Epigenome-wide DNA methylation profiling of periprostatic adipose tissue in prostate cancer patients with excess adiposity—a pilot study

Prostate cancer (PCa) is one of the most frequent malignancies in men and the second leading cause of cancer-related death in the North America and most western European countries [1, 2]. Epidemiological studies support obesity or excess adiposity as an important environmental risk factor for PCa, being primarily associated with advanced disease and death [3]. Periprostatic adipose tissue (PPAT), a white fat depot surrounding the prostate capsular-like structure, has been recognized to have the potential to exert pro-tumoral endocrine and paracrine influences on prostate cancer cell’s biological phenotypes [4]. There is now evidence that obesity and overweight result in excess fat deposit at PPAT [5], altered fatty acid profile [6], migration of tumor cells [7], secretion of a variety of adipokines, such as interleukin-1 beta (IL-1b), osteopontin, leptin, tumor necrosis factor alpha (TNF-a), and decreased adiponectin, thus contributing to a tumor microenvironment that ultimately facilitates PCa aggressiveness [7, 8].

DNA methylation is a well-known epigenetic mechanism resulting from the interaction between environmental factors and the genome [9]. DNA methylation with variation of CpG sites is associated with tissue-specific gene modulation and involved in phenotype transmission and in the development of diseases [10]. Excess adiposity, as a consequence of environmental factors such as excessive food consumption or inactive lifestyle, has been identified as a regulator of epigenetic modification in adipose tissue. Recent findings from experimental studies suggested that modification of DNA methylation pattern in adipose tissue and adipocytes was related with development of cancer, type 2 diabetes, and cardiovascular diseases through influencing metabolism and inflammation [11, 12, 13]. Additionally, several studies reported altered DNA methylation in PCa cells as compared with adjacent benign tissue, and some significantly methylated CpG sites and genes were found to be responsible for the occurrence and progression of PCa [14, 15, 16]. Nevertheless, the epigenome-wide DNA methylation profile of PPAT from excess adiposity PCa patients is currently unknown despite its potential mechanistic involvement in obesity association with PCa.

The aim of this study was to perform a epigenetic-wide association study (EWAS) in order to evaluate DNA methylation profile of PPAT obtained from obese/overweight (OB/OW) in comparison with normal weight (NW) PCa patients and identify differentially methylated sites. We also explored the consequential potential biological functions that account for the effect of PPAT from OB/OW subjects in PCa molecular mechanisms.

This pilot study revealed significant differences of DNA methylation profiles between the PPATs from OB/OW versus NW PCa patients. Variations in global DNA methylation demonstrated that excess adiposity played an important role in DNA methylation level of PPAT tissues in prostate cancer patients, which provide an opportunity to explore the effect of obesity on PPAT epigenetic modification and subsequently on prostate cancer. These findings reported for the first time in PPAT depot are in concordance with previous works reporting that excess adiposity and BMI activate DNA methylation in adipose tissue [27, 28, 29]. Thus, considering the present understanding of the potential causal relationship between excess adiposity and cancer [30], diabetes [11], and cardiovascular disease [31], our results provide methylated candidate genes, which might foster research on the potential biological mechanisms underlying epigenetic regulation of PPAT by excess adiposity and prostate cancer.

Given that DNA methylation of CpGs located at promoters and islands are associated with gene transcription silencing, we performed a strict filtering of DMCs and explored the biological functions of all promoter hypermethylated genes, aiming to find the critical methylated CpGs in the PPAT between the obese and normal weight PCa patients. Bioinformatic analysis showed that the enriched pathways were mostly involved in metabolic disorders, particularly fatty acid degradation and glycerolipid and choline metabolism. These pathways are known to mediate the pro-tumoral effect of white adipose tissue in tumors, thus contributing to tumorigenesis and metastasis [32, 33], particularly in prostate cancer [5]. Findings from other oncological models highlight excess adiposity-associated impact in methylation markers known to associate with potential effect in the cancer microenvironment (e.g., aromatase, prostaglandin E₂ receptor in breast cancer) [34, 35]. Obesity has also been shown to associate with methylation of cancer-related genes (E-cadherin, p16, and RAR-β(2)) directly in malignant breast cells [36, 37].

Pathway enrichment analysis showed a strong association between promoter hypermethylation of CPT1B, CPT1C, ACADM, and FADS1, with fatty acid metabolism. CPT1B (carnitine palmitoyltransferase 1B) and CPT1C (carnitine palmitoyltransferase 1C) genes encode rate-limiting enzymes in fatty acid degradation and play critical roles in long-chain fatty acid (LCFA) β-oxidation by controlling transportation of long-chain fatty acyl-CoAs from the cytoplasm across the outer mitochondria membrane [38]. Maple et al. reported that increased methylation of specific CpGs in the CPT1B promoter was correlated with decreased CPT1B transcripts in the skeletal muscle after lipid oversupply in severe obesity, which resulted in obese individual’s incapacity to increase fat oxidation, contributing to metabolic inflexibility [39]. Although the biochemical function of CPT1C has been verified to be necessary for the regulation of energy homeostasis in CPT1C knockout mouse brain [40], the study of CPT1C methylation was absent. CPT1B and CPT1C were previously reported to be highly expressed in the muscle, brain, and many other normal tissues including adipocytes [41]. Taken together, these findings suggest that methylation of specific CpG sites in the CPT1B and CPT1C promoters likely result in gene expression silencing, thus consequently contributing to fatty acid accumulation in adipocytes by decreasing long-chain fatty acid β-oxidation in the mitochondria (Fig. 4).

Open image in new window

LCFA and ACADM genes (aliases MCAD, medium-chain acyl-CoA dehydrogenase) coding for metabolic enzymes presented increased methylation in the PPAT of the OB/OW group. ACADM is the critical enzyme of the initial step of β-oxidation and controls the medium-chain fatty acid (MCFA) metabolism by catalyzing the dehydrogenation of medium-chain Acyl-CoA, which is the common middle product of MCFA and LCFA, in the mitochondria. Mutations in ACADM cause MCAD deficiency, which resulted in fatty acid oxidation disorder leading to disease or infantile death [42, 43, 44]. Greco et al. [45] reported inverse association between ACADM transcript abundance with fat content in the human liver. Our findings suggest that the hypermethylated ACADM found in the PPAT of OB/OW PCa patients might fail to generate medium-chain acyl-CoA β-oxidation and result in MCFA and LCFA accumulation in adipose tissue, providing a favorable tumor microenvironment for PCa cell aggressiveness (Fig. 4). Additional functional studies are required to confirm this assumption.

The hypermethylation of the FADS1 (fatty acid desaturase 1) promoter, whose transcriptional activity was significantly decreased in OB/OW PCa patients in agreement with our previous study [4], has been described as correlated with polyunsaturated fatty acid (PUFA) metabolism by catalyzing the biosynthesis of highly unsaturated fatty acids (HUFA) from the catalysis of dihomo-gamma-linoleic acid (DGLA, 20:3 n-6) and eicosatetraenoic acid (ETA, 20:4 n-3) desaturation, in order to generate arachidonic acid (AA, 20:4 n-6) and eicosapentaenoic acid (EPA, 20:5 n-3) [46]. Genetic variants in the FADS1 and FADS2 gene clusters have been associated with altered (n-6) and (n-3) PUFA metabolism [47, 48], whereas metabolic disorder in PUFA exerted effects on PCa by mediating the formation of eicosanoid inflammatory mediators (prostaglandins, leukotrienes, thromboxanes, and lipoxins), angiogenesis, immune cell regulation, and membrane structure and function [49, 50]. These results illustrated that the epigenetic modifications of FADS1 may play important roles in the regulation of fatty acid metabolic genes on PPAT in response to excess adiposity (Fig. 4).

Besides abnormal fatty acid metabolism, DMC-related genes identified in our study were also correlated with glycerolipid metabolism. MOGAT1 and MOGAT3 encode the monoacylglycerol O-acyltransferase (MOGAT) and catalyze the formation of diacylglycerol (DAG) from monoacylglycerol (MAG), which is the precursor of phosphatidylcholine, phosphatidylethanolamine, and triacylglycerol (TAG), by transferring fatty acyl-CoA to 2-monoacylglycerol [51]. While human MGAT1 (aliases for MOGAT1) is involved in intestinal dietary fat absorption and TAG synthesis in the liver, its function in adipose tissue has yet to be elucidated. The expression of MGAT1 was increased in the liver of diet-induced obese mice with nonalcoholic fatty liver disease (NAFLD), but, interestingly, there was increased DAG accumulation and no inflammatory injury reduction in hepatocytes after MGAT1 knockdown. Similarly, MOGAT3 was mostly expressed in the human intestine and liver and maintained a significant DGAT (diacylglycerol O-acyltransferase) activity. Although results indicate that the metabolic mechanism of lipid regulation by MGAT1 and MOGAT3 was altered, evidence of association between lipid metabolic disorders caused by aberrant expression of MGAT1/MOGAT3 and PCa are lacking. Our data indicate the methylation of MOGAT1 and 3 genes in PPAT may play important roles in response to excess adiposity by modulating glycerolipid metabolism (Fig. 4).

Choline metabolic disorder might be caused by epigenetic regulation of SLC44A2 (solute carrier family 44 member 2), which encodes choline transporter-like protein 2 (CTL2) and is mainly expressed on blood plasma and mitochondrial membrane of different organisms and cell types. This transporter is a rate-limiting step in choline metabolism by transporting extracellular choline into cell and mitochondria. Choline is essential for synthesizing membrane phospholipid and neurotransmitter acetylcholine and used as a donor of methyl groups via choline oxidized in mitochondria [52]. The choline transporter has been associated with choline metabolic disorders, thus playing an important role in regulating immune response, inflammation, and oxidation [53, 54]. Concordantly, abnormal choline metabolism emerged as a metabolic hallmark, associated with oncogenesis and tumor progression in prostate cancer and other malignancies [55, 56, 57]. The increased uptake of choline by the cancer cell was important to meet the needs of phosphatidylcholine synthesis [58]. We hypothesize that hypermethylated SLC44A2 in adipocytes might be associated with lower uptake and oxidation of extracellular choline, resulting in choline accumulation in PPAT extracellular media (Fig. 4) and increasing the availability of choline for PCa cell metabolism.

Besides metabolic modifications, altered immune regulation pathways were also enriched in DMC-related genes. TAPBP (alias tapasin) encodes a transmembrane glycoprotein, which mediates the interaction between MHC class I molecules and a transport protein TAP (transporter associated with antigen processing), being responsible for antigen processing and presentation. This mechanism occurs via mediating TAP to translocate endo/exogenous antigen peptides from the cytoplasm into the endoplasmic reticulum and deliver the antigen peptides to MHC class I molecules. The cancer cell’s survival depends on successful escape to immune surveillance. Loss of MHC class I has been described as a major immune evasion strategy for cancer cells. Downregulation of antigen-presenting MHC class I pathway in tumor cells was a common mechanism for tumor cells escaped from specific immune responses, which can be associated with coordinated silencing of antigen-presenting machinery genes, such as TAPBP [59]. Cross-presentation is the ability of certain antigen-presenting cells to take up, process, and present extracellular antigens with MHC class I molecules to CD8⁺ T cells. This process is necessary for immunity against most tumors. Recent studies revealed that TAPBP is a major target for cancer immune evasion mechanisms and decreased TAPBP expression in cancer was associated with reduced CD8⁺ T cell-mediated killing of the tumor cells, lowered immune responses, and enhanced tumor metastases via downregulation of antigen presentation the MHC class I pathway [60, 61]. Our results showed that TAPBP promoter hypermethylation in the PPAT of obese PCa subjects likely reduced the expression or activity of TAPBP, downregulating tumor cell’s antigen presentation of immune cells in PPAT, leading to impaired CD8⁺ T cell activation (Fig. 4). This indicates that methylation of TAPBP might be a mechanism by which prostate cancer cells escape the immune surveillance and provide an appropriate microenvironment for tumor aggressiveness, allowing prostatic cancer cells’ transfer, spread, and growth. The significant DMR identified with eight DMCs located in the TAPBP promoter further supported its role in prostate cancer.

From the PPI analysis, the network which was connected through ubiquitin C is characterized, suggesting UBC played a significant biological function with the methylated genes in PPAT between OB/OW and NW patients and somehow was correlated with the methylation. Ubiquitin is much known with the functions including roles in protein degradation, DNA repair, cell cycle regulation, kinase modification, and cell signaling pathways [62]. Recent reports expressed that the ubiquitin-proteasome system was associated with the progression and metastasis of prostate cancers [63, 64]. And long-term silencing of the UBC was found to be correlated with DNA methylation at the promoters [65]. Additional studies are needed to clarify whether the protein network for methylated genes impacts prostate cancer and if this difference is associated with ubiquitin C.

Although we present the first report on periprostatic adipose tissue methylation profile in association with excess adiposity measured by BMI, our results should be interpreted in the context of several potential limitations. This study is limited by small sample size, even though representative groups of OB/OW and NW are likely to be selected following the strict inclusion/exclusion criteria and between-group match by clinicopathological and demographic variables. Although we matched patients by clinicopathological characteristics between adiposity groups, tobacco smoking was more frequent among OB/OW compared with NW patients. Actually, albeit we cannot exclude an effect of smoking status on the presumably adiposity-associated findings presented herein, due to a known effect of tobacco on overall DNA methylation, data from previous reports indicate that methylation profiles are tissue-specific [66, 67] and that adiposity-associated DNA methylation occurs independently of tobacco smoking [68, 69]. Future studies will benefit from the confirmation of these results in larger sample sizes, determination of correspondence to matched prostate tumor methylation patterns, investigation of interactome at the interface between tumor and PPAT, and prospective investigations on the value of PPAT epigenetic modifications on cancer recurrence and survival. Future validation and replication are important to establish the accuracy and generalizability of the reported associations.

In summary, we observed differences in PPAT methylation between NW and OB individuals at several loci known to be involved in the metabolism of choline (SLC44A2), fatty acids (CPT1B, CPT1C, ACADM, FADS1), and glycerolipid (MOGAT1, MOGAT3) and in the regulation of exogenous tumor antigen presentation (TAPBP). These findings suggest a relationship of adiposity status with the methylation profile, which ultimately modulates tumor microenvironment and may influence PCa behavior.

In this preliminary study, we report DNA methylation changes in PPAT underlying the association between excess adiposity and PCa. Whole epigenome methylation profiling of PPAT of PCa patients revealed significant differences in OB/OW versus normal weight subjects. Epigenetic imprinting in association with excess adiposity expressed the methylated modifications in genes functionally related with lipid metabolism and immune function, which could ultimately contribute to an unfavorable tumor microenvironment and decreased immune surveillance for prostate tumors. This association analyses provided us novel insights into how prostate cancer patients with excess adiposity differ from those of patients with normal weight in epigenome. Findings from this study warrant confirmation in PPAT samples from larger number of patients.