow: Response to oxidative stress).

Figure 3:

Pathway analysis of human diabetic and Peyronie’s (control) corpora cavernosal tissue. (A) ShinyGO analysis presented as hierarchical clustering trees for GO Biological Process, GO Cellular Component, GO Molecular Function, KEGG, Wikipathway and Reactome. Enrichment terms with many shared genes are clustered together. Significant changes are high lighted with black dots. FDR values shown with larger dots indicate more significant p values. (B) WebGestalt analysis of GO Biological Process, GO Molecular Function, GO Cellular Component, KEGG, and DisGeNET. (C) Metascape and STRING analysis. (1) Metascape analysis. (2) The main Metascape enrichments are shown as a bar graph (-log₁₀ P scale). (3) The Metascape derived enrichment analysis of DEG associated pathways is depicted as a bar graph (-log₁₀ P scale). (4) STRING protein-protein network analysis presented with the nodes highlighted with selected GO Biological Processes. (Red: Tube Development; Blue: Response to organonitrogen compound; Green: Response to cAMP; Purple: Regulation of immune system process; Yellow: Response to oxidative stress).

Figure 3:

Pathway analysis of human diabetic and Peyronie’s (control) corpora cavernosal tissue. (A) ShinyGO analysis presented as hierarchical clustering trees for GO Biological Process, GO Cellular Component, GO Molecular Function, KEGG, Wikipathway and Reactome. Enrichment terms with many shared genes are clustered together. Significant changes are high lighted with black dots. FDR values shown with larger dots indicate more significant p values. (B) WebGestalt analysis of GO Biological Process, GO Molecular Function, GO Cellular Component, KEGG, and DisGeNET. (C) Metascape and STRING analysis. (1) Metascape analysis. (2) The main Metascape enrichments are shown as a bar graph (-log₁₀ P scale). (3) The Metascape derived enrichment analysis of DEG associated pathways is depicted as a bar graph (-log₁₀ P scale). (4) STRING protein-protein network analysis presented with the nodes highlighted with selected GO Biological Processes. (Red: Tube Development; Blue: Response to organonitrogen compound; Green: Response to cAMP; Purple: Regulation of immune system process; Yellow: Response to oxidative stress).

Figure 4:

Summary diagram of pathways altered in diabetic ED patients.

Figure 5:

Pathway analysis of differentially enriched gene expression in corpora cavernosa from BB/WOR diabetic and diabetes resistant (control) rats presented on hierarchical clustering trees with their FDR values. (A) Correlation of significant pathways was performed using GO Biological Process, GO Cellular Component, GO Molecular Function, and KEGG. Pathways with many shared genes are clustered together. Larger dots indicate more significant p values. (B) Metscape analysis is shown with the main enrichments depicted as a bar graph (-log₁₀ P scale). The Metascape derived DisGeNET enrichment bar graph ((-log₁₀ P scale) depicting the DEG associated disease enrichments, are also shown.

Figure 5:

Pathway analysis of differentially enriched gene expression in corpora cavernosa from BB/WOR diabetic and diabetes resistant (control) rats presented on hierarchical clustering trees with their FDR values. (A) Correlation of significant pathways was performed using GO Biological Process, GO Cellular Component, GO Molecular Function, and KEGG. Pathways with many shared genes are clustered together. Larger dots indicate more significant p values. (B) Metscape analysis is shown with the main enrichments depicted as a bar graph (-log₁₀ P scale). The Metascape derived DisGeNET enrichment bar graph ((-log₁₀ P scale) depicting the DEG associated disease enrichments, are also shown.

Figure 6:

Pathway analysis for differentially enriched gene expression in penis of Streptozotocin-induced diabetic rats. ShinyGO analysis is presented as hierarchical clustering trees with their FDR values. (A) Correlation among significant pathways was examined using GO Biological Process, GO Cellular Component, GO Molecular Function, and KEGG. Pathways with many shared genes are clustered together. Larger dots indicate more significant P-values. (B) Also shown is the clustered Metascape analysis, with a bar chart of the main enrichment terms with their corresponding P-values. The main Metascape derived DisGeNET disease enrichments are shown as a bar graph with their corresponding P-values (-log₁₀ P scale).

Figure 6:

Pathway analysis for differentially enriched gene expression in penis of Streptozotocin-induced diabetic rats. ShinyGO analysis is presented as hierarchical clustering trees with their FDR values. (A) Correlation among significant pathways was examined using GO Biological Process, GO Cellular Component, GO Molecular Function, and KEGG. Pathways with many shared genes are clustered together. Larger dots indicate more significant P-values. (B) Also shown is the clustered Metascape analysis, with a bar chart of the main enrichment terms with their corresponding P-values. The main Metascape derived DisGeNET disease enrichments are shown as a bar graph with their corresponding P-values (-log₁₀ P scale).