A human colorectal cancer (CRC) cell line, HCT116, was purchased from the Korean Cell Line Bank and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in a humidified atmosphere with 5% CO₂ at 37°C. BIX01294 (ab141407) was purchased from Abcam (Cambridge, United Kingdom).

For siRNA transfection, HCT116 cells were seeded in plates and incubated for 24 hours. The targeting or control siRNAs (Bioneer, Daejeon, Korea) were transfected into HCT116 cells at 100 nM using RNAiMax (Invitrogen, Waltham, MA, USA) for 48 hours. The sequences of siRNAs used were as follows: siCont (5' AUGAACGUGAAUUGCUCAATT 3', 5' UUGAGCAAUUCACGUUCAUTT 3'), siEHMT2 (5' GCAAAUAUUUCACCUGCCATT 3', 5' UGGCAGGUGAAAUAUUUGCTT 3'), siCDKN1A (5' CUGUACUGUUCUGUGUCUU 3', 5' AAGACACAGAACAGUACAG 3').

A Cell Counting Kit-8 (CCK-8) (E-CK-A361; Elabscience, Houston, TX, USA) was used to conduct the cell viability assays. HCT116 cells were seeded in 6-well plates at 2×10⁵ cells/well and incubated for 24 hours. After siRNA transfection for 48 hours, the CCK-8 solution and RPMI-1640 medium with 10% FBS were added to each well, and the plates were incubated under 5% CO₂ at 37°C for 5 minutes. The absorbance was assessed using a microplate reader at 450 nm. For crystal violet staining, the cells were fixed with methanol for 5 minutes and stained with 0.1% crystal violet after siRNA transfection for 48 hours.

Total RNA was isolated from HCT116 cells using a Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RNA aliquots of 1 µg were then reverse-transcribed using PrimeScript RT Master Mix (RR036A; Takara, Kusatsu, Japan) according to standard protocols. Quantitative real-time polymerase chain reaction (RT-PCR) was performed using the AriaMx Real-Time PCR System (Agilent Technologies, Santa Clara, CA, USA) following the manufacturer’s instructions. Quantitative RT- PCR was performed on cDNA samples using a SensiFAST SYBR Lo-ROX Kit (Meridian Bioscience Inc., Cincinnati, OH, USA), and the signal was detected by an AriaMx Real-Time PCR System (Agilent Technologies). The fluorescence threshold value was calculated using Agilent Aria 1.6 software. The PCR primers used were as follows: EHMT2 (forward, 5' GAGAACATCTGCCTGCACTG 3' and reverse, 5' GTTGACAGCATGGAGGTCAC 3'), CDKN1A (forward, 5' CTTTGTCACCGAGACACCAC 3' and reverse, 5' CAGGTCCACATGGTCTTCCT 3'), CDH1 (forward, 5' CGAGAGCTACACGTTCACGG 3' and reverse, 5' GGGTGTCGAGGGAAAAATAGG 3'), ACTB (forward, 5' ACTCTTCCAGCCTTCCTTCC 3' and reverse, 5' CAATGCCAGGGTACATGGTG 3').

ChIP was performed with a Simple ChIP Plus Sonication Chromatin IP Kit (#56383S; Cell Signaling Technology, Danvers, MA, USA) following the manufacturers’ instructions. HCT116 cells transfected with siCont and siEHMT2 for 48 hours were crosslinked with 1% formaldehyde (Sigma-Aldrich, St. Louis, MO, USA) for 10 minutes at room temperature and quenched with 1X glycine for 5 minutes at room temperature. Then, the cells were washed with cold 1X PBS (containing 1X protease inhibitor cocktail), and the cells were lysed in 1X cell lysis buffer (containing a 1X protease inhibitor cocktail). Then, after nuclear extraction, the chromatin solution was sonicated using a Bioruptor Pico sonication device (B01060010; Diagenode, Danville, NJ, USA) with 10 cycles of 30 seconds on and 30 seconds off to obtain 200 to 1,000 bp chromatin fragments. Sheared chromatin (approximately 5 to 10 µg) was incubated with 2 μg of anti-H3K9me2 ChIP grade antibody (ab1220; Abcam) at 4°C overnight. Then, complexes were incubated with 30 μL of ChIP-Grade Protein G Magnetic Beads for 2 hours at 4°C. Then, the complexes were washed after each step, incubated with ChIP elution buffer for 30 minutes at 65°C and incubated with proteinase K for 2 hours at 65°C. After DNA purification using spin columns, the samples were analyzed by quantitative RT- PCR using CDKN1A primers. The primers used for the ChIP assay were as follows: primer 1 Primer 1 (P1) forward, 5’-TCCACCTTTCACCATTCCCC-3’ and reverse, 5’-GTTGGGACATGTTCCTGACG-3’; Primer 2 (P2) forward, 5’-CACGCCCGGCCAGTATATA-3’ and reverse, 5’-C TCCCGAGTAGCTGGGATTAC-3’; Primer 3 (P3) forward, 5’-CTCTTCGAGGCCAGCATTG-3’ and reverse, 5’-GATGTACATGCTCCCGCAGA-3’.

To perform spheroid culture of HCT116 cells, ultralow attachment (ULA) microplates were used (7007; Corning, Corning, NY, USA). After siRNA transfection for 48 hours, cells were seeded into plates at concentrations of 5×10⁴ cells per well. The spheroids were cultured for 2 days and observed under a microscope (CKX53; Olympus Corporation, Tokyo, Japan). To check the effect of BIX01294, HCT116 cells were seeded into plates at concentrations of 5×10⁴ cells per well and incubated for 24 hours. Subsequently, cells were treated with 20 µM BIX01294 and cultured for 2 days. Then, the spheroids were observed under a microscope (CKX53; Olympus Corporation).

Using a TruSeq RNA Sample Preparation Kit V2, purification and library construction were conducted with total RNA, and Illumina HiSeq 2500 machines (Illumina, San Diego, CA, USA) were used for sequencing with a read length of 2×10² bases. FastQC ver. 0.11.4 was used to assess the quality of the paired-end reads. Cutadapt version 1.15 and Sickle version 1.33 were used to filter low-quality reads and adaptors. Cufflinks version 2.2.1 was used to calculate fragments per kilobase of transcripts per million mapped reads values. Cuffdiff was used to select differentially expressed genes (DEGs). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using DAVID ver. 6.8.

In The Cancer Genome Atlas (TCGA) portal, the expression of EHMT2 was significantly increased in colon cancer (n=286) compared to normal tissues (n=41) (Fig. 1A). To assess the relationship between EHMT2 and colon cancer, we designed EHMT2-specific siRNA (siEHMT2) and negative control siRNA (siCont). After treatment of HCT116 cells with siEHMT2, we clearly found growth suppression in CCK-8 and crystal violet analyses (Fig. 1B and 1C). Thus, we suggest that the overexpression of EHMT2 is associated with colon cancer growth. Next, to identify the functions of EHMT2 in colon cancer, we performed total RNA-seq analysis after treatment with siEHMT2 and siCont. Using 1505 DEGs, we performed GO and KEGG pathway enrichment analyses with DAVID ver. 6.8. Fig. 1D shows that the GO terms and KEGG pathways were enriched in apoptosis and cell cycle. Thus, we hypothesized that EHMT2 expression is closely related to colon cancer proliferation.

To identify the direct target of EHMT2 for colon cancer growth, we focused on genes upregulated by EHMT2 knockdown, because EHMT2 was previously shown to be involved in the downregulation of gene expression via the formation of the heterochromatin region by H3K9 methylation [11]. We finally selected the CDKN1A gene for further analysis based on the RNA-seq results (Fig. 2A). CDKN1A is a cell cycle regulator that suppresses the cell cycle [12]. To validate CDKN1A upregulation by EHMT2 knockdown, we performed quantitative RT-PCR analysis and found induction of CDKN1A expression by EHMT2 knockdown compared to siCont (Fig. 2B). Next, we performed a ChIP assay to validate CDKN1A as a direct target of EHMT2 using an anti-H3K9 dimethylation antibody. As shown in Fig. 2C, we designed ChIP primers on the promoter regions of CDKN1A and observed that H3K9 dimethylation was significantly decreased by EHMT2 knockdown, implying that CDKN1A is a direct target of EHMT2 in colon cancer. In the TCGA portal, CDKN1A expression was dramatically reduced in colon cancer tissues, and the expression patterns of EHMT2 and CDKN1A showed a reverse correlation (Fig. 2D and 2E). Together, overexpression of EHMT2 directly suppressed CDKN1A expression, thereby inducing colon cancer growth.

Three-dimensional culture models, such as cell spheroids and organoids, show more similar biological responses to in vivo models than 2-dimensional (2D) culture systems [13-15]. Thus, to confirm EHMT2 regulation in more detail, we constructed a 3D spheroid model with HCT116 cells. After treatment with siEHMT2, we cultured HCT116 cells on a ULA plate to construct cell spheroids. As shown in Fig. 3A, although the cell spheroids in the siCont group showed more strong cell aggregation, the spheroids in the siEHMT2 treatment group showed loose cell aggregation over 2 days, implying that EHMT2 knockdown clearly affected cell junctions and cell death. In the quantitative RT-PCR analysis with a 3D spheroid model, the expression level of CDKN1A was increased by siEHMT2 treatment, as shown in the 2D culture model. Moreover, the expression of CDH1, which is an EMT marker, clearly decreased in the siEHMT2 spheroid group (Fig. 3B), suggesting that EHMT2 knockdown suppressed cell aggregation in the ULA plate via suppression of cell growth.

Next, we again validated CDKN1A regulation by EHMT2 in a 3D spheroid model. After cotreatment with siEHMT2 and siCDKN1A, we assessed cell aggregation and CDH1 expression levels in spheroid models. As shown in Fig. 3C, the aggregation of spheroids was loosened by EHMT2 knockdown compared to transfection with siCont; however, cotreatment with siEHMT2 and siCDKN1A recovered the loosened spheroid phenotype induced by siEHMT2. Additionally, the quantitative RT-PCR analysis showed that the reduction in CDH1 expression by EHMT2 knockdown was recovered by cotreatment with siEHMT2 and siCDKN1A (Fig. 3D). Thus, we suggest that CDKN1A regulation by EHMT2 is associated with colon cancer growth.

BIX01294 is an EHMT2-specific inhibitor that suppresses H3K9 methylation [4]. As shown by EHMT2 knockdown, BIX01294 inhibited cell aggregation in ULA plates compared to DMSO treatment as a negative control (Fig. 4A). Moreover, after BIX01294 treatment of spheroids, CDKN1A expression was significantly induced, and CDH1 expression decreased, suggesting that the inhibition of EHMT2 enzyme activity also suppressed colon cancer growth (Fig. 4B). Thus, an EHMT2-specific inhibitor should be developed for colon cancer treatment.