Obesity related methylation changes in DNA of peripheral blood leukocytes

Obesity is the epidemic of our time, with sharply and steadily rising rates [1, 2]. The major adverse consequences of obesity including type 2 diabetes, atherosclerosis and essential hypertension, when added together, account for a large number of disease related deaths [3, 4]. If the obesity-related cancer cases are added to this number, obesity-related mortality by far exceeds that of other common diseases [5]. The latter indicates the urgent need to develop novel efficient therapeutic modalities for this condition.

The common denominator in the pathogenesis of the co-morbidities of obesity is the presence of an active, low-grade inflammatory process [6]. Despite evidence linking obesity to alterations in inflammatory response, little is known about the specific effects of obesity on the immune system. Recently, there has been a greater appreciation of the role of epigenetics, meiotically and mitotically heritable changes in gene expression that are not coded in the DNA sequence itself, in the immune and inflammatory responses [7‐9]. Therefore, we hypothesize that DNA methylation changes play a role in obesity induced immune dysfunction. The goal of this study was to characterize DNA methylation profile in peripheral blood leukocytes in obese versus lean subjects using a genome wide approach. Identification of methylation changes in specific genes will provide important targets for further study into the mechanisms of obesity's effect on the immune system and the potential to develop new therapies to treat multiple obesity comorbidities.

Table 1 displays the general characteristics of the screening sample. At this age range, obese subjects already showed significantly higher insulin resistance. Figure 1 is a volcano plot showing the raw P-values for all CpG sites versus mean methylation difference between the obese group and the lean group. We did not observe any CpG sites survive multiple testing correction with the most significant CpG site with a FDR of 0.14 and a raw P-value of 5 × 10^-6. Table 2 lists the top 10 most significant CpG sites and one additional CpG site which showed the largest methylation difference between cases and controls (-27.1% for CREB3L3). Since at least two CpG sites were included for the majority of the genes in this genome wide methylation chip, the P-value for the second CpG site in this gene was also listed in Table 2. Out of the 11 CpG sites listed in Table 2 (under the column "more significant CpG site"), we selected six CpG sites for validation in the replication cohort using the following inclusion criteria: (1) the most significant differentially methylated CpG site, which is the CpG site locating at -156 of the UBASH3A (ubiquitin-associated and SH3 domain-containing protein A) gene; or (2) the CpG sites in those genes of which the second CpG site also showed a raw P < 0.05. These include the CpG sites from the TRIM3 (Tripartite motif-containing 3), CTSZ (Cathepsin Z preproprotein), HIPK3 (Homeodomain interacting protein kinase 3), CDH5 (Cadherin 5 type 2 preproprotein) and CREB3L3 (cAMP responsive element binding protein 3-like 3) genes. The validation was conducted on the more significant CpG site (with Illumina ID 13517, 17029, 23241, 25599, 8829 and 23739) in each gene by pyrosequencing. The general characteristics of the replication cohort are displayed in Table 3. Since the DNA from one obese subject failed on four pyrosequencing assays, the data from this subject was excluded from the analysis. Table 4 lists the results in the replication cohort. The findings from the genome wide methylation chip were validated for the CpG sites in the UBASH3A, TRIM3, HIPK3 and CDH5 genes. The methylation status of the CpG site in the UBASH3A gene was 3.3% higher (P = 0.0019) and the methylation status of the CpG site in the TRIM3, HIPK3 and CDH5 genes was 1.21% (P = 0.0004), 7.31% (P = 0.0135) and 3.09% (P = 0.045) lower in the obese versus the lean group. These significant findings persisted after adjustment of age for the association between obesity and methylation status of the CpG site in the UBASH3A gene (P = 0.008) and TRIM3 gene (P = 0.001), but changed to borderline significance for HIPK3 gene (P = 0.05), and disappeared for the CpG site in the CDH5 gene (P = 0.095). The results of the meta-analysis for these six CpG sites on the data from the genome wide step and the replication step are shown in Additional file 2

Although the pyrosequencing assay was designed to target one specific CpG site for each gene, some of the assays covered several surrounding CpG sites. The results of these CpG sites are also listed in Table 4. The correlation within samples among the multiple CpG sites measured within a gene is listed in Additional file 3.

Gene Ontology analysis was performed to test whether some common functional trends in molecular functions and biological processes were associated with the genes exhibiting differences between obese cases and lean controls in the genome wide chip. We included those genes with a raw P ≤ 0.01 to the first list (n = 298) and included all the other genes in the second list. As expected from a pilot study in 14 subjects, we did not observe any GO categories survive multiple testing. Table 5 lists the GO categories with raw P-value less than 0.05. Interestingly, we observed enriched functional processes that are potentially relevant for inflammatory response with immune response (GO:0006955), cell activation (GO:0001775), cytokine production (GO:0001816), response to biotic stimulus (GO:0009607) and antigen binding (GO:0003823) among the top GO categories. The results support that obesity might lead to methylation changes in genes involved in inflammatory pathways.

In this study we aimed to identify obesity related methylation changes in peripheral blood leukocytes using a genome wide approach in youth and young adults who are free of obesity comorbidities. The primary findings of this study are increased methylation levels at one CpG site in the UBASH3A gene and decreased methylation level at one CpG site in TRIM3 gene in obese subjects compared with lean controls.

The protein encoded by the UBASH3A gene is the "ubiquitin-associated and SH3 domain-containing protein A", which was previously also known as T-cell ubiquitin ligand (TULA) and suppressor of T-cell signaling 2 (Sts-2). It is expressed predominantly in T-cells, where it has a suppressing effect on T-cell signaling and activation. It acts in part by inhibiting c-CBL-mediated downregulation of protein tyrosine kinases (PTKs) that are activated upon T-cell receptor stimulation [19]. UBASH3A is also capable of promoting T-cell apoptosis [20]. UBASH3A and UBASH3B (another protein in this family) gene double knock-out mice were hyperresponsive to T-cell receptor stimulation with increased cytokine secretion, although mice lacking UBASH3A were normal in all respects including T-cell function [21]. However, whether the expression of the UBASH3A gene is regulated by DNA methylation is unknown. No CpG island exists in this gene and the CpG site showing higher methylation levels in obese subjects from this study locates at the promoter region with a distance of 156 bp to the transcription start site. There is a possibility that methylation of this CpG site or other CpG sites with methylation levels correlated with this CpG site inhibits the interactions between DNA sequence and nuclear proteins, resulting in decreased gene expression. In-silico analysis of the region of this CpG site using TFSEARCH software [22] did not find this CpG site located at any known transcription factor binding sites. However, methylation of this CpG site may suppress gene transcription by recruiting methylcytosine-binding proteins that in turn associate with large protein complexes containing corepressors and histone deacetylases. The binding of these complexes to DNA may lead to a change in the chromatin structure from an active to an inactive form [23]. This speculation needs to be confirmed.

TRIM3 is one member of the TRIM protein family. These proteins share a conserved tripartite architecture and have a variety of cellular functions including cell proliferation, differentiation, oncogenesis, apoptosis, immune signaling and have been implicated in autoimmune diseases [24]. In peripheral blood leukocytes, TRIM3 is mainly expressed in macrophages and can be up-regulated by viral infection [25, 26]. Similar to UBASH3A, the role of methylation in TRIM3 gene expression has not been reported yet. The TRIM3 gene has 1 CpG island located in the promoter and exon 1 (nucleotides -446 to 576), one CpG island spanning intron 2 and exon 3 (nucleotides 7759 to 8334), and one CpG island in exon 13 and the downstream region (nucleotides 24883 to 25610). The CpG site showing lower methylation levels in obese subjects in both the scanning cohort and the replication cohort locates in the promoter CpG island (348 bp upstream to the transcription start site). The methylation status of the other CpG site which showed significant differential methylation levels between obese cases and lean controls in the genome wide methylation analysis also locates in this CpG island (331 bp upstream to the transcription start site). Because the pyrosequencing assay also measured the methylation levels of the -331 CpG site (Illumina ID: 19207), we were able to show that also for this site obese subjects had lower methylation levels in comparisons with the lean subjects (age adjusted P = 0.022 as shown in Table 4). Similar to the CpG site in the UBASH3A gene, these two CpG sites are not located at any known transcription factor binding sites. Future studies will be needed to test whether the density of the methylation or the methylation of specific CpG sites in this CpG island has effect on TRIM3 gene expression.

The observed DNA methylation differences between obese cases and controls were relatively small. They were 10.1% and 2.54% in the genome wide step and 3.3% and 1.2% in the replication step for the UBASH3A gene and the TRIM3 gene, respectively. This modest level of differences raises an important question: what is the biological significance of changes in methylation on this degree? In this study, we used the DNA from leukocytes, which represent different cell populations with distinct epigenetic profiles. It is plausible that only the change in the epigenetic profile of one specific cell type is related to obesity. In this regard, the actual epigenetic differences might be more substantial than reported here but only present in this specific blood leukocyte cell type. Future studies on epigenetic profiling of various types of cell populations in the leukocytes are warranted to gain a greater understanding of the epigenetic dysregulation in obesity. Furthermore, this kind of study will be able to identify the cells specifically reflecting obesity-associated methylation changes, which is of great interest in itself. Transcriptional profiling studies will be very valuable in understanding whether the DNA methylation status differences are associated with differences in gene expression. Unfortunately, cellular RNA is not available for the samples used in the current study, which were selected from several existing cohorts.

There are several strengths of this study. First, we selected obese cases and controls with extreme phenotypes, which maximizes the power to make discoveries. Second, we focused on youth and young adults with the distinct advantage that the results are not confounded by obesity comorbidities or use of medication, both of which are very common in adult subjects with obesity. Third, a hypothesis-free genome wide approach was used. This approach supersedes the limitations imposed by candidate gene methylation studies and allows searching the whole genome in an unbiased manner.

Interpretation of these data is also limited by several constraints. First, in this study we hypothesize that obesity will lead to methylation changes in the DNA of leukocytes, which further lead to obesity related co-morbidities. However, our study design cannot determine whether the identified methylation changes are the cause or the consequence of obesity. Future studies on subjects who changed their body size status will be needed to clarify the causal directions. On the other hand, epigenetic regulation is tissue specific. In this regard, the target tissue to identify epigenetic variations responsible for obesity should be the hypothalamus of the brain rather than peripheral leukocytes. Second, in the current study, the Infinium HumanMethylation27 Beadchip was selected because of its quantitative measure at each CpG site. However, the limited coverage of this genome wide chip will restrict the findings to certain CpG sites within certain genes. Future studies should use chips with more complete coverage of the genome such as the recently released 450K Infinium Methylation BeadChip from Illumina. Third, the current study is a small pilot study with the genome wide step conducted only in seven obese and seven lean subjects. Future studies with much larger sample size are warranted to discover a more complete profile of obesity related methylation changes. Fourth, although all the obese cases in the genome-wide step were free of clinical diseases such as CVD and diabetes, some of them already showed evidence of insulin resistance. It is possible that insulin resistance may be the factor triggering methylation changes and not the obesity per se. Recent literature supports the postulate that glucose restriction alters gene expression through epigenetic mechanisms [27]. We conducted further analysis within the obese cases and did not observe that any of the CpGs showed methylation difference between cases with impaired fasting glucose (n = 3) versus those with normal glucose levels (n = 3), with the most significant CpG site having an adjusted P-value of 0.85. However, the limited sample size prevents us from drawing any conclusion based on this analysis. Unfortunately, fasting glucose was not measured in the replication cohort. Future studies with large sample sizes and more detailed phenotypes will be needed to clarify this question.

In this study, we identified several reproducible changes in DNA methylation of peripheral blood leukocytes between obese cases and lean controls. This study provides evidence that obesity is associated with methylation changes in blood leukocyte DNA. Further studies are warranted to determine the causal direction of this relationship as well as whether such methylation changes can lead to immune dysfunction. Such studies will have the ability to identify new insight into disease etiology and provide new targets for prevention of obesity related diseases such as cardiovascular diseases and type 2 diabetes.