Intergenic sequences harboring potential enhancer elements contribute to Axenfeld-Rieger syndrome by regulating PITX2

Recent studies have uncovered that noncoding sequence variants may relate to Axenfeld-Rieger syndrome (ARS), a rare developmental anomaly with genetic heterogeneity. However, how these genomic regions are functionally and structurally associated with ARS is still unclear. In this study, we performed genome-wide linkage analysis and whole-genome sequencing in a Chinese family with ARS and identified a heterozygous deletion of about 570 kb (termed LOH-1) in the intergenic sequence between paired-like homeodomain transcription factor 2 (PITX2) and family with sequence similarity 241 member A. Knockout of LOH-1 homologous sequences caused ARS phenotypes in mice. RNA-Seq and real-time quantitative PCR revealed a significant reduction in Pitx2 gene expression in LOH-1–/– mice, while forkhead box C1 expression remained unchanged. ChIP-Seq and bioinformatics analysis identified a potential enhancer region (LOH-E1) within LOH-1. Deletion of LOH-E1 led to a substantial downregulation of the PITX2 gene. Mechanistically, we found a sequence (hg38 chr4:111,399,594–111,399,691) that is on LOH-E1 could regulate PITX2 by binding to RAD21, a critical component of the cohesin complex. Knockdown of RAD21 resulted in reduced PITX2 expression. Collectively, our findings indicate that a potential enhancer sequence that is within LOH-1 may regulate PITX2 expression remotely through cohesin-mediated loop domains, leading to ARS when absent.


Introduction
In recent years, with the development of sequencing technology, the rate of genetic diagnosis of Mendelian diseases has substantially increased.However, the proportion of unsolved exomes was much higher than expected, and a significant subset of disorders may relate to noncoding regions of the genome (1).Although the functional relevance of most noncoding variants is not known, a number of diseases have been shown to be associated with noncoding variants (2).
One example of such a Mendelian disorder is Axenfeld-Rieger syndrome (ARS), a rare developmental anomaly with an incidence of approximately 1:50,000-100,000 in newborns (3).The clinical manifestations of ARS include mainly ocular symptoms: the presence of posterior embryotoxon (manifested by a prominent, anteriorly displaced Schwalbe's line near the posterior corneal limbus), iridocorneal adhesions, and iris hypoplasia, corectopia, and/or polycoria.Typically inherited in an autosomal dominant manner, 2 causative genes for ARS have been identified: paired-like homeodomain transcription factor 2 (PITX2) at 4q25 (4) and forkhead box C1 (FOXC1) at 6p25 (5).Phillips et al. also linked ARS to 13q14 (RIEG2, MIM: 601499) (6), but so far no pathogenic gene has been clearly identified at this locus.Additionally, mutations in CYP1B1, PRDM5, COL4A1, and PAX6 have been associated with ARS in a few reports (7)(8)(9)(10).Currently, only 40% of ARS cases are linked to mutations in PITX2 and FOXC1, leaving a significant portion of ARS cases with an unknown genetic cause (3).
Recent studies have uncovered that noncoding sequence variants may relate to Axenfeld-Rieger syndrome (ARS), a rare developmental anomaly with genetic heterogeneity.However, how these genomic regions are functionally and structurally associated with ARS is still unclear.In this study, we performed genome-wide linkage analysis and whole-genome sequencing in a Chinese family with ARS and identified a heterozygous deletion of about 570 kb (termed LOH-1) in the intergenic sequence between paired-like homeodomain transcription factor 2 (PITX2) and family with sequence similarity 241 member A. Knockout of LOH-1 homologous sequences caused ARS phenotypes in mice.RNA-Seq and real-time quantitative PCR revealed a significant reduction in Pitx2 gene expression in LOH-1 -/-mice, while forkhead box C1 expression remained unchanged.ChIP-Seq and bioinformatics analysis identified a potential enhancer region (LOH-E1) within LOH-1.Deletion of LOH-E1 led to a substantial downregulation of the PITX2 gene.Mechanistically, we found a sequence (hg38 chr4:111,399,594-111,399,691) that is on LOH-E1 could regulate PITX2 by binding to RAD21, a critical component of the cohesin complex.Knockdown of RAD21 resulted in reduced PITX2 expression.Collectively, our findings indicate that a potential enhancer sequence that is within LOH-1 may regulate PITX2 expression remotely through cohesin-mediated loop domains, leading to ARS when absent.

Results
Clinical characteristics.In this family with ARS, household members underwent ocular examination (Figure 1A).Different degrees of corneal embryotoxon, iridocorneal adhesion, iris hypoplasia, and iris corectopia were present in 14 eyes of all affected patients (Figure 1B).Iris polycoria was manifest in patients II:6, III:3, and III:7 (Figure 1B).Glaucoma was found in patients II:5, II:6, II:8, III:3, and III-5 but not obvious in patients III:6 and III:7, who were at the age of 8 and 12, respectively, at the time of examination.Teeth and umbilicus were examined by the physicians, and no abnormalities were demonstrated in this family (Figure 1B).The electrocardiographic results in patients II:5 and II:6 did not suggest significant cardiac abnormalities such as atrial fibrillation.The clinical characteristics were summarized in Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.177032DS1.
LOH-1 deletion in the intergenic sequence upstream of PITX2 gene cosegregates with ARS.First, we considered the possibility that variants in the known ARS genes PITX2 and FOXC1 may be responsible for the observed clinical manifestations.Exonic regions of PITX2 and FOXC1 were examined by Sanger sequencing in proband III:3 from the ARS family and no disease-related variant was found.Then, all collected samples within the ARS family underwent linkage analysis.Parametric multipoint linkage analysis of the ARS family revealed a highly linked locus of approximately 21.2 Mb on 4q22.1-q26, with a maximum logarithm of odds (LOD) score of 3.258, surrounding the marker rs1354680 (Figure 1C and Supplemental Table 2).Interestingly, this linkage region contained the known ARS pathogenic gene PITX2.
Furthermore, 1 patient (II:6) was selected for WGS.Copy number variation analysis showed no large deletions/duplications in the PITX2 gene-coding regions and intron regions, while a heterozygous deletion of about 570 kb was detected approximately 200 kb upstream of PITX2 (Supplemental Figure 1A), situated between PITX2 and family with sequence similarity 241 member A (FAM241A) (we named the deletion LOH-1).
We used Integrative Genomics Viewer (IGV) software to visualize the WGS reads and to search for sequencing abundance anomalies near the ends of the region from 110,868,844 to 111,438,844 on chromosome 4 (hg38).The sequence information of the reads at the sequencing abundance anomaly was obtained.The precise location of the LOH-1 deletion was inferred to be between 110,869,880 and 111,437,100 (Supplemental Figure 1B).Sanger sequencing verified the putative LOH-1 location and revealed a complete cosegregation of LOH-1 with the ARS phenotype (Figure 1, A and D).
LOH-1-knockout mice have ARS phenotypes.To investigate the relationship between LOH-1 deletion and ARS, we knocked out the homologous sequence of LOH-1 (mm10 chr3:128,507,082-128,912,082) in mice by a CRISPR/Cas9-targeted strategy.Mendelian ratio analysis of heterozygous mating outcomes revealed a nearly 50% prebirth lethality rate in LOH-1 -/-(KO) mice while LOH-1 +/-(HET) mice appeared unaffected.Examination of prebirth lethal LOH-1 -/-embryos at E12.5 revealed ventral body wall defect and evisceration (Supplemental Figure 2A), which is consistent with previous reports (14,15).Observation of dorsal images of mice at 3 weeks old revealed that LOH-1 -/-mice were reduced in size compared with wildtype (WT) mice while LOH-1 +/-mice did not exhibit obvious abnormality (Figure 2A).By examining the body weight of mice from 3 to 8 weeks after birth, we found a marked weight reduction in LOH-1 -/-mice compared with WT mice, with no difference observed in LOH-1 +/-mice (Figure 2B).Slit lamp examination and anterior segment optical coherence tomography (OCT) showed clear corneas and normal deep anterior chambers in both LOH-1 +/-and WT mice.In contrast, LOH-1 -/-mice exhibited opacified corneas and lacked anterior chambers (Figure 2C).Histopathological findings verified normal corneas and angle structures in the WT and LOH-1 +/-mice.Thickened corneas with disorganized epithelial and stromal cells, missing anterior chambers, and closed angle structures were detected in the LOH-1 -/-mice (Figure 2D).By measuring the corneal thickness and corneal epithelial thickness with the histopathological slides of the mice, we identified a significant increase in whole corneal thickness but a decrease in corneal epithelial thickness in LOH-1 -/-mice, compared with WT, while LOH-1 +/-mice showed no such changes (Figure 2E).Atomic force microscopy showed increased corneal roughness in LOH-1 -/-mice compared with WT mice (Supplemental Figure 2, B and C).However, fundus retinal images and histopathologic results revealed no retinal abnormalities in LOH-1 -/-, LOH-1 +/-, and WT mice (Supplemental Figure 2, D and E).
LOH-1 deletion results in dramatically decreased expression of Pitx2 gene.To further explore the molecular mechanism of LOH-1 -/-mice, we analyzed the transcriptional profiles of 8-week-old LOH-1 -/-and WT mouse eyes using RNA-sequencing (RNA-Seq) analysis.As shown in Figure 2F, 97 genes were downregulated and 77 genes were upregulated (Supplemental Table 3).Among them, the expression of the ARS causative gene Pitx2 was significantly reduced.However, the expression levels of another causative gene, Foxc1, and genes located on either side of Pitx2, i.e., Enpep and Fam241a, were not significantly changed.

R E S E A R C H A R T I C L E
JCI Insight 2024;9(9):e177032 https://doi.org/10.1172/jci.insight.177032RT-qPCR analysis further corroborated these RNA-Seq findings.The expression of the 3 isoforms of Pitx2, Pitx2a, Pitx2b, and Pitx2c, showed marked reductions in the eyes of LOH-1 -/-mice, with moderate differences observed in LOH-1 +/-mice (Figure 2G).In addition, we examined the expression of Pitx2 by RT-qP-CR in the eye, heart, kidney, stomach, and skeletal muscle of WT and LOH-1 -/-embryos at E18.5.The results revealed that Pitx2 was significantly decreased in these tissues of LOH-1 -/-embryos (Supplemental Figure 2F).These findings indicate that the homozygous knockout of LOH-1 in mice leads to a decrease in Pitx2 gene expression.
Identification of a potential enhancer region in LOH-1.Given the aforementioned findings, we postulated the presence of cis-regulatory elements such as enhancers within LOH-1 that modulate the expression of Pitx2.To identify these elements in LOH-1, we analyzed the enhancer-associated histone modification (monomethylation at histone H3 lysine 4, H3K4Me1; and acetylation of histone H3 at lysine 27, H3K27Ac) ChIP-Seq data of 15-week-old human embryonic sclera.Numerous prominent peaks of H3K4Me1 and H3K27Ac signals were observed within this region (Figure 3A; Supplemental Figure 4, A  and B; and Supplemental Tables 4 and 5).Meanwhile, we investigated DNase I hypersensitivity in human embryonic eye, retina, and heart tissues and assessed the H3K4Me1 in human embryonic heart.We also reanalyzed data sets from the UCSC Genome Browser (16), sourced from the Encyclopedia of DNA Elements (ENCODE) (17,18), to evaluate the enrichment of H3K4Me1, trimethylation at histone H3 lysine 4 (H3K4Me3), and H3K27Ac within the LOH-1 locus across 7 cell lines.The 100 vertebrates Basewise Conservation by PhyloP and the public H3K27Ac ChIP-Seq data (human retina and heart) from the Cistrome Data Browser (19,20) were also used for reference.The results from the above data suggest that there may be multiple enhancer regions in LOH-1.
Subsequently, we selected a potential enhancer region named LOH-E1 (hg38 chr4:111,397,892-111,402,926) in LOH-1, which contains the specific epigenetic signals.The pronounced enrichment of H3K4Me1 and H3K27Ac signals in the LOH-E1 indicated potential enhancer activity (Figure 3B).We also examined the epigenetic signals of mouse homologous regions for LOH-1 and LOH-E1 by the UCSC genome browser from the ENCODE data sets (Supplemental Figure 3A).The results supported our speculation.
Additionally, according to the GeneHancer database (21), LOH-E1 appears to interact with the PITX2 gene (Supplemental Figure 3B), suggesting that PITX2 could be a potential target gene for the LOH-E1 enhancer region.
Deletion of LOH-E1 markedly reduces PITX2 expression.To investigate the mechanism of action of the LOH-E1, we employed the CRISPR/Cas9 system to specifically knock out the LOH-E1 region in the human embryonic kidney cell line (HEK293) (Supplemental Figure 5A).Sequencing verified the generation of homozygous LOH-E1 deletion clones (HEK293-KO).RT-qPCR analysis revealed that the expression of PITX2A and PITX2B was strongly downregulated in the HEK293-KO group compared with the WT cells.However, the expression of PITX2C remained unchanged (Figure 4B).Western blot also verified that the expression of PITX2B was downregulated in the HEK293-KO group (Figure 4C).
Deletion of LOH-E1 suppresses cell proliferation.Observations on LOH-E1-depleted cells showed no significant morphological changes (Figure 4A).Cell counting kit-8 (CCK-8) assay indicated a significant reduction in the proliferation rate of the HEK293-KO group compared with WT cells (Figure 4D).Cell cycle analysis revealed a notable decrease in the G 1 phase and an increase in the S phase for LOH-E1 -/-cells compared with control cells (Figure 4E).In the scratch assay, 2-well culture slices produced uniform gaps in the confluent monolayer, and wound healing was imaged at various times (Supplemental Figure 5B).The results showed a slower wound-healing rate in the KO group compared with the control group (Supplemental Figure 5C).In the apoptosis assay, early apoptosis was significantly higher in the HEK293-KO group compared with the control cells, with no significant changes in late apoptosis (Supplemental Figure 5D).
P2, which is on LOH-E1, can regulate PITX2 expression by binding to RAD21.We sought to elucidate the molecular mechanism by which LOH-E1 specifically modulates PITX2 expression.Enhancers can recruit transcription factors (TFs) and coactivators to alter chromatin spatial structure and improve transcription of target genes.To identify potential TFs, we analyzed the LOH-E1 and PITX2 core promoter region that are bound to TFs, focusing on regions enriched in H3K4Me1, H3K4Me3, and H3K27Ac.This analysis was conducted using the UCSC Genome Browser, which is sourced from ChIP-Seq of the ENCODE data sets, following the method of Deng et al. (22).Among the 340 TFs in the ENCODE data set, we identified 12 TFs associated with LOH-E1 and overlapped with regions enriched in H3K4Me1 and H3K27Ac (Supplemental Figure 6A).Meanwhile, 26 TFs were found to bind to the core promoter region of PITX2, JCI Insight 2024;9(9):e177032 https://doi.org/10.1172/jci.insight.177032overlapping with regions enriched in both H3K4Me3 and H3K27Ac (Supplemental Figure 6B).The Venn diagram revealed 2 overlapping TFs, RAD21 and CREB1(Figure 5A).
To identify the specific TFs recruited by the LOH-E1 enhancer region, we first analyzed DNase I hypersensitivity (human embryonic eye, retina, and heart) and the H3K4Me1 data (HEK293 and human heart) from the ENCODE data set as well as the human eye-and heart-related single-cell ATAC-Seq data from the CATlas database (23).This analysis identified 2 core regions within LOH-E1 with significant epigenetic signals (Supplemental Figure 7), located at hg38 chr4:111,398,770-111,400,269 (1,500 bp) and hg38 chr4:111,400,720-111,401,749 (1,030 bp).Based on these regions, we synthesized a double-stranded DNA probe labeled with biotin at the 5′ end, which was purified by incubation with nuclear proteins from HEK293 cells for liquid chromatography mass spectrometry (LC-MS) detection (Supplemental Figure 8A).Among the results of the enriched differential proteins, we found that RAD21, which is a critical component of the cohesin complex, may be the main transcription cofactor affected by LOH-E1 (Figure 5B and Supplemental Table 6).

R E S E A R C H A R T I C L E
JCI Insight 2024;9(9):e177032 https://doi.org/10.1172/jci.insight.177032 As LOH-E1 correlates with the expression of PITX2, to further investigate the relationship between RAD21 and PITX2, we performed a correlation analysis using the Gene Expression Profiling Interactive Analysis 2 database (24).The results showed a positive correlation between these 2 genes in heart and muscle (Supplemental Figure 8B), which are both gene-specific tissues in the Genotype-Tissue Expression database (25).We also visualized high-throughput chromosome conformation capture and circular chromosomal conformation capture data in human adrenal gland online in the 3D Genome Browser (26), and the results suggest a possible long-range interaction between LOH-E1 and PITX2 (Supplemental Figure 8, C and D).
Considering the potential role of RAD21 in the LOH-E1-mediated regulation of PITX2, we hypothesized that downregulation of RAD21 would reduce the expression of the target gene PITX2.We designed small interfering RNAs (siRNAs) targeting RAD21 (Figure 5C).Subsequently, we selected RAD21_si2 and RAD21_si3 to detect the expression of target genes PITX2A, -B, and -C in HEK293 cells and found that the mRNA expression of PITX2A, -B, and -C was significantly reduced (Figure 5D).
A recent study suggests that RAD21 N-terminal tail binds DNA to guide it through the kleisin gate and finally through entry into the cohesin ring (27).To scan the binding sequence of RAD21 in the LOH-E1 core enhancer region, we used the TF binding site prediction function in AnimalTFDB 3.0 database (28,29) and finally selected a 41 bp sequence named JJ (hg38 chr4:111,399,619-111,399,659) (Supplemental Table 7).According to sequence comparison, 2 regions in JJ sequence can match the RAD21 motif (Figure 5E), so we hypothesized that RAD21 could bind to JJ sequence.To verify this speculation, we performed ChIP-qPCR on the JJ sequence-containing region P2 (hg38 chr4:111,399,594-111,399,691) and surrounding regions P1 (hg38 chr4:111,398,648-111,398,756) and P3 (hg38 chr4:111,400,348-111,400,409).After that, we found a stronger enrichment of RAD21 binding to P2 compared with P1 and P3 in the HEK293 cell line, when the RAD21 antibody was compared with the control antibody immunoglobulin G (IgG) (Figure 5F).This evidence suggested a direct interaction between RAD21 and P2.

Discussion
In this study, we identified a heterozygous deletion, LOH-1, in the upstream intergenic region of the PITX2 gene by genome-wide linkage analysis and WGS, which cosegregated with ARS in a Chinese ARS family.Knockout of LOH-1 homologous sequence in mice revealed that LOH-1 -/-mice developed ARS-associated phenotypes and that Pitx2 gene expression level was substantially decreased.These results suggest that deletion of noncoding intergenic sequence LOH-1 can induce ARS in both humans and mice.
Several studies have previously highlighted the presence of functional elements in the intergenic noncoding sequences upstream of the PITX2 gene, which are linked to genetic diseases.However, the pathogenic mechanisms involved are still unclear.Combined with our genetic results, we hypothesized that intergenic sequences upstream of PITX2 should play a regulatory role in PITX2 gene expression.We then performed a series of bioinformatics analyses of epigenetic data in public databases, along with ChIP-Seq to target a potential enhancer region LOH-E1 in LOH-1.In the HEK293 cell line, cell proliferation, cell cycle, apoptosis, and migration were all affected by LOH-E1 deletion, and the expression levels of PITX2 were downregulated, verifying that LOH-E1 may be a candidate enhancer region that regulates PITX2 expression.We summarize the deletions for the 4q25 locus that have been reported and the deletions that we found (Supplemental Figure 9).Among them, ARS_Volkmann, Aniridia_Ansari, and ARS_Protas are explicitly shown to overlap with LOH-E1.Through data analysis by relevant experiments and public databases, RAD21 emerged as a potential protein binding to P2 sequence, which is on LOH-E1.As the core subunit of cohesin, RAD21 is the only physical linkage between the SMC1/SMC3 heterodimer and the STAG subunit, which regulates the binding or dissociation of cohesin from chromatin and is involved in regulating gene transcriptional expression (30,31).Whether DNA loading is successful or results in loop extrusion might be determined by a conserved RAD21 N-terminal tail that guides the DNA through the kleisin gate (27).The above evidence demonstrated that P2, which is on LOH-E1, binds to RAD21 to regulate the expression of PITX2.
However, we found that the expression of the 3 isoforms of PITX2 was not completely consistent in LOH-E1-KO cells and LOH-1-KO mice, as reflected in PITX2C.Among the 3 isoforms of Pitx2, the Pitx2a and -b splice variants share the same promoter and are expressed bilaterally in some tissues, whereas Pitx2c is transcribed from a separate promoter and is expressed asymmetrically (32).We hypothesized that LOH-E1 may act only on the PITX2A and -B promoter and not the PITX2C promoter.LOH-E1-KO mice were constructed subsequently to further clarify the mechanism of LOH-E1 with respect to Pitx2.Epigenetic data analysis further suggested that LOH-E1 might not be the only PITX2 enhancer within LOH-1, hinting at the existence of other enhancer active regions that warrant further exploration.
PITX2 plays a crucial role early in embryonic development to regulate the left-right asymmetric development of internal organs such as the gut, heart, liver, and stomach (e.g., gut rotation) (33).Human and mouse embryonic development have different requirements for the dose of PITX2/Pitx2, with human development being more sensitive to the appropriate dose of PITX2 (11).This may explain the fact that LOH-1 heterozygous deletion causes ARS in humans.In previous studies, Pitx2-KO homozygous mice died prematurely because of various developmental defects, while heterozygous mice were usually described as normal (14,15,34,35).In our study, the LOH-1 +/-mice exhibited diminished expression of Pitx2 compared with WT but were able to reproduce and survive normally and did not have a significant ARS phenotype.Moreover, the LOH-1 -/-mice were not completely lethal, with about 50% surviving to adulthood.Surviving KO mice displayed pronounced reduced expression of Pitx2 and a distinct ARS phenotype.Therefore, the LOH-1-KO mice present an ideal model for delving deeper into ARS mechanisms.
In summary, our study unveiled an enhancer region P2, which is on LOH-1, regulating the expression of PITX2 by binding to RAD21, and elucidated its significance in ARS.In the absence of LOH-1, RAD21 fails to successfully guide DNA into the cohesin ring, leading to the pathogenicity of ARS.This work helps improve the understanding of intergenic sequence variants, enhances the diagnosis of related genetic diseases, and offers potential avenues for the prevention and treatment of ARS.

Figure 1 .
Figure 1.Clinical features and genetic studies of the family with ARS.(A) Pedigree of a family with ARS.The arrow in the pedigree indicates the proband."□" and "○" symbols represent healthy male and female individuals; "■" and "•" elements stand for male and female patients.Samples selected for whole-genome sequencing (WGS) were marked with "*"."+" stands for WT allele, and "-" refers to LOH-1 deletion allele.(B) Clinical manifestations of patients III:3, III:6, and III:7 from the ARS family.Top left: An anterior segment photograph of the right eye (III:3) shows posterior embryotoxon (arrow) at the temporal corneal limbus and corectopia toward the temporal side.Top center: Pseudopolycoria and atrophic iris in the left eye (III:3).Top right: Gonioscopic photographs demonstrating iridocorneal attachment (arrow) and corectopia in the left eye (III:6).Bottom left: Extensions of the peripheral iris to Schwalbe's line are seen on gonioscopy in the right eye (III:6).Bottom center: Normal teeth are shown for affected individual III:7.Bottom right: Normal umbilicus is shown for affected individual III:7.(C) Genome-wide multipoint linkage analysis of ARS family.(D) Sanger sequencing identifies the precise location of the LOH-1 deletion as hg38 chr4:110,869,880-111,437,100.

Figure 2 .
Figure 2. Phenotypic observations and expression alterations of ARS-related genes in LOH-1-KO mice.(A) Size of male LOH-1 -/-(KO), LOH-1 +/-(HET), and wild-type (WT) mice at 3 weeks of age.The scale bar represents 1 cm.(B) Body weight of male (left) and female (right) KO mice was compared with HET and WT mice from 3 to 8 weeks of age.n = 23 for each group.Data were analyzed using 2-way ANOVA.(C) Slit lamp and optical coherence tomography examinations of the eyes from different groups showed normal anterior segment structure in WT and HET groups while opacified cornea and missing anterior chamber were seen in the KO group.(D) H&E staining and histopathologic findings demonstrated normal eyeball structures, wide angles, and regularly arranged corneal layers in WT and HET groups.However, a disorganized anterior segment and closed angle as well as thick corneal stromal layer were seen in the KO group.Scale bars represent 500 μm in eye, 100 μm in anterior chamber angle, and 20 μm in cornea.(E) The whole corneal thickness and corneal epithelial thickness were compared among the groups.n = 3 for each group.Data were analyzed using 1-way ANOVA.(F) Volcano plot displaying the differentially expressed genes (KO vs. WT) with 97 downregulated genes and 77 upregulated genes with |log 2 (fold-change)| ≥ 1 and padj ≤ 0.05.Among them, the expression of Pitx2 was significantly reduced.However, the expression levels of Foxc1, Enpep, and Fam241a were not significantly changed.(G) Real-time quantitative PCR (RT-qPCR) detection of relative Pitx2a, Pitx2b, and Pitx2c mRNA expression in the WT, HET, and KO mice.n = 3 for each group.Data were analyzed using 1-way ANOVA.All data are represented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 3 .
Figure 3. Identification of the enhancer LOH-E1 within LOH-1 region.(A) Overview of the DNase-Seq of retina, eye, and heart; the H3K4Me1 ChIP-Seq of heart; the H3K4Me1, H3K4Me3, and H3K27Ac marks on 7 cell lines; the 100 vertebrates Basewise Conservation; and the H3K27Ac ChIP-Seq of heart and retina in LOH-1 region of human supported by the UCSC Genome Browser.H3K4Me1 and H3K27Ac ChIP-Seq results from human embryonic 15-week-old sclera are indicated by the red box.LOH-1 is indicated by the green rectangle.PITX2 gene is shown to be downstream of the LOH-1.(B) Determination of the LOH-E1 enhancer location based on markers of DNase I, H3K4Me1, and H3K27Ac.LOH-E1 is indicated by the yellow rectangle.H3K4Me1 and H3K27Ac ChIP-Seq results from human embryonic 15-week-old sclera are indicated by the red box.

Figure 5 .
Figure 5. P2, which is within LOH-E1, could bind RAD21 to regulate the expression of PITX2.(A) Venn diagram presents transcription factors (TFs) that bind to both the LOH-E1 enhancer region and the PITX2 core promoter region.(B) Peptide spectrum match (PSM) of peptide from RAD21 protein by liquid chromatography mass spectrometry (LC-MS) in the DNA pulldown.(C) RT-qPCR detection of RAD21 expression following its downregulation in HEK293 cells through siRNAs.For detection of RAD21 expression level, RAD21_small interfering RNA 2 (RAD21_si2) or RAD21_si3 compared with small interfering RNA negative control (siNC), respectively.n = 3 for each group.Data were analyzed using 1-way ANOVA.(D) RT-qPCR detection of PITX2A, PITX2B, and PITX2C expression after the downregulation of RAD21 in HEK293 cells through RAD21_si2 and RAD21_si3.n = 3 for each group.Data were analyzed using 1-way ANOVA.(E) Identification of the potential RAD21 binding sequence within the LOH-E1 region; the position of the red line can match the RAD21 motif.(F) ChIP-qPCR assay for the binding of RAD21 to the P2 and surrounding regions.Quantitative PCR detection for the enrichment of the P1, P2, and P3 regions within LOH-E1 upon anti-RAD21 ChIP and IgG in the HEK293 cell line.P1 and P3 are the upstream and downstream random regions of P2.JJ is indicated by the red rectangle.n = 3 for each group.All data are represented as mean ± SEM.