Genome-wide distribution of 5hmC in the dental pulp of mouse molars and incisors

Abstract

The dental pulp is critical for the production of odontoblasts to create reparative dentin. In recent years, dental pulp has become a promising source of mesenchymal stem cells that are capable of differentiating into multiple cell types. To elucidate the transcriptional control mechanisms specifying the early phases of odontoblast differentiation, we analysed the DNA demethylation pattern associated with 5-hydroxymethylcytosine (5hmC) in the primary murine dental pulp. 5hmC plays an important role in chromatin accessibility and transcriptional control by modelling a dynamic equilibrium between DNA methylation and demethylation. Our research revealed 5hmC enrichment along genes and non-coding regulatory regions associated with specific developmental pathways in the genome of mouse incisor and molar dental pulp. Although the overall distribution of 5hmC is similar, the intensity and location of the 5hmC peaks significantly differs between the incisor and molar pulp genome, indicating cell type-specific epigenetic variations. Our study suggests that the differential DNA demethylation pattern could account for the distinct regulatory mechanisms underlying the tooth-specific ontogenetic programs.

Graphical Abstract

Graphical Abstract.

Introduction

Studies of tissue-specific epigenetic modifications have revealed that 5-hydroxymethylcytosine (5hmC) is a stable DNA modification mark important for embryonic development and cell differentiation (1). 5hmC is generated by oxidation of 5-methylcytosine (5mC), a process catalysed by the ten-eleven translocation (TET) enzymes (1,2). The TET enzymes affect transcriptional regulation by converting 5mC to 5hmC along genes and non-coding regions. A growing body of research has revealed that TET proteins govern cell fate decisions during the development of various cell types by activating a cell-specific gene expression program (3,4). Research has shown that the tissue-specific distribution and density of 5hmC influence chromatin accessibility at the regulatory elements and impact the transcription of cell identity genes (5). Moreover, emerging evidence suggests that DNA hydroxymethylation plays a fundamental role during lineage commitment, cancer and ageing (1,2,6).

The dental pulp is a highly specialized tissue for creating dentin that provides the tooth with nutrition and is involved in inflammatory responses against invading pathogens (7,8). Over the past few years, the dental pulp-derived stem cells (DPSCs) have emerged as a possible solution for repairing or regenerating damaged tissues. Stem cells derived from dental pulp exhibit the capacity to differentiate into osteogenic, angiogenic, myogenic and chondrogenic cells (9–11). However, the precise mechanisms that determine lineage specification are still largely unknown. A number of studies have suggested that epigenetic marks play a pivotal role in the control of lineage differentiation of mesenchymal stem cells from DPSCs (9,12). Hence, functionally elucidating the epigenetic mechanisms has become increasingly important for understanding the gene regulatory networks underlying differentiation of pulp progenitor cells (9–11).

Previous studies have implicated modification of 5hmC mediated by TET enzymes in the regulation of cellular and developmental processes in dental pulp. Rao et al. (13) revealed that TET1 facilitates dental pulp repair and regeneration. The functional inactivation of TET1 suppresses odontoblast differentiation by regulating FAM20C (14). Another study suggested that TET2-dependent DNA demethylation might play an important role in dental pulp inflammation (15). Moreover, DNA demethylation appears to increase histone acetylation and histone methylation in DPSCs (16).

In order to clarify the biological function of 5hmC in the mouse dental pulp we used hydroxymethylated DNA immunoprecipitation sequencing (hMeDIP-seq), which combines hMeDIP with massive parallel DNA sequencing. We demonstrated that 5hmC is highly enriched at the transcribed regions of genes (gene bodies) and the non-coding regulatory elements in the incisor and molar pulp genome.

Materials and Methods

Primary dental pulp isolation and genome-wide sequencing

Primary pulp from mouse incisors and molars was prepared from 5- to 6-day-old mice according to the previously described procedures (17). hMeDIP-seq and RNA-seq services were provided by Active Motif (Carlsbad, CA). Briefly, genomic DNA was isolated using the Monarch Genomic DNA Isolation kit (New England Biolabs, Ipswich, MA) following the manufacturer’s instructions, DNA was sonicated to ~150–300 bp and Illumina adaptors were ligated to the DNA ends. To generate genome-wide maps of 5hmC, Active Motif performed hMeDIP-seq experiments using the antibody AM39791 to 5hmC. The input DNA was used as a control. Finally, immunoprecipitated (IP) DNA and input DNA that did not go through the IP step were processed into sequencing libraries and sequenced using the Illumina platform (NextSeq 500, 75-nt single-end). RNA isolation was performed using the RNAeasy Mini/Midi kit (Qiagen, Germantown, MD). RNA sequencing was performed using an Illumina NextDeq 500 to generate 42-nt paired-end sequences.

Computational analyses

The 75-nt single-end (SE75) sequence reads generated by Illumina sequencing (using NextSeq 500) were mapped to the mouse genome (mm10) using the BWA algorithm with default settings. Alignment information for each read was stored in the BAM format. For the analysis, only uniquely mapped reads without duplicates were used, and tag numbers were normalized to the lowest number among the samples (by down sampling), which was 19.7 million. Methylated regions (peak intervals) were identified using the MACS peak finding algorithm (18) with default cutoff of P = 1e-7. The genomic locations of the peak intervals were annotated with the genes in their proximity (10 kb) to understand the genomic distribution of methylated regions in both incisors and molar samples as illustrated in Fig. 1B. The 5hmC tag distribution around the genes (promoters, transcription start sites (TSSs), gene bodies, transcription end sites (TESs) and downstream areas) were determined and presented as average plots (average of values for all target regions) in Fig. 1C. A set of genes with 5hmC peak intervals on their promoter region (−2 kb to 500 bp from the TSS) was identified for both incisors and molars. Similarly, a set of genes with methylation peaks on their gene bodies was identified for both dental pulp studies. GO enrichment analysis and KEGG pathway analysis were performed on these four sets of genes via the NCBI DAVID web application (19–21). To understand the relationship between 5hmC enrichment and the gene expression profile, we used the approach described by Wang et al. (22) where genes were separated into five groups (10–20%, 20–40%, 40–60%, 60–80% and 80–100%) based on their mRNA expression levels. Box plots of average 5hmC signal values were generated for each group of genes in promoters and gene bodies for studying both incisors and molars.

Fig. 1.

Global distribution of 5hmC in the mouse dental pulp genome. (A) High-resolution map of 5hmC enrichment across the mouse chromosome 2. Genome-browser view of hMeDIP-seq in the mouse dental pulp genome. The molar and incisor hMeDIP-seq datasets along with input DNA controls are shown. Each track is represented as normalized density of reads/bp/million uniquely mapped reads. (B) 5hmC is primarily enriched at the intergenic regions and gene bodies of both the incisor and molar genomes. (C) The regulatory regions flanking the gene body were defined as the upstream promoter regions with an interval from −2 kb to 500 bp from the TSS and the downstream regulatory regions with an interval from 500 bp from the TES of the gene until the 2 kb downstream region. The average plot revealed that 5hmC signal is low around the TSS, and the overall intensity of 5hmC increases in promoters and gene bodies.

The RNA-seq fragments were mapped to the reference genome mm10 using the STAR aligner (23). The fragment assignment step was carried out to count the number of fragments overlapping the genomic sequence. Only the read pairs that have both ends aligned at the same chromosome and the same DNA strand were considered for subsequent analysis. The feature count (FPKM assignment to genes) was performed using the Subread package (24). Gene annotations were originally from the NCBI RefSeq database and then adapted by merging the overlapping exons from the same gene to form a set of disjoint exons for each gene. After obtaining the same table containing the fragment (or read) of genes, differential analysis was performed using DESeq2 (25).

Results

Global distribution of 5hmC in the mouse dental pulp genome

We used the hMeDIP-seq method to examine the genome-wide landscape of 5hmC in the dental pulp of mouse incisors and molars. Figure 1A shows a genome-browser view of the hMeDIP-seq datasets along with the input DNA control. Our analysis revealed that 5hmC is primarily enriched in the gene body regions (41.3% in incisors and 44.6% in molars) and the intergenic regions (39.7% in incisors and 38.7% in molars). By contrast, the spread of 5hmC across the upstream and downstream regions was relatively modest, with 9.4% and 9.6% in the incisor genome and 8.1% and 8.6% in the molar genome, respectively (Fig. 1B). These results suggest that, in the dental pulp genome 5hmC is mainly enriched at intergenic regions and gene bodies.

Genic distribution of 5hmC in the mouse dental pulp genome

We defined the regulatory regions flanking the gene body as the upstream promoter regions with an interval from −2 kb to 500 bp from the TSS and the downstream regulatory regions with an interval from 500 bp from the TES of the gene until the 2 kb downstream region (Fig. 1C). Average plots revealed that 5hmC signal is lowest around the TSS, and the overall intensity of 5hmC increases at promoters and gene bodies (Fig. 1C). We performed KEGG (Fig. 2) and gene ontology (GO, data not shown) pathway analyses via the NCBI DAVID web application. In the incisor pulp genome, 5hmC was enriched at promoters of genes associated with cell adhesion, cytotoxicity, cytokine receptor interaction and cell signalling pathways (Fig. 2A). In the molar pulp genome, we detected a relatively modest enrichment of 5hmC signal in the promoter regions of genes involved in immune response, signalling and cell adhesion (Fig. 2B). Finally, within the gene body region in both incisors and molars, 5hmC was enriched at active genes involved in development, differentiation, cell adhesion and cell signalling (Fig. 2A and B).

Fig. 2.

KEGG analysis indicates that 5hmC is enriched at promoters and gene bodies of key developmental genes. (A) In the incisor pulp genome, 5hmC is enriched at promoters of genes associated with cell adhesion, cytotoxicity, cytokine receptor interaction and cell signalling pathways. (B) In the molar pulp genome, 5hmC signal is enriched at promoters of genes involved in cell adhesion. In the gene body region in both incisors and molars, 5hmC is enriched at active genes involved in development, differentiation, cell adhesion and cell signalling.

Relationship between 5hmC enrichment and gene expression

To analyse the relationship between 5hmC enrichment and gene expression, we used the approach described by Wang et al. (22). We separated genes into five groups according to their expression levels (from low to high: 0–20%, 20–40%, 40–60%, 60–80% and 80–100%) and plotted the 5hmC enrichment profiles across the promoter and gene body regions. We discovered that in the molar pulp genome, deposition of 5hmC is relatively low at promoters of highly expressed genes (80–100% group) (Fig. 3A, left). We also detected a relatively low level of 5hmC in the 60–80% and 40–60% gene groups. In contrast, 5hmC signal was higher in the 20–40% and 0–20% gene groups. However, within the gene bodies 5hmC was more enriched in the 40–60%, 20–40% and 0–20% groups (Fig. 3A, right). 5hmC signal was relatively modest in the 80–100% and 60–80% gene groups. Interestingly, the intensity of 5hmC was higher at the gene bodies of the 20–40% group genes (Fig. 3A, right). We found that 5hmC showed similar distribution patterns in the incisor pulp genome (Fig. 3B). Collectively, these findings suggest a low correlation between the intensity of 5hmC and gene expression in the primary pulp genome.

Fig. 3.

The relationship between 5hmC enrichment and gene expression. (A) Genes were separated into five groups according to their RNA expression levels (from low to high: 0–20%, 20–40%, 40–60%, 60–80% and 80–100%) and 5hmC enrichment profiles were plotted across the promoter and gene body regions. In the molar pulp genome, 5hmC signal is relatively low at promoters of highly expressed genes (80–100% group) (A, left). A relatively low level of 5hmC is observed in the 60–80% and 40–60% gene groups. In contrast, 5hmC signal is higher in the 20–40% and 0–20% gene groups. Within the gene bodies, 5hmC is more enriched in the 40–60%, 20–40% and 0–20% groups (A, right). 5hmC signal is relatively modest in the 80–100% and 60–80% gene groups. Interestingly, the intensity of the 5hmC signal is higher at the gene bodies of the 20–40% group. (B) We found that 5hmC showed similar distribution patterns in the incisor pulp genome. The distribution of average FPKM values in dental pulp ranged from 0 to 19,388.6 for molars and from 0 to 6,056.69 for incisors from a total of 16,331 genes.

Differential enrichment of 5hmC in the mouse incisor and molar pulp genome

We observed significant differences in the intensity and localization of 5hmC signal along the non-coding regions and gene bodies between the incisor and molar dental pulp genomes. In the incisor pulp genome, 5hmC was enriched at genes engaged in lineage commitment and specification. In the mouse incisor dental pulp, the 5hmC signal was significantly higher at the promoter of Sp7, at the end of the gene body region of Twist2 and along the non-coding regulatory regions and gene bodies of Dspp and Dmp1 (Fig. 4). Sp7 is required for the proliferation and differentiation of ameloblasts and odontoblasts (26). Studies have reported that Twist1, a close paralog of Twist2, is essential for tooth morphogenesis and odontoblast differentiation (27). Moreover, Dspp and Dmp1 significantly affect bone mineralization and activate integrin signalling during tooth development (28).

Fig. 4.

The differential enrichment of 5hmC in the regulatory regions and gene bodies of mouse dental pulp. Within the incisor pulp genome, 5hmC is strongly associated with genes engaged in lineage commitment and specification. 5hmC signal is significantly higher (A) at the promoter of Sp7, (B) at the end of the gene body region of Twist2, (C) and along the non-coding regulatory regions and gene bodies of Dspp and Dmp1.

Previous studies have shown that 5hmC-rich genomic sites mark specific genes and non-coding regulatory elements associated with development or disease pathogenesis (29). Therefore, our results revealed significant differences between the incisor and molar dental pulp genomes in the location and intensity of 5hmC at the key genes associated with tooth development.

Discussion

In the presented study, we examined the global distribution of 5hmC in the mouse dental pulp genome. Our data demonstrated that 5hmC is enriched in the genic and intergenic regions in both the incisor and molar pulp genomes. Consistent with several previously reported studies of different vertebrate tissues, our analysis showed a low level of 5hmC signal around TSS. Conversely, 5hmC levels were significantly higher at the gene bodies and non-coding regulatory regions. A recent study reported the preferential enrichment of 5hmC on the tissue-specific gene bodies and enhancers suggesting the involvement of 5hmC in lineage-specific gene regulation (29). Reports have estimated that approximately one-third of all 5hmC peaks are engaged in tissue-specific gene expression (30). These DNA hydroxymethylated regions are also enriched in tissue-specific transcription factors that could rewire distinct gene regulatory networks.

Although our research revealed 5hmC enrichment at the gene body regions and promoters, this DNA demethylation pattern did not correlate with gene expression. The purpose of 5hmC is likely more complicated than mere involvement in gene activation. DNA demethylation is a highly coordinated enzymatic process and 5hmC is one of the intermediate marks engaged in specific interactions with other components of chromatin remodelling such as histone methylation and histone acetylation. It is possible that in the mouse dental pulp genome 5hmC pre-marks poised enhancers and downstream genes for later activation. Interestingly, in the placental genome, 5hmC marks the poised enhancers and is depleted at the active chromatin domains that are enriched in H3K27ac and H3K4me1 (31). However, gene activation could require additional epigenetic modifications such as histone acetylation or histone methylation. For example, in the intestinal progenitor cells a low level of 5hmC does not correlate with gene expression, yet, upon differentiation, an increase in 5hmC can overlap with the active histone modification marks H3K36me3, H3K27ac and H3K4me1 (32). In mouse neurons, colocalization of 5hmC with H3K4me2 is associated with control of the active chromatin state (33). Another recent report showed that changes in DNA demethylation in concert with the bivalent histone marks H3K27me3 and H3K4me3 could initiate reprogramming and regeneration of the embryonic retina (34).

In human embryonic stem cells (ESCs), deposition of 5hmC triggers an ESC-specific regulatory program by cooperating with the active histone modification marks H4K8ac and H3K4me1 (35). Surprisingly, in mouse ESCs, 5hmC acts as a repressor at specific regulatory elements known as ‘silenced enhancers’ (36). In human ESCs, TET1 works in concert with QSER1 to safeguard chromatin from DNA methylation (37). QSER1 is a relatively unknown factor that preferentially protects bivalent promoters and poised enhancers. The interaction between TET1 and QSER1 ensures efficient exclusion of the DNA methylation enzymes DNMT3A/3B from binding to chromatin and repressing genes (37). Based on these findings, it is possible that 5hmC defines both the poised and active chromatin and thus plays a dual role in the dental pulp genome.

Surprisingly, we found that 5hmC intensity and its distribution along the promoter and gene body regions are significantly different in the dental pulp genome of mouse molars and incisors. The specific differences in the DNA demethylation status could play an instrumental role in the transcriptional control mechanisms underlying incisor and molar formation. The functional significance of 5hmC in the differential gene regulation in the mouse pulp genome requires further investigation. In conclusion, we provide the first evidence of the genome-wide distribution of the DNA demethylation mark 5hmC in the mouse dental pulp.