Promoter methylation of the MGAT3 and BACH2 genes correlates with the composition of the immunoglobulin G glycome in inflammatory bowel disease

Promoter methylation of the MGAT3 and BACH2 genes correlates with the composition of the... Background: Many genome- and epigenome-wide association studies (GWAS and EWAS) and studies of promoter methylation of candidate genes for inflammatory bowel disease (IBD) have demonstrated significant associations between genetic and epigenetic changes and IBD. Independent GWA studies have identified genetic variants in the BACH2, IL6ST, LAMB1, IKZF1,and MGAT3 loci to be associated with both IBD and immunoglobulin G (IgG) glycosylation. Methods: Using bisulfite pyrosequencing, we analyzed CpG methylation in promoter regions of these five genes from peripheral blood of several hundred IBD patients and healthy controls (HCs) from two independent cohorts, respectively. Results: We found significant differences in the methylation levels in the MGAT3 and BACH2 genes between both Crohn’s disease and ulcerative colitis when compared to HC. The same pattern of methylation changes was identified for both genes in CD19 B cells isolated from the whole blood of a subset of the IBD patients. A correlation analysis was performed between the MGAT3 and BACH2 promoter methylation and individual IgG glycans, measured in the same individuals of the two large cohorts. MGAT3 promoter methylation correlated significantly with galactosylation, sialylation, and bisecting GlcNAc on IgG of the same patients, suggesting that activity of the GnT-III enzyme, encoded by this gene, might be altered in IBD. The correlations between the BACH2 promoter methylation and IgG glycans were less obvious, since BACH2 is not a glycosyltransferase and therefore may affect IgG glycosylation only indirectly. Conclusions: Our results suggest that epigenetic deregulation of key glycosylation genes might lead to an increase in pro-inflammatory properties of IgG in IBD through a decrease in galactosylation and sialylation and an increase of bisecting GlcNAc on digalactosylated glycan structures. Finally, we showed that CpG methylation in the promoter of the MGAT3 gene is altered in CD3 T cells isolated from inflamed mucosa of patients with ulcerative colitis from a third smaller cohort, for which biopsies were available, suggesting a functional role of this glyco-gene in IBD pathogenesis. * Correspondence: vzoldos@biol.pmf.hr Marija Klasić and Dora Markulin contributed equally to this work. Department of Biology, Division of Molecular Biology, Faculty of Science, University of Zagreb, Horvatovac 102a, 10000 Zagreb, Croatia Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Klasić et al. Clinical Epigenetics (2018) 10:75 Page 2 of 14 Background proteins, including components of the immune system Inflammatory bowel disease (IBD) is a chronic intestinal [27]. Aberrant protein glycosylation is implicated in inflammatory condition classified in two major forms— virtually every human complex disease, including inflam- Crohn’s disease (CD) and ulcerative colitis (UC)—which mation [28–31]. Previous studies have suggested that exhibit etiologically and clinically distinct features. Now- N-glycosylation of secreted and membrane proteins adays, IBD affects 2.5–3 million people in Europe and might be regulated epigenetically and that aberrant gly- causes considerable morbidity [1]. Despite numerous cosylation profiles in disease can arise through aberrant clinical, genetic, and other experimental studies, our epigenetics [32–38]. A comprehensive review about the understanding of IBD development and progression role of protein glycosylation in IBD has been given remains incomplete. recently [39]. N-glycosylation of serum-circulating pro- It is generally accepted that IBD represents an aber- teins (such as the acute phase proteins; immunoglobulin rant immune response to gut microbiota in genetically G, IgG; and immunoglobulin A, IgA) or whole plasma susceptible individuals [2]. Genome-wide association N-glycome (i.e., N-glycans present on all plasma pro- studies (GWAS) have identified over 200 genetic suscep- teins) has been the focus of IBD biomarker discovery tibility loci, the majority of which were associated with [36, 40–43]. In addition, our partners from IBD consor- both forms of IBD in genome-wide meta-analysis [3–7]. tium and others established that altered glycosylation of However, common genetic variants account only for 8.2 IgG, which is a key effector of the humoral immune and 13.1% heritability of UC and CD, respectively [7]. system, has a role in balancing inflammation at the sys- Interactionofanindividual’s gut microbiome, im- temic level [42–46]. mune system, genetic background, and environmental GWA studies indicated associations of IBD with sev- factors, such as smoking, diet, drugs, and physical eral loci involved in protein glycosylation [47, 48]. More activity [2, 8–10], makes IBD a complex etiopatho- recently, the first GWAS of IgG glycosylation identified genic entity. The challenge is therefore to identify 16 loci specifically associated with changes in IgG glyco- additional factors involved in the development and sylation [49]. Interestingly, five of these loci showed plei- progression of this disease, especially given its rapidly otropy with IBD: MGAT3, a glyco-gene encoding for a increasing incidence. It is probable that epigenetics glycosyltransferase, GnT-III; LAMB1, a member of play a key role in the interactions between environ- transmembrane glycoprotein family of extracellular mental, microbial, and genetic factors that participate matrix; the IL6ST, a signal transducer shared by many in IBD development and progression. These include cytokines; IKZF1; and BACH2, transcription factors in- DNA methylation and histone modifications, as well volved in B cell differentiation, activation, and matur- as some other epigenetic mechanisms [11–13]; for a ation. Only the MGAT3 is a classical glyco-gene with a review, see [14, 15]. known function in IgG glycosylation, while the exact DNA methylation remains the most studied epigenetic functional roles for other four GWAS hits in IgG glyco- modification, readily assayed in a large number of indi- sylation or IBD remain unknown. viduals/samples. Hypermethylation of gene promoters is In this study, we investigated promoter methylation generally associated with gene silencing, while promoter differences in these five genes, associated with both IBD hypomethylation is associated with gene activation [16]. and IgG glycosylation, in peripheral whole blood of Environmentally changed DNA methylation pattern may several hundred IBD patients from two independent contribute to the development of many complex diseases cohorts. We also correlated promoter methylation data by mediating the interplay between external and internal with IgG glycosylation data analyzed previously for the factors and the gene expression [17–21]. There are also same IBD patients by our partners from the IBD consor- data to suggest that the aforementioned environmen- tium [43, 46, 50]. Peripheral blood was used for DNA tal modifiers of IBD can also affect DNA methylation methylation analysis and serum or plasma was used for [17–19, 22]. Epigenetic component of IBD has been glycan analysis, since one of our goals was the search for addressed in many studies, mostly by whole genome potential IBD biomarkers. As peripheral whole blood is methylation analysis performed on peripheral blood a heterogeneous cell mixture with specific methylation mononuclear cells (PBMCs) or mucosal tissue, reveal- pattern for each of the cell types [51], we also analyzed ing regions differentially methylated between the promoter methylation of our candidate genes in CD19 diseaseand healthystate,aswellasbetween CD and B cells and CD3 T cells isolated from peripheral blood UC [11–13, 23–26]. mononuclear cells (PBMCs). B cells were of our particu- The majority of eukaryotic proteins are modified by lar interest since these cells produce IgG on their mem- addition of complex oligosaccharides (glycans) through brane and are precursors of plasma cells which secrete the process of glycosylation. Therefore, glycans are an IgG. We have further explored if aberrant promoter integral part of nearly all membrane and secreted methylation recorded in peripheral whole blood of IBD Klasić et al. Clinical Epigenetics (2018) 10:75 Page 3 of 14 patients can be a proxy for epigenetic events occurring allowed them to clot at 4 °C for 60 min, and then centri- in the inflamed mucosa. To address this question, we fuged at 2500×g for 15 min. The serum was aliquoted analyzed DNA methylation from PBMCs, CD3 T cells off and stored at − 80 °C until further analysis. isolated from PBMCs, and CD3 T cells isolated from A subset of patients and controls recruited in Edinburgh inflamed colonic mucosa of UC patients from the third (Additional file 1: Table S3) underwent immunomagnetic smaller cohort, for which biopsies were available. cell separation to obtain CD19 B cells. The methods have previously been detailed elsewhere [13]. Venepuncture Methods using 9-ml K3 EDTA vacuette (Greiner) tubes was Patient selection and ethics performed to obtain between of 18 and 36 ml of Patients were recruited prospectively from Edinburgh, EDTA-buffered blood. An initial Ficoll (Ficoll-Paque, GE UK, and Florence, Italy, as a part of the IBD-BIOM pro- Healthcare, Bucks, UK) density gradient centrifugation ject. The recruitment of patients from Edinburgh has been was performed to obtain peripheral blood mononuclear described elsewhere [13, 43]. Briefly, we recruited IBD cells. Cells labelled with antibody-coated microbeads + + 7 patients prospectively as close as possible to the date of (human CD8 and CD19 microbeads, 20 μlper 1×10 diagnosis from gastroenterology outpatient and endoscopy cells) were immunomagnetic separated using the auto- appointments between 2012 and 2015. We recruited MACs Pro cell separator (Miltenyi, Germany). CD19 symptomatic controls from gastroenterology clinics dur- separations were performed following an initial CD8 ing the same period. In these individuals, we had excluded depletion step. Nucleic acids were extracted using AllPrep IBD and other organic bowel pathology following bio- (Qiagen, Hilden, Germany) according to the manufac- chemical and/or endoscopic investigations. We recruited turer’s guidance and stored at − 80 °C. a further healthy volunteer cohort with no gastrointestinal Colonic biopsies from controls and UC patients with symptoms. IBD patients were stratified by disease type inactive and active form of disease were mechanically (ulcerative colitis, UC, and Crohn’s disease, CD). Detailed dissociated to prepare single-cell suspensions using genetic, phenotypic, and other data regarding IBD cases Hanks’ balanced salt solution modified medium, without are given in Additional file 1: Tables S1 and S2. Florence calcium chloride and magnesium sulfate (HBSS) (Sigma), cohort was collected through the network of the Italian with penicillin/streptomycin and gentamicin. PBMCs Group for IBD (IG-IBD) since the beginning of 2001 and were obtained by density gradient centrifugation using first described in 2005 [1] following an internal validation Lymphoprep. CD3 T cells (from biopsies and blood) of phenotyping. Subsequently, longitudinal update has were magnetically sorted by using the EasySep™ Human been performed on a yearly basis. T Cell Enrichment Kit (STEMCELL) following the man- Ethical approvals were obtained from Tayside Com- ufacturer’s instructions. Following cell isolation, DNA mittee on Medical Ethics B, and all patients and controls extraction was performed using the Invisorb Spin Tissue provided written, informed consent (LREC 06/S1101/16, Mini Kit (Stratec Molecular) following the manufac- LREC 2000/4/192). turer’s instructions. Florence recruitment details DNA methylation analysis IBD patients were prospectively recruited as close as We analyzed promoter methylation of the candidate possible to the date of diagnosis from gastroenterology genes in the DNA from whole blood, as well as from the outpatient and endoscopy appointments between years separated CD19 B cells. In addition, for the MGAT3, 2012 and 2015 in different tertiary referral centers in which is a glycosyltransferase with direct and known San Giovanni Rotondo, Rome, Rozzano (Milan), Padua, function in IgG glycosylation [52], we analyzed promoter and Florence, Italy. Symptomatic controls were recruited methylation—in DNA from PBMCs, CD3 T cells iso- in the same centers (gastroenterology clinics) during the lated from PBMCs, and CD3 T cells isolated from the same period. In these individuals, IBD and other organic colonic mucosa of healthy controls and UC patients bowel pathology were excluded by biochemical and/or (classified according to active and inactive form of the endoscopic investigations. IBD patients were stratified disease) of the third independent smaller subcohort col- by disease type (ulcerative colitis, UC, and Crohn’s lected by the Gastroenterology Department of Centro disease, CD). Samples were obtained with the same Hospitalar do Porto-Hospital de Santo António, Portugal methodology (see further) and centrally collected at San (Additional file 1: Table S4). All specimens were sub- Giovanni Rotondo, Italy. jected to histological examination and classification. All participants gave informed consent about all clinical Sample collection procedures, and research protocols were approved by We collected whole blood at the time of patient recruit- the ethics committee of CHP/HSA, Portugal (233/ ment into 9-ml serum Z-clot activator tubes (Greiner), 12(179-DEFI/177-CES). Klasić et al. Clinical Epigenetics (2018) 10:75 Page 4 of 14 For DNA methylation analysis, 500 ng of DNA from primers for the BACH2 and the MGAT3 genes are listed whole blood was bisulfite converted using EZ-96 DNA in Additional file 1: Table S5. EpiTect PCR Control DNA Methylation Gold kit (Zymo Research, Freiburg, Set (methylated and unmethylated bisulfite-converted Germany), and 100 ng of DNA from CD19 B cells, human DNA, Qiagen) was used as a control for PCR and PBMCs, and T cells was converted using EZ DNA pyrosequencing reactions. Methylation Gold kit (Zymo Research, Freiburg, Germany) according to the manufacturer’s protocol. Statistical analysis Two to six pyrosequencing assays were developed for The nonparametric Mann-Whitney U test was used to promoter regions of each of the five candidate genes compare the methylation status of CpG sites encom- (BACH2, MGAT3, IL6ST, IKZF1, and LAMB1). The se- passed by the pyrosequencing assays in the MGAT3 and lection of analyzed CpG sites was random for assays 2–5 BACH2 genes between the two independent groups: HC of the MGAT3 gene. CpG sites within the MGAT3 assay compared to each of CD or UC. Significance threshold 1 were selected based on the GEO (Gene Expression was set at p < 0.05 with additional Bonferroni correction Omnibus) database where methylation data were for multiple testing. Given that age was our primary obtained using Illumina HumanMethylation450 Bead- concern as a potential confounder, we visualized the age Chip v1.1 technology. For the BACH2 gene, assays were in the three groups (CD, UC, and HC) for the samples selected based on location of differentially methylated included in each analysis as violin plots (Additional file 3: CpGs in different cell lines tested by ENCODE project, Figure S2) and assured there was no significant differ- using Illumina HumanMethylation450 BeadChip v1.1 ence between the age groups (p > 0.05) using the technology (a newer version, the Infinium MethylationE- Mann-Whitney U test. This was done to assure the PIC 850K was not available at the time). We used validity and strengthen the rationale for the selection of traditional bisulfite-based protocols which cannot statistical methods. discriminate between 5-methylcytosine (5-mC) and For thedataofthe MGAT3 promoter methylation 5-hydroxymethylcytosine (5-hmC) as oxidative bisulfite from PBMCs, CD3 T cells isolated from blood, and (oxBS-450K) method can [53]. However, recent studies CD3 T cells isolated from inflamed colonic mucosa have shown that global DNA hydroxymethylation is very (the Porto cohort), the Mann-Whitney U test was low in blood cells [54, 55]. Furthermore, hydroxymethy- applied with Bonferroni correction accounting for 15 lation is significantly depleted from promotors and CpG CpG sites. islands, while enriched in the gene bodies [53, 56]. Based on the estimated statistical power, we did initial Glycan analysis screening on 60 patients for each pyrosequencing assay, Glycans present on IgG were analyzed from serum of after which we excluded those genes (pyrosequencing over 1000 IBD (UC and CD) patients and healthy con- assays) that did not show any statistically significant dif- trols in the Edinburgh cohort using ultra performance ferences between IBD patients and healthy controls. liquid chromatography (UPLC) [43, 50]. In the Florence Pyrosequencing assays for LAMB1, IL6ST, and IKZF1 cohort, plasma samples of 3500 IBD patients and healthy are shown in Additional file 2: Figure S1. We continued controls was used for analysis of IgG glycopeptides by to analyze promoter methylation only in the BACH2 and liquid chromatography coupled to mass spectrometry MGAT3 genes. Specific regions were amplified using (LC-MS) [46]. The data for IgG glycosylation analysis PyroMark PCR kit (Qiagen, Hilden, Germany). The cyc- were used in this work for correlation analysis with pro- ling conditions for the BACH2 gene were as follows: ini- moter methylation data of MGAT3 and BACH2 genes, tial polymerase activation step for 15 min at 95 °C with matching samples from the very same patients and followed by 50 cycles of 30 s denaturation at 95 °C, pri- healthy controls. mer annealing for 30 s at primer-specific temperatures (Additional file 1: Table S5), and 30 s at 72 °C, with final Isolation of IgG from blood plasma extension at 72 °C for 10 min. The cycling protocol used IgG has been isolated from blood plasma by affinity for amplification of the MGAT3 gene fragments was de- chromatography using CIM Protein G 96-well plate scribed previously [35], with the annealing temperature (BIA Separations, Ajdovščina, Slovenia) and vacuum adjusted to 55 °C for the fragment 1 performed on DNA manifold (Pall Corporation, Port Washington, NY, USA) from CD19 B cells. For quantitative measurement of as previously described [57, 58]. In short, plasma sam- DNA methylation level at specific CpG sites, ples (50–90 μl) were diluted with 1 × PBS, pH 7.4 in the PCR-amplified bisulfite-converted DNA was sequenced ratio 1:7. All samples were filtered through 0.45 and using the PyroMark Q24 Advanced pyrosequencing sys- 0.2-μm AcroPrep GHP filter plates (Pall Corporation) tem (Qiagen) according to the manufacturer’s recom- using vacuum manifold and immediately applied to pre- mendations. Sequences of PCR and pyrosequencing conditioned Protein G plate. After washing of the Klasić et al. Clinical Epigenetics (2018) 10:75 Page 5 of 14 −1 Protein G plate, IgG was eluted with 0.1 mol L formic individual patients were matched with their correspond- acid and immediately neutralized with ammonium bicar- ing glycan profiles. Sizes of datasets and patient classes bonate to pH 7.0. Protein G plate was regenerated and obtained after including complete records (i.e., both stored at 4 °C. methylation data and glycan profiles present) are shown in Additional file 1: Table S6. Individual glycan struc- IgG glycosylation analysis using ultra-performance liquid tures were represented as relative abundances and chromatography batch-corrected. Percentage of structures with bisecting N-glycans from isolated IgG in the Edinburgh cohort N-acetylglucosamine was calculated for each cohort as a were released with PNGase F after drying 300 μl of each derived trait at this point. Glycan structures identified by IgG elution fraction, labeled with 2-aminobenzamide each method were translated to Oxford notation, and and excess of regents removed by clean-up using hydro- only the 13 structures present in both the Edinburgh philic interaction liquid chromatography solid phase and Florence datasets were considered for correlation. extraction (HILIC-SPE). Fluorescently labeled and puri- We used IgG1 data from the Florence cohort, as this fied N-glycans were separated by HILIC-UPLC using isoform was the most abundant. Three additional de- Acquity UPLC instrument (Waters, Milford, MA, USA) rived traits were calculated: ratios of FA2B to FA2, as previously described [43]. Samples were separated FA2BG1 to FA2G1, and FA2BG2 to FA2G2. into 24 peaks [57], and the amount of N-glycans in each Pearson correlation between CpG methylation data chromatographic peak was expressed as a percentage of and 17 glycan features (13 structures and 4 derived total integrated area (% area). traits) was calculated. Significance threshold was set at p < 0.05 with additional Bonferroni correction for IgG glycosylation analysis using liquid chromatography 17-fold multiple testing. coupled to mass spectrometry Methylation of assayed CpG sites in promoters of the In the Florence cohort, Fc-specific IgG glycopeptides BACH2 and MGAT3 genes was correlated with mea- were analyzed after IgG purification, overnight trypsin sured glycan structures. Pearson correlation coefficient digestion at 37 °C, and reverse-phase purification on along with the associated p value was calculated between Chromabond C18 beads using vacuum manifold as average CpG methylation (for all genes/assays) and each described [46, 59]. Samples were analyzed using nanoli- measured IgG glycan structure. Calculation was done on quid chromatography coupled to mass spectrometry pairwise complete observations. Only correlations with (nanoLC-MS), on a nanoACQUITY UPLC system the p value below 0.01 were considered further. Next, (Waters, Milford Massachusetts, USA) coupled to correlation coefficients for all CpG assays were calcu- quadrupole-TOF-MS (Compact; Bruker Daltonics, lated, which was used to rank glycan structures accord- Bremen, Germany) equipped with a sheath-flow ESI ing to regulation by the assayed region. Glycan sprayer (capillary electrophoresis ESI-MS sprayer; structures with the strongest correlation (either positive Agilent Technologies, Santa Clara, USA) as previously or negative) to CpG methylation were then used to described [46]. The nanoACQUITY UPLC system and explain regulatory effects. All calculations and data visu- the Bruker Compact Q-TOF-MS were operated under alizations were done in R language and environment for HyStar software version 3.2. statistical computing (R Foundation for Statistical Com- Data was processed as described previously [46, 60]. puting, Vienna, Austria). Visualization of correlations This resulted in the extraction of 16 IgG1, 16 IgG2/3, was done using the R package “corrplot.” and 11 IgG4 glycoforms. The tryptic Fc-glycopeptides for IgG2 and IgG3 subclasses have identical peptide Results moieties in the Caucasian population and are therefore Promoter methylation of the candidate genes in whole not distinguishable with this methodology. Annotation blood and B cells of IBD patients of the spectra was done based on accurate mass accord- In order to assess the level of methylation in CpG ing to the relevant literature [40, 57]. islands of the five candidate genes (BACH2, MGAT3, IKZF1, LAMB1,and IL6ST), associated with both IBD Correlation analysis and IgG glycosylation by GWAS, we developed sev- Methylation data for the BACH2 (assay 2) and the eral pyrosequencing assays for each of the genes MGAT3 (assays 1 and 2) genes (obtained for the two (Fig. 1 and Additional file 2:FigureS1).Weper- large cohorts) were filtered according to the peak quality formed initial screening of the pyrosequencing assays by rejecting peaks marked as “failed” by the pyrose- on 60 patients. Overall cytosine methylation levels quencing software. Average methylation across all were very low for LAMB1 (average value per group < assayed CpG sites was calculated for each pyrosequenc- 8%), for IL6ST (< 3.5%), and for IKZF1 (< 4%) in the ing assay in each cohort. Methylation results for assayed portion of their promoters; therefore, we Klasić et al. Clinical Epigenetics (2018) 10:75 Page 6 of 14 Fig. 1 Positions of the BACH2 and MGAT3 genes in the human genome and relative positions of the fragments analyzed for methylation level (pyrosequencing assays) within these genes. For each pyrosequencing assay (A1–A5), the region amplified by PCR is shown. Positions of the genes on the chromosomes are shown using chromosome models (red vertical lines). Coordinates are relative to the hg19 human genome assembly. The genes are displayed in the direction corresponding to their reading frames. Annotations (CpG islands and pyrosequencing assays) are to scale. TSS transcription start site could not identify differential methylation. We then methylation level was high, with all CpG sites show- excluded these genes from further analysis. ing a reproducibly significant difference between HC MGAT3 and BACH2 promoter methylation was ana- andbothCDand UC.CpG sites2,13, and15 lyzed in several hundred IBD patients and healthy con- showed significant differences only for CD but not for trols from two independent cohorts (Additional file 1: UC. Direction of change was different for the two Tables S1 and S2). In these genes, we analyzed methyla- genes—differentially methylated CpG sites within the tion level at 47 CpGs covered by five pyrosequencing as- BACH2 promoter were hypomethylated, while those says in the BACH2 gene: 21 CpG sites were in the for the MGAT3 gene were hypermethylated in disease promoter region, 1 CpG site was in first exon, and 25 compared to healthy individuals. CpG sites were located in the first intron of the gene. A These results were confirmed on CD19 Bcells total of 32 CpG sites, covered by five pyrosequencing isolated from peripheral whole blood of the independ- assays, was analyzed for MGAT3: 18 CpG sites were ent, smaller patient sample from the Edinburgh located in the promoter region and 14 CpG sites in the cohort (67 samples). The CpG sites 1–5 and 12–13 in first intron (Fig. 1). Most of those CpG sites were the MGAT3 promoter were differentially methylated located within CpG islands of the both genes. We found between CD and HC. Only CpG site 5 within the differential CpG methylation between IBD patients and assay A2 of the BACH2 gene showed change in the HC within the assay A2, located at 213–368 bp up- methylation level between HC and CD in CD19 B stream (relative to the gene orientation) of the TSS in cells (Fig. 2b). There were no differences in the the BACH2 promoter and within the assays A1 and A2, methylation level of the same CpG sites within located in the CpG island 1 of the MGAT3 gene. The assayed fragments of the BACH2 and MGAT3 genes same pattern of differential methylation at these CpG between UC and HC. sites was observed in whole blood of patients and HC It is worth noting that the same pattern of CpG from two large independent cohorts (Fig. 2). CpG methylation differences was observed in PBMCs of the methylation level was generally low (up to 20%) in the IBD patients and HC from both large independent assayed portion of the BACH2 promoter; however, sig- cohorts, and most of the CpG sites within the assayed nificant differences between HC and CD methylation portion of the MGAT3 promoter were also differentially level were recorded at CpG sites 4, 5, 6, and 8 (Fig. 2a). methylated in CD19 B cells from the subset of IBD For the assayed portion of the MGAT3 promoter, general patients from the Edinburgh cohort (Fig. 2a, b). Klasić et al. Clinical Epigenetics (2018) 10:75 Page 7 of 14 Fig. 2 (See legend on next page.) Klasić et al. Clinical Epigenetics (2018) 10:75 Page 8 of 14 (See figure on previous page.) Fig. 2 Box plot of CpG methylation in peripheral whole blood for the BACH2 and MGAT3 genes in the Edinburgh and Florence cohorts and in B cells from a subset of patients from Edinburgh cohort. Groups were compared using the Mann-Whitney U test with significance threshold of p = 0.05, corrected for multiple testing using the Bonferroni method. a Methylation levels were generally low in the assayed portion of the BACH2 gene promoter, with significant differences between HC and CD methylation at CpG sites 4, 5, 6, and 8 (replicated in both cohorts). For the MGAT3 gene, general methylation level was high, with all CpG sites showing a reproducibly significant difference between HC and both CD and UC, except for CpG sites 2, 13, and 15 for which reproducible significant differences were found only between HC and CD. b In B cells, isolated from PBMCs of a subset of the patients from the Edinburg cohort, differential methylation was found at the CpG position 5 of the BACH2 gene (assay 2) between HC and CD, while for the MGAT3 gene, differentially methylated were CpG sites 1–5, 12, and 13 between HC and CD. CD Crohn’s disease, UC ulcerative colitis, HC healthy controls Promoter methylation of the MGAT3 gene in CD3 T cells inflamed colonic mucosa in comparison with healthy from PBMCs and inflamed colonic mucosa of UC patients mucosa. We included in our investigation biopsy samples of UC patients from an independent cohort from the Gastro- Correlation between the MGAT3 and BACH2 promoter enterology Department of Centro Hospitalar do methylation and IgG glycosylation Porto-Hospital de Santo António, Portugal. Given the There were statistically significant correlations that repli- technical challenges in obtaining DNA and RNA from a cated across assays and cohorts between the MGAT3 small number of purified cells from inflamed colonic promoter methylation and glycan structures FA2, mucosa, a subset of patients with active and inactive FA2G2, FA2BG2, and FA2G2S1, as well as the derived phase of UC was selected for methylation analysis from trait of the ratio of FA2B to FA2 (Fig. 4a). All correla- three sources: (1) PBMCs, (2) CD3 T cells isolated from tions except with FA2 were negative. No reproducible PBMCs, and (3) CD3 T cells isolated from colonic significant correlations could be found between BACH2 mucosa (see also Additional file 1: Table S4). promoter methylation and the glycan structures Inter-individual variation of MGAT3 methylation level (Fig. 4a). measured from PBMCs and from CD3 T cells isolated In order to infer a mechanistic pathway of the from PBMCs was quite large—it varied from 47 to 94% observed correlations, we mapped them to the glycan and from 26 to 90%, respectively. Therefore, we could biosynthesis pathways (Fig. 4b). The ratio of bisecting not find any difference in CpG methylation level glycans to FA2 was taken as an indicator of MGAT3 between UC patients and HC in assayed fragments of (GnT-III) activity. This interpretation allowed us to infer the MGAT3 promoter, neither in PBMCs nor in CD3 T lower GnT-III enzymatic activity when the promoter of cells isolated from PBMCs. However, we recorded a total the MGAT3 gene was methylated. Increase in MGAT3 of 7 (out of 15) differentially methylated CpG sites in promoter methylation correlated with a decrease in cer- CD3 T cells isolated from colonic mucosa of UC tain galactosylated and sialylated structures (Fig. 4b). In patients with active disease compared with HC (Fig. 3). addition to the decreased levels of bisecting GlcNAc on Overall, the methylation level of CpGs within assayed non-galactosylated glycans (B/FA2), the most significant fragments of the MGAT3 promoter was high in CD3 T effect of the MGAT3 promoter methylation on IgG gly- cells from healthy colonic mucosa (between 77 and come composition was a decrease of IgG galactosylation. 98%). When compared to inflamed mucosa of UC patients with active phase of the disease, the same CpG Discussion sites were hypomethylated, with the highest difference at Results from this study strongly indicate that the −5 the CpG position 10 (13.24%; p = 5.08 × 10 ; Fig. 3). In MGAT3 and BACH2 genes play an important role in inactive UC, no significant differences could be found IBD pathogenesis and suggest a possible disease pathway after Bonferroni correction for multiple testing. mediated by the pro-inflammatory properties of IgG It is worth noting that the methylation pattern in antibodies acquired by alterations in Fc glycosylation. CD3 T cells isolated from inflamed colonic mucosa dif- Our recent study, performed on a large cohort of over fered from the methylation patterns in PBMCs and for 1000 IBD patients, reported a significant difference in CD3 T cells isolated from PBMCs. The latter two were IgG glycome composition in both UC and CD compared very similar and had much lower methylation levels than to healthy controls [43, 46]. We found a decrease in that measured for CD3 T cells from inflamed colonic quantity of galacosylated glycans in both CD and UC, as mucosa (Fig. 3). Also, MGAT3 methylation level was well as a decrease in sialylated glycans and an increase increased in UC compared with HC when measured of bisecting GlcNAc on digalactosylated glycan struc- from PBMCs or CD3 T cells isolated from PBMCs tures on IgG in CD. Indeed, alternative N-glycosylation (hypermethylation), while it was decreased (hypomethy- of an IgG molecule influences its function—pro-inflam- lation) when measured from CD3 T cells isolated from matory and anti-inflammatory activity depends on the Klasić et al. Clinical Epigenetics (2018) 10:75 Page 9 of 14 Fig. 3 Box plot of CpG methylation level in the MGAT3 gene promoter (assays A1 and A2) analyzed from PBMCs (a), CD3 T cells isolated from PBMCs (b), and CD3 T cells isolated from inflamed colonic mucosa (c) from the independent cohort of Porto. Changes between UC patients with active disease and HC were statistically significant only in CD3 T cells isolated from inflamed colonic mucosa at CpG positions 3 and 7–12 (p < 0.05 after Bonferroni correction for 15 hypotheses). PBMC peripheral blood mononuclear cells, UC ulcerative colitis, HC healthy controls glycans added on the Cy2 domain of its Fc region [29]. These glycans are of a biantennary complex type with or without bisecting GlcNAc, core fucose, galactose, and sialic acid residues [61]. Recently, this was confirmed in a large multi-centric study of IgG glycome in IBD [46]. Therefore, glycan changes observed on IgG in peripheral blood of UC and CD patients are obviously associated with increased inflammatory potential of IgG, suggesting functional relevance of IgG glycosylation for IBD. Here, we propose a possible mechanism underlying the aberrant IgG glycosylation pattern observed in IBD [43, 46]. Out of five candidate genes analyzed in this work, the MGAT3, a glycosyltransferase which partici- pates in synthesis of IgG glycans, and the BACH2,a transcription factor and a master regulator of a network of genes relevant for B cell integrity [62, 63], showed dif- ferential methylation in peripheral blood of both CD and UC patients when compared to healthy individuals. Even though we identified changes in methylation level for both UC and CD compared to HC, the differences were more pronounced for CD. This is concordant with other studies that explored either whole genome methy- lation or promoter methylation of candidate genes in IBD [11]. The extent of the change in IgG glycome com- position was also consistently higher in CD than UC compared to HC [43, 46]. The protein encoded by the MGAT3 gene (N-acetyl- glucosaminyltransferase III, GnT-III) is responsible for significant functional alteration of glycans on the Fc region of an IgG antibody. The GnT-III adds N-acetyl- glucosamine (GlcNAc) on β1,4-linked mannose in the three-mannose core of N-glycans, producing bisecting GlcNAc structures. In the same CD patients, who showed changed MGAT3 promoter methylation level in peripheral blood cells, a significant increase in the per- centage of bisecting GlcNAc on glycans of circulating IgG antibodies was recorded, too. The association of the MGAT3 with both IgG N-glycosylation [49] and Crohn’s disease [4, 5] suggests that N-glycans with bisecting GlcNAc could be involved in CD pathogenesis through functional effect on IgG antibody. Correlations between BACH2 and MGAT3 promoter methylation and glycan structures have given further insight into the changes of IgG glycosylation pattern me- diated by those two genes (Fig. 4b). The MGAT3 Klasić et al. Clinical Epigenetics (2018) 10:75 Page 10 of 14 Fig. 4 (See legend on next page.) Klasić et al. Clinical Epigenetics (2018) 10:75 Page 11 of 14 (See figure on previous page.) Fig. 4 Correlations between CpG methylation in the BACH2 and MGAT3 gene promoters and glycan structures measured from the same individuals of the Edinburgh and Florence cohorts, mapped to the glycan biosynthesis pathways. a Correlation coefficients between average CpG methylation in the assayed gene promoter fragments and glycan structure percentages are shown as blue (positive) or red (negative correlation) circles with their size and shade proportional to the correlation coefficient. Correlations without statistical significance (p > 0.05 after Bonferroni correction for multiple testing) are crossed. Columns represent 13 individual glycan structures and four derived traits (beige box). EDI Edinburgh cohort, FLO Florence cohort, Bisecting, percentage of all structures with bisecting N-acetylglucosamine, B/FA2 ratio of FA2B to FA2 structures, B/FA2G1 ratio of FA2BG1 to FA2G1 structures, B/ FA2G2 ratio of FA2BG2 to FA2G2 structures. b Glycan biosynthesis pathways with the glycan structures, labels, and the enzymes mapped to correlation results for the MGAT3 gene. Light blue rectangles indicate positive, while light red rectangles indicate negative correlation between the glycan structures or traits and CpG methylation levels. Only correlations replicated across assays and/or cohorts are shown. The red rectangle around the MGAT3 enzyme reflects the negative correlation between CpG methylation and the derived trait B/FA2, which effectively measures enzyme activity at this step. MGAT3 N-acetilglucosaminyltransferase III (GnT-III), FUT8 fucosyltransferase 8, GalT1 galactosyltranserase 1, ST6GalT1 Beta-galactoside alpha-2,6-sialyltransferase 1 promoter methylation probably led to decreased GnT-III (Additional file 1: Tables S7 and S8). Our present efforts enzymatic activity, as revealed by negative correlation are focused on functional studies with hope to reveal a between methylation and total bisecting glycans to FA2 more complete view of the BACH2 role in IgG ratio. Namely, GnT-III adds a bisecting GlcNAc to FA2. glycosylation. A further proof is the positive correlation between the Since DNA methylation pattern is tissue-specific, our MGAT3 promoter methylation and FA2, since it is not goal was to ascertain if CpG methylation from blood surprising that substrate accumulates when enzyme ac- could be a proxy for CpG methylation of the same tivity is decreased. More complex effects were observed candidate gene in the tissue where the inflammation is on galactosylation and sialylation. The negative correl- taking place. In fact, IBD is an immune-mediated dis- ation between MGAT3 methylation and galactosylation order in which T cells are actively implicated in develop- of both, glycans with (FA2BG2) and without bisecting ment of gut-mucosa inflammation [68]. Previous GlcNAc (FA2G2), suggests that the effect of increased evidence has suggested that N-glycosylation of intestinal galactosylation is not caused only by steric effects of T cells is associated with UC pathogenesis and disease bisecting GlcNAc, but also through some indirect effects severity [50, 69]. Therefore, we analyzed MGAT3 pro- of MGAT3 expression on galactosyltransferase activity. It moter methylation in CD3 T cells isolated from PBMCs seems as though galactosylation and sialylation are and from intestinal mucosa of UC patients with active co-regulated with the addition of a bisecting GlcNAc and inactive form of the disease and compared with catalyzed by GnT-III. Furthermore, Dekkers and MGAT3 promoter methylation from PBMCs of the same co-workers recently reported that transfection of cells patients from the Porto cohort. We found 7 out of 15 with MGAT3 causes an increase of IgG galactosyla- differentially methylated CpG sites in CD3 T cells tion [64]. isolated from colonic mucosa of UC patients with active Much weaker correlation was observed between BACH2 form of the disease compared to CD3 T cells from mu- methylation and IgG glycosylation. This was expected since cosa of healthy individuals. On the other hand, there BACH2 is not a glycosyltransferase and thus is not directly were no differences in MGAT3 promoter methylation involved in glycan biosynthetic pathways. However, weak between patients with either active or inactive form of positive correlations with A2G2, FA2G2, and FA2G2S1 UC and healthy controls neither in PBMCs nor in CD3 structures, which involve galactosylation and sialylation, T cells isolated from PBMCs. This could be due to high were observed, as well as weak negative correlation with dispersion in the methylation level when measured from fucosylated bianntenary structure FA2. This is interesting PBMCs (47–94%) and CD3 T cells isolated from because GWA studies associated BACH2 with IgG galacto- PBMCs (26–90%), probably due to small sample size sylation [49] as well as with various immune and inflam- and dispersion in age. Namely, cell composition changes matory diseases including IBD [4, 5, 65–67]inwhich IgG across age in whole blood, and it can explain dispersion acquires pro-inflammatory properties through decrease in of CpG methylation level observed in our sample [70]. galactosylation, sialylation, and fucosylation [43, 46]. Since Considering much smaller dispersion in values of methy- BACH2 is orchestrating a gene regulatory network in B lation level (65–97%) measured from CD3 T cells from cells [62], we believe that some glyco-genes are also regu- lamina propria, the differences could be used as a signa- lated by this transcription factor. Indeed, our in silico ture for inflammation. analysis identified several glyco-genes, mostly galactosyl- transferases (including B4GALT1 and B4GALT2), to pos- Conclusions sess putative AP-1 and NFE2 binding sites for BACH2 Taken together, our results suggest that the aberrant transcription factor [63], suggesting that these galactosyl- methylation observed in the MGAT3 gene in CD3 T transferases could be controlled by BACH2 cells from intestinal mucosa of UC patients, B cells from Klasić et al. Clinical Epigenetics (2018) 10:75 Page 12 of 14 peripheral blood, and the whole peripheral blood in UC Acknowledgements The authors would like to thank Stephanie Scott for her organizational and and CD patients is a possible mechanism underlying in- administrational contribution. The study has been funded by the EU FP7 flammation due to a change in the immune system—ei- grant European Commission IBD-BIOM (contract # 305479), EU FP7 Regional ther through the change of glycans on Fc region of IgGs Potential Grant INTEGRA-Life (contract # 315997), European Structural and Investment Funds grant for the Croatian National Centre of Research or by modulating the glycosylation profile of glycopro- Excellence in Personalized Healthcare (contract # KK.01.1.1.01.0010), and teins on intestinal T cells. Others [24] have shown that Croatian Science Foundation grant EpiGlycoIgG (contract # 3361). Financial some of their candidate genes changed promoter methy- support from Portugal (PI: SSP): FEDER—Fundo Europeu de Desenvolvimento Regional funds through the COMPETE 2020—Operacional Programme for lation level in whole biopsies, while some of the genes Competitiveness and Internationalisation (POCI), Portugal 2020, and by showed changes only in some cell types of the heteroge- Portuguese funds through FCT—Fundação para a Ciência e a Tecnologia/ neous cell population from the epithelial and Ministério da Ciência, Tecnologia e Inovação in the framework of the project (POCI-01/0145-FEDER-016601; PTDC/DTP-PIC/0560/2014) was received. SSP non-epithelial cells, pointing out the importance of cell also acknowledges the European Crohn’s and Colitis Organization (ECCO) separation from mucosal biopsies. Interestingly, one of and the “Broad Medical Research program at Crohn’s and Colitis Foundation the genes that showed differential methylation in the of America-CCFA” for funding. SSP acknowledges the Portuguese Group of Study on IBD (GEDII) for funding. A.M.D. [PD/BD/105982/2014] also acknowledges non-epithelial fraction, representing immune and stro- FCT for funding. mal cells, was FUT7, the fucosyltransferase involved in IBD-BIOM consortium: Daniel Kolarich (Department of Biomolecular Systems, sialyl Lewis X synthesis, a ligand in selectin-mediated Max Planck Institute of Colloids and Interfaces, Potsdam, Germany), Manfred Wuhrer (Center for Proteomics and Metabolomics, Leiden University Medical adhesion of leukocytes to activated endothelium. Fur- Center, Leiden, The Netherlands; Division of BioAnalytical Chemistry, VU thermore, Dias and collaborators proposed a molecular University Amsterdam, Amsterdam, the Netherlands), Dermot P. B. McGovern mechanism in IBD involving another glyco-gene, the (F. Widjaja Family Foundation Inflammatory Bowel and Immunobiology Research Institute, Cedars-Sinai Medical Center, Los Angeles), Iain K. Pemberton MGAT5 (GnT-V), responsible for branching of N-gly- (IP Research Consulting SAS, Paris, France), Daniel IR Spencer (Ludger Ltd., cans. They showed decreased expression of branched Culham Science Centre, Oxford, UK, Daryl L. Fernandes (Ludger Ltd., Culham N-glycans on T cell receptor (TCR) of lamina propria Science Centre, Oxford, UK), Rahul Kalla, Kate O’Leary, Alex T Adams, Hazel Drummond, Elaine Nimmo, Ray Boyapati, David C Wilson (Centre for Genetics associated with disease severity in patients with active and Molecular Medicine, University of Edinburgh, Edinburgh, UK), Ray Doran UC [50]. Dysregulation of N-glycan branching on TCR (Ludger Ltd., Culham Science Centre, Oxford, UK), Igor Rudan (all, Centre for contributes to a decreased threshold of T cell activation Population Health Sciences, University of Edinburgh, Edinburgh, UK), Paolo Lionetti (Paediatric Gastroenterology Unit, AOU Meyer, Viale Pieraccini, Florence, leading to a hyper-immune response which is a feature Italy), Natalia Manetti (Department of Medical and Surgical Sciences, Division of of UC patients. Taken together, our results and those of Gastroenterology, University Hospital Careggi, Florence, Italy), Fabrizio Bossa others suggest an important role of aberrant protein (Department of Medical Sciences, Division of Gastroenterology, IRCCS-CSS Hospital, Viale Cappuccini, Rotondo, Italy), Paola Cantoro, Anna Kohn (Division glycosylation (partly through epigenetic mechanisms) in of Gastroenterology, S. Camillo Hospital, Rome, Italy), Giancarlo Sturniolo IBD through dysregulation of the immune system. Also, (Gastrointestinal Unit, University of Padua, Padua, Italy), Silvio Danese (IBD Unit, in IBD diagnosis and treatment, it is important to find a Humanitas Research Institute, Rozzano, Milan, Italy), Mariek Pierik (Maastricht University Medical Centre (MUMC), Maastricht, the Netherlands), and David C. non-invasive, specific, and clinically useful biomarkers in Wilson (Centre for Genetics and Molecular Medicine, University of Edinburgh, order to identify high-risk patients. Using MGAT3 Edinburgh, UK). hypermethylation together with the glycan traits as markers from peripheral blood of IBD patients seems Funding promising in the disease identification. This independent research was generously supported by the following grants: EU FP7 research grant IBD-BIOM (contract # 305479) to JS, VA, GL, and VZ; EU FP7 Regional Potential Grant INTEGRA-Life (contract # 315997) to GL and VZ; Additional files European Structural and Investment Funds grant for the Croatian National Centre of Research Excellence in Personalized Healthcare (contract # KK.01.1.1.01.0010) to Additional file 1: Supplementary Tables 1-8. Demographics of IBD GL and VZ; Croatian Science Foundation grant EpiGlycoIgG (contract # 3361) to patients and healthy controls (1-4), PCR primers (5), number of samples VZ; FEDER COMPETE 2020 POCI, Portugal 2020, and Portuguese funds through per analysis (6) and in silico analysis of transcription factor binding sites FCT (contracts # POCI-01/0145-FEDER-016601 and PTDC/DTP-PIC/0560/2014) to in gene promoters (7, 8). (DOCX 70 kb) SP; and FTC (contract # PD/BD/105982/2014) to AMD. Additional file 2: Figure S1. Position of the pyrosequencing assays for the genes LAMB1, IL6ST, and IKZF1 in the genome relative to CpG islands, annotated promoters, and exons. (PDF 523 kb) Availability of data and materials All original data can be obtained from the authors upon request. Additional file 3: Figure S2. Violin plots showing the age distribution in IBD patients (CD, UC) and healthy controls (HC). The groups were well matched by age, which was shown by Mann-Whitney U test: no significant Authors’ contributions differences between groups were found at the level p = 0.05. (PDF 704 kb) Study design was conceived by VZ, AV, GL, and SP. Sample provision was provided by JS, NTV, NAK, ERN, VA, RD’I, and SP. Blood and biopsy processing Abbreviations was conducted by AMD and AL. DNA methylation analyses were carried out by CD: Crohn’s disease; EWAS: Epigenome-wide association study; GlcNAc: N- MK, DM, PD, IS, and IB. Glycan analysis was carried out by IT, JŠ,MŠ,and GR. acetylglucosamine; GWAS: Genome-wide association study; HC: Healthy Statistical and correlation analyses were carried out by AV, MK, and DM. control; IBD: Inflammatory bowel disease; IgG: Immunoglobulin G; LC- Drafting of the manuscript was carried out by VZ, AV, MK, DM, and IT-A. All MS: Liquid chromatography coupled to mass spectrometry; authors were involved in critical review, editing, revision, and approval of the PBMCs: Peripheral blood mononuclear cells; UC: Ulcerative colitis final manuscript. Klasić et al. Clinical Epigenetics (2018) 10:75 Page 13 of 14 Ethics approval and consent to participate 12. Nimmo ER, Prendergast JG, Aldhous MC, Kennedy NA, Henderson P, For the Edinburgh and Florence cohorts, ethical approvals were obtained Drummond HE, Ramsahoye BH, Wilson DC, Semple CA, Satsangi J. Genome- from Tayside Committee on Medical Ethics B, and all patients and controls wide methylation profiling in Crohnʼs disease identifies altered epigenetic provided written, informed consent (LREC 06/S1101/16, LREC 2000/4/192). regulation of key host defense mechanisms including the Th17 pathway. For the patients from Portugal, all clinical procedures and research protocols Inflamm Bowel Dis. 2012;18:889–99. were approved by the ethics committee of CHP/HSA, Portugal (233/12(179- 13. Ventham NT, Kennedy NA, Adams AT, Kalla R, Heath S, O'Leary KR, DEFI/177-CES); all participants gave their informed consent. Drummond H, consortium IB, consortium IC, Wilson DC, et al: Integrative epigenome-wide analysis demonstrates that DNA methylation may mediate Competing interests genetic risk in inflammatory bowel disease. Nat Commun 2016, 7:13507. The authors declare that they have no competing interests. 14. Low D. DNA methylation in inflammatory bowel disease and beyond. World J Gastroenterol. 2013;19:5238. 15. Ventham NT, Kennedy NA, Nimmo ER, Satsangi J. Beyond gene discovery in Publisher’sNote inflammatory bowel disease: the emerging role of epigenetics. Springer Nature remains neutral with regard to jurisdictional claims in Gastroenterology. 2013;145:293–308. published maps and institutional affiliations. 16. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet. 2012;13:484–92. Author details 17. Ambatipudi S, Cuenin C, Hernandez-Vargas H, Ghantous A, Le Calvez-Kelm Department of Biology, Division of Molecular Biology, Faculty of Science, F, Kaaks R, Barrdahl M, Boeing H, Aleksandrova K, Trichopoulou A, et al. University of Zagreb, Horvatovac 102a, 10000 Zagreb, Croatia. Tobacco smoking-associated genome-wide DNA methylation changes in Gastrointestinal Unit, Centre for Genomics and Molecular Medicine, the EPIC study. Epigenomics. 2016;8:599–618. University of Edinburgh, Edinburgh EH4 6XU, UK. Genos Glycoscience 18. Barrès R, Yan J, Egan B, Treebak Jonas T, Rasmussen M, Fritz T, Caidahl K, Krook Research Laboratory, Borongajska cesta 83h, 10000 Zagreb, Croatia. Faculty A, O’Gorman Donal J, Zierath Juleen R. Acute exercise remodels promoter of Pharmacy and Biochemistry, University of Zagreb, Zagreb, Croatia. IBD methylation in human skeletal muscle. Cell Metab. 2012;15:405–11. Pharmacogenetics, University of Exeter, Exeter, UK. Institute of Molecular 19. Olszak T, An D, Zeissig S, Vera MP, Richter J, Franke A, Glickman JN, Siebert R, Pathology and Immunology of the University of Porto (IPATIMUP), Porto, Baron RM, Kasper DL, Blumberg RS. Microbial exposure during early life has Portugal. Department of Medical and Surgical Sciences, Division of persistent effects on natural killer T cell function. Science. 2012;336:489–93. Gastroenterology, University Hospital Careggi, Florence, Italy. Department of 20. Symonds ME, Sebert SP, Budge H. The impact of diet during early life Medical Sciences, Division of Gastroenterology, IRCCS-CSS Hospital, Viale and its contribution to later disease: critical checkpoints in Cappuccini, Rotondo, Italy. Gastrointestinal Unit, University of Padua, Padua, development and their long-term consequences for metabolic health. Italy. Translational Gastroenterology Unit, Nuffield Department of Medicine, Proc Nutr Soc. 2009;68:416. University of Oxford, Oxford, UK. 21. Vaissiere T, Hung RJ, Zaridze D, Moukeria A, Cuenin C, Fasolo V, Ferro G, Paliwal A, Hainaut P, Brennan P, et al. Quantitative analysis of DNA Received: 28 February 2018 Accepted: 22 May 2018 methylation profiles in lung cancer identifies aberrant DNA methylation of specific genes and its association with gender and cancer risk factors. Cancer Res. 2009;69:243–52. References 22. Milagro FI, Mansego ML, De Miguel C, Martinez JA. Dietary factors, 1. Burisch J, Pedersen N, Čuković-Čavka S, Brinar M, Kaimakliotis I, Duricova D, epigenetic modifications and obesity outcomes: progresses and Shonová O, Vind I, Avnstrøm S, Thorsgaard N, et al. East–West gradient in perspectives. Mol Asp Med. 2013;34:782–812. the incidence of inflammatory bowel disease in Europe: the ECCO-EpiCom 23. Adams AT, Kennedy NA, Hansen R, Ventham NT, OʼLeary KR, Drummond HE, inception cohort. Gut. 2013;63:588–97. Noble CL, El-Omar E, Russell RK, Wilson DC, et al. Two-stage genome-wide 2. Xavier RJ, Podolsky DK. Unravelling the pathogenesis of inflammatory bowel methylation profiling in childhood-onset Crohnʼs disease implicates disease. Nature. 2007;448:427–34. epigenetic alterations at the VMP1/MIR21 and HLA loci. Inflamm Bowel Dis. 3. Anderson CA, Boucher G, Lees CW, Franke A, D'Amato M, Taylor KD, Lee JC, 2014;20:1784–93. Goyette P, Imielinski M, Latiano A, et al. Meta-analysis identifies 29 24. Cooke J, Zhang H, Greger L, Silva A-L, Massey D, Dawson C, Metz A, Ibrahim additional ulcerative colitis risk loci, increasing the number of confirmed A, Parkes M. Mucosal genome-wide methylation changes in inflammatory associations to 47. Nat Genet. 2011;43:246–52. bowel disease. Inflamm Bowel Dis. 2012;18:2128–37. 4. Franke A, Balschun T, Sina C, Ellinghaus D, Häsler R, Mayr G, Albrecht M, 25. Harris RA, Nagy-Szakal D, Mir SAV, Frank E, Szigeti, Kaplan JL, Bronsky J, Wittig M, Buchert E, Nikolaus S, et al. Genome-wide association study for Opekun A, Ferry GD, Winter H, Kellermayer R. DNA methylation-associated ulcerative colitis identifies risk loci at 7q22 and 22q13 (IL17REL). Nat Genet. colonic mucosal immune and defense responses in treatment-naïve 2010;42:292–4. pediatric ulcerative colitis. Epigenetics. 2014;9:1131–7. 5. Franke A, McGovern DP, Barrett JC, Wang K, Radford-Smith GL, Ahmad T, 26. Karatzas PS, Gazouli M, Safioleas M, Mantzaris GJ. DNA methylation changes Lees CW, Balschun T, Lee J, Roberts R, et al. Genome-wide meta-analysis in inflammatory bowel disease. Ann Gastroenterol. 2014;27:125–32. increases to 71 the number of confirmed Crohn’s disease susceptibility loci. 27. Hart GW, Copeland RJ. Glycomics hits the big time. Cell. 2010;143:672–6. Nat Genet. 2010;42:1118–25. 28. Böhm S, Kao D, Nimmerjahn F. Sweet and sour: the role of glycosylation for 6. Jostins L, Ripke S, Weersma RK, Duerr RH, McGovern DP, Hui KY, Lee JC, the anti-inflammatory activity of immunoglobulin G. In: Fc Receptors. Basel: Schumm LP, Sharma Y, Anderson CA, et al. Host-microbe interactions have Springer International Publishing; 2014. p. 393–417. 393–417. shaped the genetic architecture of inflammatory bowel disease. Nature. 29. Kaneko Y. Anti-inflammatory activity of immunoglobulin G resulting from Fc 2012;491:119–24. sialylation. Science. 2006;313:670–3. 7. Liu JZ, van Sommeren S, Huang H, Ng SC, Alberts R, Takahashi A, Ripke S, 30. National Research Council (U.S.). Committee on Assessing the Importance Lee JC, Jostins L, Shah T, et al. Association analyses identify 38 susceptibility and Impact of Glycomics and Glycosciences., National Research Council (U.S. loci for inflammatory bowel disease and highlight shared genetic risk across ). Board on Chemical Sciences and Technology., National Research Council populations. Nat Genet. 2015;47:979–86. (U.S.). Board on Life Sciences.: Transforming glycoscience: a roadmap for the 8. Bonaz BL, Bernstein CN. Brain-gut interactions in inflammatory bowel future. Washington, D.C.: National Academies Press; 2012. disease. Gastroenterology. 2013;144:36–49. 31. Stowell SR, Ju T, Cummings RD. Protein glycosylation in cancer. Annu Rev 9. Pituch-Zdanowska A, Banaszkiewicz A, Albrecht P. The role of dietary fibre in Pathol: Mech Dis. 2015;10:473–510. inflammatory bowel disease. Gastroenterol Rev. 2015;3:135–41. 10. Rubin DT, Hanauer SB. Smoking and inflammatory bowel disease. Eur J 32. Horvat T, Deželjin M, Redžić I, Barišić D, Herak Bosnar M, Lauc G, Zoldoš V. Gastroenterol Hepatol. 2000;12:855–62. Reversibility of membrane N-Glycome of HeLa cells upon treatment with 11. McDermott E, Ryan EJ, Tosetto M, Gibson D, Burrage J, Keegan D, Byrne K, epigenetic inhibitors. PLoS One. 2013;8:e54672. Crowe E, Sexton G, Malone K, et al. DNA methylation profiling in 33. Horvat T, Mužinić A, Barišić D, Bosnar MH, Zoldoš V. Epigenetic modulation inflammatory bowel disease provides new insights into disease of the HeLa cell membrane N-glycome. Biochim Biophys Acta Gen Subj. pathogenesis. J Crohn’s Colitis. 2015;10:77–86. 2012;1820:1412–9. Klasić et al. Clinical Epigenetics (2018) 10:75 Page 14 of 14 34. Horvat T, Zoldoš V, Lauc G. Evolutional and clinical implications of the 53. Stewart SK, Morris TJ, Guilhamon P, Bulstrode H, Bachman M, epigenetic regulation of protein glycosylation. Clin Epigenetics. 2011;2: Balasubramanian S, Beck S. oxBS-450K: a method for analysing 425–32. hydroxymethylation using 450K BeadChips. Methods. 2015;72:9–15. 35. Klasić M, Krištić J, Korać P, Horvat T, Markulin D, Vojta A, Reiding KR, Wuhrer 54. Kroeze LI, Aslanyan MG, van Rooij A, Koorenhof-Scheele TN, Massop M, M, Lauc G, Zoldoš V. DNA hypomethylation upregulates expression of the Carell T, Boezeman JB, Marie JP, Halkes CJ, de Witte T, et al. Characterization MGAT3 gene in HepG2 cells and leads to changes in N-glycosylation of of acute myeloid leukemia based on levels of global hydroxymethylation. secreted glycoproteins. Sci Rep. 2016;6:24363. Blood. 2014;124:1110–8. 55. Sanchez-Guerra M, Zheng Y, Osorio-Yanez C, Zhong J, Chervona Y, Wang S, 36. Saldova R, Dempsey E, Pérez-Garay M, Mariño K, Watson JA, Blanco- Chang D, McCracken JP, Diaz A, Bertazzi PA, et al. Effects of particulate Fernández A, Struwe WB, Harvey DJ, Madden SF, Peracaula R, et al. 5-AZA- matter exposure on blood 5-hydroxymethylation: results from the Beijing 2′-deoxycytidine induced demethylation influences N-glycosylation of truck driver air pollution study. Epigenetics. 2015;10:633–42. secreted glycoproteins in ovarian cancer. Epigenetics. 2011;6:1362–72. 56. Yu M, Hon GC, Szulwach KE, Song CX, Zhang L, Kim A, Li X, Dai Q, Shen Y, 37. Vojta A, Samaržija I, Bočkor L, Zoldoš V. Glyco-genes change expression in Park B, et al. Base-resolution analysis of 5-hydroxymethylcytosine in the cancer through aberrant methylation. Biochim Biophys Acta Gen Subj. 2016; mammalian genome. Cell. 2012;149:1368–80. 1860:1776–85. 57. Pucic M, Knezevic A, Vidic J, Adamczyk B, Novokmet M, Polasek O, Gornik O, 38. Zoldoš V, Horvat T, Novokmet M, Cuenin C, Mužinić A, Pučić M, Huffman JE, Supraha-Goreta S, Wormald MR, Redzic I, et al. High throughput isolation Gornik O, Polašek O, Campbell H, et al. Epigenetic silencing of HNF1A and glycosylation analysis of IgG-variability and heritability of the IgG associates with changes in the composition of the human plasma N- glycome in three isolated human populations. Mol Cell Proteomics. 2011; glycome. Epigenetics. 2012;7:164–72. M111(010090):10. 39. Theodoratou E, Campbell H, Ventham NT, Kolarich D, Pučić-Baković M, 58. Trbojevic-Akmacic I, Ugrina I, Lauc G. Comparative analysis and validation of Zoldoš V, Fernandes D, Pemberton IK, Rudan I, Kennedy NA, et al. The role different steps in glycomics studies. Methods Enzymol. 2017;586:37–55. of glycosylation in IBD. Nat Rev Gastroenterol Hepatol. 2014;11(10):588–600. 59. Selman MH, Derks RJ, Bondt A, Palmblad M, Schoenmaker B, Koeleman CA, 40. Arnold JN, Wormald MR, Sim RB, Rudd PM, Dwek RA. The impact of van de Geijn FE, Dolhain RJ, Deelder AM, Wuhrer M. Fc specific IgG glycosylation on the biological function and structure of human glycosylation profiling by robust nano-reverse phase HPLC-MS using a immunoglobulins. Annu Rev Immunol. 2007;25:21–50. sheath-flow ESI sprayer interface. J Proteome. 2012;75:1318–29. 41. Miyahara K, Nouso K, Saito S, Hiraoka S, Harada K, Takahashi S, Morimoto Y, 60. Jansen BC, Falck D, de Haan N, Hipgrave Ederveen AL, Razdorov G, Lauc G, Kobayashi S, Ikeda F, Miyake Y, et al. Serum glycan markers for evaluation of Wuhrer M. LaCyTools: a targeted liquid chromatography-mass spectrometry disease activity and prediction of clinical course in patients with ulcerative data processing package for relative quantitation of glycopeptides. J colitis. PLoS One. 2013;8:e74861. Proteome Res. 2016;15:2198–210. 42. Shinzaki S, Iijima H, Nakagawa T, Egawa S, Nakajima S, Ishii S, Irie T, Kakiuchi 61. Fujii S, Nishiura T, Nishikawa A, Miura R, Taniguchi N. Structural Y, Nishida T, Yasumaru M, et al. IgG oligosaccharide alterations are a novel heterogeneity of sugar chains in immunoglobulin G. Conformation of diagnostic marker for disease activity and the clinical course of immunoglobulin G molecule and substrate specificities of inflammatory bowel disease. Am J Gastroenterol. 2008;103:1173–81. glycosyltransferases. J Biol Chem. 1990;265:6009–18. 43. Trbojevic Akmacic I, Ventham NT, Theodoratou E, Vuckovic F, Kennedy NA, 62. Igarashi K, Ochiai K, Itoh-Nakadai A, Muto A. Orchestration of plasma cell Kristic J, Nimmo ER, Kalla R, Drummond H, Stambuk J, et al. Inflammatory differentiation by Bach2 and its gene regulatory network. Immunol Rev. bowel disease associates with proinflammatory potential of the 2014;261:116–25. immunoglobulin G glycome. Inflamm Bowel Dis. 2015;21:1237–47. 63. Kuwahara M, Suzuki J, Tofukuji S, Yamada T, Kanoh M, Matsumoto A, 44. Plomp R, Ruhaak LR, Uh HW, Reiding KR, Selman M, Houwing-Duistermaat Maruyama S, Kometani K, Kurosaki T, Ohara O, et al. The Menin–Bach2 axis JJ, Slagboom PE, Beekman M, Wuhrer M. Subclass-specific IgG glycosylation is critical for regulating CD4 T-cell senescence and cytokine homeostasis. is associated with markers of inflammation and metabolic health. Sci Rep. Nat Commun. 2014;5:3555. 2017;7:12325. 64. Dekkers G, Plomp R, Koeleman CA, Visser R, von Horsten HH, Sandig V, 45. Schwab I, Nimmerjahn F. Intravenous immunoglobulin therapy: how does Rispens T, Wuhrer M, Vidarsson G. Multi-level glyco-engineering techniques IgG modulate the immune system? Nat Rev Immunol. 2013;13:176–89. to generate IgG with defined Fc-glycans. Sci Rep. 2016;6:36964. 46. Simurina M, de Haan N, Vuckovic F, Kennedy NA, Stambuk J, Falck D, 65. Chu X, Pan C-M, Zhao S-X, Liang J, Gao G-Q, Zhang X-M, Yuan G-Y, Li C-G, Trbojevic-Akmacic I, Clerc F, Razdorov G, Khon A, et al: Glycosylation of Xue L-Q, Shen M, et al. A genome-wide association study identifies two immunoglobulin G associates with clinical features of inflammatory bowel new risk loci for Graves’ disease. Nat Genet. 2011;43:897–901. diseases. Gastroenterology. 2018;154(5):1320–1333.e10. 66. Dubois PCA, Trynka G, Franke L, Hunt KA, Romanos J, Curtotti A, Zhernakova 47. Barrett JC, Lee JC, Lees CW, Prescott NJ, Anderson CA, Phillips A, Wesley E, A, Heap GAR, Ádány R, Aromaa A, et al. Multiple common variants for celiac Parnell K, Zhang H, Drummond H, et al. Genome-wide association study of disease influencing immune gene expression. Nat Genet. 2010;42:295–302. ulcerative colitis identifies three new susceptibility loci, including the HNF4A 67. Sawcer S, Hellenthal G, Pirinen M, Spencer CC, Patsopoulos NA, Moutsianas region. Nat Genet. 2009;41:1330–4. L, Dilthey A, Su Z, Freeman C, Hunt SE, et al. Genetic risk and a primary role 48. McGovern DPB, Jones MR, Taylor KD, Marciante K, Yan X, Dubinsky M, for cell-mediated immune mechanisms in multiple sclerosis. Nature. 2011; Ippoliti A, Vasiliauskas E, Berel D, Derkowski C, et al. Fucosyltransferase 2 476:214–9. (FUT2) non-secretor status is associated with Crohn’s disease. Hum Mol 68. Macdonald TT, Monteleone G. Immunity, inflammation, and allergy in the Genet. 2010;19:3468–76. gut. Science. 2005;307:1920–5. 49. Lauc G, Huffman JE, Pucic M, Zgaga L, Adamczyk B, Muzinic A, Novokmet 69. Vossenkamper A, Hundsrucker C, Page K, van Maurik A, Sanders TJ, Stagg AJ, M, Polasek O, Gornik O, Kristic J, et al. Loci associated with N-glycosylation Das L, MacDonald TT. A CD3-specific antibody reduces cytokine production of human immunoglobulin G show pleiotropy with autoimmune diseases and alters phosphoprotein profiles in intestinal tissues from patients with and haematological cancers. PLoS Genet. 2013;9:e1003225. inflammatory bowel disease. Gastroenterology. 2014;147:172–83. 50. Dias AM, Dourado J, Lago P, Cabral J, Marcos-Pinto R, Salgueiro P, Almeida 70. Jaffe AE, Irizarry RA. Accounting for cellular heterogeneity is critical in CR, Carvalho S, Fonseca S, Lima M, et al. Dysregulation of T cell receptor N- epigenome-wide association studies. Genome Biol. 2014;15:R31. glycosylation: a molecular mechanism involved in ulcerative colitis. Hum Mol Genet. 2014;23:2416–27. 51. Reinius LE, Acevedo N, Joerink M, Pershagen G, Dahlen SE, Greco D, Soderhall C, Scheynius A, Kere J. Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS One. 2012;7:e41361. 52. Krištić J, Zoldoš V, Lauc G. Complex genetics of protein N-glycosylation. In: Endo T, Seeberger PH, Hart GW, Wong C-H, Taniguchi N, editors. Glycoscience: biology and medicine. Tokyo: Springer Japan; 2014. p. 1–7. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clinical Epigenetics Springer Journals
Loading next page...
 
/lp/springer_journal/promoter-methylation-of-the-mgat3-and-bach2-genes-correlates-with-the-YX5H7WyL7F
Publisher
BioMed Central
Copyright
Copyright © 2018 by The Author(s).
Subject
Biomedicine; Human Genetics; Gene Function
ISSN
1868-7075
eISSN
1868-7083
D.O.I.
10.1186/s13148-018-0507-y
Publisher site
See Article on Publisher Site

Abstract

Background: Many genome- and epigenome-wide association studies (GWAS and EWAS) and studies of promoter methylation of candidate genes for inflammatory bowel disease (IBD) have demonstrated significant associations between genetic and epigenetic changes and IBD. Independent GWA studies have identified genetic variants in the BACH2, IL6ST, LAMB1, IKZF1,and MGAT3 loci to be associated with both IBD and immunoglobulin G (IgG) glycosylation. Methods: Using bisulfite pyrosequencing, we analyzed CpG methylation in promoter regions of these five genes from peripheral blood of several hundred IBD patients and healthy controls (HCs) from two independent cohorts, respectively. Results: We found significant differences in the methylation levels in the MGAT3 and BACH2 genes between both Crohn’s disease and ulcerative colitis when compared to HC. The same pattern of methylation changes was identified for both genes in CD19 B cells isolated from the whole blood of a subset of the IBD patients. A correlation analysis was performed between the MGAT3 and BACH2 promoter methylation and individual IgG glycans, measured in the same individuals of the two large cohorts. MGAT3 promoter methylation correlated significantly with galactosylation, sialylation, and bisecting GlcNAc on IgG of the same patients, suggesting that activity of the GnT-III enzyme, encoded by this gene, might be altered in IBD. The correlations between the BACH2 promoter methylation and IgG glycans were less obvious, since BACH2 is not a glycosyltransferase and therefore may affect IgG glycosylation only indirectly. Conclusions: Our results suggest that epigenetic deregulation of key glycosylation genes might lead to an increase in pro-inflammatory properties of IgG in IBD through a decrease in galactosylation and sialylation and an increase of bisecting GlcNAc on digalactosylated glycan structures. Finally, we showed that CpG methylation in the promoter of the MGAT3 gene is altered in CD3 T cells isolated from inflamed mucosa of patients with ulcerative colitis from a third smaller cohort, for which biopsies were available, suggesting a functional role of this glyco-gene in IBD pathogenesis. * Correspondence: vzoldos@biol.pmf.hr Marija Klasić and Dora Markulin contributed equally to this work. Department of Biology, Division of Molecular Biology, Faculty of Science, University of Zagreb, Horvatovac 102a, 10000 Zagreb, Croatia Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Klasić et al. Clinical Epigenetics (2018) 10:75 Page 2 of 14 Background proteins, including components of the immune system Inflammatory bowel disease (IBD) is a chronic intestinal [27]. Aberrant protein glycosylation is implicated in inflammatory condition classified in two major forms— virtually every human complex disease, including inflam- Crohn’s disease (CD) and ulcerative colitis (UC)—which mation [28–31]. Previous studies have suggested that exhibit etiologically and clinically distinct features. Now- N-glycosylation of secreted and membrane proteins adays, IBD affects 2.5–3 million people in Europe and might be regulated epigenetically and that aberrant gly- causes considerable morbidity [1]. Despite numerous cosylation profiles in disease can arise through aberrant clinical, genetic, and other experimental studies, our epigenetics [32–38]. A comprehensive review about the understanding of IBD development and progression role of protein glycosylation in IBD has been given remains incomplete. recently [39]. N-glycosylation of serum-circulating pro- It is generally accepted that IBD represents an aber- teins (such as the acute phase proteins; immunoglobulin rant immune response to gut microbiota in genetically G, IgG; and immunoglobulin A, IgA) or whole plasma susceptible individuals [2]. Genome-wide association N-glycome (i.e., N-glycans present on all plasma pro- studies (GWAS) have identified over 200 genetic suscep- teins) has been the focus of IBD biomarker discovery tibility loci, the majority of which were associated with [36, 40–43]. In addition, our partners from IBD consor- both forms of IBD in genome-wide meta-analysis [3–7]. tium and others established that altered glycosylation of However, common genetic variants account only for 8.2 IgG, which is a key effector of the humoral immune and 13.1% heritability of UC and CD, respectively [7]. system, has a role in balancing inflammation at the sys- Interactionofanindividual’s gut microbiome, im- temic level [42–46]. mune system, genetic background, and environmental GWA studies indicated associations of IBD with sev- factors, such as smoking, diet, drugs, and physical eral loci involved in protein glycosylation [47, 48]. More activity [2, 8–10], makes IBD a complex etiopatho- recently, the first GWAS of IgG glycosylation identified genic entity. The challenge is therefore to identify 16 loci specifically associated with changes in IgG glyco- additional factors involved in the development and sylation [49]. Interestingly, five of these loci showed plei- progression of this disease, especially given its rapidly otropy with IBD: MGAT3, a glyco-gene encoding for a increasing incidence. It is probable that epigenetics glycosyltransferase, GnT-III; LAMB1, a member of play a key role in the interactions between environ- transmembrane glycoprotein family of extracellular mental, microbial, and genetic factors that participate matrix; the IL6ST, a signal transducer shared by many in IBD development and progression. These include cytokines; IKZF1; and BACH2, transcription factors in- DNA methylation and histone modifications, as well volved in B cell differentiation, activation, and matur- as some other epigenetic mechanisms [11–13]; for a ation. Only the MGAT3 is a classical glyco-gene with a review, see [14, 15]. known function in IgG glycosylation, while the exact DNA methylation remains the most studied epigenetic functional roles for other four GWAS hits in IgG glyco- modification, readily assayed in a large number of indi- sylation or IBD remain unknown. viduals/samples. Hypermethylation of gene promoters is In this study, we investigated promoter methylation generally associated with gene silencing, while promoter differences in these five genes, associated with both IBD hypomethylation is associated with gene activation [16]. and IgG glycosylation, in peripheral whole blood of Environmentally changed DNA methylation pattern may several hundred IBD patients from two independent contribute to the development of many complex diseases cohorts. We also correlated promoter methylation data by mediating the interplay between external and internal with IgG glycosylation data analyzed previously for the factors and the gene expression [17–21]. There are also same IBD patients by our partners from the IBD consor- data to suggest that the aforementioned environmen- tium [43, 46, 50]. Peripheral blood was used for DNA tal modifiers of IBD can also affect DNA methylation methylation analysis and serum or plasma was used for [17–19, 22]. Epigenetic component of IBD has been glycan analysis, since one of our goals was the search for addressed in many studies, mostly by whole genome potential IBD biomarkers. As peripheral whole blood is methylation analysis performed on peripheral blood a heterogeneous cell mixture with specific methylation mononuclear cells (PBMCs) or mucosal tissue, reveal- pattern for each of the cell types [51], we also analyzed ing regions differentially methylated between the promoter methylation of our candidate genes in CD19 diseaseand healthystate,aswellasbetween CD and B cells and CD3 T cells isolated from peripheral blood UC [11–13, 23–26]. mononuclear cells (PBMCs). B cells were of our particu- The majority of eukaryotic proteins are modified by lar interest since these cells produce IgG on their mem- addition of complex oligosaccharides (glycans) through brane and are precursors of plasma cells which secrete the process of glycosylation. Therefore, glycans are an IgG. We have further explored if aberrant promoter integral part of nearly all membrane and secreted methylation recorded in peripheral whole blood of IBD Klasić et al. Clinical Epigenetics (2018) 10:75 Page 3 of 14 patients can be a proxy for epigenetic events occurring allowed them to clot at 4 °C for 60 min, and then centri- in the inflamed mucosa. To address this question, we fuged at 2500×g for 15 min. The serum was aliquoted analyzed DNA methylation from PBMCs, CD3 T cells off and stored at − 80 °C until further analysis. isolated from PBMCs, and CD3 T cells isolated from A subset of patients and controls recruited in Edinburgh inflamed colonic mucosa of UC patients from the third (Additional file 1: Table S3) underwent immunomagnetic smaller cohort, for which biopsies were available. cell separation to obtain CD19 B cells. The methods have previously been detailed elsewhere [13]. Venepuncture Methods using 9-ml K3 EDTA vacuette (Greiner) tubes was Patient selection and ethics performed to obtain between of 18 and 36 ml of Patients were recruited prospectively from Edinburgh, EDTA-buffered blood. An initial Ficoll (Ficoll-Paque, GE UK, and Florence, Italy, as a part of the IBD-BIOM pro- Healthcare, Bucks, UK) density gradient centrifugation ject. The recruitment of patients from Edinburgh has been was performed to obtain peripheral blood mononuclear described elsewhere [13, 43]. Briefly, we recruited IBD cells. Cells labelled with antibody-coated microbeads + + 7 patients prospectively as close as possible to the date of (human CD8 and CD19 microbeads, 20 μlper 1×10 diagnosis from gastroenterology outpatient and endoscopy cells) were immunomagnetic separated using the auto- appointments between 2012 and 2015. We recruited MACs Pro cell separator (Miltenyi, Germany). CD19 symptomatic controls from gastroenterology clinics dur- separations were performed following an initial CD8 ing the same period. In these individuals, we had excluded depletion step. Nucleic acids were extracted using AllPrep IBD and other organic bowel pathology following bio- (Qiagen, Hilden, Germany) according to the manufac- chemical and/or endoscopic investigations. We recruited turer’s guidance and stored at − 80 °C. a further healthy volunteer cohort with no gastrointestinal Colonic biopsies from controls and UC patients with symptoms. IBD patients were stratified by disease type inactive and active form of disease were mechanically (ulcerative colitis, UC, and Crohn’s disease, CD). Detailed dissociated to prepare single-cell suspensions using genetic, phenotypic, and other data regarding IBD cases Hanks’ balanced salt solution modified medium, without are given in Additional file 1: Tables S1 and S2. Florence calcium chloride and magnesium sulfate (HBSS) (Sigma), cohort was collected through the network of the Italian with penicillin/streptomycin and gentamicin. PBMCs Group for IBD (IG-IBD) since the beginning of 2001 and were obtained by density gradient centrifugation using first described in 2005 [1] following an internal validation Lymphoprep. CD3 T cells (from biopsies and blood) of phenotyping. Subsequently, longitudinal update has were magnetically sorted by using the EasySep™ Human been performed on a yearly basis. T Cell Enrichment Kit (STEMCELL) following the man- Ethical approvals were obtained from Tayside Com- ufacturer’s instructions. Following cell isolation, DNA mittee on Medical Ethics B, and all patients and controls extraction was performed using the Invisorb Spin Tissue provided written, informed consent (LREC 06/S1101/16, Mini Kit (Stratec Molecular) following the manufac- LREC 2000/4/192). turer’s instructions. Florence recruitment details DNA methylation analysis IBD patients were prospectively recruited as close as We analyzed promoter methylation of the candidate possible to the date of diagnosis from gastroenterology genes in the DNA from whole blood, as well as from the outpatient and endoscopy appointments between years separated CD19 B cells. In addition, for the MGAT3, 2012 and 2015 in different tertiary referral centers in which is a glycosyltransferase with direct and known San Giovanni Rotondo, Rome, Rozzano (Milan), Padua, function in IgG glycosylation [52], we analyzed promoter and Florence, Italy. Symptomatic controls were recruited methylation—in DNA from PBMCs, CD3 T cells iso- in the same centers (gastroenterology clinics) during the lated from PBMCs, and CD3 T cells isolated from the same period. In these individuals, IBD and other organic colonic mucosa of healthy controls and UC patients bowel pathology were excluded by biochemical and/or (classified according to active and inactive form of the endoscopic investigations. IBD patients were stratified disease) of the third independent smaller subcohort col- by disease type (ulcerative colitis, UC, and Crohn’s lected by the Gastroenterology Department of Centro disease, CD). Samples were obtained with the same Hospitalar do Porto-Hospital de Santo António, Portugal methodology (see further) and centrally collected at San (Additional file 1: Table S4). All specimens were sub- Giovanni Rotondo, Italy. jected to histological examination and classification. All participants gave informed consent about all clinical Sample collection procedures, and research protocols were approved by We collected whole blood at the time of patient recruit- the ethics committee of CHP/HSA, Portugal (233/ ment into 9-ml serum Z-clot activator tubes (Greiner), 12(179-DEFI/177-CES). Klasić et al. Clinical Epigenetics (2018) 10:75 Page 4 of 14 For DNA methylation analysis, 500 ng of DNA from primers for the BACH2 and the MGAT3 genes are listed whole blood was bisulfite converted using EZ-96 DNA in Additional file 1: Table S5. EpiTect PCR Control DNA Methylation Gold kit (Zymo Research, Freiburg, Set (methylated and unmethylated bisulfite-converted Germany), and 100 ng of DNA from CD19 B cells, human DNA, Qiagen) was used as a control for PCR and PBMCs, and T cells was converted using EZ DNA pyrosequencing reactions. Methylation Gold kit (Zymo Research, Freiburg, Germany) according to the manufacturer’s protocol. Statistical analysis Two to six pyrosequencing assays were developed for The nonparametric Mann-Whitney U test was used to promoter regions of each of the five candidate genes compare the methylation status of CpG sites encom- (BACH2, MGAT3, IL6ST, IKZF1, and LAMB1). The se- passed by the pyrosequencing assays in the MGAT3 and lection of analyzed CpG sites was random for assays 2–5 BACH2 genes between the two independent groups: HC of the MGAT3 gene. CpG sites within the MGAT3 assay compared to each of CD or UC. Significance threshold 1 were selected based on the GEO (Gene Expression was set at p < 0.05 with additional Bonferroni correction Omnibus) database where methylation data were for multiple testing. Given that age was our primary obtained using Illumina HumanMethylation450 Bead- concern as a potential confounder, we visualized the age Chip v1.1 technology. For the BACH2 gene, assays were in the three groups (CD, UC, and HC) for the samples selected based on location of differentially methylated included in each analysis as violin plots (Additional file 3: CpGs in different cell lines tested by ENCODE project, Figure S2) and assured there was no significant differ- using Illumina HumanMethylation450 BeadChip v1.1 ence between the age groups (p > 0.05) using the technology (a newer version, the Infinium MethylationE- Mann-Whitney U test. This was done to assure the PIC 850K was not available at the time). We used validity and strengthen the rationale for the selection of traditional bisulfite-based protocols which cannot statistical methods. discriminate between 5-methylcytosine (5-mC) and For thedataofthe MGAT3 promoter methylation 5-hydroxymethylcytosine (5-hmC) as oxidative bisulfite from PBMCs, CD3 T cells isolated from blood, and (oxBS-450K) method can [53]. However, recent studies CD3 T cells isolated from inflamed colonic mucosa have shown that global DNA hydroxymethylation is very (the Porto cohort), the Mann-Whitney U test was low in blood cells [54, 55]. Furthermore, hydroxymethy- applied with Bonferroni correction accounting for 15 lation is significantly depleted from promotors and CpG CpG sites. islands, while enriched in the gene bodies [53, 56]. Based on the estimated statistical power, we did initial Glycan analysis screening on 60 patients for each pyrosequencing assay, Glycans present on IgG were analyzed from serum of after which we excluded those genes (pyrosequencing over 1000 IBD (UC and CD) patients and healthy con- assays) that did not show any statistically significant dif- trols in the Edinburgh cohort using ultra performance ferences between IBD patients and healthy controls. liquid chromatography (UPLC) [43, 50]. In the Florence Pyrosequencing assays for LAMB1, IL6ST, and IKZF1 cohort, plasma samples of 3500 IBD patients and healthy are shown in Additional file 2: Figure S1. We continued controls was used for analysis of IgG glycopeptides by to analyze promoter methylation only in the BACH2 and liquid chromatography coupled to mass spectrometry MGAT3 genes. Specific regions were amplified using (LC-MS) [46]. The data for IgG glycosylation analysis PyroMark PCR kit (Qiagen, Hilden, Germany). The cyc- were used in this work for correlation analysis with pro- ling conditions for the BACH2 gene were as follows: ini- moter methylation data of MGAT3 and BACH2 genes, tial polymerase activation step for 15 min at 95 °C with matching samples from the very same patients and followed by 50 cycles of 30 s denaturation at 95 °C, pri- healthy controls. mer annealing for 30 s at primer-specific temperatures (Additional file 1: Table S5), and 30 s at 72 °C, with final Isolation of IgG from blood plasma extension at 72 °C for 10 min. The cycling protocol used IgG has been isolated from blood plasma by affinity for amplification of the MGAT3 gene fragments was de- chromatography using CIM Protein G 96-well plate scribed previously [35], with the annealing temperature (BIA Separations, Ajdovščina, Slovenia) and vacuum adjusted to 55 °C for the fragment 1 performed on DNA manifold (Pall Corporation, Port Washington, NY, USA) from CD19 B cells. For quantitative measurement of as previously described [57, 58]. In short, plasma sam- DNA methylation level at specific CpG sites, ples (50–90 μl) were diluted with 1 × PBS, pH 7.4 in the PCR-amplified bisulfite-converted DNA was sequenced ratio 1:7. All samples were filtered through 0.45 and using the PyroMark Q24 Advanced pyrosequencing sys- 0.2-μm AcroPrep GHP filter plates (Pall Corporation) tem (Qiagen) according to the manufacturer’s recom- using vacuum manifold and immediately applied to pre- mendations. Sequences of PCR and pyrosequencing conditioned Protein G plate. After washing of the Klasić et al. Clinical Epigenetics (2018) 10:75 Page 5 of 14 −1 Protein G plate, IgG was eluted with 0.1 mol L formic individual patients were matched with their correspond- acid and immediately neutralized with ammonium bicar- ing glycan profiles. Sizes of datasets and patient classes bonate to pH 7.0. Protein G plate was regenerated and obtained after including complete records (i.e., both stored at 4 °C. methylation data and glycan profiles present) are shown in Additional file 1: Table S6. Individual glycan struc- IgG glycosylation analysis using ultra-performance liquid tures were represented as relative abundances and chromatography batch-corrected. Percentage of structures with bisecting N-glycans from isolated IgG in the Edinburgh cohort N-acetylglucosamine was calculated for each cohort as a were released with PNGase F after drying 300 μl of each derived trait at this point. Glycan structures identified by IgG elution fraction, labeled with 2-aminobenzamide each method were translated to Oxford notation, and and excess of regents removed by clean-up using hydro- only the 13 structures present in both the Edinburgh philic interaction liquid chromatography solid phase and Florence datasets were considered for correlation. extraction (HILIC-SPE). Fluorescently labeled and puri- We used IgG1 data from the Florence cohort, as this fied N-glycans were separated by HILIC-UPLC using isoform was the most abundant. Three additional de- Acquity UPLC instrument (Waters, Milford, MA, USA) rived traits were calculated: ratios of FA2B to FA2, as previously described [43]. Samples were separated FA2BG1 to FA2G1, and FA2BG2 to FA2G2. into 24 peaks [57], and the amount of N-glycans in each Pearson correlation between CpG methylation data chromatographic peak was expressed as a percentage of and 17 glycan features (13 structures and 4 derived total integrated area (% area). traits) was calculated. Significance threshold was set at p < 0.05 with additional Bonferroni correction for IgG glycosylation analysis using liquid chromatography 17-fold multiple testing. coupled to mass spectrometry Methylation of assayed CpG sites in promoters of the In the Florence cohort, Fc-specific IgG glycopeptides BACH2 and MGAT3 genes was correlated with mea- were analyzed after IgG purification, overnight trypsin sured glycan structures. Pearson correlation coefficient digestion at 37 °C, and reverse-phase purification on along with the associated p value was calculated between Chromabond C18 beads using vacuum manifold as average CpG methylation (for all genes/assays) and each described [46, 59]. Samples were analyzed using nanoli- measured IgG glycan structure. Calculation was done on quid chromatography coupled to mass spectrometry pairwise complete observations. Only correlations with (nanoLC-MS), on a nanoACQUITY UPLC system the p value below 0.01 were considered further. Next, (Waters, Milford Massachusetts, USA) coupled to correlation coefficients for all CpG assays were calcu- quadrupole-TOF-MS (Compact; Bruker Daltonics, lated, which was used to rank glycan structures accord- Bremen, Germany) equipped with a sheath-flow ESI ing to regulation by the assayed region. Glycan sprayer (capillary electrophoresis ESI-MS sprayer; structures with the strongest correlation (either positive Agilent Technologies, Santa Clara, USA) as previously or negative) to CpG methylation were then used to described [46]. The nanoACQUITY UPLC system and explain regulatory effects. All calculations and data visu- the Bruker Compact Q-TOF-MS were operated under alizations were done in R language and environment for HyStar software version 3.2. statistical computing (R Foundation for Statistical Com- Data was processed as described previously [46, 60]. puting, Vienna, Austria). Visualization of correlations This resulted in the extraction of 16 IgG1, 16 IgG2/3, was done using the R package “corrplot.” and 11 IgG4 glycoforms. The tryptic Fc-glycopeptides for IgG2 and IgG3 subclasses have identical peptide Results moieties in the Caucasian population and are therefore Promoter methylation of the candidate genes in whole not distinguishable with this methodology. Annotation blood and B cells of IBD patients of the spectra was done based on accurate mass accord- In order to assess the level of methylation in CpG ing to the relevant literature [40, 57]. islands of the five candidate genes (BACH2, MGAT3, IKZF1, LAMB1,and IL6ST), associated with both IBD Correlation analysis and IgG glycosylation by GWAS, we developed sev- Methylation data for the BACH2 (assay 2) and the eral pyrosequencing assays for each of the genes MGAT3 (assays 1 and 2) genes (obtained for the two (Fig. 1 and Additional file 2:FigureS1).Weper- large cohorts) were filtered according to the peak quality formed initial screening of the pyrosequencing assays by rejecting peaks marked as “failed” by the pyrose- on 60 patients. Overall cytosine methylation levels quencing software. Average methylation across all were very low for LAMB1 (average value per group < assayed CpG sites was calculated for each pyrosequenc- 8%), for IL6ST (< 3.5%), and for IKZF1 (< 4%) in the ing assay in each cohort. Methylation results for assayed portion of their promoters; therefore, we Klasić et al. Clinical Epigenetics (2018) 10:75 Page 6 of 14 Fig. 1 Positions of the BACH2 and MGAT3 genes in the human genome and relative positions of the fragments analyzed for methylation level (pyrosequencing assays) within these genes. For each pyrosequencing assay (A1–A5), the region amplified by PCR is shown. Positions of the genes on the chromosomes are shown using chromosome models (red vertical lines). Coordinates are relative to the hg19 human genome assembly. The genes are displayed in the direction corresponding to their reading frames. Annotations (CpG islands and pyrosequencing assays) are to scale. TSS transcription start site could not identify differential methylation. We then methylation level was high, with all CpG sites show- excluded these genes from further analysis. ing a reproducibly significant difference between HC MGAT3 and BACH2 promoter methylation was ana- andbothCDand UC.CpG sites2,13, and15 lyzed in several hundred IBD patients and healthy con- showed significant differences only for CD but not for trols from two independent cohorts (Additional file 1: UC. Direction of change was different for the two Tables S1 and S2). In these genes, we analyzed methyla- genes—differentially methylated CpG sites within the tion level at 47 CpGs covered by five pyrosequencing as- BACH2 promoter were hypomethylated, while those says in the BACH2 gene: 21 CpG sites were in the for the MGAT3 gene were hypermethylated in disease promoter region, 1 CpG site was in first exon, and 25 compared to healthy individuals. CpG sites were located in the first intron of the gene. A These results were confirmed on CD19 Bcells total of 32 CpG sites, covered by five pyrosequencing isolated from peripheral whole blood of the independ- assays, was analyzed for MGAT3: 18 CpG sites were ent, smaller patient sample from the Edinburgh located in the promoter region and 14 CpG sites in the cohort (67 samples). The CpG sites 1–5 and 12–13 in first intron (Fig. 1). Most of those CpG sites were the MGAT3 promoter were differentially methylated located within CpG islands of the both genes. We found between CD and HC. Only CpG site 5 within the differential CpG methylation between IBD patients and assay A2 of the BACH2 gene showed change in the HC within the assay A2, located at 213–368 bp up- methylation level between HC and CD in CD19 B stream (relative to the gene orientation) of the TSS in cells (Fig. 2b). There were no differences in the the BACH2 promoter and within the assays A1 and A2, methylation level of the same CpG sites within located in the CpG island 1 of the MGAT3 gene. The assayed fragments of the BACH2 and MGAT3 genes same pattern of differential methylation at these CpG between UC and HC. sites was observed in whole blood of patients and HC It is worth noting that the same pattern of CpG from two large independent cohorts (Fig. 2). CpG methylation differences was observed in PBMCs of the methylation level was generally low (up to 20%) in the IBD patients and HC from both large independent assayed portion of the BACH2 promoter; however, sig- cohorts, and most of the CpG sites within the assayed nificant differences between HC and CD methylation portion of the MGAT3 promoter were also differentially level were recorded at CpG sites 4, 5, 6, and 8 (Fig. 2a). methylated in CD19 B cells from the subset of IBD For the assayed portion of the MGAT3 promoter, general patients from the Edinburgh cohort (Fig. 2a, b). Klasić et al. Clinical Epigenetics (2018) 10:75 Page 7 of 14 Fig. 2 (See legend on next page.) Klasić et al. Clinical Epigenetics (2018) 10:75 Page 8 of 14 (See figure on previous page.) Fig. 2 Box plot of CpG methylation in peripheral whole blood for the BACH2 and MGAT3 genes in the Edinburgh and Florence cohorts and in B cells from a subset of patients from Edinburgh cohort. Groups were compared using the Mann-Whitney U test with significance threshold of p = 0.05, corrected for multiple testing using the Bonferroni method. a Methylation levels were generally low in the assayed portion of the BACH2 gene promoter, with significant differences between HC and CD methylation at CpG sites 4, 5, 6, and 8 (replicated in both cohorts). For the MGAT3 gene, general methylation level was high, with all CpG sites showing a reproducibly significant difference between HC and both CD and UC, except for CpG sites 2, 13, and 15 for which reproducible significant differences were found only between HC and CD. b In B cells, isolated from PBMCs of a subset of the patients from the Edinburg cohort, differential methylation was found at the CpG position 5 of the BACH2 gene (assay 2) between HC and CD, while for the MGAT3 gene, differentially methylated were CpG sites 1–5, 12, and 13 between HC and CD. CD Crohn’s disease, UC ulcerative colitis, HC healthy controls Promoter methylation of the MGAT3 gene in CD3 T cells inflamed colonic mucosa in comparison with healthy from PBMCs and inflamed colonic mucosa of UC patients mucosa. We included in our investigation biopsy samples of UC patients from an independent cohort from the Gastro- Correlation between the MGAT3 and BACH2 promoter enterology Department of Centro Hospitalar do methylation and IgG glycosylation Porto-Hospital de Santo António, Portugal. Given the There were statistically significant correlations that repli- technical challenges in obtaining DNA and RNA from a cated across assays and cohorts between the MGAT3 small number of purified cells from inflamed colonic promoter methylation and glycan structures FA2, mucosa, a subset of patients with active and inactive FA2G2, FA2BG2, and FA2G2S1, as well as the derived phase of UC was selected for methylation analysis from trait of the ratio of FA2B to FA2 (Fig. 4a). All correla- three sources: (1) PBMCs, (2) CD3 T cells isolated from tions except with FA2 were negative. No reproducible PBMCs, and (3) CD3 T cells isolated from colonic significant correlations could be found between BACH2 mucosa (see also Additional file 1: Table S4). promoter methylation and the glycan structures Inter-individual variation of MGAT3 methylation level (Fig. 4a). measured from PBMCs and from CD3 T cells isolated In order to infer a mechanistic pathway of the from PBMCs was quite large—it varied from 47 to 94% observed correlations, we mapped them to the glycan and from 26 to 90%, respectively. Therefore, we could biosynthesis pathways (Fig. 4b). The ratio of bisecting not find any difference in CpG methylation level glycans to FA2 was taken as an indicator of MGAT3 between UC patients and HC in assayed fragments of (GnT-III) activity. This interpretation allowed us to infer the MGAT3 promoter, neither in PBMCs nor in CD3 T lower GnT-III enzymatic activity when the promoter of cells isolated from PBMCs. However, we recorded a total the MGAT3 gene was methylated. Increase in MGAT3 of 7 (out of 15) differentially methylated CpG sites in promoter methylation correlated with a decrease in cer- CD3 T cells isolated from colonic mucosa of UC tain galactosylated and sialylated structures (Fig. 4b). In patients with active disease compared with HC (Fig. 3). addition to the decreased levels of bisecting GlcNAc on Overall, the methylation level of CpGs within assayed non-galactosylated glycans (B/FA2), the most significant fragments of the MGAT3 promoter was high in CD3 T effect of the MGAT3 promoter methylation on IgG gly- cells from healthy colonic mucosa (between 77 and come composition was a decrease of IgG galactosylation. 98%). When compared to inflamed mucosa of UC patients with active phase of the disease, the same CpG Discussion sites were hypomethylated, with the highest difference at Results from this study strongly indicate that the −5 the CpG position 10 (13.24%; p = 5.08 × 10 ; Fig. 3). In MGAT3 and BACH2 genes play an important role in inactive UC, no significant differences could be found IBD pathogenesis and suggest a possible disease pathway after Bonferroni correction for multiple testing. mediated by the pro-inflammatory properties of IgG It is worth noting that the methylation pattern in antibodies acquired by alterations in Fc glycosylation. CD3 T cells isolated from inflamed colonic mucosa dif- Our recent study, performed on a large cohort of over fered from the methylation patterns in PBMCs and for 1000 IBD patients, reported a significant difference in CD3 T cells isolated from PBMCs. The latter two were IgG glycome composition in both UC and CD compared very similar and had much lower methylation levels than to healthy controls [43, 46]. We found a decrease in that measured for CD3 T cells from inflamed colonic quantity of galacosylated glycans in both CD and UC, as mucosa (Fig. 3). Also, MGAT3 methylation level was well as a decrease in sialylated glycans and an increase increased in UC compared with HC when measured of bisecting GlcNAc on digalactosylated glycan struc- from PBMCs or CD3 T cells isolated from PBMCs tures on IgG in CD. Indeed, alternative N-glycosylation (hypermethylation), while it was decreased (hypomethy- of an IgG molecule influences its function—pro-inflam- lation) when measured from CD3 T cells isolated from matory and anti-inflammatory activity depends on the Klasić et al. Clinical Epigenetics (2018) 10:75 Page 9 of 14 Fig. 3 Box plot of CpG methylation level in the MGAT3 gene promoter (assays A1 and A2) analyzed from PBMCs (a), CD3 T cells isolated from PBMCs (b), and CD3 T cells isolated from inflamed colonic mucosa (c) from the independent cohort of Porto. Changes between UC patients with active disease and HC were statistically significant only in CD3 T cells isolated from inflamed colonic mucosa at CpG positions 3 and 7–12 (p < 0.05 after Bonferroni correction for 15 hypotheses). PBMC peripheral blood mononuclear cells, UC ulcerative colitis, HC healthy controls glycans added on the Cy2 domain of its Fc region [29]. These glycans are of a biantennary complex type with or without bisecting GlcNAc, core fucose, galactose, and sialic acid residues [61]. Recently, this was confirmed in a large multi-centric study of IgG glycome in IBD [46]. Therefore, glycan changes observed on IgG in peripheral blood of UC and CD patients are obviously associated with increased inflammatory potential of IgG, suggesting functional relevance of IgG glycosylation for IBD. Here, we propose a possible mechanism underlying the aberrant IgG glycosylation pattern observed in IBD [43, 46]. Out of five candidate genes analyzed in this work, the MGAT3, a glycosyltransferase which partici- pates in synthesis of IgG glycans, and the BACH2,a transcription factor and a master regulator of a network of genes relevant for B cell integrity [62, 63], showed dif- ferential methylation in peripheral blood of both CD and UC patients when compared to healthy individuals. Even though we identified changes in methylation level for both UC and CD compared to HC, the differences were more pronounced for CD. This is concordant with other studies that explored either whole genome methy- lation or promoter methylation of candidate genes in IBD [11]. The extent of the change in IgG glycome com- position was also consistently higher in CD than UC compared to HC [43, 46]. The protein encoded by the MGAT3 gene (N-acetyl- glucosaminyltransferase III, GnT-III) is responsible for significant functional alteration of glycans on the Fc region of an IgG antibody. The GnT-III adds N-acetyl- glucosamine (GlcNAc) on β1,4-linked mannose in the three-mannose core of N-glycans, producing bisecting GlcNAc structures. In the same CD patients, who showed changed MGAT3 promoter methylation level in peripheral blood cells, a significant increase in the per- centage of bisecting GlcNAc on glycans of circulating IgG antibodies was recorded, too. The association of the MGAT3 with both IgG N-glycosylation [49] and Crohn’s disease [4, 5] suggests that N-glycans with bisecting GlcNAc could be involved in CD pathogenesis through functional effect on IgG antibody. Correlations between BACH2 and MGAT3 promoter methylation and glycan structures have given further insight into the changes of IgG glycosylation pattern me- diated by those two genes (Fig. 4b). The MGAT3 Klasić et al. Clinical Epigenetics (2018) 10:75 Page 10 of 14 Fig. 4 (See legend on next page.) Klasić et al. Clinical Epigenetics (2018) 10:75 Page 11 of 14 (See figure on previous page.) Fig. 4 Correlations between CpG methylation in the BACH2 and MGAT3 gene promoters and glycan structures measured from the same individuals of the Edinburgh and Florence cohorts, mapped to the glycan biosynthesis pathways. a Correlation coefficients between average CpG methylation in the assayed gene promoter fragments and glycan structure percentages are shown as blue (positive) or red (negative correlation) circles with their size and shade proportional to the correlation coefficient. Correlations without statistical significance (p > 0.05 after Bonferroni correction for multiple testing) are crossed. Columns represent 13 individual glycan structures and four derived traits (beige box). EDI Edinburgh cohort, FLO Florence cohort, Bisecting, percentage of all structures with bisecting N-acetylglucosamine, B/FA2 ratio of FA2B to FA2 structures, B/FA2G1 ratio of FA2BG1 to FA2G1 structures, B/ FA2G2 ratio of FA2BG2 to FA2G2 structures. b Glycan biosynthesis pathways with the glycan structures, labels, and the enzymes mapped to correlation results for the MGAT3 gene. Light blue rectangles indicate positive, while light red rectangles indicate negative correlation between the glycan structures or traits and CpG methylation levels. Only correlations replicated across assays and/or cohorts are shown. The red rectangle around the MGAT3 enzyme reflects the negative correlation between CpG methylation and the derived trait B/FA2, which effectively measures enzyme activity at this step. MGAT3 N-acetilglucosaminyltransferase III (GnT-III), FUT8 fucosyltransferase 8, GalT1 galactosyltranserase 1, ST6GalT1 Beta-galactoside alpha-2,6-sialyltransferase 1 promoter methylation probably led to decreased GnT-III (Additional file 1: Tables S7 and S8). Our present efforts enzymatic activity, as revealed by negative correlation are focused on functional studies with hope to reveal a between methylation and total bisecting glycans to FA2 more complete view of the BACH2 role in IgG ratio. Namely, GnT-III adds a bisecting GlcNAc to FA2. glycosylation. A further proof is the positive correlation between the Since DNA methylation pattern is tissue-specific, our MGAT3 promoter methylation and FA2, since it is not goal was to ascertain if CpG methylation from blood surprising that substrate accumulates when enzyme ac- could be a proxy for CpG methylation of the same tivity is decreased. More complex effects were observed candidate gene in the tissue where the inflammation is on galactosylation and sialylation. The negative correl- taking place. In fact, IBD is an immune-mediated dis- ation between MGAT3 methylation and galactosylation order in which T cells are actively implicated in develop- of both, glycans with (FA2BG2) and without bisecting ment of gut-mucosa inflammation [68]. Previous GlcNAc (FA2G2), suggests that the effect of increased evidence has suggested that N-glycosylation of intestinal galactosylation is not caused only by steric effects of T cells is associated with UC pathogenesis and disease bisecting GlcNAc, but also through some indirect effects severity [50, 69]. Therefore, we analyzed MGAT3 pro- of MGAT3 expression on galactosyltransferase activity. It moter methylation in CD3 T cells isolated from PBMCs seems as though galactosylation and sialylation are and from intestinal mucosa of UC patients with active co-regulated with the addition of a bisecting GlcNAc and inactive form of the disease and compared with catalyzed by GnT-III. Furthermore, Dekkers and MGAT3 promoter methylation from PBMCs of the same co-workers recently reported that transfection of cells patients from the Porto cohort. We found 7 out of 15 with MGAT3 causes an increase of IgG galactosyla- differentially methylated CpG sites in CD3 T cells tion [64]. isolated from colonic mucosa of UC patients with active Much weaker correlation was observed between BACH2 form of the disease compared to CD3 T cells from mu- methylation and IgG glycosylation. This was expected since cosa of healthy individuals. On the other hand, there BACH2 is not a glycosyltransferase and thus is not directly were no differences in MGAT3 promoter methylation involved in glycan biosynthetic pathways. However, weak between patients with either active or inactive form of positive correlations with A2G2, FA2G2, and FA2G2S1 UC and healthy controls neither in PBMCs nor in CD3 structures, which involve galactosylation and sialylation, T cells isolated from PBMCs. This could be due to high were observed, as well as weak negative correlation with dispersion in the methylation level when measured from fucosylated bianntenary structure FA2. This is interesting PBMCs (47–94%) and CD3 T cells isolated from because GWA studies associated BACH2 with IgG galacto- PBMCs (26–90%), probably due to small sample size sylation [49] as well as with various immune and inflam- and dispersion in age. Namely, cell composition changes matory diseases including IBD [4, 5, 65–67]inwhich IgG across age in whole blood, and it can explain dispersion acquires pro-inflammatory properties through decrease in of CpG methylation level observed in our sample [70]. galactosylation, sialylation, and fucosylation [43, 46]. Since Considering much smaller dispersion in values of methy- BACH2 is orchestrating a gene regulatory network in B lation level (65–97%) measured from CD3 T cells from cells [62], we believe that some glyco-genes are also regu- lamina propria, the differences could be used as a signa- lated by this transcription factor. Indeed, our in silico ture for inflammation. analysis identified several glyco-genes, mostly galactosyl- transferases (including B4GALT1 and B4GALT2), to pos- Conclusions sess putative AP-1 and NFE2 binding sites for BACH2 Taken together, our results suggest that the aberrant transcription factor [63], suggesting that these galactosyl- methylation observed in the MGAT3 gene in CD3 T transferases could be controlled by BACH2 cells from intestinal mucosa of UC patients, B cells from Klasić et al. Clinical Epigenetics (2018) 10:75 Page 12 of 14 peripheral blood, and the whole peripheral blood in UC Acknowledgements The authors would like to thank Stephanie Scott for her organizational and and CD patients is a possible mechanism underlying in- administrational contribution. The study has been funded by the EU FP7 flammation due to a change in the immune system—ei- grant European Commission IBD-BIOM (contract # 305479), EU FP7 Regional ther through the change of glycans on Fc region of IgGs Potential Grant INTEGRA-Life (contract # 315997), European Structural and Investment Funds grant for the Croatian National Centre of Research or by modulating the glycosylation profile of glycopro- Excellence in Personalized Healthcare (contract # KK.01.1.1.01.0010), and teins on intestinal T cells. Others [24] have shown that Croatian Science Foundation grant EpiGlycoIgG (contract # 3361). Financial some of their candidate genes changed promoter methy- support from Portugal (PI: SSP): FEDER—Fundo Europeu de Desenvolvimento Regional funds through the COMPETE 2020—Operacional Programme for lation level in whole biopsies, while some of the genes Competitiveness and Internationalisation (POCI), Portugal 2020, and by showed changes only in some cell types of the heteroge- Portuguese funds through FCT—Fundação para a Ciência e a Tecnologia/ neous cell population from the epithelial and Ministério da Ciência, Tecnologia e Inovação in the framework of the project (POCI-01/0145-FEDER-016601; PTDC/DTP-PIC/0560/2014) was received. SSP non-epithelial cells, pointing out the importance of cell also acknowledges the European Crohn’s and Colitis Organization (ECCO) separation from mucosal biopsies. Interestingly, one of and the “Broad Medical Research program at Crohn’s and Colitis Foundation the genes that showed differential methylation in the of America-CCFA” for funding. SSP acknowledges the Portuguese Group of Study on IBD (GEDII) for funding. A.M.D. [PD/BD/105982/2014] also acknowledges non-epithelial fraction, representing immune and stro- FCT for funding. mal cells, was FUT7, the fucosyltransferase involved in IBD-BIOM consortium: Daniel Kolarich (Department of Biomolecular Systems, sialyl Lewis X synthesis, a ligand in selectin-mediated Max Planck Institute of Colloids and Interfaces, Potsdam, Germany), Manfred Wuhrer (Center for Proteomics and Metabolomics, Leiden University Medical adhesion of leukocytes to activated endothelium. Fur- Center, Leiden, The Netherlands; Division of BioAnalytical Chemistry, VU thermore, Dias and collaborators proposed a molecular University Amsterdam, Amsterdam, the Netherlands), Dermot P. B. McGovern mechanism in IBD involving another glyco-gene, the (F. Widjaja Family Foundation Inflammatory Bowel and Immunobiology Research Institute, Cedars-Sinai Medical Center, Los Angeles), Iain K. Pemberton MGAT5 (GnT-V), responsible for branching of N-gly- (IP Research Consulting SAS, Paris, France), Daniel IR Spencer (Ludger Ltd., cans. They showed decreased expression of branched Culham Science Centre, Oxford, UK, Daryl L. Fernandes (Ludger Ltd., Culham N-glycans on T cell receptor (TCR) of lamina propria Science Centre, Oxford, UK), Rahul Kalla, Kate O’Leary, Alex T Adams, Hazel Drummond, Elaine Nimmo, Ray Boyapati, David C Wilson (Centre for Genetics associated with disease severity in patients with active and Molecular Medicine, University of Edinburgh, Edinburgh, UK), Ray Doran UC [50]. Dysregulation of N-glycan branching on TCR (Ludger Ltd., Culham Science Centre, Oxford, UK), Igor Rudan (all, Centre for contributes to a decreased threshold of T cell activation Population Health Sciences, University of Edinburgh, Edinburgh, UK), Paolo Lionetti (Paediatric Gastroenterology Unit, AOU Meyer, Viale Pieraccini, Florence, leading to a hyper-immune response which is a feature Italy), Natalia Manetti (Department of Medical and Surgical Sciences, Division of of UC patients. Taken together, our results and those of Gastroenterology, University Hospital Careggi, Florence, Italy), Fabrizio Bossa others suggest an important role of aberrant protein (Department of Medical Sciences, Division of Gastroenterology, IRCCS-CSS Hospital, Viale Cappuccini, Rotondo, Italy), Paola Cantoro, Anna Kohn (Division glycosylation (partly through epigenetic mechanisms) in of Gastroenterology, S. Camillo Hospital, Rome, Italy), Giancarlo Sturniolo IBD through dysregulation of the immune system. Also, (Gastrointestinal Unit, University of Padua, Padua, Italy), Silvio Danese (IBD Unit, in IBD diagnosis and treatment, it is important to find a Humanitas Research Institute, Rozzano, Milan, Italy), Mariek Pierik (Maastricht University Medical Centre (MUMC), Maastricht, the Netherlands), and David C. non-invasive, specific, and clinically useful biomarkers in Wilson (Centre for Genetics and Molecular Medicine, University of Edinburgh, order to identify high-risk patients. Using MGAT3 Edinburgh, UK). hypermethylation together with the glycan traits as markers from peripheral blood of IBD patients seems Funding promising in the disease identification. This independent research was generously supported by the following grants: EU FP7 research grant IBD-BIOM (contract # 305479) to JS, VA, GL, and VZ; EU FP7 Regional Potential Grant INTEGRA-Life (contract # 315997) to GL and VZ; Additional files European Structural and Investment Funds grant for the Croatian National Centre of Research Excellence in Personalized Healthcare (contract # KK.01.1.1.01.0010) to Additional file 1: Supplementary Tables 1-8. Demographics of IBD GL and VZ; Croatian Science Foundation grant EpiGlycoIgG (contract # 3361) to patients and healthy controls (1-4), PCR primers (5), number of samples VZ; FEDER COMPETE 2020 POCI, Portugal 2020, and Portuguese funds through per analysis (6) and in silico analysis of transcription factor binding sites FCT (contracts # POCI-01/0145-FEDER-016601 and PTDC/DTP-PIC/0560/2014) to in gene promoters (7, 8). (DOCX 70 kb) SP; and FTC (contract # PD/BD/105982/2014) to AMD. Additional file 2: Figure S1. Position of the pyrosequencing assays for the genes LAMB1, IL6ST, and IKZF1 in the genome relative to CpG islands, annotated promoters, and exons. (PDF 523 kb) Availability of data and materials All original data can be obtained from the authors upon request. Additional file 3: Figure S2. Violin plots showing the age distribution in IBD patients (CD, UC) and healthy controls (HC). The groups were well matched by age, which was shown by Mann-Whitney U test: no significant Authors’ contributions differences between groups were found at the level p = 0.05. (PDF 704 kb) Study design was conceived by VZ, AV, GL, and SP. Sample provision was provided by JS, NTV, NAK, ERN, VA, RD’I, and SP. Blood and biopsy processing Abbreviations was conducted by AMD and AL. DNA methylation analyses were carried out by CD: Crohn’s disease; EWAS: Epigenome-wide association study; GlcNAc: N- MK, DM, PD, IS, and IB. Glycan analysis was carried out by IT, JŠ,MŠ,and GR. acetylglucosamine; GWAS: Genome-wide association study; HC: Healthy Statistical and correlation analyses were carried out by AV, MK, and DM. control; IBD: Inflammatory bowel disease; IgG: Immunoglobulin G; LC- Drafting of the manuscript was carried out by VZ, AV, MK, DM, and IT-A. All MS: Liquid chromatography coupled to mass spectrometry; authors were involved in critical review, editing, revision, and approval of the PBMCs: Peripheral blood mononuclear cells; UC: Ulcerative colitis final manuscript. Klasić et al. Clinical Epigenetics (2018) 10:75 Page 13 of 14 Ethics approval and consent to participate 12. Nimmo ER, Prendergast JG, Aldhous MC, Kennedy NA, Henderson P, For the Edinburgh and Florence cohorts, ethical approvals were obtained Drummond HE, Ramsahoye BH, Wilson DC, Semple CA, Satsangi J. Genome- from Tayside Committee on Medical Ethics B, and all patients and controls wide methylation profiling in Crohnʼs disease identifies altered epigenetic provided written, informed consent (LREC 06/S1101/16, LREC 2000/4/192). regulation of key host defense mechanisms including the Th17 pathway. For the patients from Portugal, all clinical procedures and research protocols Inflamm Bowel Dis. 2012;18:889–99. were approved by the ethics committee of CHP/HSA, Portugal (233/12(179- 13. Ventham NT, Kennedy NA, Adams AT, Kalla R, Heath S, O'Leary KR, DEFI/177-CES); all participants gave their informed consent. Drummond H, consortium IB, consortium IC, Wilson DC, et al: Integrative epigenome-wide analysis demonstrates that DNA methylation may mediate Competing interests genetic risk in inflammatory bowel disease. Nat Commun 2016, 7:13507. The authors declare that they have no competing interests. 14. Low D. DNA methylation in inflammatory bowel disease and beyond. World J Gastroenterol. 2013;19:5238. 15. Ventham NT, Kennedy NA, Nimmo ER, Satsangi J. Beyond gene discovery in Publisher’sNote inflammatory bowel disease: the emerging role of epigenetics. Springer Nature remains neutral with regard to jurisdictional claims in Gastroenterology. 2013;145:293–308. published maps and institutional affiliations. 16. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet. 2012;13:484–92. Author details 17. Ambatipudi S, Cuenin C, Hernandez-Vargas H, Ghantous A, Le Calvez-Kelm Department of Biology, Division of Molecular Biology, Faculty of Science, F, Kaaks R, Barrdahl M, Boeing H, Aleksandrova K, Trichopoulou A, et al. University of Zagreb, Horvatovac 102a, 10000 Zagreb, Croatia. Tobacco smoking-associated genome-wide DNA methylation changes in Gastrointestinal Unit, Centre for Genomics and Molecular Medicine, the EPIC study. Epigenomics. 2016;8:599–618. University of Edinburgh, Edinburgh EH4 6XU, UK. Genos Glycoscience 18. Barrès R, Yan J, Egan B, Treebak Jonas T, Rasmussen M, Fritz T, Caidahl K, Krook Research Laboratory, Borongajska cesta 83h, 10000 Zagreb, Croatia. Faculty A, O’Gorman Donal J, Zierath Juleen R. Acute exercise remodels promoter of Pharmacy and Biochemistry, University of Zagreb, Zagreb, Croatia. IBD methylation in human skeletal muscle. Cell Metab. 2012;15:405–11. Pharmacogenetics, University of Exeter, Exeter, UK. Institute of Molecular 19. Olszak T, An D, Zeissig S, Vera MP, Richter J, Franke A, Glickman JN, Siebert R, Pathology and Immunology of the University of Porto (IPATIMUP), Porto, Baron RM, Kasper DL, Blumberg RS. Microbial exposure during early life has Portugal. Department of Medical and Surgical Sciences, Division of persistent effects on natural killer T cell function. Science. 2012;336:489–93. Gastroenterology, University Hospital Careggi, Florence, Italy. Department of 20. Symonds ME, Sebert SP, Budge H. The impact of diet during early life Medical Sciences, Division of Gastroenterology, IRCCS-CSS Hospital, Viale and its contribution to later disease: critical checkpoints in Cappuccini, Rotondo, Italy. Gastrointestinal Unit, University of Padua, Padua, development and their long-term consequences for metabolic health. Italy. Translational Gastroenterology Unit, Nuffield Department of Medicine, Proc Nutr Soc. 2009;68:416. University of Oxford, Oxford, UK. 21. Vaissiere T, Hung RJ, Zaridze D, Moukeria A, Cuenin C, Fasolo V, Ferro G, Paliwal A, Hainaut P, Brennan P, et al. Quantitative analysis of DNA Received: 28 February 2018 Accepted: 22 May 2018 methylation profiles in lung cancer identifies aberrant DNA methylation of specific genes and its association with gender and cancer risk factors. Cancer Res. 2009;69:243–52. References 22. Milagro FI, Mansego ML, De Miguel C, Martinez JA. Dietary factors, 1. Burisch J, Pedersen N, Čuković-Čavka S, Brinar M, Kaimakliotis I, Duricova D, epigenetic modifications and obesity outcomes: progresses and Shonová O, Vind I, Avnstrøm S, Thorsgaard N, et al. East–West gradient in perspectives. Mol Asp Med. 2013;34:782–812. the incidence of inflammatory bowel disease in Europe: the ECCO-EpiCom 23. Adams AT, Kennedy NA, Hansen R, Ventham NT, OʼLeary KR, Drummond HE, inception cohort. Gut. 2013;63:588–97. Noble CL, El-Omar E, Russell RK, Wilson DC, et al. Two-stage genome-wide 2. Xavier RJ, Podolsky DK. Unravelling the pathogenesis of inflammatory bowel methylation profiling in childhood-onset Crohnʼs disease implicates disease. Nature. 2007;448:427–34. epigenetic alterations at the VMP1/MIR21 and HLA loci. Inflamm Bowel Dis. 3. Anderson CA, Boucher G, Lees CW, Franke A, D'Amato M, Taylor KD, Lee JC, 2014;20:1784–93. Goyette P, Imielinski M, Latiano A, et al. Meta-analysis identifies 29 24. Cooke J, Zhang H, Greger L, Silva A-L, Massey D, Dawson C, Metz A, Ibrahim additional ulcerative colitis risk loci, increasing the number of confirmed A, Parkes M. Mucosal genome-wide methylation changes in inflammatory associations to 47. Nat Genet. 2011;43:246–52. bowel disease. Inflamm Bowel Dis. 2012;18:2128–37. 4. Franke A, Balschun T, Sina C, Ellinghaus D, Häsler R, Mayr G, Albrecht M, 25. Harris RA, Nagy-Szakal D, Mir SAV, Frank E, Szigeti, Kaplan JL, Bronsky J, Wittig M, Buchert E, Nikolaus S, et al. Genome-wide association study for Opekun A, Ferry GD, Winter H, Kellermayer R. DNA methylation-associated ulcerative colitis identifies risk loci at 7q22 and 22q13 (IL17REL). Nat Genet. colonic mucosal immune and defense responses in treatment-naïve 2010;42:292–4. pediatric ulcerative colitis. Epigenetics. 2014;9:1131–7. 5. Franke A, McGovern DP, Barrett JC, Wang K, Radford-Smith GL, Ahmad T, 26. Karatzas PS, Gazouli M, Safioleas M, Mantzaris GJ. DNA methylation changes Lees CW, Balschun T, Lee J, Roberts R, et al. Genome-wide meta-analysis in inflammatory bowel disease. Ann Gastroenterol. 2014;27:125–32. increases to 71 the number of confirmed Crohn’s disease susceptibility loci. 27. Hart GW, Copeland RJ. Glycomics hits the big time. Cell. 2010;143:672–6. Nat Genet. 2010;42:1118–25. 28. Böhm S, Kao D, Nimmerjahn F. Sweet and sour: the role of glycosylation for 6. Jostins L, Ripke S, Weersma RK, Duerr RH, McGovern DP, Hui KY, Lee JC, the anti-inflammatory activity of immunoglobulin G. In: Fc Receptors. Basel: Schumm LP, Sharma Y, Anderson CA, et al. Host-microbe interactions have Springer International Publishing; 2014. p. 393–417. 393–417. shaped the genetic architecture of inflammatory bowel disease. Nature. 29. Kaneko Y. Anti-inflammatory activity of immunoglobulin G resulting from Fc 2012;491:119–24. sialylation. Science. 2006;313:670–3. 7. Liu JZ, van Sommeren S, Huang H, Ng SC, Alberts R, Takahashi A, Ripke S, 30. National Research Council (U.S.). Committee on Assessing the Importance Lee JC, Jostins L, Shah T, et al. Association analyses identify 38 susceptibility and Impact of Glycomics and Glycosciences., National Research Council (U.S. loci for inflammatory bowel disease and highlight shared genetic risk across ). Board on Chemical Sciences and Technology., National Research Council populations. Nat Genet. 2015;47:979–86. (U.S.). Board on Life Sciences.: Transforming glycoscience: a roadmap for the 8. Bonaz BL, Bernstein CN. Brain-gut interactions in inflammatory bowel future. Washington, D.C.: National Academies Press; 2012. disease. Gastroenterology. 2013;144:36–49. 31. Stowell SR, Ju T, Cummings RD. Protein glycosylation in cancer. Annu Rev 9. Pituch-Zdanowska A, Banaszkiewicz A, Albrecht P. The role of dietary fibre in Pathol: Mech Dis. 2015;10:473–510. inflammatory bowel disease. Gastroenterol Rev. 2015;3:135–41. 10. Rubin DT, Hanauer SB. Smoking and inflammatory bowel disease. Eur J 32. Horvat T, Deželjin M, Redžić I, Barišić D, Herak Bosnar M, Lauc G, Zoldoš V. Gastroenterol Hepatol. 2000;12:855–62. Reversibility of membrane N-Glycome of HeLa cells upon treatment with 11. McDermott E, Ryan EJ, Tosetto M, Gibson D, Burrage J, Keegan D, Byrne K, epigenetic inhibitors. PLoS One. 2013;8:e54672. Crowe E, Sexton G, Malone K, et al. DNA methylation profiling in 33. Horvat T, Mužinić A, Barišić D, Bosnar MH, Zoldoš V. Epigenetic modulation inflammatory bowel disease provides new insights into disease of the HeLa cell membrane N-glycome. Biochim Biophys Acta Gen Subj. pathogenesis. J Crohn’s Colitis. 2015;10:77–86. 2012;1820:1412–9. Klasić et al. Clinical Epigenetics (2018) 10:75 Page 14 of 14 34. Horvat T, Zoldoš V, Lauc G. Evolutional and clinical implications of the 53. Stewart SK, Morris TJ, Guilhamon P, Bulstrode H, Bachman M, epigenetic regulation of protein glycosylation. Clin Epigenetics. 2011;2: Balasubramanian S, Beck S. oxBS-450K: a method for analysing 425–32. hydroxymethylation using 450K BeadChips. Methods. 2015;72:9–15. 35. Klasić M, Krištić J, Korać P, Horvat T, Markulin D, Vojta A, Reiding KR, Wuhrer 54. Kroeze LI, Aslanyan MG, van Rooij A, Koorenhof-Scheele TN, Massop M, M, Lauc G, Zoldoš V. DNA hypomethylation upregulates expression of the Carell T, Boezeman JB, Marie JP, Halkes CJ, de Witte T, et al. Characterization MGAT3 gene in HepG2 cells and leads to changes in N-glycosylation of of acute myeloid leukemia based on levels of global hydroxymethylation. secreted glycoproteins. Sci Rep. 2016;6:24363. Blood. 2014;124:1110–8. 55. Sanchez-Guerra M, Zheng Y, Osorio-Yanez C, Zhong J, Chervona Y, Wang S, 36. Saldova R, Dempsey E, Pérez-Garay M, Mariño K, Watson JA, Blanco- Chang D, McCracken JP, Diaz A, Bertazzi PA, et al. Effects of particulate Fernández A, Struwe WB, Harvey DJ, Madden SF, Peracaula R, et al. 5-AZA- matter exposure on blood 5-hydroxymethylation: results from the Beijing 2′-deoxycytidine induced demethylation influences N-glycosylation of truck driver air pollution study. Epigenetics. 2015;10:633–42. secreted glycoproteins in ovarian cancer. Epigenetics. 2011;6:1362–72. 56. Yu M, Hon GC, Szulwach KE, Song CX, Zhang L, Kim A, Li X, Dai Q, Shen Y, 37. Vojta A, Samaržija I, Bočkor L, Zoldoš V. Glyco-genes change expression in Park B, et al. Base-resolution analysis of 5-hydroxymethylcytosine in the cancer through aberrant methylation. Biochim Biophys Acta Gen Subj. 2016; mammalian genome. Cell. 2012;149:1368–80. 1860:1776–85. 57. Pucic M, Knezevic A, Vidic J, Adamczyk B, Novokmet M, Polasek O, Gornik O, 38. Zoldoš V, Horvat T, Novokmet M, Cuenin C, Mužinić A, Pučić M, Huffman JE, Supraha-Goreta S, Wormald MR, Redzic I, et al. High throughput isolation Gornik O, Polašek O, Campbell H, et al. Epigenetic silencing of HNF1A and glycosylation analysis of IgG-variability and heritability of the IgG associates with changes in the composition of the human plasma N- glycome in three isolated human populations. Mol Cell Proteomics. 2011; glycome. Epigenetics. 2012;7:164–72. M111(010090):10. 39. Theodoratou E, Campbell H, Ventham NT, Kolarich D, Pučić-Baković M, 58. Trbojevic-Akmacic I, Ugrina I, Lauc G. Comparative analysis and validation of Zoldoš V, Fernandes D, Pemberton IK, Rudan I, Kennedy NA, et al. The role different steps in glycomics studies. Methods Enzymol. 2017;586:37–55. of glycosylation in IBD. Nat Rev Gastroenterol Hepatol. 2014;11(10):588–600. 59. Selman MH, Derks RJ, Bondt A, Palmblad M, Schoenmaker B, Koeleman CA, 40. Arnold JN, Wormald MR, Sim RB, Rudd PM, Dwek RA. The impact of van de Geijn FE, Dolhain RJ, Deelder AM, Wuhrer M. Fc specific IgG glycosylation on the biological function and structure of human glycosylation profiling by robust nano-reverse phase HPLC-MS using a immunoglobulins. Annu Rev Immunol. 2007;25:21–50. sheath-flow ESI sprayer interface. J Proteome. 2012;75:1318–29. 41. Miyahara K, Nouso K, Saito S, Hiraoka S, Harada K, Takahashi S, Morimoto Y, 60. Jansen BC, Falck D, de Haan N, Hipgrave Ederveen AL, Razdorov G, Lauc G, Kobayashi S, Ikeda F, Miyake Y, et al. Serum glycan markers for evaluation of Wuhrer M. LaCyTools: a targeted liquid chromatography-mass spectrometry disease activity and prediction of clinical course in patients with ulcerative data processing package for relative quantitation of glycopeptides. J colitis. PLoS One. 2013;8:e74861. Proteome Res. 2016;15:2198–210. 42. Shinzaki S, Iijima H, Nakagawa T, Egawa S, Nakajima S, Ishii S, Irie T, Kakiuchi 61. Fujii S, Nishiura T, Nishikawa A, Miura R, Taniguchi N. Structural Y, Nishida T, Yasumaru M, et al. IgG oligosaccharide alterations are a novel heterogeneity of sugar chains in immunoglobulin G. Conformation of diagnostic marker for disease activity and the clinical course of immunoglobulin G molecule and substrate specificities of inflammatory bowel disease. Am J Gastroenterol. 2008;103:1173–81. glycosyltransferases. J Biol Chem. 1990;265:6009–18. 43. Trbojevic Akmacic I, Ventham NT, Theodoratou E, Vuckovic F, Kennedy NA, 62. Igarashi K, Ochiai K, Itoh-Nakadai A, Muto A. Orchestration of plasma cell Kristic J, Nimmo ER, Kalla R, Drummond H, Stambuk J, et al. Inflammatory differentiation by Bach2 and its gene regulatory network. Immunol Rev. bowel disease associates with proinflammatory potential of the 2014;261:116–25. immunoglobulin G glycome. Inflamm Bowel Dis. 2015;21:1237–47. 63. Kuwahara M, Suzuki J, Tofukuji S, Yamada T, Kanoh M, Matsumoto A, 44. Plomp R, Ruhaak LR, Uh HW, Reiding KR, Selman M, Houwing-Duistermaat Maruyama S, Kometani K, Kurosaki T, Ohara O, et al. The Menin–Bach2 axis JJ, Slagboom PE, Beekman M, Wuhrer M. Subclass-specific IgG glycosylation is critical for regulating CD4 T-cell senescence and cytokine homeostasis. is associated with markers of inflammation and metabolic health. Sci Rep. Nat Commun. 2014;5:3555. 2017;7:12325. 64. Dekkers G, Plomp R, Koeleman CA, Visser R, von Horsten HH, Sandig V, 45. Schwab I, Nimmerjahn F. Intravenous immunoglobulin therapy: how does Rispens T, Wuhrer M, Vidarsson G. Multi-level glyco-engineering techniques IgG modulate the immune system? Nat Rev Immunol. 2013;13:176–89. to generate IgG with defined Fc-glycans. Sci Rep. 2016;6:36964. 46. Simurina M, de Haan N, Vuckovic F, Kennedy NA, Stambuk J, Falck D, 65. Chu X, Pan C-M, Zhao S-X, Liang J, Gao G-Q, Zhang X-M, Yuan G-Y, Li C-G, Trbojevic-Akmacic I, Clerc F, Razdorov G, Khon A, et al: Glycosylation of Xue L-Q, Shen M, et al. A genome-wide association study identifies two immunoglobulin G associates with clinical features of inflammatory bowel new risk loci for Graves’ disease. Nat Genet. 2011;43:897–901. diseases. Gastroenterology. 2018;154(5):1320–1333.e10. 66. Dubois PCA, Trynka G, Franke L, Hunt KA, Romanos J, Curtotti A, Zhernakova 47. Barrett JC, Lee JC, Lees CW, Prescott NJ, Anderson CA, Phillips A, Wesley E, A, Heap GAR, Ádány R, Aromaa A, et al. Multiple common variants for celiac Parnell K, Zhang H, Drummond H, et al. Genome-wide association study of disease influencing immune gene expression. Nat Genet. 2010;42:295–302. ulcerative colitis identifies three new susceptibility loci, including the HNF4A 67. Sawcer S, Hellenthal G, Pirinen M, Spencer CC, Patsopoulos NA, Moutsianas region. Nat Genet. 2009;41:1330–4. L, Dilthey A, Su Z, Freeman C, Hunt SE, et al. Genetic risk and a primary role 48. McGovern DPB, Jones MR, Taylor KD, Marciante K, Yan X, Dubinsky M, for cell-mediated immune mechanisms in multiple sclerosis. Nature. 2011; Ippoliti A, Vasiliauskas E, Berel D, Derkowski C, et al. Fucosyltransferase 2 476:214–9. (FUT2) non-secretor status is associated with Crohn’s disease. Hum Mol 68. Macdonald TT, Monteleone G. Immunity, inflammation, and allergy in the Genet. 2010;19:3468–76. gut. Science. 2005;307:1920–5. 49. Lauc G, Huffman JE, Pucic M, Zgaga L, Adamczyk B, Muzinic A, Novokmet 69. Vossenkamper A, Hundsrucker C, Page K, van Maurik A, Sanders TJ, Stagg AJ, M, Polasek O, Gornik O, Kristic J, et al. Loci associated with N-glycosylation Das L, MacDonald TT. A CD3-specific antibody reduces cytokine production of human immunoglobulin G show pleiotropy with autoimmune diseases and alters phosphoprotein profiles in intestinal tissues from patients with and haematological cancers. PLoS Genet. 2013;9:e1003225. inflammatory bowel disease. Gastroenterology. 2014;147:172–83. 50. Dias AM, Dourado J, Lago P, Cabral J, Marcos-Pinto R, Salgueiro P, Almeida 70. Jaffe AE, Irizarry RA. Accounting for cellular heterogeneity is critical in CR, Carvalho S, Fonseca S, Lima M, et al. Dysregulation of T cell receptor N- epigenome-wide association studies. Genome Biol. 2014;15:R31. glycosylation: a molecular mechanism involved in ulcerative colitis. Hum Mol Genet. 2014;23:2416–27. 51. Reinius LE, Acevedo N, Joerink M, Pershagen G, Dahlen SE, Greco D, Soderhall C, Scheynius A, Kere J. Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS One. 2012;7:e41361. 52. Krištić J, Zoldoš V, Lauc G. Complex genetics of protein N-glycosylation. In: Endo T, Seeberger PH, Hart GW, Wong C-H, Taniguchi N, editors. Glycoscience: biology and medicine. Tokyo: Springer Japan; 2014. p. 1–7.

Journal

Clinical EpigeneticsSpringer Journals

Published: Jun 4, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off