Identification of potential molecular pathways involved in prostate carcinogenesis in offspring exposed to maternal malnutrition

Abstract

The developmental origins of health and disease concept links adult diseases with early-life exposure to inappropriate environmental conditions. Intrauterine and postnatal malnutrition may lead to an increased incidence of type 2 diabetes, obesity, and cardiovascular diseases. Maternal malnutrition (MM) has also been associated with prostate carcinogenesis. However, the molecular mechanisms associated with this condition remain poorly understood. Using a proteomic analysis, we demonstrated that MM changed the levels of proteins associated with growth factors, estrogen signaling, detoxification, and energy metabolism in the prostate of both young and old rats. These animals also showed increased levels of molecular markers of endoplasmic reticulum function and histones. We further performed an in silico analysis that identified commonly deregulated proteins in the ventral prostate of old rats submitted to MM with a mouse model and patients with prostate cancer. In conclusion, our results demonstrated that estrogenic signaling pathways, endoplasmic reticulum functions, energy metabolism, and molecular sensors of protein folding and Ca2+ homeostasis, besides histone, and RAS-GTPase family appear to be involved in this process. Knowledge of these factors may raise discussions regarding the role of maternal dietary intervention as a public policy for the lifelong prevention of chronic diseases.

Keywords: DOHaD; mass spectrometry; maternal exposure to low protein diet; prostate diseases.

Figure 1

Body weight (A, B) and hormonal levels (C, D) of male offspring on PND 21 and 540. All data are expressed as mean±SD. Asterisks (*) represent statistical differences between experimental groups with p<0.05. CTR = control; GLLP = gestational and lactational low protein; PND = postnatal day; VP = ventral prostate.

Figure 2

Representative histological sections of the VP lobes from the CTR and GLLP groups on PND 21 and 540, stained with hematoxylin-eosin (HE). Glandular growth in the GLLP group on PND 21 was impaired compared to the CTR. At PND 540, the carcinoma in situ was highlighted by the dashed circle. S = Stroma, L = Lumen, E = Epithelium, Scale bar: 500 μm.

Figure 3

Venn diagram. (A) Shared proteins between CTR and GLLP groups in PND 21 and 540. (B) Shared proteins differentially expressed between the CTR and GLLP groups on PND 21 and 540. CTR = Control; GLLP = Gestational and lactational low protein; PND = Postnatal day; up = upregulated proteins; down = downregulated proteins.

Figure 4

(A) Ontological enrichment of upregulated (red) and downregulated (blue) proteins on PND 21 by KOBAs 3.0. All data were expressed as -Log10 (p-value). (B) Circus plot graphic identifying the top 10 enriched terms and the DEP associated with each term. The numbers in front of the bars mean the number of proteins that enriched each term.

Figure 5

(A) Ontological enrichment of upregulated (red) and downregulated (blue) proteins on PND 540 using the KOBAs 3.0 tool. All data were expressed as -Log10 (p-value). (B) The Circus plot graphic identifying the top 10 enriched terms and the DEP associated with each term.

Figure 6

(A) Clustergram generated by Enrichr using upregulated proteins on PND 21. The red cells in the matrix indicate the genes associated with each term. It was demonstrated the top 10 enriched terms with p-value <0.05. (B) Clustergram generated by Enrichr using downregulated proteins on PND 21. The red cells in the matrix indicate the genes associated with each term. It was demonstrated the enriched terms with p-value <0.05 (top 10).

Figure 7

(A) Clustergram generated by Enrichr using upregulated proteins on PND 540. The red cells in the matrix indicate the genes associated with each term. It was demonstrated the enriched terms with p-value <0.05 (top 10). (B) Clustergram generated by Enrichr using downregulated proteins on PND 540. The red cells in the matrix indicate the genes associated with each term. It was demonstrated the enriched terms with p-value <0.05 (top 10).

Figure 8

Protein-protein interaction network between commonly upregulated proteins on both PND 21 and 540. Interactions of the identified proteins were mapped by searching the STRING database version 9.0 with a confidence cut-off of 0.7. In the resulting protein association network, proteins are presented as nodes that are connected by lines whose thickness represents the confidence level (0.7-0.9).

Figure 9

(A) Heatmap showing the commonly differentially targets from our set of DEP and RNA-seq data from ventral, dorsal, lateral, and anterior prostate lobes in the mice model of PCa (PB-Cre/PtenloxP/loxP). The percentage of commonly deregulated targets increases as the prostatic disorders worsen (PIN to Medium to Advanced PCa). (B) The commonly deregulated targets between our DEP and those extracted from RNA-seq data by GEPIA. The percentage of commonly deregulated targets increases in the prostate of older offspring. (C) Immunostaining of normal and prostate tumor samples for five commonly upregulated targets in our samples and GEPIA database (http://gepia.cancerpku.cn/) using immunohistochemical data available at the Human Protein Atlas database (https://proteinatlas.org/). PND: Postnatal day; VP: Ventral prostate; LP: Lateral prostate; DP: Dorsal prostate; AP: Anterior prostate PIN: Prostate intraepithelial neoplasia; PCa: Prostate cancer.

Figure 10

Representative immunohistochemistry reaction for Calreticulin (CALR) in the VP lobes from the CTR and GLLP groups on PND 21 (A–D) and 540 (E–H). RT-qPCR reaction for CALR in the VP lobes from CTR and GLLP groups on PND 21 (I) and 540 (J). CTR = control; GLLP = gestational and lactational low protein; PND = postnatal day. Data are expressed as fold change {plus minus} SD. Asterisks (*) means the statistical difference between experimental groups with p < .05. Scale bar: 500 μm, and detail 50 μm.