The mechanisms underlying congenital impairment in DS are unclear. Accumulating evidence suggests that DS phenotypes, including various pathological changes, occur by alterations of the complex regulation of many genes on and outside HSA21. The alterations could affect the roles of the upstream effector genes, located on HSA21, and the downstream target genes that are affected by DNA methylation, which are distributed throughout the genome of DS individuals. Therefore, investigating the epigenomic changes that contribute to the various phenotypes of DS may improve our understanding of the pathophysiology of DS.
In the current study, we performed DNA methylation profiling at single-base resolution in the chorionic villi cells of human DS and euploid fetuses at an early developmental stage. Global DNA hypermethylation across the whole epigenome was evident in DS fetuses, with methylation differing in 4439 DMCs compared to that in euploid fetuses. Most DMCs were hypermethylated in DS and only approximately 4% of DMCs were hypomethylated. These DMCs were distributed in all chromosomes, regardless of the methylation pattern in DS. Hypermethlyated DMCs were pronounced in the exonic regions in intragenic regions and the promoters in intergenic regions. Additionally, two genes (CPT1B and TNXB) were commonly hypermethylated in fetal brain and chorionic villi in DS. We identified 141 DMRs that ranged from three to 79 DMCs. All DMRs, except one DMR, were hypermethylated in DS. In a functional annotation of identified DMR genes, we found that these genes with hypermethylated DMRs were significantly related to the various phenotypes and pathophysiology including chromosomal aberration and congenital abnormalities of DS. This finding suggests the possibility of fetal epigenetic therapy by inhibiting the global DNA methylation in DS. Furthermore, chorionic villi cells (e.g., trophoblasts) are the major origin of cell-free fetal DNA in maternal blood [21]. Therefore, profiling DNA methylation in chorionic villi cells of fetal DS could lead to potential clinical application to develop novel approaches for the noninvasive prenatal test of DS in the first trimester of pregnancy by analysis of fetal DS-specific DNA methylation in maternal blood.
Global hypermethylation in DS was conserved in different tissue types and developmental stages [13,14,15,16,17,18,19]. However, the number and location of DMCs differed in the whole genome depending on tissue types and developmental stage. In peripheral blood leukocytes of DS adults, eight genes (TMEM131, TCF7, CD3Z/CD247, SH3BP2, EIF4E, PLD6, SUMO3, and CPT1B) were hypermethylated [18]. In the fetal cortex of DS, nine genes (C1orf35, CPT1B, DECR2, FAM83H, GLI4, LRRC14, LRRC24, STK19, and TNXB) were hypermethylated [14]. In the fetal placenta of DS, hypermethylation of CPT1B, TCF7, and FAM62C was evident [15]. We also found hypermethylation in the 5′UTR of CPT1B in chorionic villi cells at an early developmental stage of fetal DS. These results suggest that hypermethylation of CPT1B is conserved in various tissue samples of the different developmental origin of DS and may be a common epigenetic characteristic in DS. CPT1B on chromosome 22 is one of the carnitine palmitoyltransferase 1 genes [22]. This enzyme is required for the net transport of long-chain fatty acyl-CoAs from the cytoplasm to the mitochondria [14, 22]. Mitochondrial dysfunction has been found in the brains of patients with various psychiatric disorders, including DS [14]. Placental mitochondrial dysfunction may also be critical in a range of gestational disorders, which has important implications for maternal and fetal/offspring health. Therefore, these results provide the new insight that the hypermethylation of CPT1B may be associated with mitochondrial dysfunction in the developing DS.
In this study, the methylation degree of ZNF837, C11orf95, KIAA1875, SHROOM1, and CASZ1 with more than 40 hypermethylated DMCs were confirmed by methylation-specific real-time PCR. The results were consistent with the MC-seq results. Among these hypermethylated DMR genes, SHROOM1 and CASZ1 appear important potentials in pathophysiology of complications related to DS. SHROOM1 on chromosome 5 encode Shroom Family Member 1. SHROOM family members play diverse roles in the development of the nervous system and other tissues and may be involved in the assembly of microtubule arrays during cell elongation [23]. Diseases associated with SHROOM1 include atrial septal defect 2 and atrial heart septal defect. GO annotations related to this gene include actin filament binding. Ectopic expression of SHROOM1 elicited apical accumulation of gamma-tubulin in naive epithelial cells, consistent with a role for SHROOM1 in governing microtubule architecture [24]. It was recently reported that patients with knee osteoarthritis had marked hypermethylation status in all SHROOM1 differentially methylated sites between damaged and non-damaged tissues [25]. The authors suggested hypermethylated SHROOM1 as a promising candidate for functional studies of osteoarthritis pathology [25]. In another study, hypermethylation of SHROOM1 was present in DS fetuses with or without congenital heart defects, as well as in fetuses with heart malformations [16]. Congenital heart defects are the most common malformation occurring in DS. There is also an increased risk of skeletal abnormalities, including osteoarthritis, in DS. There certainly is a higher incidence of heart and joint problems in DS. However, these disorders are thought to have a multifactorial etiology and the main causes are largely unknown. Therefore, hypermethylation of SHROOM1 in DS may be potentially associated with the pathophysiology of the congenital heart defects and osteoarthritis related to DS. Additionally, CASZ1 on chromosome 1 is a protein-coding gene of castor zinc finger 1, which is a zinc finger transcription factor. The single-nucleotide polymorphisms in this gene are associated with blood pressure variation. Alternative splicing yields multiple transcript variants encoding different protein isoforms [26]. They are involved in vascular assembly and morphogenesis through direct transcriptional regulation of EGFL7 [27]. Diseases associated with CASZ1 include retroperitoneal sarcoma and retroperitoneum carcinoma. Interestingly, Casz1 is expressed in a number of developing tissues, including neural, endothelial, and cardiac tissues. Improper regulation of CASZ1 could potentially lead to the defects that are collectively observed in DS, such as cognitive defects, congenital heart defects, and hypertension [28, 29]. CASZ1 also has an established role in regulating cellular adhesion in multiple tissues. In previous studies of CASZ1 function, HSA21 resident-congenital heart disease 5 protein (CHD5) was demonstrated as being interactive and essential protein for CASZ1 function, with the CHD5-CASZ1 interaction being necessary for cardiac morphogenesis [30, 31]. Therefore, a potential mechanism that could link the misregulation of CASZ1 to one or more pathologies related to DS has been suggested. However, the precise mechanism by which CASZ1 is regulated by its interaction with CHD5 has yet to be determined. In this study, we firstly found that exonic region of Casz1 is hypermethylated in DS. This epigenetic change of Casz1 in DS may provide additional insight into the phenotypes generated by its misregulation. Furthermore, ZNF837 on chromosome 19 encodes zinc finger protein 837 and is commonly hypermethylated in the adult and fetal brain, and the fetal placenta in DS [1]. In this study, ZNF837 was identified as the DMR gene with the largest DMCs in the chorionic villi cells in the early developmental stage of DS fetuses. Although this gene may be involved in transcriptional regulation, since ZNF837 is located in the nucleus, an obvious function is still unclear. C11orf95 (Chromosome 11 Open Reading Frame 95) is associated with ependymoma as tumors of the brain and spinal cord [32]. C11orf95–RELA fusion protein translocates spontaneously to the nucleus to activate nuclear factor-kappa B target genes, and rapidly transform neural stem cells—the originating cells of ependymoma—to form these tumors [33]. However, ependymoma in DS has only been reported so far in only two cases, and the association of ependymoma with DS is a rarity [34, 35]. Finally, KIAA1875 (previous HGNC symbols for WDR97 and encode WD Repeat Domain 97) has an unknown function and protein-coding potential [36]. Further research is needed to understand the epigenetic function and mechanisms of known and still-undiscovered DMR genes. Our results have potential utility value as basic information for further studies of these DMR genes in DS.
Several studies have compared the DNA methylation profiles between normal and DS patients and have demonstrated common epigenetic characteristics. These include global hypermethylation and a lack of enrichment on HSA21. This DNA methylation pattern may be tissue and developmental stage-specific in DS. Moreover, DS as a congenital aneuploidy may involve continuous epigenetic changes in whole genome from embryos early in development. These early epigenetic changes may be the basis of the developmental defects in DS. Therefore, it is important to investigate epigenetic changes in the early development of DS fetuses to clarify the pathogenesis of DS. Recently, global hypermethylation was demonstrated in the fetal cortex and placenta of human DS [14, 15]. We also found the same methylation pattern in early chorionic villi cells of fetal sex and gestational age-matched DS and normal fetuses. However, the percentage of hypermethylated DMCs identified in each study varied from 1.85 to 4%. The variation likely reflects the impact of various factors, such as tissue type, gestational age at sampling, criteria for DMC selection, and methods used. In this study, we used MC-Seq for high-throughput methylome profiling in DS. This is a next-generation sequencing approach that targets specific loci in the epigenome. MC-seq utilizes target- specific bait sequences, circumventing this restriction through bait design and allows for epigenome-wide surveying of specific genomic loci of physiological and clinical interest. Thus, it presents an attractive, cost-effective alternative to uncover novel disease-associated genomic loci. Moreover, MC-seq overcomes the limitations of lower genome coverage in a methylation microarray, high cost, and processing time in whole-genome bisulfite sequencing, while avoiding overrepresentation in next-generation sequencing platforms, such as reduced representation bisulfite sequencing and methylated DNA immunoprecipitation sequencing [20]. Furthermore, in this study, the MC-seq with a depth exceeding 300× was performed to overcome the technical variation in detecting small inter-individual/inter-group differences. Therefore, we were available to interrogate DNA methylation more precisely in clinical samples at the single-base resolution. However, many of our results were based on databases of bioinformatics tools. Although these in silico bioinformatic tools are useful to predict multi-genes associated with the pathophysiology of disorder, the in silico results are not strong enough to justify the functional significance of genes. Moreover, due to the very small amount of fetal chorionic cells that were obtained, we could not gain more confidence through further studies of gene expression in DS. Therefore, the functional roles of the potential pathways in our study need further characterization, possibly in cell lines or tissues relevant to specific DS phenotypes. These potential pathways with epigenetic changes in DS remain to be further elucidated. Furthermore, this study was limited by its small sample size and no adjustment for confounders. DNA methylation could be affected by various environmental and clinical factors. Therefore, future, larger studies are needed to substantiate our findings. Despite the limitations of this study, our findings warrant further studies addressing DNA methylation of the whole genome associated with the pathogenesis of DS. Various further studies might support the functional significance of our new insight into the pathophysiology of DS.