2020 Volume 45 Issue 8 Pages 435-447
The imbalance of testosterone to estradiol ratio has been related to the development of prostate diseases. Although rat models of prostate diseases induced by endocrine-disrupting chemicals (EDCs) and/or hormone exposure are commonly used to analyze gene expression profiles in the prostate, most studies utilize a single endpoint. In this study, microarray analysis was used for gene expression profiling in rat prostate tissue after exposure to EDCs and sex hormones over multiple time points (prepubertal through adulthood). We used dorsolateral prostate tissues from Sprague-Dawley rats (male offspring) and postnatally administered estradiol benzoate (EB) on postnatal days (PNDs) 1, 3, and 5, followed by treatment with additional hormones [estradiol (E) and testosterone (T)] on PNDs 90–200, as described by Ho et al. Microarray analysis was performed for gene expression profiling in the dorsolateral prostate, and the results were validated via qRT-PCR. The genes in cytokine-cytokine receptor interaction, cell adhesion molecules, and chemokines were upregulated in the EB+T+E group on PNDs 145 and 200. Moreover, early-stage downregulation of anti-inflammatory gene: bone morphogenetic protein 7 gene was observed. These findings suggest that exposure to EB, T, and E activates multiple pathways and simultaneously downregulates anti-inflammatory genes. Interestingly, these genes are reportedly expressed in prostate cancer tissues/cell lines. Further studies are required to elucidate the mechanism, including analyses using human prostate tissues.
Prostate diseases (i.e., prostatitis, enlarged prostate, and cancer) are prevalent in a specific age-group of males. Prostatitis, chronic inflammation of the prostate, is likely to occur in young or middle-aged people. Acute inflammation is often caused by bacterial infections. Chronic inflammation of the prostate can also be caused by bacteria, although most cases result from unknown causes (Nickel, 2012). The risk of an enlarged prostate [e.g., benign prostatic hyperplasia (BPH)] is higher after 40 years of age, when 20% of men in their fifties and 60% of men in their sixties have the condition. Prostate cancer is the most common of all new cancer cases (23%) in the United States (http://www.cancer.gov/cancertopics/types/commoncancers). Although the causes of prostate diseases have yet to be determined in detail, recently, increased estrogen levels in the serum or prostate have been thought to be a key factor in the progression of prostate carcinogenesis, as well as BPH development (Krieg et al., 1993; Bosland, 2005; Ho et al., 2011; Nicholson and Ricke, 2011). These studies suggest that prostate diseases are likely to occur when prostate estrogen levels increase due to decreasing serum testosterone levels brought on by aging or endocrine-disrupting chemicals (EDCs).
Early postnatal exposure to EDCs is specifically influential on prostate development and testosterone production (vom Saal et al., 1997; Prins et al., 2006; Anway and Skinner, 2008; Hofkamp et al., 2008; Prins, 2008; Diamanti-Kandarakis et al., 2009). In addition, testosterone levels in men peak at 25–30 years of age and then decline gradually. The testosterone levels show a 30–60% decrease at age 60 and older (Kaufman and Vermeulen, 2005). In contrast, there is no substantial age-related change in male serum estrogen levels (Jasuja et al., 2013); however, there is a significant age-related increase in estrogen levels in normal human prostates (Krieg et al., 1993). Further, the ratio of estrogen to testosterone levels significantly increased with aging (Suzuki et al., 1992).
Rat models are still widely used as animal models of prostate cancer and/or BPH (Noble, 1977; Pollard, 1982; Bosland et al., 1995; Russell and Voeks, 2003), including gene expression profiling (Thompson et al., 2002; Ho et al., 2006). Prins and colleagues developed rat models of prostate disease (i.e., neoplasia and inflammation), induced by EDCs supplemented with additional testosterone and estradiol treatments, to examine gene expression profiles during prostate development through adulthood (Gilleran et al., 2003; Huang et al., 2004). However, most studies have elucidated the mechanisms underlying these disease-inducing treatments at a later stage (one endpoint). There are some studies for gene expression or methylation profiling during the development of prostate disease (Tang et al., 2012; Wong et al., 2015; Cheong et al., 2016; Lam et al., 2016; Prins et al., 2017).
Previously, we studied gene expression profiling in Sprague-Dawley rats (Hsd:SD) dosed with estradiol benzoate (EB), testosterone (T), and 17-β-estradiol (E) as described by Ho et al. (2006). Pups at postnatal days (PNDs) 1, 3, and 5 were injected with EB. From PND 90 through PND 200, they were dosed with T and E via silastic tube implants in the subcutaneous region. Similar to Gilleran et al. (2003), we found the occurrence of chronic inflammation only in the dorsolateral prostate of rats dosed with EB, T, and E on PNDs 145 and 200 (Nakamura et al., in press), which can be associated with gene alteration.
The purpose of this study was to examine gene expression profiling in rat prostate dosed with EB, T, and E during prepubertal through adult ages using rat prostates To perform this study, we performed microarray analysis using rat prostates dosed with EB, T, and E and verified these results by qRT-PCR.
All reagents were purchased from Fisher Scientific (Pittsburgh, PA, USA) and Sigma-Aldrich (St. Louis, MO, USA), unless otherwise indicated.
Animals and treatmentsThe dorsolateral prostates were obtained from Sprague-Dawley (SD) rats (Nakamura et al., in press). Treatment of male offspring was carried out following the method described by Ho et al. (2006). Figure 1 shows the experimental design. Briefly, male offspring were obtained from fifty 11–13-week-old time-mated female SD rats purchased from Envigo (Indianapolis, IN, USA) and delivered to the National Center for Toxicological Research (NCTR) on gestation day (GD) 3 (day of birth = PND 0). Animals were housed individually and maintained under a 12:12-hr light-dark cycle with controlled room temperature (23 °C ± 3 °C) and humidity (50% ± 20%). Their diet upon arrival consisted of a low-phytoestrogen 5K96 chow (Purina Mills, St. Louis, MO, USA). Water was provided ad libitum. All animal procedures were approved by the NCTR Institutional Animal Care and Use Committee and followed the guidelines set forth by the National Research Council’s Guide for the Care and Use of Laboratory Animals (National Research Council, 2011).
Scheme of experimental design. Male SD:Hsd offspring were injected with 2.5 mg/kg body weight (BW) EB subcutaneously at PNDs 1, 3, and 5 for the treated group. Male offspring in the control group were injected with a vehicle (i.e., corn oil). On PND 90, the animals in T+E only and EB+T+E groups were treated with testosterone and estradiol packed-silastic tube implants, while the animals in the control and EB only groups were implanted with empty silastic tubes. After 8 weeks (on PND 146) the animals received new implants.
The male offspring were divided into two groups: untreated (n =148) and estradiol benzoate (EB)-treated (n =148). In the EB-treated group, male pups were injected subcutaneously with 2.5 mg/kg body weight (BW) EB (Sigma-Aldrich, E8515) at PNDs 1, 3, and 5. Male pups in the untreated group (control) were injected with a vehicle (i.e., tocopherol-stripped corn oil; #0290141584-400; ICN Biomedicals, Inc., Aurora, OH, USA). At PND 90, all groups (untreated/control and EB-treated groups) were divided into two groups: control, testosterone (T) and estradiol (E) only, EB only, and EB+T+E groups (n = 8/group/collection point). All animals were then subcutaneously implanted with several tubes. Animals in the control and EB only groups were implanted with three empty silastic tubing inserts (two 2-cm tubes and one 1-cm tube), whereas animals in the T+E only and EB+T+E groups were implanted with two silastic tubing inserts (Dow Corning, ID; 1.47 mm; absorbance, 1.95 mm) packed with T powder (two 2-cm tubes; Sigma-Aldrich) and one tube packed with E (one 1-cm tube; Sigma-Aldrich) for additional testosterone and 17-β-estradiol treatment until PND 200. Eight weeks after the first surgery on PND 90 (on PND 146), 32 animals underwent a second surgery to replace the hormone-containing silastic tubing implants. These animals were sacrificed on PND 200.
Microarray experimentProstate RNA samples from PNDs 30, 100, and 145 rats (n = 3–4 litters/group) were extracted using a miRNeasy Mini Kit (Qiagen, Valencia, CA, USA). After extraction, individual RNA concentrations were determined via a NanoDrop 2000c spectrophotometer (version 3.0.1, Thermo Fisher Scientific, Inc., Wilmington, DE, USA). One-color microarray-based gene expression analysis was performed using Agilent QuickAmp Kits following the manufacturer’s instructions (Agilent Technologies, Santa Clara, CA, USA). The total RNA (~500 ng) was labeled with Cy3, and the cRNA (Cy3-labeled RNA) was purified with a RNeasy Kit (Qiagen). cRNA concentrations were determined with a Nanodrop ND 1000 spectrophotometer (Thermo Fisher Scientific). Equal amounts (600 ng) of purified cRNA were hybridized to Agilent SurePrint G3 rat 8 × 60K microarrays (Agilent Technologies) for 17 hr at 65°C in a hybridization oven. After washing, hybridized microarray slides were scanned with an Agilent DNA Microarray Scanner C, and the images were further analyzed using Agilent’s Feature Extraction software (version 10.7.3).
Analysis of microarray dataThe Agilent rat whole-genome (8 × 60K) arrays contain 62,976 spots for probe localization. Data processing, normalization, and statistical analyses were performed on all transcripts of prostates from all groups of PNDs 30, 100, and 145 rats using the SAS 9.4 (SAS, Cary, NC, USA). The 33 raw microarray data files and preprocessed normalized data (33 files) are accessible at NCBI’s Gene Expression Omnibus (GEO) website, and the GEO accession number is GSE145917 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE145917). A generalized linear model procedure (proc glm) was used to measure statistical significance (p < 0.05) between the control and exposed groups. A relative change in the abundance of each transcript was calculated as a difference in the average log2 intensity values. Genes with expression changes greater than 1.5 folds with p-values less than 0.05 were considered differentially expressed. These differentially expressed genes (DEGs) were used for further pathway analysis utilizing ArrayTrack software (Tong et al., 2003).
cDNA synthesisGene expression changes identified in the microarray analysis were confirmed by qRT-PCR using cDNA synthesized with 1 µg RNA from the microarray samples using a Super Script IV VILO Master Mix (Thermo Fisher Scientific).
Quantitative PCR (qPCR)The qPCR analyses were performed using the ABI PRISM 7900HT Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) with PowerUp SYBR Green Master Mix (Thermo Fisher Scientific) per the manufacturer’s guidelines. The cDNA, synthesized as described above, served as a template in 10-μL reaction mixtures. The PowerUp SYBR Mix reaction conditions were as follows: initial denaturation steps at 50°C for 2 min and 95°C for 2 min, followed by 40–45 amplification cycles (95°C for 15 sec, 60°C for 1 min). The relative steady-state transcript levels were calculated using threshold cycle (Ct) values and the following equation: relative quantity = 2-ΔΔCt (Livak and Schmittgen, 2001). The transcript levels were normalized using Rn18s as an internal control for each sample. The relative ratios of the transcript levels in each sample were calculated by setting the values for controls at each collection time (PND 30, 100, or 145) to one. qPCR was performed in triplicate for each sample (n = 5/group). Supplemental Table 1 shows the specific primer pairs used in this study.
Statistical analysisData are presented as the mean ± standard error of the mean (SEM). Statistical analysis of qPCR data was performed using t-tests (PNDs 30 and 90) or one-way ANOVA (PNDs 100, 145, and 200) with Bonferroni adjustment (Hochberg, 1988; Shaffer, 1995). Log-transformation (Hellemans and Vandesompele, 2011) was performed on qPCR data with higher valiances before the statistical analysis. A p-value of < 0.05 was considered statistically significant.
Microarray analyses of prostate gene expression patterns from rats showed significant differential expression of 2,084, 5,207, and 12,447 genes on PNDs 30, 100, and 145, respectively, compared to controls (p < 0.05, fold change (FC) > 1.5; ANOVA analysis on PNDs 100 and 145). A total of 219 genes overlapped in the prostate among PNDs 30, 100, and 145 (Supplemental Fig. 1).
The genes with treatment-specific changes in expression levels were identified in T+E only, EB only, or EB+T+E groups on PND 100 or 145 (FC > 5.0, p < 0.01) compared to the controls on PND 100 or 145 (Supplemental Tables 2 and 3) and in the EB-treated on PND 30 compared to the controls on PND 30 (Supplemental Table 4). To identify unique DEGs in the EB-treated on PND 30 or EB only groups on PNDs 100 and 145, we found the genes (FC > 5.0, P < 0.01) on PND 30 (EB only exposure) were not overlapped to the genes on the EB only on PNDs 100 and 145 via microarray analysis. Several genes in the EB-treated group on PND 30 were overlapped to the genes on the T+E only on PNDs 100 and 145, respectively (Supplemental Fig. 2). Thus, most of the genes in the group at the different collection points were unique.
Principal component analysis (PCA)Using DEGs data, PCA was conducted by comparing the overall expression profiles of the untreated and EB-treated prostates on PND 30, and of the control, T+E only, EB only, and EB+T+E groups’ prostates on PNDs 100 and 145 (Figs. 2A, B, and C) to determine the association among the two or four groups. Principal Component 1 (PC1) captured 73.3%, PC2 captured 7.2%, and PC3 captured 5.7% of the variance in the data on PND 30 (Fig. 2A). Together, 86.2% of the variance in these samples was captured. PC1 captured 50.7% and 60.4%, PC2 captured 18.5% and 15.3%, and PC3 captured 10.9% and 6.6% of the variance on PNDs 100 (Fig. 2B) and 145 (Fig. 2C), respectively. Together, 80.1% and 82.2% of the total variance on PNDs 100 and 145, respectively, were captured by the first three components. PC1 represents a treatment-related distinct separation that exists among the control and exposed groups, suggesting EB only, T+E only, or EB+T+E exposure has different effects.
Principal component analysis (PCA) of rat dorsolateral prostates. PCA was performed using ArrayTrack software (Tong et al., 2003). (A) PCA between the control and EB-treated groups on PND 30 rat prostate. (B) PCA among the control, T+E only, EB only and EB+T+E groups at PND 100 rat prostate. (C) PCA among the control, T+E only, EB only and EB+T+E groups at PND 145 rat prostate.
ArrayTrack was used to determine the affected genetic pathways in the prostate following EB only (PND 30) or EB, T, and E exposure (PNDs 100 and 145). In all, 63 KEGG pathways in the prostate were significantly altered among the groups on PNDs 30, 100, or 145 compared to the control collection date (Table 1). Of the 63 pathways, the significantly altered pathways in any groups on PNDs 30, 100, and 145 were eight pathways [i.e., cell adhesion molecules (CAMs) (rno04514); HTLV-I infection (rno05166); focal adhesion (rno04510); PI3K-Akt signaling pathway (rno04151) pathways in cancer (rno05200)].
| Map title | PND 30 | PND 100 | PND 145 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| EB-treated | T+E only | EB only | EB+T+E | T+E only | EB only | EB+T+E | |||
| Fisher P value | Fisher P value | Fisher P value | Fisher P value | Fisher P value | Fisher P value | Fisher P value | |||
| Cytokine-cytokine receptor interaction (rno04060) | N.S. | 0.0191 | N.S. | 0.0000538 | 0.00216 | 0.00000462 | 1.22E-20 | ||
| Staphylococcus aureus infection (rno05150) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.00424 | 3.40E-11 | ||
| Chemokine signaling pathway (rno04062) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0000249 | 9.91E-11 | ||
| Cell adhesion molecules (CAMs) (rno04514) | 0.0224 | N.S. | N.S. | 0.0458 | 0.00308 | 0.0000115 | 2.13E-10 | ||
| Osteoclast differentiation (rno04380) | N.S. | N.S. | N.S. | N.S. | 0.0372 | 0.0301 | 3.21E-09 | ||
| ECM-receptor interaction (rno04512) | 0.0183 | 0.0373 | N.S. | N.S. | 0.000514 | N.S. | 1.41E-08 | ||
| Leishmaniasis (rno05140) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0033 | 3.45E-08 | ||
| Amoebiasis (rno05146) | 0.0174 | N.S. | 0.00634 | 0.0148 | 0.000203 | 0.0269 | 9.24E-08 | ||
| TNF signaling pathway (rno04668) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.007 | 1.38E-07 | ||
| Antigen processing and presentation (rno04612) | N.S. | N.S. | N.S. | N.S. | 0.000515 | N.S. | 1.42E-07 | ||
| Phagosome (rno04145) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0311 | 3.35E-07 | ||
| Pertussis (rno05133) | N.S. | N.S. | 0.017 | 0.00103 | N.S. | N.S. | 9.10E-07 | ||
| Chagas disease (American trypanosomiasis) (rno05142) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0375 | 0.0000012 | ||
| Hematopoietic cell lineage (rno04640) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0113 | 0.00000128 | ||
| HTLV-I infection(rno05166) | 0.0445 | N.S. | N.S. | 0.00908 | N.S. | 0.00743 | 0.00000192 | ||
| Rheumatoid arthritis (rno05323) | 0.0198 | N.S. | N.S. | N.S. | N.S. | 0.00764 | 0.00000632 | ||
| NF-kappa B signaling pathway (rno04064) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.00163 | 0.0000132 | ||
| Protein processing in endoplasmic reticulum (rno04141) | 0.0363 | N.S. | N.S. | N.S. | N.S. | 0.0312 | 0.0000147 | ||
| Focal adhesion (rno04510) | 0.00561 | 0.02 | 0.0387 | N.S. | 0.000103 | N.S. | 0.0000295 | ||
| Natural killer cell mediated cytotoxicity (rno04650) | N.S. | 0.00491 | N.S. | 0.032 | N.S. | 0.00988 | 0.0000307 | ||
| Primary immunodeficiency (rno05340) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.000277 | 0.0000407 | ||
| Graft-versus-host disease (rno05332) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0212 | 0.0000553 | ||
| Type I diabetes mellitus (rno04940) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0244 | 0.0000778 | ||
| Endocytosis (rno04144) | 0.0341 | N.S. | N.S. | N.S. | 0.0203 | 0.0165 | 0.000096 | ||
| Toxoplasmosis (rno05145) | N.S. | N.S. | N.S. | N.S. | 0.00105 | N.S. | 0.000101 | ||
| Platelet activation (rno04611) | 0.00388 | N.S. | N.S. | N.S. | 0.00556 | N.S. | 0.000216 | ||
| Allograft rejection (rno05330) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0112 | 0.000303 | ||
| Measles (rno05162) | N.S. | N.S. | N.S. | N.S. | 0.0302 | 0.00875 | 0.00034 | ||
| Thyroid hormone synthesis (rno04918) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0126 | 0.00042 | ||
| PI3K-Akt signaling pathway (rno04151) | 0.00796 | 0.022 | N.S. | N.S. | 0.000252 | N.S. | 0.00129 | ||
| Protein digestion and absorption (rno04974) | 0.0248 | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0025 | ||
| Axon guidance (rno04360) | 0.016 | N.S. | N.S. | N.S. | 0.0461 | 0.0399 | 0.00273 | ||
| Propanoate metabolism (rno00640) | N.S. | N.S. | N.S. | N.S. | 0.0268 | N.S. | 0.004 | ||
| Arginine and proline metabolism (rno00330) | 0.033 | N.S. | N.S. | N.S. | N.S. | 0.0252 | 0.00456 | ||
| Viral myocarditis (rno05416) | N.S. | N.S. | N.S. | N.S. | 0.0498 | 0.0298 | 0.00483 | ||
| Glutathione metabolism (rno00480) | N.S. | N.S. | N.S. | N.S. | 0.00694 | N.S. | 0.00581 | ||
| Fc gamma R-mediated phagocytosis (rno04666) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0117 | 0.00601 | ||
| Autoimmune thyroid disease (rno05320) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0269 | 0.00618 | ||
| Glycine, serine and threonine metabolism (rno00260) | N.S. | N.S. | N.S. | N.S. | 0.0267 | N.S. | 0.00789 | ||
| MicroRNAs in cancer (rno05206) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.00477 | 0.00987 | ||
| Rap1 signaling pathway (rno04015) | 0.00788 | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0122 | ||
| Pathways in cancer (rno05200) | 0.0000813 | 0.0256 | N.S. | N.S. | 0.0219 | N.S. | 0.0138 | ||
| Cytosolic DNA-sensing pathway (rno04623) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0419 | 0.0195 | ||
| Transcriptional misregulation in cancers (rno05202) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.00519 | 0.0208 | ||
| Small cell lung cancer(rno05222) | N.S. | N.S. | N.S. | N.S. | 0.000733 | N.S. | 0.0236 | ||
| Glycosphingolipid biosynthesis - ganglio series (rno00604) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0482 | 0.0274 | ||
| GABAergic synapse (rno04727) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0228 | 0.0279 | ||
| Cholinergic synapse (rno04725) | N.S. | N.S. | N.S. | N.S. | N.S. | 0.0175 | 0.0359 | ||
| Pentose phosphate pathway (rno00030) | 0.0232 | N.S. | N.S. | 0.0334 | N.S. | 0.0141 | N.S. | ||
| Complement and coagulation cascades (rno04610) | N.S. | N.S. | N.S. | N.S. | 0.0267 | 0.0178 | N.S. | ||
| Melanogenesis (rno04916) | 0.0000969 | N.S. | N.S. | N.S. | N.S. | 0.0209 | N.S. | ||
| Ovarian Steroidogenesis (rno04913) | 0.00762 | N.S. | N.S. | N.S. | N.S. | 0.0376 | N.S. | ||
| Metabolic pathways(rno01100) | 0.00117 | 0.0000295 | 0.0141 | N.S. | 0.00417 | N.S. | N.S. | ||
| Glycosaminoglycan biosynthesis - keratan sulfate (rno00533) | 0.00235 | N.S. | N.S. | N.S. | 0.0128 | N.S. | N.S. | ||
| Hedgehog signaling pathway(rno04340) | 0.000145 | N.S. | N.S. | N.S. | 0.0227 | N.S. | N.S. | ||
| Glycosphingolipid biosynthesis - lacto and neolacto series (rno00601) | N.S. | N.S. | N.S. | N.S. | 0.0233 | N.S. | N.S. | ||
| Arachidonic acid metabolism (rno00590) | 0.00397 | N.S. | N.S. | N.S. | 0.048 | N.S. | N.S. | ||
| Sphingolipid metabolism (rno00600) | N.S. | 0.0124 | N.S. | 0.0429 | N.S. | N.S. | N.S. | ||
| Starch and sucrose metabolism (rno00500) | 0.0426 | 0.0188 | 0.0233 | 0.00794 | N.S. | N.S. | N.S. | ||
| Aldosterone-regulated sodium reabsorption (rno04960) | 0.0413 | 0.0197 | N.S. | N.S. | N.S. | N.S. | N.S. | ||
| ErbB signaling pathway (rno04012) | 0.0248 | 0.02 | N.S. | N.S. | N.S. | N.S. | N.S. | ||
| Steroid hormone biosynthesis (rno00140) | 0.00397 | 0.0248 | N.S. | N.S. | N.S. | N.S. | N.S. | ||
| Serotonergic synapse (rno04726) | 0.0107 | 0.0307 | N.S. | N.S. | N.S. | N.S. | N.S. | ||
N.S.: No significance. The significant differences are shown orange-highlighted. PND 30: the comparison between untreated vs EB-treated groups via Student t-tests. PNDs 100 and 145: the comparison between CTRL vs T+E only, EB only, or EB+T+E groups via Student t-test. Sorted by p-Value in the EB+T+E group on PND145 as primary.
The higher statistical significances were observed in the cytokine-cytokine receptor interaction (rno04060) pathway in the T+E only and EB+T+E groups on PND 100 and all treated groups on PND 145 (i.e., those with p-values < 0.05). Also, they were observed in the chemokine signaling (rno0462) pathway related to the immune system in the EB only and EB+T+E groups. The KEGG pathways were specifically altered by the EB-treated group on PND 30 (Table 2). Interestingly, the estrogen signaling and Wnt signaling pathways were significantly altered on only the EB-treated group on PND 30. Further, Table 3 shows the pathways affected by exposure to only T and E by comparing the controls on PND 30 vs. the T+E only group on PNDs 100 and 145. Of the top 20, the pathways related to the cell proliferation/growth (e.g., cell cycle and DNA replication) were identified. The KEGG pathways that were specifically altered in the T+E only, EB only, or EB+T+E on PNDs 100 or 145 were examined (Supplemental Tables 5 and 6).
| Map title | Category | PND 30 |
|---|---|---|
| Fisher P value | ||
| Basal cell carcinoma(rno05217) | Cancers/Human Diseases | 2.57E-05 |
| Hippo signaling pathway(rno04390) | Signal transduction/Environmental Information Processing | 0.000113 |
| Retinol metabolism(rno00830) | Metabolism of cofactors and vitamins/Metabolism | 0.000401 |
| Acute myeloid leukemia(rno05221) | Cancers/Human Diseases | 0.0026 |
| Estrogen signaling pathway(rno04915) | Endocrine system/Organismal Systems | 0.00279 |
| Oxytocin signaling pathway(rno04921) | Endocrine system/Organismal Systems | 0.00465 |
| Morphine addiction(rno05032) | Substance dependence/Human Diseases | 0.00469 |
| Calcium signaling pathway(rno04020) | Signal transduction/Environmental Information Processing | 0.0049 |
| Gastric acid secretion(rno04971) | Digestive system/Organismal Systems | 0.00496 |
| Collecting duct acid secretion(rno04966) | Excretory system/Organismal Systems | 0.00569 |
| Circadian entrainment(rno04713) | Environmental adaptation/Organismal Systems | 0.00675 |
| Wnt signaling pathway(rno04310) | Signal transduction/Environmental Information Processing | 0.00753 |
| Chemical carcinogenesis(rno05204) | Cancers/Human Diseases | 0.00786 |
| Lysine degradation(rno00310) | Amino acid metabolism/Metabolism | 0.00856 |
| Vascular smooth muscle contraction(rno04270) | Circulatory system/Organismal Systems | 0.00992 |
N.S. No Significance. The significant differences are shown orange-highlighted. PND 30: the comparison between untreated vs EB-treated groups via Student t-tests.
| Map title | Category | PND 100 | PND 145 |
|---|---|---|---|
| Fisher P value | Fisher P value | ||
| Cell cycle(rno04110) | Cell growth and death/Cellular Processes | 3.79E-07 | 0.00159 |
| Circadian entrainment(rno04713) | Environmental adaptation/Organismal Systems | 0.00000197 | 0.00654 |
| Cholinergic synapse(rno04725) | Nervous system/Organismal Systems | 0.0000102 | N.S. |
| Pathways in cancer(rno05200) | Cancers/Human Diseases | 0.0000183 | 0.0463 |
| DNA replication(rno03030) | Replication and repair/Genetic Information Processing | 0.0000237 | 5.81E-08 |
| Morphine addiction(rno05032) | Substance dependence/Human Diseases | 0.0000488 | 0.0123 |
| Dopaminergic synapse(rno04728) | Nervous system/Organismal Systems | 0.000095 | 0.00692 |
| Glutamatergic synapse(rno04724) | Nervous system/Organismal Systems | 0.000145 | 0.000758 |
| Chemokine signaling pathway(rno04062) | Immune system/Organismal Systems | 0.00019 | N.S. |
| Synaptic vesicle cycle(rno04721) | Nervous system/Organismal Systems | 0.000461 | 0.00189 |
| Ras signaling pathway(rno04014) | Signal transduction/Environmental Information Processing | 0.000477 | N.S. |
| GABAergic synapse(rno04727) | Nervous system/Organismal Systems | 0.000528 | 0.0168 |
| Serotonergic synapse(rno04726) | Nervous system/Organismal Systems | 0.000683 | N.S. |
| Pancreatic cancer(rno05212) | Cancers/Human Diseases | 0.000738 | N.S. |
| MicroRNAs in cancer(rno05206) | Cancers/Human Diseases | 0.000765 | N.S. |
| Adrenergic signaling in cardiomyocytes(rno04261) | Circulatory system/Organismal Systems | 0.00133 | N.S. |
| Insulin secretion(rno04911) | Endocrine system/Organismal Systems | 0.00138 | N.S. |
| Alcoholism(rno05034) | Substance dependence/Human Diseases | 0.00172 | N.S. |
| Gap junction(rno04540) | Cellular commiunity/Cellular Processes | 0.00197 | 0.0396 |
| African trypanosomiasis(rno05143) | Infectious diseases/Human Diseases | 0.00245 | N.S. |
N.S. No Significance. The comparison between CTRL on PND 30 vs T+E only on PNDs 100 and 145 via Student t-tests.
Several of the top ten genes on the list were of unknown function in the prostate. Therefore, we selected genes from the cytokine-cytokine receptor interaction (rno04060), chemokine signaling (rno0462), and CAM (rno04514) pathways and evaluated their fuction in the prostate during various treatments (Supplemental Table 7). These genes were also included in the list of significantly altered genes in the EB+T+E group on PND 145 compared to the control group (Supplemental Table 3), or their function had already been reported in the prostate.
qPCR validation Upregulated genesThe expression of genes related to inflammation [e.g., C-C motif chemokine ligand 2 (Ccl2), C-X-C motif chemokine ligand 13 (Cxcl13), interleukin 7 receptor (Il7r), C-X-C motif chemokine receptor 6 (Cxcr6), C-C motif chemokine ligand 19 (Ccl19), C-C motif chemokine ligand 28 (Ccl28), neutrophil cytosolic factor 1 (Ncf)] was significantly higher in the EB+T+E group on PNDs 145 and 200 compared to the control group. The transcription levels of Ccl2 and Ccl19 genes were significantly higher in the EB+T+E group compared to the control group on PND 100 as well (Fig. 3A). The transcripts levels of HCK proto-oncogene, Src family tyrosine kinase (Hck) gene were significantly increased in the EB+T+E groups on PNDs 100, 145, and 200 compared to the control group (Fig. 3B). Integrin alpha 4 (Itga4) and neural cell adhesion molecule 1 (Ncam1) genes, which belong to the CAM pathway, showed interesting changes in their expression levels (Fig. 3B). The transcript levels of Itga4 gene were significantly increased in the EB+T+E groups on PNDs 145 and 200 and in the EB only on PND 145. In contrast, the transcript levels of the Ncam1 gene were lower in all groups on PND 100 and in the EB only and EB+T+E group on PND 200 compared to the control group, with a significant difference in the EB+T+E group on PND200. On PND 145, its levels were higher in all groups than in the control group.
Relative transcript levels of genes related to inflammation (Ccl2, Cxcl13, Il7r, Cxcr6, Ccl19, Ccl28, and Ncf1), proto-oncogene (Hck), and cell adhesion molecule (Itga4 and Ncam1) in the dorsolateral prostate of rats dosed postnatally with EB, T and E. Data are expressed as the mean fold change ± SEM (n = 5 per group). *p < 0.05 compared to the control group each endpoint (PND 30, 90, 100, 145 or 200) in the prostate. The relative ratios of the transcript levels in each sample were calculated by setting the values for controls at each collection time (PND 30, 100, or 145) to one.
The transcript levels of the bone morphogenetic protein 7 (Bmp7) gene were downregulated in most groups on PNDs 30, 100, 145, and 200 except PND 90. Significant decreases were observed in the EB-treated group on PND 30 and in the T+E only and EB+T+E groups on PND 100. In addition, the transcript levels of Bmp7 were lower in the T+E only and EB+T+E groups compared to the control group, in which changes in the EB+T+E group on PNDs 145 and 200 were statistically significant (Fig. 4).
Relative transcript levels of down-regulated gene (Bmp7) in dorsolateral prostate of rats dosed postnatally with EB, T and E. Data are expressed as the mean fold change ± SEM (n = 5 per group). *p < 0.05 compared to the control group each endpoint (PND 30, 90, 100, 145 or 200) in the prostate. The relative ratios of the transcript levels in each sample were calculated by setting the values for controls at each collection time (PND 30, 100, or 145) to one.
This study examined gene expression profiling at multiple time points in rat dorsolateral prostate dosed early postnatally with EB only (PND 30) and/or with T+E in adult age to determine what effect exposure to EB and/or T and E had on gene expression profiles in the rat prostate. Most of the DEGs and pathways in the dorsolateral prostate were identified in the EB+T+E group on PND 145; in addition, some pathways (i.e., estrogen signaling and Wnt signaling) were altered on PND 30 (EB exposure) specifically and did not overlap the pathways after T and E exposure. These results are consistent with other researchers’ findings (Huang et al., 2004, 2009; Ho et al., 2011).
DEGs induced by EB exposure had few overlapping genes at the EB only group on PNDs 100 and 145. This finding suggests unique DEGs on PND 30, 100, or 145 identified in this study may be caused by aging and/or different mechanisms induced by EB or T and E exposure.
Cheong et al. (2016) examined gene expression profiling in the dorsal prostate dosed neonatally with EB using microarray or RNA-seq. Of the 15 differentially methylated genes they identified, only progestin and adipoQ Receptor family member 4 gene were identified in the EB+T+E group on PND 145. There was a lack of concurrence of gene expression profiling between studies despite folling their experimental design (animal strain, dosing chemical, and concentration). This discrepancy may be caused by difference in collection points (PND 30 vs PND 90) and/or prostate lobes (dorsolateral vs dorsal prostate). This study did not examine gene expression profiling using rat prostates on PND 90. Further analyses are necessary to clarify this matter.
This research selected and used known-function genes in prostates for qPCR validation. The genes (Ccl2, Ccl19, Ccl28, Cxcl13, and Cxcr6) are known to be inflammatory chemokines that work during injury (Mélik-Parsadaniantz and Rostène, 2008). The proteins encoded by the Cxcl13 gene are expressed in prostate cancer cell lines, neoplasia, and inflamed tissues and are thought to be related to prostate cancer development. The Cxcr6 gene is related to tumor growth and invasion in prostate cancer cells (Salazar et al., 2013). NFκB signaling regulates Cxcl13 and Ccl19 in prostate cancer and monocyte-derived dendritic cells (Pietilä et al., 2007; Garg et al., 2017). NFκB also regulates Il7r gene expression stably in naïve T-cellsthathave an anti-inflammatory role (Miller et al., 2014).
The proteins encoded by the Hck gene belong to the Src family of tyrosine kinases (Summy and Gallick, 2006; Tatarov and Edwards, 2007). Interactions between the Src family kinases and proteins in the MAPK and PI3K pathways may induce cell proliferation. The Src family is thought to be related to prostate cancer development or tumor growth (Fizazi, 2007; Varkaris et al., 2014).
The protein encoded by the Bmp7 gene is a member of the TGF-β superfamily and plays a role in prostate development. Expression of BMP7 was highest in the normal prostate glandular tissue, whereas it tended to be lower during the development and progression of prostate cancer (Masuda et al., 2004; Morrissey et al., 2010). BMP7 is also known to have an anti-inflammation function (Li et al., 2015; Tsujimura et al., 2016).
The genes related to CAMs have an important function in prostate development and cancer development (Hsieh, 2001). The NCAM protein is related to prostate cancer invasion (Li et al., 2003); the transcript levels of Igta4 gene were reported to be downregulated in prostate cancer (Saramäki et al., 2006).
Comparing the DEGs obtained via microarray in the EB+T+E group on PND 145 to the EB only group in various time points (PND 30, 100 or 145; Supplemental Tables 2–4), showed that fewer genes were overlapped in the EB only group in the various time points (FC > 5.0, P < 0.01) (Supplemental Fig. 2). The Ccl28 gene was overlapped when compared between EB+T+E group on PND 145 vs EB only PND 100, but there were no observed statistically significancant variations in transcription levels of Ccl28 gene in the EB only group at PND 100.
Of the genes (Ccl2, Ccl19, Ccl28, Cxcl13, Cxcr6, Il7r, Ncf1, Hck, Ncam1, and Itga4) selected for qPCR validation, all except Ccl2 and Ncam1 genes were significancantly elevated in the EB+T+E group on PNDs 145 and 200 although the transcript levels of these genes were not increased in the T+E group each point.
These findings suggest, up-regulation of the Ccl19, Cxcl13, Cxcr6, Il7r, Ncf1, Hck, and Itga4 genes requires both early neonatal exposure to EB and by T+E exposure in adult age.
However, regarding the Ccl2, Ccl19, Ccl28, Cxcl13, Cxcr6, Il7r, Ncf1, and Hck genes, Prins, (1992) and Bosland et al. (1995) reported that due to inflammation mononuclear cells invaded into the prostate of rats dosed with EB or T and E. The elevated expression levels of these cytokineor chemokine-related genes may be caused by the prostate tissues, including mononuclear cells. Some genes were reported to have changes of expression in prostate cancer tissues. Further studies, including analyses of protein expression and their localization, are necessary to confirm qPCR results.
Figure 5 summarizes this study. The EB exposure on PNDs 1, 3, and 5 significantly changed the estrogen receptor and Wnt signaling pathways, specifically. In addition, the CAM pathway was altered on PND 30. After T and E exposure in adult age, cytokine-cytokine receptor interaction/chemokine signaling pathways were altered, as were CAMs. Itga4 and Hck genes and the genes related to cytokine-cytokine receptor interaction/chemokine signaling pathways (Ccl2, Ccl19, Ccl28, Cxcl13, Cxcr6, Il7r, and Ncf1) were altered after T and E exposure. The expression of proto-oncogene (e.g., Hck gene), which belongs to the Src superfamily, was increased after T and E exposure. We have also identified the Bmp7 gene, which has a function of anti-inflammation. The transcript levels of Bmp7gene are downregulated via EB and/or T and E exposure.
Summary of gene and pathway changes in the dorsolateral prostate of rats dosed with EB and/or T and E. EB exposure on PNDs 1, 3 and 5 changed estrogen receptor/Wnt signaling pathways specifically on PND 30. Cell adhesion molecule pathways were significantly altered by early neonatal exposure to EB on PND 30 and T and E exposure on PNDs 100 and 145. T and E exposure significantly altered cytokine-cytokine receptor interaction /chemokine signaling. The proto-oncogenes (Hck) and genes that related to chemokines and cytokine signaling (Cxcl13, Il7r, Ccl2) were increased. Bmp7 gene were down-regulated by neonatal EB exposure only and exposure to EB, T and E. Red arrows show an increased activity; green arrows show down-activity of the genes and pathways. Dotted lines indicate possible interactions. T: testosterone, E: estrogen, EB: estradiol benzoate.
Interestingly, the genes identified in this study were previously identified as expressed in prostate cancer cell lines or human prostate cancer. However, further studies are needed in determining whether the change in gene expression levels are related to inflammation caused by mononuclear cell invasion.
In conclusions, this study found that most of the differentially expressed genes/pathways induced by EB and/or T and E exposure were unique at different collection points (PND 30, 100, or 145). Some pathways changes were time-specific occuring at PND 30, 100, or 145. The gene (Bmp7) related to anti-inflammation was downregulated by both neonatal EB exposure and later exposure to EB and T and E. The upregulated genes (Itga4, Hck, Ccl2, Ccl19, Ccl28, Cxcl13, Cxcr6, Il7r, and Ncf1) occured after T and E exposure in the prostate of rats dosed with EB on PNDs 1, 3, and 5. Additional studies are required to determine whether the changes observed in the rat prostates correspond with human prostate samples and whether they are specific to only the prostate.
The authors thank Drs. Richard Beger, Tao Chen, and Pierre Alusta for their suggestions on the manuscript, and Dr. John K. Leighton (CDER/FDA) for his suggestion on this project proposal and manuscript. The authors also recognize the contributions of Dr. Tao Han for supervising microarray procedure and Dr. Davis Kelly (TPA) for histological evaluation of prostates. Special thanks are given to Drs. Guangxu Zhou (DBB), Feng Qian (DBB), and Ryan Curtis (IT) for assisting with the analysis using ArrayTrack software, and Dr. Paul Rogers (DBB) for helping log-transformation of qPCR data. This study was funded by the NCTR/FDA (E0759701).The views expressed are those of the authors and do not represent the views of the Food and Drug Administration.
Conflict of interestThe authors declare that there is no conflict of interest.
