Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: Mol Carcinog. 2021 Apr 13;60(6):391–402. doi: 10.1002/mc.23299

DNA methylome, transcriptome and prostate cancer prevention by phenethyl isothiocyanate (PEITC) in TRAMP mice

Renyi Wu 1,#, Shanyi Li 1,#, Davit Sargsyan 1, Ran Yin 1, Hsiao-Chen Kuo 1, Rebecca Peter 1, Lujing Wang 1, Rasika Hudlikar 1, Xia Liu 1, Ah-Ng Kong 1,*
PMCID: PMC8201649  NIHMSID: NIHMS1702237  PMID: 33848375

Abstract

Epigenetics/epigenomics has been shown to be involved in carcinogenesis. However, how the epigenome would be altered in TRAMP prostate cancer (PCa) model and the effect of cancer chemopreventive phytochemical phenethyl isothiocyanate (PEITC) on the epigenome in TRAMP mice are not known. PEITC has been reported to reduce the risk of many cancers including PCa. In this study, male TRAMP mice were fed control diet or diet containing 0.05% PEITC from 8 weeks to 16 weeks. The tumor incidence was reduced in PEITC diet (0/6) as compared to control diet (6/7). RNA-seq analyses on non-tumor and tumor prostatic tissues revealed several pathways like cell cycle/Cdc42 signaling, inflammation and cancer-related signaling, were activated in prostate tissues of TRAMP mice but were reversed or attenuated in TRAMP mice fed with PEITC diet. DNA CpG methyl-seq analyses showed that global methylation patterns of prostate samples from TRAMP mice were hugely different from those of wildtype mice. Dietary PEITC partially reversed the global methylation changes during prostatic carcinogenesis. Integration of RNA-seq and DNA methyl-seq analyses identified a list of genes, including Adgrb1, and Ebf4, with inverse regulatory relationship between their RNA expression and CpG methylation. In summary, our current study demonstrates that alteration of global epigenome in TRAMP prostate tumor and PEITC administration suppresses PCa carcinogenesis, impacts global CpG epigenome and transcriptome, and attenuates carcinogenic pathways like cell cycle arrest and inflammation. These results may provide insights and epigenetic markers/targets for PCa prevention and treatment in human PCa patients.

Keywords: PEITC, Prostate cancer, epigenetics, DNA methylation, TRAMP

1. Introduction

Prostate cancer (PCa) is the most frequently diagnosed form of cancer and the second leading cancer-related death among men in the USA 1. According to statistics and prediction data from National Cancer Institute, one in 6 American men will be diagnosed with PCa in their lifetime. Currently, the is no cure for metastatic PCa and the median survival time is 2~3 years. PCa a multi-step molecular pathogenesis including genetic and epigenetic changes that alter the cellular balance in proliferation, apoptosis, differentiation, senescence, among others. It initially responds well to hormone ablation but eventually becomes hormone resistant 2. The development and formation of prostate carcinoma is a long process. Progression of prostate carcinoma from prostatic intraepithelial neoplasia (PIN) to androgen-independent invasive carcinoma can take years to decades. Therefore, to decrease the incidence of advance PCa, to delay the neoplastic development, or to slow the progression of prostate cancer, it is logical to intervene with long-term prevention strategy using relatively nontoxic dietary cancer chemopreventive agents.

Many studies on the chemopreventive properties of various plant source bioactive compounds indicate that cruciferous vegetables have superior chemopreventive potentials over those of other plant families 3. The plants of Cruciferae (a.k.a. Brassicaceae) family include broccoli, horseradish, cabbage, cauliflower, and watercress. Cruciferous vegetables are rich in glucosinolates, which can be further degraded to isothiocyanates (ITCs) such as sulforaphane and phenethyl isothiocyanate (PEITC) during food preparation, cooking, chewing, and intestinal microbial metabolism 4,5. Numerous epidemiological studies have shown that dietary consumption of cruciferous vegetables is inversely related to the risk of developing various cancers 6. Mechanistic studies have suggested that remarkable chemopreventive activities of ITCs stem from their biological effects such as apoptosis via the caspases pathways 7,8 or the p53-dependent pathway 9–11, detoxification of carcinogens (Phase II enzyme activation), blocking carcinogen activation (Phase I enzyme inhibition), inhibition of the IKK-NF-kB signaling pathway, and inducing cell cycle arrest 12,13. Many studies have shown that PEITC exerts promising anti-cancer properties in PCa cell lines and xenograft and transgenic animal models of PCa 13,14. More recently, emerging evidence has linked its anti-cancer effects to its ability to modulate epigenetic modifications, including regulation of both histone deacetylases (HDACs) and DNA methyltransferases (DNMTs), which may play critical roles in carcinogenesis.

Genetically engineered mouse models have emerged as powerful tools in understanding many aspects of cancer and have provided opportunities for studying cancer preventive effects of phytochemicals at various stages of disease progression 15,16. The transgenic adenocarcinoma of the mouse prostate (TRAMP) model was developed on the C57BL/6 mouse using minimal rat probasin promoter to drive expression of the SV40 early genes specifically in the epithelial cells of the dorsolateral and ventral lobes of the prostate. This model develops prostate cancer in a typical fashion, from a neoplastic intraepithelial lesion (prostatic intraepithelial neoplasia; PIN), to invasive carcinoma that mimics the various stages in human prostate carcinogenesis 17–19. Since this model recapitulates many salient features of the progressive forms of human PCa, we evaluated the cancer prevention effects of PEITC in this model, focusing on the epigenome and transcriptome changes following PEITC dietary administration at various stages of PCa development. Our results demonstrated that PEITC could prevent tumor formation by regulating the epigenome and transcriptome resulting in a series of signaling pathways involving cell cycle arrest and inflammation. We also identified a list of genes whose regulation of RNA expression in prostate carcinogenesis was inversely related to their regulation of DNA CpG methylation.

2. Materials and Methods

2.1. Animals and sample preparation

Wildtype male C57BL/6 mice and female hemizygous for the TRAMP transgene line probasin (PB) Tag 8247NG, were purchased from the Jackson Laboratory (Bar Harbor, ME), as described in previous studies 20,21. The mice were bred on the same genetic background and housed at the Rutgers Animal Facility, maintained under 12 hours light and dark cycles, and provided ad libitum access to food and water. The F1 offspring were genotyped by PCR and agarose gel electrophoresis on the genomic DNA. Male hemizygous for the TRAMP genes were used for the TRAMP groups and littermate wildtype males were used for the control group. After weaning, mice were randomly divided and housed in the animal facility for at least one week for acclimation. Starting at the age of 8 weeks, wildtype mice were fed control AIN93M diet and TRAMP mice were fed AIN93M diet with or without 0.05% PEITC. At the age of 16 weeks and 24 weeks, mice were sacrificed, and prostate samples were measured and saved for future analysis. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC; protocol number: PROTO999900171) of Rutgers University.

At the designated time points, mice were sacrificed followed by immediately extracting the solid tumors and the ventral, and dorsolateral lobes of the prostate. For RNA-seq and DNA CpG Methyl-Seq experiments, 2 biological replicates of the tissues were used. RNA and DNA extractions were performed using the AllPrep DNA/RNA Mini Kit (QIAGEN Cat. No. 80204) as described previously 22–26.

2.2. DNA CpG Methyl-Seq library preparation

DNA CpG Methyl-seq library preparation was performed using Agilent SureSelect Methyl-Seq kit (Cat. No. G9651A) as described in the manufacturer’s protocol (Methyl-seq protocol, Version X, August 2017). 3 μg of DNA was used in the library preparation. 550 ng of adaptor ligated DNA was used for the hybridization capture and the final concentration of indexed library was 8 to 15 nM. The following changes were made to minimize the loss of DNA in the process of enzymatic reactions: AMPure XP beads (Cat. No. A63880, Beckman Coulter, USA) were incubated with DNA reaction mix for 10 min at room temperature prior to pelleting by magnetization. AMPure XP beads were then washed twice with 80% ethanol and dried at 37°C for 5 min. DNA was dissolved for 10 min at 37°C. AMPure beads were retained in the solution for adenylation and end-repair reaction. An equal volume of binding buffer was added to this reaction mix at 1:1 ratio to enable the AMPure XP beads to rebind to DNA, incubated at room temperature for 10 min. DNA was further purified as described above. Concentration and size of DNA fragments were determined by Agilent Bioanalyzer 2100.

2.3. Bisulfite conversion and next-generation sequencing

Bisulfite conversion was performed using EZ DNA Methylation-Gold kit (Zymo Research, USA) as described in the manufacturer’s protocol. Sequencing was performed on an Illumina HiSeq 2500 platform with 150 bp paired end reads as described in the manufacturer’s protocol.

2.4. Bioinformatics analyses of SureSelect Methyl-seq

Bioinformatics analyses of CpG Methyl-seq and RNA-seq were performed as we have reported previously 24,27. The reads were aligned to the in silico C-T converted mouse genome (mm10) with the Bismark (version 0.15.0) alignment algorithm 28. After alignment, DMRfinder (version 0.1) was used to extract methylation counts and cluster CpG sites into DMRs 29. Each DMR contains at least three CpG sites. Methylation differences greater than 0.10 and with a P value smaller than 0.05 were considered significant. Genomic annotation was performed with ChIPseeker (version 1.10.3) in R (version 3.4.0) 30.

2.5. RNA extraction, library preparation, and next-generation sequencing

Total RNA was extracted from snap-frozen tissue and/or tumor samples from the control and TRAMP groups using the AllPrep DND/RNA Mini Kit (Qiagen, Valencia, CA, USA). The quality and quantity of the extracted RNA samples were determined with an Agilent 2100 Bioanalyzer. The library was constructed using the Illumina TruSeq RNA preparation kit (Illumina, San Diego, CA, USA) according to the manufacturer’s manual. Samples were sequenced on an Illumina HiSeq 2500 instrument with 150 bp paired-end reads, to a minimum depth of 40 million reads per sample.

2.6. Quantitative polymerase chain reaction (qPCR)

The mRNA expression of selected genes for selected mouse samples was determined using qPCR analysis. First-strand cDNA was synthesized from 500 ng of total RNA using TaqMan® Reverse Transcription reagents (ThermoFisher, Waltham, MA, USA) and then subjected to qPCR using a QuantStudio 5 Real-Time PCR System (ThermoFisher) with SYBR Green PCR Master Mix (ThermoFisher). Primers were designed by and purchased from Integrated DNA Technologies (IDT, Coralville, Iowa, USA) and their sequences are: Arhgap40 (forward), AGGAGACACCGAATCCCAGT; Arhgap40 (reverse), CCTTGACGTGAAGACCCCA; Ebf4 (forward), TCATCGACTTCGTGGAAAAGG; Ebf4 (reverse), GGCGCACGTAGAGGTCTTG; Kcng4 (forward), AGTGCTATTACATCTTCGTGGTG; Kcng4 (reverse), GCGACACATAGTATGGGGAGA; Papln (forward), CAGGGTCTTGGGCTCGAAATG; Paplb (reverse), GTCTCTCCTCTGGGAGTAGCA; The mRNA expression was calculated as fold change with normalization to the expression of Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) using the ΔΔCT method.

2.7. Computational analyses of RNA-seq data

The reads were aligned to the mouse genome (mm10) with Hisat2 31. Reads were deduplicated and mapped to genomic features with featureCounts. Raw counts were suppled to R/Bioconductor package DESeq2 for deferentially expression analysis 32.

2.8. Ingenuity pathway analysis (IPA)

Genes that exhibited a log2 fold change greater than 1 and a P value less than 0.05 were subjected to Ingenuity Pathway Analysis (IPA 4.0, Ingenuity Systems, www.Ingenuity.com). The input genes were mapped to IPA’s knowledge base, and the relevant biological functions, networks, and pathways related to the TRAMP and PEITC treatment were identified.

2.9. Statistical analysis

For body weight measurement, the data are presented as means with upper and lower limits. For significance comparisons of genitourinary apparatus among multiple groups, Student’s t-test was used. For RNA-seq and methyl-seq data analysis, a P value less than 0.05 was considered statistically significant unless otherwise indicated.

3. Results

3.1. PEITC suppresses prostate carcinogenesis in TRAMP mice.

The animal study was carried out according to the scheme shown in Fig. 1a. 7-week-old male wildtype C57BL6J mice or TRAMP mice on the same background were divided into three groups: Wildtype, TRAMP, and TRAMP+PEITC. Starting at week 8, mice in the first two groups were fed AIN93M control diet while mice in the last group were fed AIN93M diet with 0.05% PEITC (w/w). Body weight were measured weekly and animals in the TRAMP+PEITC group had a slightly smaller average body weight than the other two groups (Figure 1b). Mice were then sacrificed at designated time points. Number of mice for each group and time point is shown in Figure 1c. At week 16, no mice in any group had developed palpable tumors. At week 24, 6 out 7 mice (85.7%) in the TRAMP group had developed tumors of various size (Figure 1c). The wet weight of the genitourinary apparatus for mice in the TRAMP group was significantly greater than that in the other two groups (Figure 1d). The genitourinary apparatus with tumor of the 6 mice in the TRAMP group (week 24) are shown in Figure 1e. The genitourinary apparatus of mice from other groups or time points had similar normal appearance and are not shown.

Figure 1.

Figure 1.

Open in a new tab

PEITC diet blocked prostate carcinogenesis in TRAMP mice. (a) Experimental design of the animal study. The TRAMP transgenic model was generated on a C57BL6 background and is purchased from Jackson Laboratory. Hemizygotes were used in TRAMP and TRAMP+PEITC groups and non-carriers were used in the Wildtype group. Food was provided ad libitum. (b) Average body weight of the mice. Mice were individually weighed weekly. The error bars denote upper and lower limits of measurements. (c) Prostate tumor incidences of the three groups at two time points (16 versus 24 weeks). Only the TRAMP group at week 24 had noticeable tumors at the time of sacrifice. (d) Weight of genitourinary apparatus of the mice at week 24. The measuring was performed right after the wet genitourinary apparatus was dissected from the body. Genitourinary apparatus weights at week 24 were not significantly different and are not shown here. (e) Representative images of genitourinary apparatus and/or prostate tumors from mice in the TRAMP group at week 24. Mice in other groups or at other time point did not develop solid tumors and are not shown here. Ruler unit is in centimeter (cm).

3.2. Gene expression changes in prostate carcinogenesis of TRAMP mice and the effects of PEITC diet on prostate carcinogenesis.

To identify the transcriptomic profile of TRAMP mice and the effects of PEITC on the transcriptome profile, RNA-seq was performed on the normal prostate samples versus prostate tumor samples for comparison. For normal prostate samples, both ventral and dorsolateral lobes were used as they have the characteristics of the TRAMP transgenic model and are more likely to develop tumor 19,33. Principal component analysis (PCA) reveals that the tumor sample is clustered separately from normal tissue samples, and among normal tissue samples, samples of wildtype mice are separated from those of the TRAMP mice. However, normal tissue samples of the TRAMP mice are not separated by diet or timepoint (Figure 2a). The heatmap and its dendrogram clustering in Figure 2b also shows this pattern of separation. Hence, for the purpose of comparing transcriptional profiles, samples from the two different time points were pooled within each group and the comparisons were performed between normal tissue samples (excluding the tumor sample). When comparing RNA expression of the TRAMP (control diet) group to the Wildtype (control diet) group, a list of 2,598 genes was obtained with at least 2-fold change in RNA expression and with P value less than 0.05. While comparing of TRAMP (PEITC diet) to TRAMP (control diet), only 131 genes met the same criteria (Figure 2c). There are 56 genes in both comparisons and their relative RNA expression levels are shown in a heatmap in Figure 2d. Most of the overlapping genes were downregulated in the TRAMP (control diet) group but were upregulated to some degree in the TRAMP (PEITC diet) group. There are also some genes that were upregulated in the TRAMP group but then downregulated in the TRAMP+PEITC group. The RNA expression pattern suggests that PEITC dietary administration reverses RNA expression of some pathways/genes in prostate carcinogenesis of TRAMP mice but not all the pathways/genes, as we have seen in other mouse cancer models 24,27. We also performed Ingenuity Pathway Analysis (IPA, Qiagen) with the 2,598 differentially expressed genes in TRAMP (control diet) group versus Wildtype (control diet) group, and identified many signaling pathways are altered in the TRAMP group, including GP6 Signaling Pathway, CD28 Signaling in T Helper Cells, PKCθ Signaling in T Lymphocytes, Role of NFAT in Regulation of the Immune Response, Dendritic Cell Maturation, Fcγ Receptor-mediated Phagocytosis in Macrophages and Monocytes, Cell Cycle: G2/M DNA Damage Checkpoint Regulation, and Cell Cycle: G1/S Checkpoint Regulation, among others (Table S1). Importantly, when comparing TRAMP (PEITC diet) group versus TRAMP (control diet) group, several of these signaling pathways related to cellular functions like cell cycle regulation, inflammatory NF-kB, IL-8, pancreatic adenocarcinoma, glioblastoma multiforme, and colorectal cancer metastasis were reversed by PEITC dietary administration (Figure 2e). For instance, G2/M DNA damage checkpoint regulation was suppressed in TRAMP (control diet) mice, potentially due to sequestering and inactivation of p53. However, PEITC dietary administration reactivated this pathway (activation z score of 0.333). For the pathways listed in Fig. 2e, the regulatory pathways of the involved genes are highly consistent with the predictions by the IPA knowledge base (Table S2). For instance, IPA made a determination based on 27 genes that the pathway Pancreatic Adenocarcinoma Signaling was activated in the TRAMP group versus the Wildtype group. 22 of the 27 genes (81%) were regulated (up or down) in a direction that is consistent with activation of the pathway, while only 5 of the 27 genes were regulated in a direction that is not consistent with the activation of the pathway.

Figure 2.

Figure 2.

Open in a new tab

Gene expression profiles and cellular pathway and function analysis. (a) Principal Component Analysis (PCA) on the RNA-seq samples. TRAMP mice, regardless of diet, are clustered together but distant from the wildtype mice, suggesting genotype has prevailing effects over diet. (b) Dendrogram clustering by Euclidean distance and heatmap showing top 2000 expressed genes among all samples. The dendrogram shows that the samples are first clustered on tissue type (tumor vs adjacent tissue) and then by genotype, which is consistent with the PCA in Figure A. (c) Venn Diagram showing numbers of significantly expressed genes with at least 2-fold change in the comparisons of TRAMP vs Wildtype and TRAMP+PEITC vs TRAMP. Consistent with Figures A and B, only a small number of genes were regulated by PEITC diet administration. P < 0.05 is considered to be significant. (d) Heatmap showing the RNA expression of the 56 overlapping genes as defined in Figure 2c. FPKM values were first log transformed then mean-centered on row. (e) Top 9 regulated pathways in TRAMP compared to Wildtype. Pathways were identified by Ingenuity Pathway Analysis (IPA, Qiagen) with a list of 2,598 regulated genes as discussed in Figure 2c.

3.3. DNA methylation alterations in prostate carcinogenesis of TRAMP mice and reverse effects of PEITC dietary administration.

To identify DNA CpG methylation changes in prostate carcinogenesis of TRAMP mice and how PEITC dietary administration would affect the CpG methylation profile of TRAMP mice, we performed single base-pair resolution methyl-seq on the same samples as with the RNA-seq above. The genomic DNA samples were subjected to Agilent SureSelect Mouse Methyl-seq enrichment and library preparation then sequenced on an Illumina HiSeq 2000 platform. Sequencing reads were aligned to in silico C-T converted mouse genome (mm10) and deduplicated. Individual CpG sites were clustered into DNA methylation regions (DMRs) according to the default settings in the DMRfinder package 29. Specifically, each DMR has at least three CpG sites and a maximum length of 500 bp with no more than 100 bp between any two CpG sites. Average methylation ratio for each DMR was calculated as aggregated counts of 5-mC (as C in bisulfite converted sequencing) divided by the aggregated counts of 5-mC and C (as both C and T in bisulfite converted sequencing) for all CpG dinucleotides in that DMR. We then collected DNA methylation data for all 12 samples with a total of 177,552 DMRs. These DMRs were further annotated with gene features using ChIPseeker (v1.15). As shown in Figure 3a, most of the DMRs are located in the promoter or distal intergenic regions (> 3 kb upstream transcription start site [TSS] or downstream 3’ untranslated region [UTR]). Principal component analysis (PCA) shows that, similar to the clustering of the RNA-seq samples, TRAMP samples and wildtype samples are clustered separately, and the effects of aging were minimal. We also identified an outlier from the TRAMP (control diet) group at week 16, which was excluded from further analysis (Figure 3b). Heatmap and dendrogram clustering by Euclidean distance also show the similar findings for these samples (Figure 3c). Hence, as in the RNA-seq analysis, samples from the 2 different time points were pooled within each group. Next, pairwise comparison of DNA methylation between any two of the three groups were performed. In comparison of TRAMP (control diet) vs Wildtype (control diet), several DMRs were identified (Figure 4a). However, in the comparison of TRAMP (PEITC diet) vs TRAMP (control diet), there were only a small number of significant differences of DMRs between the 2 groups (Figure 4b). On the other hand, in the comparison of TRAMP (PEITC diet) vs Wildtype (control diet), there are a high number of significant DMRs (Figure 4c). When a cutoff of 0.1 was used for DNA methylation ratio and 0.05 for P values, the numbers of DMRs in the promoter region in the comparisons of TRAMP (control diet) vs Wildtype (control diet) and TRAMP (PEITC diet) vs TRAMP (control diet), 5,540 DMRs in the first comparison and only 732 in the second comparison, respectively. The number of overlapping DMRs in both comparisons was 253 (Figure 4d). When examining the DMRs at all genomic locations except distal regions of the genes, the number of overlapping DMRs increased from 253 to 878. Interestingly, for these DMRs, the methylation pattern for the TRAMP (PEITC diet) group was similar to, that of the Wildtype (control diet) group, suggesting PEITC dietary administration could reverse the methylation alterations during prostate carcinogenesis of TRAMP mice. The absolute methylation ratios in all three groups and their methylation changes in the two comparisons are shown as heatmaps in Figures 4e and 4f.

Figure 3.

Figure 3.

Open in a new tab

DNA methylation profiles of all samples in the study. (a) Distribution of annotated differentially methylated regions (DMRs) by gene feature. Each DMR has at least 3 CpG sites. (b) Principal component analysis (PCA) on the methylation profiles of the samples. One sample from the TRAMP group (week 16), shown on the upper right corner of the PCA plot, was an outlier and was excluded from further analyses. (c) Dendrogram clustering by Euclidean distance and heatmap showing the methylation profiles of all samples. In contrast to the gene expression profile of these samples, the methylation profiles were first clustered by genotype (TRAMP vs wildtype) and then by tissue type (tumor vs adjacent tissue), while diet and time point were not major contributing factors of clustering. Consistent with results from Figure 3b PCA plot, the same outlier was also identified in the dendrogram clustering (the leftmost column).

Figure 4.

Figure 4.

Open in a new tab

DNA methylation alteration in TRAMP mice and by PEITC diet administration. (a-c) MA plots showing methylation changes between any two of the three groups. Color highlighting indicates significant change with P value smaller than 0.05. Only a small number of DMRs are in the TRAMP group as compared to the Wildtype group. (d) Venn Diagram showing numbers of significantly changed DMRs in promoter region in comparisons of TRAMP vs Wildtype and TRAMP+PEITC vs TRAMP. (e) heatmap showing absolute methylation ratio of 878 DMRs in promoter region and gene body for all three groups. (f) Heatmap showing methylation differences of the same 878 DMRs for the comparisons of TRAMP vs Wildtype and TRAMP+PEITC vs TRAMP. Most DMRs had reversal methylation patterns by PEITC diet administration. For d, e, and f, the cutoff was 0.1 for methylation ratios and 0.05 for P values.

3.4. Correlations between alterations of DNA methylation versus RNA expression in TRAMP mice and in PEITC dietary administration

We next combined DNA methylation profiles with RNA expression profiles in the comparisons of TRAMP vs Wildtype and TRAMP+PEITC vs TRAMP and obtained 4,583 and 185 DMRs, respectively. These DMRs had at least methylation ratio change of 0.1 and at least 2-fold change in their corresponding RNA expression, as shown in Figures 5a and 5b. In these two figures, each dot represents a DMR and its genomic features is indicated by a unique color. We next focused on the DMRs and genes that had an inverse regulation between TRAMP group and TRAMP+PEITC group, and inverse correlation between DNA methylation and RNA expression changes. After filtering, we obtained a list of 28 genes. The RNA-seq data and Methyl-seq data of these genes for each group are shown in Figure 5c. Of these 28 genes, 7 have increased RNA expression in the TRAMP group with corresponding decreased DNA CpG methylation and both RNA expression and DNA CpG methylation were reversed by PEITC administration. The other 21 genes have decreased RNA expression in the TRAMP group with corresponding increased methylation and both RNA expression and DNA methylation were also reversed by PEITC administration. All these genes have the inverse relationship between RNA expression and DNA methylation, which would be consistent with the current dogma of suppressive effect of DNA CpG methylation on gene expression. We performed qPCR for 4 genes: Arhgap40, Ebf4, kcnq4, and Papln, and the qPCR results validate the same RNA expression change as by RNA-seq analysis.

Figure 5.

Figure 5.

Open in a new tab

Correlations between RNA expression and DNA methylation. (a-b) Scatter plots showing DMRs with cutoff of 0.1 for DNA methylation ratio and 2-fold change for RNA expression in the comparisons of TRAMP vs Wildtype and TRAMP+PEITC vs TRAMP. DMR locations (gene features) are indicated by colors. (c) A list of genes that were differentially expressed and methylated in prostate carcinogenesis of TRAMP mice and were also reversed by PEITC diet administration. Absolute methylation radio is shown on the left Y-axis and RNA expression (in FPKM) is shown on the right Y-axis. (d) qPCR validation of 4 of the 28 genes in c.

4. Discussion.

Previous work from our laboratory has shown that several dietary phytochemicals including curcumin, tocopherols, and PEITC significantly decrease the incidence of prostatic tumorigenesis and inhibit high-grade PIN in TRAMP mice via regulation of a series of cellular pathways including proliferation, apoptosis, cell cycle arrest, Akt signaling, and Nrf2/ARE-mediated antioxidative stress and detoxification 20,34–40. However, how the epigenome would be impacted during TRAMP PCa tumorigenesis and importantly how PEITC would affect the epigenome remain unknown. Here, we focused on exploring the transcriptomic and epigenomic changes in prostate tumorigenesis of TRAMP mice and how cancer chemopreventive agent PEITC would impact or reverse these epigenomic-transcriptomic alterations. The major advantage of this transgenic model is that tumors arise from normal prostate epithelial cells in their natural tissue microenvironment and progress through multiple stages, exhibiting both histological and molecular features similar to those of human PCa.

It is well established that at approximately 6 weeks of age, TRAMP mice exhibit low-grade PIN which progresses to high grade PIN at 12~16 weeks of age. Focal adenocarcinoma develops in this time period and advance to poorly differentiated carcinoma before 24 weeks of age. At 28 weeks of age, all TRAMP mice would have metastasized cancer in liver, lymph nodes and lung. In the present study, we show that at 24 weeks of age, six out of seven mice had developed palpable tumors and PEITC would block, or at least, delay tumor formation at this time point. With the rapid development of NGS techniques, we were able to perform RNA-seq and DNA methyl-seq to study the transcriptome profiles of the samples and the corresponding DNA CpG methylation profiles of the genome at single base-pair resolution. With our bioinformatic analyses, we revealed that a large number of genes were differentially expressed in TRAMP mice as compared to their wildtype counterparts. Interestingly, only a small number of genes were differentially expressed when comparing TRAMP mice at different time points (week-24 vs week-16) or comparing TRAMP mice on PEITC diet to TRAMP mice on control diet, suggesting a few key genes are involved in prostate carcinogenesis of TRAMP mice and PEITC could reverse the expression of these genes. From the list of differentially expressed genes, we identified a list of key signaling pathways related to cell cycle arrest and inflammation regulation, which is consistent with our previous published findings 12,26,35,41. Furthermore, in the present study, we mapped the RNA expression of the key genes to their DNA CpG methylation status and identified a list of genes that could be epigenetically regulated by PEITC via DNA CpG methylation. These genes with inverse DNA methylation-RNA expression patterns would be a new discovery for the first time in an in vivo PCa setting.

In summary, we have utilized the latest RNA-seq and DNA Methyl-seq approaches to dissect the gene expression changes and epigenomic changes during prostate carcinogenesis of TRAMP mice and in PCa prevention by PEITC. Both RNA expression and DNA CpG methylation in TRAMP mice were drastically different from that in wildtype mice, suggesting tremendous impact of epigenetics/epigenomics during PCa development. In contrast, only slight difference was observed in both RNA transcription and DNA CpG methylation between 16 versus 24 weeks and between TRAMP (PEITC diet) mice and TRAMP (control diet) mice. RNA-seq data revealed a list of regulated pathways in prostate carcinogenesis of TRAMP mice and the top regulated ones include cell cycle regulation, inflammation, and cancer-related. A list of genes was also identified with concomitant epigenetic CpG modifications of these genes’ promoters and or gene body in prostate carcinogenesis and impacted by treatment with PEITC in TRAMP mice. Taken together, our findings yield important insights for future studies in targeting these pathways/genes for prevention and treatment of PCa in humans.

Supplementary Material

tableS1
tableS2

Acknowledgements

This study was supported in part by institutional funds and by R01 AT009152, from the National Center for Complementary and Integrative Health and the Office of Dietary Supplements, and R01 CA200129 from the National Cancer Institute.

Abbreviations

PEITC

phenethyl isothiocyanate

TRAMP

transgenic adenocarcinoma of the mouse prostate

PCa

prostate cancer

PIN

prostatic intraepithelial neoplasia

HDAC

histone deacetylase

DNMT

DNA methyltransferase

PCA

Principal component analysis

DMR

DNA methylation region

Footnotes

Conflicts of interest

The authors declare there are no conflicts of interest.

Accession codes

Transcript sequencing data and bisulfite methyl sequencing data have been deposited under Gene Expression Omnibus (GEO) accession GSE140314.

Supplementary data

There is 2 supplementary table to this manuscript.

References

  • 1.Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J Clin 2008;58(2):71–96. [DOI] [PubMed] [Google Scholar]
  • 2.Petrylak DP. The current role of chemotherapy in metastatic hormone-refractory prostate cancer. Urology. 2005;65(5 Suppl):3–7; discussion 7–8. [DOI] [PubMed] [Google Scholar]
  • 3.Verhoeven DT, Verhagen H, Goldbohm RA, van den Brandt PA, van Poppel G. A review of mechanisms underlying anticarcinogenicity by brassica vegetables. Chem Biol Interact 1997;103(2):79–129. [DOI] [PubMed] [Google Scholar]
  • 4.Tian S, Liu X, Lei P, Zhang X, Shan Y. Microbiota: a mediator to transform glucosinolate precursors in cruciferous vegetables to the active isothiocyanates. J Sci Food Agric 2018;98(4):1255–1260. [DOI] [PubMed] [Google Scholar]
  • 5.Narbad A, Rossiter JT. Gut Glucosinolate Metabolism and Isothiocyanate Production. Mol Nutr Food Res 2018;62(18):e1700991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Hecht SS. Chemoprevention of cancer by isothiocyanates, modifiers of carcinogen metabolism. J Nutr 1999;129(3):768S–774S. [DOI] [PubMed] [Google Scholar]
  • 7.Kim BR, Hu R, Keum YS, et al. Effects of glutathione on antioxidant response element-mediated gene expression and apoptosis elicited by sulforaphane. Cancer research. 2003;63(21):7520–7525. [PubMed] [Google Scholar]
  • 8.Hu R, Kim BR, Chen C, Hebbar V, Kong AN. The roles of JNK and apoptotic signaling pathways in PEITC-mediated responses in human HT-29 colon adenocarcinoma cells. Carcinogenesis. 2003;24(8):1361–1367. [DOI] [PubMed] [Google Scholar]
  • 9.Huang C, Ma WY, Li J, Hecht SS, Dong Z. Essential role of p53 in phenethyl isothiocyanate-induced apoptosis. Cancer research. 1998;58(18):4102–4106. [PubMed] [Google Scholar]
  • 10.Xu K, Thornalley PJ. Studies on the mechanism of the inhibition of human leukaemia cell growth by dietary isothiocyanates and their cysteine adducts in vitro. Biochem Pharmacol 2000;60(2):221–231. [DOI] [PubMed] [Google Scholar]
  • 11.Xu K, Thornalley PJ. Signal transduction activated by the cancer chemopreventive isothiocyanates: cleavage of BID protein, tyrosine phosphorylation and activation of JNK. Br J Cancer. 2001;84(5):670–673. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Huang Y, Wu R, Su ZY, et al. A naturally occurring mixture of tocotrienols inhibits the growth of human prostate tumor, associated with epigenetic modifications of cyclin-dependent kinase inhibitors p21 and p27. J Nutr Biochem 2017;40:155–163. [DOI] [PubMed] [Google Scholar]
  • 13.Zhang C, Shu L, Kim H, et al. Phenethyl isothiocyanate (PEITC) suppresses prostate cancer cell invasion epigenetically through regulating microRNA-194. Mol Nutr Food Res 2016;60(6):1427–1436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Sakao K, Vyas AR, Chinni SR, Amjad AI, Parikh R, Singh SV. CXCR4 is a novel target of cancer chemopreventative isothiocyanates in prostate cancer cells. Cancer Prev Res (Phila). 2015;8(5):365–374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Thyagarajan T, Totey S, Danton MJ, Kulkarni AB. Genetically altered mouse models: the good, the bad, and the ugly. Crit Rev Oral Biol Med 2003;14(3):154–174. [DOI] [PubMed] [Google Scholar]
  • 16.Weiss B, Shannon K. Mouse cancer models as a platform for performing preclinical therapeutic trials. Curr Opin Genet Dev 2003;13(1):84–89. [DOI] [PubMed] [Google Scholar]
  • 17.Gingrich JR, Barrios RJ, Morton RA, et al. Metastatic prostate cancer in a transgenic mouse. Cancer research. 1996;56(18):4096–4102. [PubMed] [Google Scholar]
  • 18.Kaplan-Lefko PJ, Chen TM, Ittmann MM, et al. Pathobiology of autochthonous prostate cancer in a pre-clinical transgenic mouse model. Prostate. 2003;55(3):219–237. [DOI] [PubMed] [Google Scholar]
  • 19.Greenberg NM, DeMayo F, Finegold MJ, et al. Prostate cancer in a transgenic mouse. Proceedings of the National Academy of Sciences of the United States of America. 1995;92(8):3439–3443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Yu S, Khor TO, Cheung KL, et al. Nrf2 expression is regulated by epigenetic mechanisms in prostate cancer of TRAMP mice. PloS one. 2010;5(1):e8579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Barve A, Khor TO, Nair S, et al. Gamma-tocopherol-enriched mixed tocopherol diet inhibits prostate carcinogenesis in TRAMP mice. Int J Cancer. 2009;124(7):1693–1699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Guo Y, Wu R, Gaspar JM, et al. DNA Methylome and Transcriptome Alterations and Cancer Prevention by Curcumin in Colitis-accelerated Colon Cancer in Mice. Carcinogenesis. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Yang Y, Yin R, Wu R, et al. DNA methylome and transcriptome alterations and cancer prevention by triterpenoid ursolic acid in UVB-induced skin tumor in mice. Molecular carcinogenesis. 2019;58(10):1738–1753. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Yang Y, Wu R, Sargsyan D, et al. UVB drives different stages of epigenome alterations during progression of skin cancer. Cancer Lett 2019;449:20–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Wang C, Wu R, Sargsyan D, et al. CpG methyl-seq and RNA-seq epigenomic and transcriptomic studies on the preventive effects of Moringa isothiocyanate in mouse epidermal JB6 cells induced by the tumor promoter TPA. J Nutr Biochem 2019;68:69–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Yang J, Wu R, Li W, et al. The triterpenoid corosolic acid blocks transformation and epigenetically reactivates Nrf2 in TRAMP-C1 prostate cells. Molecular carcinogenesis. 2018;57(4):512–521. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Guo Y, Wu R, Gaspar JM, et al. DNA methylome and transcriptome alterations and cancer prevention by curcumin in colitis-accelerated colon cancer in mice. Carcinogenesis. 2018;39(5):669–680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Krueger F, Andrews SR. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics. 2011;27(11):1571–1572. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Gaspar JM, Hart RP. DMRfinder: efficiently identifying differentially methylated regions from MethylC-seq data. BMC Bioinformatics. 2017;18(1):528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Yu G, Wang LG, He QY. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics. 2015;31(14):2382–2383. [DOI] [PubMed] [Google Scholar]
  • 31.Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol 2013;14(4):R36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Guo Y, Su ZY, Zhang C, et al. Mechanisms of colitis-accelerated colon carcinogenesis and its prevention with the combination of aspirin and curcumin: Transcriptomic analysis using RNA-seq. Biochemical pharmacology. 2017;135:22–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Zhang J, Wang L, Zhang Y, Li L, Higgins L, Lu J. Lobe-specific proteome changes in the dorsal-lateral and ventral prostate of TRAMP mice versus wild-type mice. Proteomics. 2011;11(12):2542–2549. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Wang C, Feng Y, Zhang C, et al. PTEN deletion drives aberrations of DNA methylome and transcriptome in different stages of prostate cancer. Faseb j 2020;34(1):1304–1318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Khor TO, Fuentes F, Shu L, et al. Epigenetic DNA methylation of antioxidative stress regulator NRF2 in human prostate cancer. Cancer Prev Res (Phila). 2014;7(12):1186–1197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Wu TY, Saw CL, Khor TO, Pung D, Boyanapalli SS, Kong AN. In vivo pharmacodynamics of indole-3-carbinol in the inhibition of prostate cancer in transgenic adenocarcinoma of mouse prostate (TRAMP) mice: involvement of Nrf2 and cell cycle/apoptosis signaling pathways. Molecular carcinogenesis. 2012;51(10):761–770. [DOI] [PubMed] [Google Scholar]
  • 37.Huang Y, Khor TO, Shu L, et al. A gamma-tocopherol-rich mixture of tocopherols maintains Nrf2 expression in prostate tumors of TRAMP mice via epigenetic inhibition of CpG methylation. J Nutr 2012;142(5):818–823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Zheng X, Cui XX, Khor TO, et al. Inhibitory Effect of a gamma-Tocopherol-Rich Mixture of Tocopherols on the Formation and Growth of LNCaP Prostate Tumors in Immunodeficient Mice. Cancers. 2011;3(4):3762–3772. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Saw CL, Wu TY, Paredes-Gonzalez X, Khor TO, Pung D, Kong AN. Pharmacodynamics of fish oil: protective effects against prostate cancer in TRAMP mice fed with a high fat western diet. Asian Pac J Cancer Prev 2011;12(12):3331–3334. [PubMed] [Google Scholar]
  • 40.Barve A, Khor TO, Reuhl K, Reddy B, Newmark H, Kong AN. Mixed tocotrienols inhibit prostate carcinogenesis in TRAMP mice. Nutr Cancer. 2010;62(6):789–794. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Yang Y, Fuentes F, Shu L, et al. Epigenetic CpG Methylation of the Promoter and Reactivation of the Expression of GSTP1 by Astaxanthin in Human Prostate LNCaP Cells. The AAPS journal. 2017;19(2):421–430. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

tableS1
NIHMS1702237-supplement-tableS1.docx (28.6KB, docx)
tableS2
NIHMS1702237-supplement-tableS2.docx (16.2KB, docx)

RESOURCES