DNA methylation profile of Aire-deficient mouse medullary thymic epithelial cells
© Wu et al.; licensee BioMed Central Ltd. 2012
Received: 23 June 2012
Accepted: 27 October 2012
Published: 2 November 2012
Skip to main content
© Wu et al.; licensee BioMed Central Ltd. 2012
Received: 23 June 2012
Accepted: 27 October 2012
Published: 2 November 2012
Medullary thymic epithelial cells (mTECs) are characterized by ectopic expression of self-antigens during the establishment of central tolerance. The autoimmune regulator (Aire), which is specifically expressed in mTECs, is responsible for the expression of a large repertoire of tissue-restricted antigens (TRAs) and plays a role in the development of mTECs. However, Aire-deficient mTECs still express TRAs. Moreover, a subset of mTECs, which are considered to be at a stage of terminal differentiation, exists in the Aire-deficient thymus. The phenotype of a specific cell type in a multicellular organism is governed by the epigenetic regulation system. DNA methylation modification is an important component of this system. Every cell or tissue type displays a DNA methylation profile, consisting of tissue-dependent and differentially methylated regions (T-DMRs), and this profile is involved in cell-type-specific genome usage. The aim of this study was to examine the DNA methylation profile of mTECs by using Aire-deficient mTECs as a model.
We identified the T-DMRs of mTECs (mTEC-T-DMRs) via genome-wide DNA methylation analysis of Aire−/− mTECs by comparison with the liver, brain, thymus, and embryonic stem cells. The hypomethylated mTEC-T-DMRs in Aire−/− mTECs were associated with mTEC-specific genes, including Aire, CD80, and Trp63, as well as other genes involved in the RANK signaling pathway. While these mTEC-T-DMRs were also hypomethylated in Aire+/+ mTECs, they were hypermethylated in control thymic stromal cells. We compared the pattern of DNA methylation levels at a total of 55 mTEC-T-DMRs and adjacent regions and found that the DNA methylation status was similar for Aire+/+ and Aire−/− mTECs but distinct from that of athymic cells and tissues.
These results indicate a unique DNA methylation profile that is independent of Aire in mTECs. This profile is distinct from other cell types in the thymic microenvironment and is indicated to be involved in the differentiation of the mTEC lineage.
Medullary thymic epithelial cells (mTECs) are highly involved in the establishment of central tolerance by ectopically expressing a variety of tissue-restricted antigens (TRAs) . Deficiency of the autoimmune regulator, Aire—which regulates a large pool of genes in mTECs—results in autoimmune polyglandular syndrome in humans and autoimmunity in mice . There are estimated to be hundreds to over one thousand TRAs that are activated by Aire in mTECs, based on microarray data from the transcriptomes of wild-type and Aire-deficient mTECs [3, 4]. However, numerous TRAs are expressed in mature mTECs in an Aire-independent pattern [3, 4]. Moreover, single-cell PCR analysis with sorted mTECs demonstrated that approximately 70% , or even fewer , of mature mTECs express Aire, and TRAs are expressed in Aire-negative mTECs . These findings indicate that Aire is necessary, but not sufficient, for the expression of TRAs in mTECs.
The role of Aire in the maturation of mTECs has been indicated by using Aire-deficient mice, which exhibit contracted compartments in the thymic medulla and an increased number of medullary cells positive for p63, whose expression is diminished in Aire-positive mTECs [6–9]. The role of Aire in the differentiation of mTECs has been reviewed by Matsumoto . However, the existence of the medullary compartment and the increased number of a small subset of K5−K8+ cells—considered to represent end-stage and terminally differentiated mTECs—in the Aire-deficient thymus indicate a limited role of Aire in the lineage determination of mTECs, in which the RANK signaling pathway is involved [11, 12]. Thus, Aire-deficient mTECs may represent a suitable model to study not only Aire-independent TRA expression but also the fundamental regulation system that governs cell lineage determination of mTECs.
Genome-wide analysis of DNA methylation has indicated that a specific DNA methylation profile exists for every cell or tissue type. This profile consists of numerous tissue-dependent and differentially methylated regions (T-DMRs) at cytosine residues in CpG dinucleotides (CpGs), which are differentially methylated among distinct cells and tissues . DNA methylation profiles change during cell differentiation, reflect developmental similarity among cell lineages [14, 15], and illuminate cell- or tissue-specific gene expression profiles [16–18]. These findings suggest that DNA methylation profiles are involved in the establishment and maintenance of the regulatory system for cell- or tissue-type-specific genome use, and could therefore be used to identify types of cells and tissues.
The DNA methylation profile responsible for the phenotype of mTECs is still largely unknown, although a limited number of genes, including Aire and 2 Aire-independent TRA genes, Csn2 (Csnb) and Gad1 (Gad67), are unmethylated in their promoter regions [19–21]. To explore the genome-wide DNA methylation of mTECs, we applied a microarray-based screening system, T-D MR profiling with re striction-tag-mediated am plification (D-REAM) , to UEA1+CD45− mTECs that were isolated from Aire-deficient (Aire−/−) mice . We identified T-DMRs in mTECs (mTEC-T-DMRs) and compared the DNA methylation status at mTEC-T-DMRs in Aire+/+ and Aire−/− mTECs, UEA1−CD45− thymic stromal cells, and athymic cell and tissues.
The mTECany2-T-DMRs mapped to within 8.5-kb regions around the transcription start sites (TSSs) of 3081 ENSTs (Ensembl transcripts; mm9) (Additional file 3: Table S1). The distribution of mTECany2-T-DMRs exhibited a similar pattern to that observed in previous analyses of somatic tissues and ES cells [16, 17], displaying a peak of T-DMRs at the 3'-downstream regions of TSSs (Additional file 4: Figure S3). Hypomethylated signals in these regions are often associated with cell- and tissue-specific highly expressed genes [16, 17].
We analyzed the expression of ENSTs associated with mTECany2-T-DMRs in mTECs using the gene expression profile of CDR1intB7-1hi CD45− mTECs from the Gene Expression Omnibus (GEO) database, and compared with those of mouse liver, and CD4+CD8+ thymocytes as a representative of the thymic cells (Additional file 5: Table S2). The differences in expression levels of genes associated with mTECany2-T-DMRs were relatively small between Aire−/− and Aire+/+ mTECs as compared with those between mTECs and the liver or thymocytes, indicating that the expression patterns of genes associated with mTECany2-T-DMRs in mTECs are not affected by functional Aire (Additional file 6: Figure S4). High gene expression in mTECs was more prominent for genes associated with mTECany2-T-DMRs in the −1.5 kb upstream and +2.5 kb downstream (proximal) regions of TSSs. Among these genes, those associated with proximal mTECu-T-DMRs showed significantly higher expression in mTECs compared with both the liver and thymocytes (Figure 1B). These data indicate that mTEC-specific genes are likely to be associated with mTECu-T-DMRs, and methylation may be responsible for gene silencing in Aire non-expressing tissues and cells. Indeed, we found that among the mTECu-T-DMR–associated genes, some had been previously reported as mTEC-specific genes, of which the majority showed Aire-independent expression in mTECs (Additional file 7: Table S3). Moreover, marker genes of mTECs, including Aire, Cd80, and Trp63, were found to be associated with proximal mTECu-T-DMRs.
We analyzed the DNA methylation status around the mTECu-T-DMR of Aire which is located downstream of its TSS, and the upstream region by using bisulfite sequencing (Figure 1C). In the −624 to −218 bp upstream region, the liver, which represents an Aire non-expressing tissue, showed a methylated status comparable with Aire+/+ and Aire−/− mTECs as well as with ES cells, which express Aire . The promoter regions of human AIRE (−295 to +65 bp) and mouse Aire (−287 to +133 bp) are both associated with a CpG island (>50% CpG content) and are unmethylated in mTECs [19, 20]. Since Aire-negative cTECs and several extrathymic tissues were also found to be unmethylated at this region , hypomethylation of the promoter is considered to be necessary, but not sufficient, for expression. In contrast, Aire+/+ and Aire−/− mTECs, but not ES cells, exhibited a hypomethylated status at the 9 CpGs in the region between +2082 and +1672 bp from the TSS, including the CpG corresponding to the mTECu-T-DMR. Among the tissue-specific genes carrying CpG islands within 1 kb of their TSSs, hypomethylated T-DMRs tend to be located downstream of TSSs [16, 17]. Thus, the mTEC-TDMR for Aire located 3'-downstream of the Aire TSS could serve as an epigenetic marker to distinguish mTECs from Aire-expressing ES cells, in which the expression level of Aire was significantly less than that in the thymus and mTECs (Additional file 1: Figure S1).
The mTECu-T-DMRs associated with CD80 and p63 (Trp63) were analyzed with combined bisulfite restriction analysis (COBRA). Both Aire+/+ and Aire−/− mTECs exhibited a hypomethylated status compared with the liver, thymus, and ES cells (Figure 1D). Notably, the mTECu-T-DMR of p63 was found to be associated with ΔNp63, the short isoform of p63, which is essential for maintenance of the progenitor status of thymic epithelial cells  and not expressed in Aire-positive mTECs [7–9]. Hypomethylation at this T-DMR in Aire+/+ mTECs suggests that DNA methylation does not determine the repression of p63 in Aire+/+ mTEC cells.
Among the analyzed genes, Fgg (fibrinogen gamma chain), which is a liver-specific gene that is expressed in mTECs in an Aire-dependent pattern [3, 4], was hypermethylated in both Aire+/+ and Aire−/− mTECs, similar to its status in ES cells and the thymus. The gene expression of TRAs in mTECs occurs in specific subpopulations and in a stochastic manner [1, 2]; therefore, DNA methylation changes may occur at specific loci in an Aire-dependent manner in subpopulations of mTECs that express a specific TRA.
During thymic organogenesis, Aire is expressed at E16.5 , while the stromal cells of the medulla and cortex are separated at embryonic day 13 (E13) . In the absence of Aire, mTECs survive Aire-induced apoptosis, reach a terminally differentiated stage , and maintain the ability to express TRAs [3, 4]. Thus, the similar DNA methylation status at T-DMRs in Aire+/+ and Aire−/− mTECs suggests the existence of an Aire-independent DNA methylation profile that serves as an intrinsic system to determine the phenotype of mTECs.
In this study, we demonstrated the DNA methylation profile for Aire−/− mTECs by genome-wide identification of T-DMRs. The DNA methylation status at dozens of mTEC-T-DMRs in Aire+/+ mTECs was similar to that in Aire−/− mTECs, while distinct from athymic cells and tissues as well as the other thymic stromal cells, indicating that the Aire-independent DNA methylation profile is involved in differentiation of the mTEC lineage and other thymic stroma cells that form the thymic microenvironment.
Aire+/− mice (B6.129S2-Airetm1.1Doi/J) were obtained from the Jackson Laboratory  and backcrossed onto C57BL/6J (B6) mice (Oriental Yeast, Japan). Aire−/− mice and their Aire+/+ littermates were used for isolation of mTECs. The mice were housed and bred under specific pathogen-free conditions.
All experiments using mice were carried out according to the institutional guidelines for the care and use of laboratory animals (Graduate School of Agricultural and Life Sciences, the University of Tokyo).
The isolation of mTECs was conducted by modifying previously described methods [16, 30]. Briefly, thymi were collected from 6- to 7-week-old male mice, finely minced, washed in RPMI 1640 medium, and sequentially digested twice with 0.2 mg/mL collagenase D for 15 min at 37°C. The pellets were incubated with 0.2 mg/mL collagenase D and 0.2 mg/mL Dispase I (Roche Applied Science) with 25 μg/mL DNase I (Takara), and the reaction was stopped with 5 mM EDTA. Subsequently, after centrifugation at 1350 × g on a Percoll gradient, supernatants between ρ = 1.06 g/mL and ρ = 1.0 g/mL were incubated with Mouse BD Fc Block (2.4G2), PE Rat Anti-Mouse CD45 (30-F11) (BD Pharmingen), and FITC-conjugated UEA-1 (Sigma) and sorted using an EPICS Altra flow cytometer (Beckman-Coulter). The fractionation of cells was monitored by RT-PCR (Additional file 10).
In D-REAM analysis, differential methylation status at HpyCH4IV loci is indicated by differential scores corresponding to fragments generated by this methylation-sensitive restriction enzyme between 2 samples [16; Additional file 2: Figure S2]. In the present study, we performed microarray experiments using mTECs from Aire−/− mice and the thymus from C57BL/6J mice. Briefly, after digestion of genomic DNA with HpyCH4IV (New England Biolabs), 50 ng of DNA was ligated with an adaptor, followed by TaqI digestion and ligation with a second adaptor. The fragments were amplified by ligation-mediated PCR to allow selective amplification of unmethylated fragments generated by HpyCH4IV digestion and then hybridized using the GeneChip Mouse Promoter 1.0R Array (Affymetrix), which comprises probes that map to the regions (−6 to +2.5 kb) around the TSSs of approximately 50,000 transcripts (ENST IDs; mm9) obtained from the UCSC genome browser. Hybridization signals were processed with MAT  and can be found at Array Express (http://www.ebi.ac.uk/arrayexpress/; accession number: E-MTAB-1113). The data from Aire−/− mTECs and the thymus were compared with D-REAM data from the liver, brain, and ES cells  to identify fragments that exhibited a difference in microarray scores, thereby indicating the DNA methylation status at corresponding HpyCH4IV sites. We set the cut-off value at 0.5% of FDR (false detecting rates) for T-DMRs. The Galaxy genome browser (https://main.g2.bx.psu.edu/)  and “R” (http://www.R-project.org)  were used for genomic fragment analysis.
To analyze the gene expression of Aire+/+ and Aire−/− mTECs, we obtained their gene expression profiles from the GEO database (http://www.ncbi.nlm.nih.gov/geo/)  and compared with those of the liver and CD4+CD8+ thymocytes on the same microarray platform (Additional file 5: Table S2). The array data were normalized using the gcrma package . The ENST IDs were converted into Affymetrix IDs by using BioMart software (http://www.biomart.org) .
mTEC cells and gDNAs from the thymus, liver, and ES cells were subjected to bisulfite treatment using the EZ DNA Methylation-Direct Kit (Zymo Research Cooperation) according to the manufacturer’s instructions. Unmethylated cytosine residues are converted into thymine residues by the sodium bisulfite reaction, while methylated cytosine residues remain unchanged. PCR was carried out using the bisulfite-converted genomic DNA with the primers listed in Additional file 11: Table S5. To evaluate the DNA methylation status at specific CpGs, COBRA was performed. The PCR products were evenly divided and incubated with or without HpyCH4IV for 5 h at 37°C and then analyzed with a microchip electrophoresis system (MultiNA, Shimadzu Biotech). DNA methylation levels were calculated as the percentage of digested fragment compared with the sum of digested and undigested fragments. To evaluate the DNA methylation status at CpGs within hundreds of base pairs, the PCR fragments were cloned into the pGEM-T Easy Vector (Promega), and more than 10 clones were sequenced using an ABI 3130 sequencer with a BigDye Terminator kit (Applied Biosystems).
Medullary thymic epithelial cells
Cortical thymic epithelial cells
Tissue-dependent and differentially methylated region
T-DMR profiling with restriction tag-mediated amplification
Ulex europaeus agglutinin I.
This research was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan (no. 21221008 to K.S.); by a grant from the Japanese Society for the Promotion of Science (to TA); and by CREST from the Japan Science and Technology Agency (JST) (to SY).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.