Hepatitis B Virus-KMT2B Integration Drives Hepatic Oncogenic Processes in a Human Gene-edited Induced Pluripotent Stem Cells-derived Model

Abstract

Background & Aims

Hepatitis B virus (HBV)-DNA integration into the host genome contributes to hepatocellular carcinoma (HCC) development. KMT2B is the second most frequent locus of HBV-DNA integration in HCC; however, its role and function remain unclear. We aimed to clarify the impact of HBV-KMT2B integration in HCC development using a human genome-edited induced pluripotent stem cell (iPSCs) model.

Methods

Based on the genetic information on HBV-KMT2B integration in HCC, we determined its complete DNA sequence and transcript variants. To exclude the effect of other oncogenic mutations, we reproduced HBV integration in healthy donor iPSCs with an intact genome and analyzed its effects using iPSC-derived hepatic progenitor cells (HPCs) and hepatocytes (iPS-Heps).

Results

The reproduced HBV-KMT2B integration significantly upregulated the proliferation of hepatic cells. Comprehensive transcriptional and epigenetic analyses revealed enhanced expression of cell cycle-related genes in hepatic cells with HBV-KMT2B integration based on perturbation of histone 3 lysine 4 tri-methylation (H3K4me3), mimicking that in the original HCC sample. Long-read RNA-sequence detected the common KMT2B transcript variants in the HCC sample and HPCs. Overexpression of the truncated variant significantly enhanced proliferation of hepatic cells, whereas HBV-KMT2B fusion transcripts did not enhance proliferation. HBV-KMT2B-integrated HPCs exhibited replication stress and DNA damage, indicating that our model initiated the process of hepatocarcinogenesis due to abnormally promoted KMT2B function.

Conclusions

Our disease model using genetically engineered iPSCs provides the first insight into both the KMT2B function in HCC development and the oncogenic processes by HBV-KMT2B integration. We clarified the novel oncogenic mechanism in HBV-related HCC due to aberrant KMT2B function.

Graphical abstract

Summary.

Our disease model revealed that hepatitis B virus (HBV) integration into KMT2B locus induces hyperproliferation of hepatic cells via perturbation of histone tri-methylation. KMT2B transcript variant induced by HBV integration drives replication stress and DNA damage in host hepatic cells.

Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death worldwide.¹ Hepatitis B virus (HBV) infection is one of the major risk factors for HCC, alongside hepatitis C virus infection, alcohol-induced liver damage, and metabolic liver disease. Recent studies on HCC using whole-genome and exome sequencing have revealed mutations in the driver genes, such as TERT, TP53, and CTNNB1.^2,3 Unlike colorectal cancer, which follows a clear multistep carcinogenic process, no such well-defined process is evident in HCC. In addition to these mutations, integrated viral genomes derived from HBV are identified and accumulate in particular gene loci, including TERT (including its promoter region), KMT2B (also known as MLL4), CCNA2, and CCNE1, in HBV-related HCC.^4,5

The second most frequent locus of HBV-DNA integration, KMT2B, codes a histone methyltransferase that forms a protein complex (the COMPASS complex) with Menin, WDR5, ASH2L, RBBP5, and DPY30.⁶ The KMT2B protein undergoes cleavage by taspase-1, resulting in 2 protein fragments, N-terminal and C-terminal fragments, that together form a functional heterodimeric complex.⁷ KMT2B-deficient mice exhibited embryonic lethality, and KMT2B knockdown in embryonic stem cells led to defective proliferation, increased apoptosis, and aberrant differentiation potential, suggesting that KMT2B function contributes to the promotion of cell proliferation.^8,9 HBV-KMT2B integrations in HCC preferentially occurred between exon 3 and exon 6.¹⁰ Compared with non-tumorous samples, an elevation in KMT2B expression has been observed in HCC, particularly when HBV integrates into the exons of the KMT2B locus.¹¹ HBV-KMT2B fusion transcripts were detected by quantitative polymerase chain reaction (qPCR)^11,12 or RNA sequencing (RNA-seq),⁴ suggesting the hypothesis that HBV-KMT2B integration functions as a dominant-negative form of KMT2B.⁴ Meanwhile, genomic insertions (including the KMT2B locus) of the adeno-associated virus 2 (AAV2) in HCC were recently identified.^13,14 Both AAV2 and HBV showed preferences for integration between exon 3 and exon 6 within the KMT2B locus. These similarities in integration sites strongly suggest a common and viral species-independent mechanism underlying the oncogenic process related to DNA integration into the KMT2B locus. However, the oncogenic molecular mechanism resulting from HBV-DNA integration into the KMT2B locus in HCC is yet to be elucidated.

Here, we report a novel oncogenic process initiated by HBV-KMT2B integration in HCC. We reproduced HBV-KMT2B integration in induced pluripotent stem cells (iPSCs) derived from a healthy donor with an intact genome by gene editing. Subsequently gene-edited HBV-integrated iPSCs differentiated into hepatic cells. This HBV-KMT2B integration model enabled us to observe the integration-specific oncogenic impact on hepatic cells without the effect of other tumor-related mutations. HBV-KMT2B integrated hepatic cells exhibited enhanced proliferation via alteration of H3K4me3 profiles by aberrant KMT2B from HBV-integrated allele, resulting in replication stress and DNA damage, as early oncogenic processes. Moreover, replication stress in KMT2B integrated HCCs was also validated using the other public clinical data sets. Our findings demonstrate that the HBV-KMT2B integrated iPSCs model unveiled a novel oncogenic process of HBV-related HCC due to HBV-DNA integration.

Results

Identification of Complete Sequences of HBV-KMT2B Integration in HBV-related HCC Samples

First, we examined HBV-human fusion sequence in our in-house HBV-related HCC database with available clinical samples suitable for further analysis. We identified 1 case of HBV-KMT2B integration that occurred in exon 3 of KMT2B, which is within the hot spot region for HBV-DNA integration (Figure 1A).^3,10 Then the complete genomic sequence of the HBV-integrated region was determined. The data revealed that 18 bp at exon 3 of KMT2B was deleted, and 2083 bp of the partial HBV genome was inserted (Figure 1B). To characterize the profile of KMT2B transcription, RNA-seq was performed on the tumor sample. The read coverages of KMT2B were elevated in the middle of the HBV integration region (Figure 1C), which was consistent with the previous report.¹¹ qPCR using 3′ terminal primer sets showed significantly elevated KMT2B expression in the tumor sample, validating the result of RNA-seq (Figure 1C and D). Furthermore, the upregulation of the KMT2B C-terminal fragment was observed in the tumor sample compared with the non-tumorous liver tissue (Figure 1E). Thus, we hypothesized that abnormal KMT2B transcription derived from an integrated allele contributes to the oncogenic mechanism.

Figure 1.

Identification of complete sequences of HBV-KMT2B integration in the tumor sample and its gene expression analyses of KMT2B. (A) Identified HBV integration into KMT2B locus in a HCC case of our HCC database. Dotted line shows HBV integration hot spot region of KMT2B. (B) Identification of complete HBV-integrated KMT2B sequences. A total of 18 bp of KMT2B (35721700-35721717) were deleted and a part of the HBV sequence (2934-1798) was integrated. (C) Coverage of HBV-KMT2B in RNA-seq of tumor tissues. PCR primers of KMT2B (5′) and KMT2B (3′) were designed to amplify the upstream and downstream sequences of KMT2B from HBV integration, respectively. (D) qPCR of KMT2B in tumor and non-tumorous tissues. Significant upregulation of KMT2B (3′) in tumor tissues by qPCR. (E) Schema of the KMT2B protein showing KMT2B domains, HBV-integrated regions, and the recognition region of the KMT2B antibody. KMT2B is cleaved by Taspase-1, and the 2 cleaved fragments associate via a FYRN and FYRC domain interaction. The recognition region of the KMT2B antibody located on C-terminal fragment, resulting detection on 80 kDa. Immunoblotting analyses of KMT2B(C) in tumor and non-tumorous tissues and significant increase of KMT2B(C) in tumor tissues by immunoblotting (right panel).

Generation of HBV-KMT2B Integrated iPSCs

To elucidate the effect of HBV-KMT2B integration, we utilized gene-edited iPSCs and their hepatic derivatives (Figure 2A). We hypothesized that specific phenotype induced by HBV-KMT2B integration could be observed in iPSCs with normal genome, as these iPSCs lack oncogenic mutations, somatic mutations, and chromosomal abnormalities, unlike cancer cell lines. We recapitulated HBV-KMT2B integration in human iPSCs (KMT2B-Int iPSCs) (Figure 2B and C). Additionally, to explore the possibility of dominant-negative effects by HBV-KMT2B integration, we engineered KMT2B homozygously knockout iPSCs (KMT2B-KO iPSCs) (Figure 3A and B). Concurrently, we generated KMT2B heterozygously knockout iPSCs (KMT2B-HT iPSCs) (Figure 3C), which were used as control cells in subsequent experiments due to the heterozygous characteristics of KMT2B-Int iPSCs (Figure 2C).

Figure 2.

Establishment of iPSC-derived HPCs that reproduce HBV-KMT2B integration in HCC. (A) Strategical schema of this study. HBV-KMT2B-integrated iPSCs (KMT2B-Int iPSCs), heterogeneously KMT2B-knockout iPSCs (KMT2B-HT iPSCs), and homozygously KMT2B-knockout iPSCs (KMT2B-KO iPSCs) were generated from genome-intact iPSCs by gene editing. These iPSCs were differentiated into the hepatic lineage, and iPSC-derived HPCs were established and then analyzed. (B) Scheme of HBV genome knock-in into KMT2B locus using CRISPR-Cas9 and Cre-loxP systems for establishing KMT2B-Int iPSCs. KMT2B-Int iPSCs were established using 2 different guide RNAs targeted to intron 3. (C) Schema of KMT2B locus in established 2 KMT2B-Int iPSC lines. HBV-KMT2B integration cassette was heterozygously inserted into host genome, and a single nucleotide is also inserted by NHEJ in the intact allele. (D) Immunoblotting analysis of KMT2B(C) in established iPSCs. (E) Immunostaining of HPCs. All established HPCs formed colonies on the feeder cells, expressing hepatic markers (AFP and ALB). Scale bars: 100 μm. (F, G) qPCR and immunoblotting analyses of KMT2B-WT HPCs, KMT2B-HT HPCs and KMT2B-Int HPCs. Significant enhancement of KMT2B in KMT2B-Int HPCs by qPCR and immunoblotting.

Figure 3.

Establishment of KMT2B-edited iPSCs, including KMT2B-HT and KO iPSCs. (A) Scheme of gene editing of KMT2B using CRISPR-Cas9 system for establishing KMT2B-HT iPSCs and KMT2B-KO iPSCs. Both iPSCs were established using 2 different guide RNAs targeted to exon1. (B) Representative schema of KMT2B locus in established 2 KMT2B-HT iPSC lines. (C) Representative schema of KMT2B locus in established 2 KMT2B-KO iPSC lines.

First, we analyzed the characteristics of established gene-edited iPSCs in an undifferentiated state. The KMT2B production was completely lost in KMT2B-KO iPSCs, confirming successful gene editing (Figure 2D). The production in KMT2B-Int and KMT2B-HT iPSCs was slightly decreased compared with that in KMT2B-WT iPSCs, consistent with the analysis of KMT2B mRNA expression (Figure 4A). Expression of pluripotent markers confirmed that the pluripotency of all iPSC lines was maintained after genome-editing manipulations (Figure 4A and B).

Figure 4.

Analysis of pluripotency in established iPSCs. (A) Quantitative PCR analyses of KMT2B (5′), KMT2B (3′), OCT3/4, and NANOG in iPSC lines. PCR primers of KMT2B (5′) and KMT2B (3′) were designed to amplify the upstream and downstream sequences of HBV-integration site, respectively (compared with KMT2B-WT iPSCs). (B) Representative data of the flowcytometric analysis for pluripotency. The TRA1-60-positive and SSEA-4-positive fractions were over 90%, indicating all iPSC lines maintained their pluripotency.

Hepatic Cells Derived From HBV-KMT2B Integrated iPSCs Mimic Transcriptional Enhancement of KMT2B in the Integrated HCC Case

To investigate the phenotypes of hepatic cells resulting from HBV-KMT2B integration, all 4 types of the established iPSCs were differentiated into iPSC-derived hepatic progenitor cells (HPCs), which were derived from CD13^high and CD133^high iPSC-derived hepatoblasts as previously described.15, 16, 17 The number of CD13^high and CD133^high cells significantly decreased in KMT2B-KO iPSC-derived hepatoblasts, indicating the limited differentiation potential of KMT2B-KO iPSCs, which is consistent with the previous report of KMT2B-deficient murine embryonic stem cells (Figure 5A).⁸ Nevertheless, HPCs were established from all types of KMT2B-edited iPSCs, including KMT2B-KO iPSCs. The successful differentiation into hepatic lineages was indicated by the immunostaining of alpha fetoprotein (AFP) and albumin (ALB) (Figure 2E), as well as by the expression of hepatic markers (HNF4α, AFP, and ALB), and a cholangiocytic/progenitor marker (EpCAM) (Figure 5B). Interestingly, KMT2B-Int HPCs demonstrated a significantly elevated expression of KMT2B (3′) compared with KMT2B-HT HPCs (Figure 2F). Immunoblot analysis revealed that KMT2B production of KMT2B-Int HPCs was significantly elevated compared with KMT2B-HT HPCs (Figure 2G). This increased transcription and production of KMT2B-3′ side (downstream sequence of HBV-DNA integration) in HPCs resembled those observed in the original HCC tissue, indicating that our established HBV-KMT2B integrated hepatic cells sufficiently mimic the feature of HBV-KMT2B integrated HCC.

Figure 5.

Establishment and qPCR of HPCs. (A) Representative data of the flowcytometric analysis of iPSC-derived hepatoblasts after 14 days of hepatic differentiation. The induction efficacy of CD13-high and CD133-high cells was significantly decreased in KMT2B-KO cells (compared with KMT2B-HT cells). (B) Gene expression of HNF4a, AFP, and ALB (hepatic marker) and EpCAM (cholangiocytic/progenitor marker) in established HPCs confirmed the phenotype of hepatic progenitor cells.

Aberrant Proliferation in HBV-KMT2B Integrated HPCs

Analyzing the effect of HBV-KMT2B integration in HPCs, we focused on cell proliferation. Our data revealed a significant up-regulation in the proliferation of KMT2B-Int HPCs compared with that of KMT2B-WT, HT, and KO HPCs (Figure 6A and B). To confirm the increased proliferation of KMT2B-Int HPCs, we performed the colony formation assay of HPCs (Figure 6C). The colony numbers of KMT2B-Int HPCs significantly increased compared with KMT2B-HT HPCs (Figure 6D). In contrast, undifferentiated KMT2B-Int iPSCs did not show any alteration of their proliferation potential relative to KMT2B-HT and KMT2B-WT iPSCs (Figure 6E), indicating that HBV-KMT2B integration affected the potential of proliferation in a hepatic lineage-specific manner.

Figure 6.

Increased cell proliferation of HPCs derived from KMT2B-Int iPSCs. (A) Cell proliferation analysis by the MTS assay. The proliferation of KMT2B-Int HPCs was significantly upregulated compared with that of KMT2B-HT HPCs. The proliferation of KMT2B-KO HPCs was downregulated. ∗∗P < .01 (compared with KMT2B-HT HPCs). (B) Time course of cell numbers. An increase in KMT2B-Int HPCs proliferation and a decrease in KMT2B-KO HPCs proliferation were observed. ∗∗P < .01 (compared with KMT2B-HT HPCs). (C) Schema of colony formation. HPCs were sorted as single cells on feeder cells (MEFs). Representative images of colony formation assay. In the lower panel, colonies were stained with DAPI. Scale bars: 500 μm. (D) Colony formation assay of HPCs. Colony numbers were significantly increased in KMT2B-Int HPCs and decreased in KMT2B-KO HPCs (compared with KMT2B-HT HPCs). (E) Cell proliferation analysis by the MTS assay in iPSCs. The proliferation of KMT2B-Int iPSCs was equal to that of KMT2B-HT iPSCs, whereas KMT2B-KO iPSCs were significantly decreased. ∗∗P < .01 (compared with KMT2B-HT iPSCs).

HBV-KMT2B Integration Does Not Induce Loss of KMT2B Function

The proliferation of both KMT2B-KO iPSCs and HPCs was significantly decreased (Figure 6A and E). Microarray analyses of all 4 types of iPSC lines clarified the different expression profiles of KMT2B-KO iPSCs compared with those of the other 3 types of iPSCs (Figure 7A). Gene set enrichment analyses (GSEA) revealed a tendency of enrichment of the apoptosis pathway in KMT2B-KO iPSCs (Figure 7B). In fact, increased cleaved caspase-3 production was detected in KMT2B-KO iPSCs (Figure 7C). These results showed that the loss of KMT2B function in iPSCs decreases proliferation due to their enhanced apoptosis. Consequently, the phenotype of KMT2B-Int cells was functionally different from that of KMT2B-KO cells, indicating that HBV-KMT2B integration did not induce loss of KMT2B function. Thus, we focused on the phenotypes of KMT2B-WT, KMT2B-HT, and KMT2B-Int HPCs (except for KMT2B-KO HPCs) in the following studies.

Figure 7.

Unique expression pattern and apoptosis of KMT2B-KO iPSCs. (A) Microarray analysis using cDNAs of iPSCs. A heatmap of the Euclidean distance matrix was shown. The cluster analysis revealed the distinct expression pattern of KMT2B-KO iPSCs. (B) GSEA using data of iPSCs microarray. GSEA revealed a tendency of enrichment of the apoptosis pathway in KMT2B-KO iPSCs. NES, Normalized enrichment score; NOM p-val, normal P value; FDR q-val, False discovery rate q-value. (C) Immunoblotting analysis of Cleaved Caspase-3 and Caspase-3 antibodies demonstrated increased Cleaved Caspase-3 production in KMT2B-KO iPSCs. (D) GSEA analysis of iPSCs microarray data showed no enrichment of cell cycle-related gene sets.

Transcriptional and Epigenetic Enrichment of Cell Cycle-related Gene Sets in Hepatic Cells With HBV-KMT2B Integration

To elucidate the mechanism of hepatic lineage-specific aberrant proliferation observed in KMT2B-Int cells, we conducted comprehensive study on both profile of gene expression and H3K4me3, which was one of the main functions of KMT2B (Figure 8A). The comprehensive microarray analyses revealed a distinct expression profile in KMT2B-Int HPCs compared with others (Figure 8B). Enrichment analysis of upregulated genes in KMT2B-Int HPCs showed that pathways related to the cell cycle and cell proliferation were significantly enriched (Figure 8C). GSEA analyses also demonstrated significant enrichment of gene sets associated with DNA replication and cell cycle in KMT2B-Int HPCs, as compared with KMT2B-HT HPCs (Figure 8D). GSEA and enrichment analyses revealed a significant enhancement in cell cycle-related gene sets in KMT2B-Int HPCs compared with KMT2B-WT HPCs, similar to findings from comparisons with KMT2B-HT HPCs (Figure 9). In contrast to HPCs, microarray analysis of iPSCs showed no significant enrichment of DNA replication and cell cycle pathways (Figure 7D), indicating that transcriptional change, as well as their proliferation potential, was upregulated after hepatic lineage commitment in cells with HBV-KMT2B integration.

Figure 8.

Both transcriptional and epigenetic comprehensive analysis of HPCs revealed enrichment of cell cycle-related gene sets in KMT2B-Int HPCs. (A) Comprehensive analyses using KMT2B-Int and KMT2B-HT HPCs. The KMT2B-HT HPCs are used as a control of KMT2B-Int HPCs. (B) The cluster analysis of the cDNA microarray of HPCs. The heatmap showed the Euclidean distance matrix and revealed the unique expression pattern of KMT2B-Int HPCs. (C) Enrichment analysis using Clusterprofiler/ReactomePA (KEGG and Reactome pathways). The KMT2B-Int HPCs showed significant enrichment in cell cycle-related pathways (red). (D) GSEA using microarray data. Gene sets of DNA replication and the cell cycle were significantly enriched in KMT2B-Int HPCs relative to KMT2B-HT HPCs. NES, Normalized enrichment score; NOM p-val, normal P-value; FDR q-value, false discovery rate q-value. (E, F) Comprehensive analysis of histone tri-methylation at lysine 4 (H3K4me3) by CUT&Tag assay. (E) Average plots and heatmaps of promoter regions showed an elevated average of fragment densities in KMT2B-Int HPCs. (F) The CUT&TAG analysis was visualized as a heatmap using the Seqplot package. The Y-axis represents gene groups that have been clustered into 5 clusters (C1-C5) along with their corresponding intensity. (G) Enrichment analysis using ChIP-Enrich (KEGG and Reactome pathways). Red letters indicate cell cycle-related pathways.

Figure 9.

Comprehensive analysis of transcription in KMT2B-Int HPCs and KMT2B-WT HPCs. (A) GSEA using microarray data. Gene sets of DNA replication and the cell cycle were significantly enriched in KMT2B-Int HPCs relative to KMT2B-WT HPCs. NES, Normalized enrichment score; NOM p-val, normal P-value; FDR q-value, false discovery rate q-value. (B) Enrichment analysis using Clusterprofiler/ReactomePA (KEGG and Reactome pathways). The KMT2B-Int HPCs showed significant enrichment in cell cycle-related pathways. Comprehensive analysis of transcription revealed enrichment of cell cycle-related gene sets in KMT2B-Int HPCs compared with KMT2B-WT HPCs.

Subsequently, we performed a global analysis of H3K4me3 using CUT&Tag assay. The CUT&Tag assay revealed a slight enhancement in the average fragment densities of H3K4me3 in promoter regions of KMT2B-Int HPCs compared with KMT2B-HT HPCs (Figure 8E and F), suggesting the functional acceleration of KMT2B. Subsequent analysis of higher peak regions in KMT2B-Int HPCs indicated enrichment in various pathways or gene sets, including those related to cell cycle (Figure 8G). These comprehensive analyses strongly suggested that HBV-KMT2B integration resulted in aberrant proliferation through both epigenetic and transcriptional alterations, especially in cell cycle-related genes.

HBV-KMT2B Integration Increases CCNA2 and CDK1 Expression via H3K4me3 Enhancement in HPCs, Similar to Phenotype in the HBV-KMT2B Integrated HCC

Exploring detailed molecular mechanisms of aberrant proliferation caused by HBV-KMT2B integration, we focused on upregulated genes in both of our comprehensive analyses. We identified 101 overlapping upregulated genes (Figure 10A; Table 1). Enrichment analysis of these 101 genes showed a targeted enhancement of cell cycle-related gene sets (Figure 10B). We then validated the expression and H3K4me3 status of CCNA2 and CDK1 in HPCs, both of which directly influence cell cycle control. Both qPCR and immunoblot analysis confirmed the significant upregulation of these 2 genes in HBV-KMT2B integrated HPCs compared with both KMT2B-WT and HT HPCs (Figure 10C). The CUT&Tag analysis visualized the elevated peaks of CCNA2 and CDK1 transcription start site (TSS) regions. Chromatin immunoprecipitation quantitative PCR (ChIP-qPCR) confirmed significant H3K4me3 enhancements of CCNA2 and CDK1 (Figure 10D). In undifferentiated iPSCs, there were no significant enhancements in mRNA, protein levels, and ChIP-qPCR of these 2 genes between KMT2B-HT and -Int iPSCs (Figure 11), consistent with the observation that proliferation of these iPSCs were not changed (Figure 6E). These data indicated that enhancements of CCNA2 and CDK1 expression via H3K4me3 perturbation are specific for hepatic lineages.

Figure 10.

HBV-KMT2B integration induced aberrant expression of CCNA2 and CDK1 via enhancement of H3K4me3, similar to the HBV-KMT2B-integrated HCC. (A) The Venn diagram identified 101 overlapped genes upregulated in both CUT&Tag assay and microarray. (B) Enrichment analysis of 101 genes using Clusterprofiler/ReactomePA. (C) Significant upregulation of CCNA2 and CDK1 in KMT2B-Int HPC by qPCR and immunoblotting. (D) H3K4me3 was elevated in the regions of CCNA2 and CDK1 in CUT&Tag analysis. ChIP-qPCR analysis confirmed that H3K4me3 was significantly enhanced in CCNA2 and CDK1 regions. (E, F) Analysis of the original tumor and non-tumorous tissues of HCC. Expression of CCNA2 and CDK1 was significantly increased. ChIP-qPCR validated significant enhancement of H3K4me3 in the CCNA2 region. (G) Inhibition of CDK1 in KMT2B-Int HPCs. Colony formation assay showed CDK1 inhibitor (Ro-3306) revert the aberrant proliferation in KMT2B-Int HPCs.

Table 1.

101 Upregulated Gene Lists in Both Microarray and CUT&Tag Assay

Gene symbol
ACTR3C	CENPI	HLA-DMB	NYNRIN	ST3GAL6
ANLN	CLEC2D	HTRA4	OAS3	STON1
ARHGAP22	CORO1A	ID4	OSBPL6	SVOPL
ASPM	CPVL	IGF2	OXR1	TACC1
ATAD5	CTNNAL1	KANK1	P2RY6	TACSTD2
ATP23	CYBA	KIF14	PARP14	TBX2
AURKB	CYP11A1	KIF23	PBX3	THSD7A
AXL	DEK	KIF2C	PDZD2	TK1
B3GALNT1	DEPDC1	KRT17	PEG3	TRIM71
BBX	DLX3	LIMA1	PLK1	TRPC1
BCAT2	DUT	MCM10	PRKAR2B	TTC7B
BIRC5	EFEMP1	MCM2	PTN	UBE2C
CCNA2	ERVH48-1	MDM4	RAD51C	VWDE
CDC20	EXO1	MEST	RAD54L	WNT7B
CDC25C	FAM13A	MIS12	RIMKLB	ZNF331
CDC6	FBLN1	MTMR7	SCARA3	ZNF433
CDCA7	FBXO4	MYBL2	SCIN	ZNF502
CDK1	FKBP5	NAV2	SHCBP1
CDO1	GNE	NEURL1	SLC16A12
CECR2	GRAMD1C	NRK	SP6
CENPF	HFE	NSUN7	SPINK2

Figure 11.

qPCR, immunoblot, and ChIP-qPCR of CCNA2 and CDK1 in iPSCs. (A) qPCR of CCNA2 and CDK1 in iPSCs. These 2 genes were not upregulated in KMT2B-Int iPSCs (compared with KMT2B-HT iPSCs). (B) No significant change in immunoblotting analysis of CCNA2 and CDK1 in iPSCs before hepatic differentiation. (C) No significant change in H3K4me3 of CCNA2 and CDK1 in iPSCs before hepatic differentiation.

To validate these H3K4me3 perturbations in cell cycle-related genes in HBV-KMT2B integrated HPCs, we analyzed the original tumor and non-tumorous samples. Expressions of CCNA2 and CDK1 were significantly upregulated in the tumor sample (Figure 10E). H3K4me3 of CCNA2 was significantly enhanced in HBV-KMT2B integrated HCC, and H3K4me3 of CDK1 was slightly upregulated (Figure 10F). This suggested that the epigenetic perturbations in HPCs caused by HBV-KMT2B integration reproduced the phenotype of original HBV-KMT2B integrated HCC. Furthermore, inhibition of CDK1 function using the CDK1 inhibitor (Ro-3306) reverted the increase of proliferation of KMT2B-Int HPCs (Figure 10G). Taken together, HBV-KMT2B integration in hepatic cells, even in the absence of other oncogenic mutations, triggered aberrant proliferation via epigenetic alterations of cell cycle-related genes.

HBV-KMT2B Integration Upregulates CCNA2 and CDK1 Expression via H3K4me3 Enhancement in iPSC-derived Liver Organoids

HBV infection and HBV integration occur in mature hepatocytes. To verify the effect of HBV-KMT2B integration in mature hepatocytes, we differentiated iPSCs into iPS-hepatocytes (iPS-Heps) and performed an organoid formation assay in 3D culture, iPSC-derived liver organoids (iPS-LOs) as previously reported¹⁸ (Figure 12A). To confirm the proliferative potential of iPS-LOs, we analyzed cell viability using the CellTiter-Glo 3D assay. Our data showed significantly enhanced proliferation of KMT2B-Int iPS-LOs compared with KMT2B-HT iPS-LOs (Figure 12B). Quantitative PCR showed elevated KMT2B (3') expression in iPS-LOs of KMT2B-Int compared with those of KMT2B-HT (Figure 12C). RNA expressions of CCNA2 and CDK1 were significantly upregulated in KMT2B-Int iPS-LOs. On the other hand, there was no significant impact on hepatic differentiation of iPSCs, because hepatic marker genes ALB and AFP were not significantly changed between KMT2B-HT and KMT2B-Int iPS-LO (Figure 12D). Furthermore, significant H3K4me3 enhancements of CCNA2 and CDK1 were also confirmed in KMT2B-Int iPS-LOs (Figure 12E). These results clearly showed that both transcriptional and epigenetic changes in KMT2B-Int HPC are also observed in mature hepatocytes.

Figure 12.

Enhanced proliferation and transcriptional/epigenetic changes in liver organoids (LOs) derived from KMT2B-Int iPS-Heps. (A) Schema of the generation of iPS-LOs. iPS-LOs were established from iPS-Heps. 3D culture by Matrigel was used for cultivating iPS-Los. (B) The representative images of KMT2B-HT and KMT2B-Int iPS-LOs. Scale bars: 500 μm (left panel). Cell proliferation analysis by the CellTiter-Glo 3D assay (right panel). The proliferation of KMT2B-Int iPS-LOs was significantly upregulated compared with that of KMT2B-HT iPS-LOs. (C) Quantitative PCR of KMT2B (3′) in iPS-LOs. Significant upregulation of KMT2B (3′) in KMT2B-Int iPS-LOs was detected compared with KMT2B-HT iPS-LOs. (D) Quantitative PCR of ALB, AFP, CCNA2 and CDK1 in iPS-LOs. Expression of ALB and AFP were not significantly changed, whereas CCNA2 and CDK1 were significantly upregulated in KMT2B-Int iPS-LOs compared with KMT2B-HT iPS-LOs. (E) ChIP-qPCR analysis confirmed that H3K4me3 was significantly enhanced in CCNA2 and CDK1 regions of KMT2B-Int iPS-LOs. These data demonstrated that the phenotype of enhanced proliferation, as well as transcriptional and epigenetic features, are observed in mature hepatocytes consistent with HPCs.

A Truncated KMT2B Transcript Variant, not HBV-KMT2B Fusion Transcripts, is Responsible for Aberrant Proliferation of HPCs

We investigated the mechanism of H3K4me3 enhancement in HBV-KMT2B integrated cells. Although in HBV-KMT2B integrated HCCs, HBV-KMT2B fusion transcripts, and abnormal KMT2B alternative splicing variants have been detected by qPCR,^11,12 there is no evidence of full-length fusion or abnormal transcripts. Thus, to determine precise full-length transcripts from the KMT2B region with HBV integration, we performed long-read RNA-seq (Iso-seq) both on the HCC sample and its replicational HBV-KMT2B integrated HPCs. Iso-seq analysis of the HCC demonstrated 34 KMT2B transcript variants and 3 HBV-KMT2B fusion transcripts (Figure 13A). The analysis of HBV-KMT2B integrated HPCs also demonstrated 3 truncated transcripts of KMT2B and no HBV-KMT2B fusion transcript (Figure 13B). Among 2 analyses of Iso-seq, all transcripts including fusion transcripts lacked the sequences from exon 1 to HBV-KMT2B integration. Notably, transcripts #20 and #33 were detected in both the HCC sample and HBV-KMT2B integrated HPCs (Figure 13A and B), leading to the hypothesis that these abnormal KMT2B proteins lacking N-terminal parts were produced and responsible for aberrant cell proliferation of hepatic cells. To validate this hypothesis, the KMT2B-transcript variants detected in Iso-seq analyses were overexpressed in KMT2B-HT HPCs (Figure 13C). We selected the common transcripts, #20 and #33, which encoded truncated KMT2B containing the PHD4 domain and a C-terminal subset after taspase-1 cleavage.

Figure 13.

Abnormally truncated KMT2B transcripts promoted proliferation of HPCs. (A) Coverage of HBV-KMT2B in RNA-seq (upper panel shown in Figure 1C) and transcripts identified by Iso-seq (lower panel). (B) Full-length transcripts identified by Iso-seq analysis in HBV-KMT2B integrated HPCs (KMT2B-Int HPCs). Three KMT2B transcript variants lacking N-terminal (5′) sequence were identified in KMT2B-Int HPCs. (C) Schema of the transcript variants (upper panel) and experimental schema of colony formation assay (lower panel). The transcript variants were overexpressed by lentiviral vectors. The KMT2B-HT HPCs after transduction were cultured after positive selection of GFP expression by FACS. (D) Overexpression of transcripts was confirmed by qPCR. (E) Colony formation assay of overexpressed HPCs. Overexpression of the #20 transcript significantly upregulated the number of colonies. In contrast, other transcripts including #33, HBx-ΔC, and #F2 significantly decreased the number of HPC colonies (compared with mock). The representative images of colony formation assay (left panel). Scale bars: 500 μm. (F) Cell proliferation analysis by the MTS assay. The proliferation of #20 was significantly upregulated compared with that of mock. The proliferation of #33 was downregulated. ∗∗P < .01 (compared with mock)

We also focused on 3 HBV-KMT2B fusion (# F1, # F2, and # F3) transcripts, which encoded C-terminal-deleted HBx protein (HBx-ΔC) with the out-of-frame sequence of 33-amino acid (# F1), partial HBs protein subsequently conjugated with the KMT2B C-terminal part (# F2), and the partial HBs protein (# F3) (Figure 13C). Translation of # F3 terminated at HBs stop codon, resulting in no conjugation of KMT2B protein. Therefore, we tested the function of 2 fusion transcripts, HBx-ΔC (# F1 without out-of-frame sequence) and the sequence encoded by # F2. The overexpression of these transcripts was confirmed by qPCR (Figure 13D).

As a result, the lentiviral overexpression of #20 significantly increased colony numbers of HPCs compared with mock-transduced HPCs. In contrast, the overexpression of other transcripts, including #33, HBx-ΔC, and #F2, did not increase the colony numbers of HPCs (Figure 13E). MTS assay validated aberrant proliferation of HPCs by #20 overexpression (Figure 13F). To test the effect of #20 on mature hepatocyte proliferation, #20 overexpression was analyzed using KMT2B-HT iPS-Heps (Figure 14A and B). The enhanced proliferation by #20 overexpression was also observed in iPS-Heps by measurement of crystal violet positive areas (Figure 14C). These results demonstrated that the transcript #20 is involved in aberrant proliferation in mature hepatocytes as well as in HPCs.

Figure 14.

Truncated KMT2B transcripts promote proliferation of iPS-Heps. (A) Schema of the transcript variants and experimental schema of proliferation assay of iPS-Heps. (B) Overexpression of #20 was confirmed by qPCR. (C) Representative images of the crystal violet staining (left panel). Scale bars: 500 μm. Overexpression of the #20 transcript significantly upregulated the proliferation of iPS-Heps (right panel).

Inhibition of KMT2B function was examined using the small molecule, MI-2, which is known to inhibit the interaction between menin and KMT2B or KMT2A (a paralog of KMT2B),¹⁹ because there is a menin interaction domain at the N-terminal region of KMT2B. KMT2A expression in KMT2B-Int HPCs was equal to that in KMT2B-HT HPCs (Figure 15A). In colony formation assays, KMT2B-Int HPCs were significantly resistant to the inhibition of menin-KMT2 interactions compared with KMT2B-HT HPCs (Figure 15B). Taken together, these results demonstrated that the N-terminal truncated KMT2B protein from the HBV-KMT2B integrated allele, not the viral protein (HBx-ΔC) and HBV-KMT2B fusion protein, is responsible for the increased cell proliferation of KMT2B-Int HPCs.

Figure 15.

Inhibition of menin-KMT2 interactions in HPCs. (A) qPCR of KMT2A in HPCs. There was no change in the expression of KMT2A. (B) Colony formation assay showed a significant resistance of KMT2B-Int HPCs to the addition of KMT2A/B-menin inhibitor (MI-2).

Both PHD4 Domain and SET Domain are Required for the Aberrant Proliferation of HPCs due to the Truncated KMT2B Transcript Variant

The transcript of #20 contains both PHD4 and SET domains of KMT2B. To confirm which domain is involved in the aberrant proliferation of HPCs, #20-ΔPHD4, which is #20 lacking PHD4 domain, and #20-ΔC, which is lacking the N-terminal region, including the FYRN and SET domains, were overexpressed in KMT2B-HT HPCs using lentiviral vectors (Figure 16A). The overexpression of these transcripts was confirmed by qPCR (Figure 16B). Overexpression of #20-ΔPHD and #20-ΔC significantly downregulated the number of colonies compared with #20 (Figure 16C).

Figure 16.

Both the PHD4 domain and the SET domain are required for the aberrant proliferation of HPCs. (A) Schema of the constructs of #20, #20-ΔPHD4, #20-ΔC, and #33 and PCR primer positions of #20-1, #20-2, and #20-3 and experimental schema of colony formation assay. (B) Validation of overexpression by qPCR. (C) Colony formation assay using KMT2B-HT HPCs overexpressing deletion variants of #20. Overexpression of the #20 transcript significantly upregulated the number of colonies. In contrast, other transcripts including #20-ΔPHD4, #20-ΔC, and #33 did not significantly increase the number of HPC colonies (compared with mock). (D, E) Schema of the N-terminal truncated KMT2B (KMT2B-N-HBV) and experimental schema of colony formation assay. (F) Overexpression of KMT2B-N-HBV was confirmed by qPCR. (G) KMT2B-N-HBV significantly decreased the number of HPC colonies compared with mock.

Considering the possibility that HBV integration produces the truncated proteins covering exons 1-3 of KMT2B containing the menin-binding domain, we also analyzed the proliferation potential of the N-terminal KMT2B protein. The putative N-terminal KMT2B protein (KMT2B-N-HBV) in our recapitulated HBV-KMT2B integration, consisting of the first 783 amino acids of KMT2B linked to 634 out-of-frame amino acids from HBV sequences, was overexpressed in KMT2B-HT HPCs using the lentiviral vector (Figure 16D–F). Colony formation assays showed that overexpression of KMT2B-N-HBV significantly down-regulated the number of colonies compared with mock-infected HPCs (Figure 16G). These results showed that the N-terminal truncated KMT2B containing both PHD4 and SET domains is essential in the aberrant proliferation of hepatic cells.

HBV-KMT2B Integration Induces Replication Stress and DNA Damage, as an Oncogenic Process, via Aberrant Cell Proliferation

To explore the impact of HBV-KMT2B integration on the oncogenic mechanism initiated by aberrant proliferation, DNA replication stress due to aberrant cell proliferation was analyzed, which leads to DNA damage. In microarray analyses of HPCs, we found the enrichment of the pathway related to replication stress in KMT2B-Int HPCs, for example, the Fanconi anemia pathway and mismatch repair in KEGG enrichment analyses (Figure 8C, upper panel) and G2/M DNA damage checkpoint in Reactome enrichment analyses (Figure 8C, lower panel). In GSEA analyses using microarray data, the gene ontology (GO) DNA repair and E2F pathway gene sets were enriched in KMT2B-Int HPCs compared with KMT2B-HT HPCs (Figure 17A). The gene set “activation of ATR in the response of replication stress” was significantly enriched in KMT2B-Int HPCs (Figure 17A). Using the replication stress signature that was constituted by 20 pathways related to the replication stress,²⁰ a gene set variation analysis (GSVA) demonstrated the high enrichment score of the pathways of the replication stress signature in KMT2B-Int HPCs (Figure 17B).

Figure 17.

HBV-KMT2B integration induced replication stress and DNA damage of HPCs as an oncogenic process. (A) GSEA using microarray data of HPCs showed significant enrichment of the DNA repair gene set and the replication stress gene set in KMT2B-Int HPCs. (B) The GSVA on microarray data in KMT2B-HT and KMT2B-Int HPCs using previously reported replication stress-related gene sets revealed high replication stress signature in KMT2B-Int HPCs. (C) Immunoblot analyses of ATR, Chk1, ATM, and Chk2. KMT2B-Int HPCs represented higher intensity of ATR, Chk1, ATM, and Chk2. (D) Immunoblotting analysis of γH2AX. KMT2B-Int HPCs represented higher intensity of γH2AX. HPCs purified using a magnetic cell sorter (removal of feeder MEFs were used as samples. (E) Immunostaining of γH2AX in KMT2B-Int and HT HPCs. Colonies of HPCs were stained with anti-HNF4α and γH2AX antibodies. After images were acquired, the intensity of γH2AX in each nucleus was measured. Feeder MEFs exhibited the high intensity of γH2AX due to MMC treatment, using them as a positive control. KMT2B-Int HPCs represented a significantly higher intensity of γH2AX compared to KMT2B-HT HPCs, indicating high DNA damage in KMT2B-Int HPCs. Scale bars: 100 μm. (F) The GSVA on RNA-seq data of GEO dataset (GSE65485) in 3 cases with HBV integration into exons of KMT2B locus (315T: exon5, 316T:exon3, 348T:exon3), 6 HBV-negative HCC cases, and 4 non-tumorous cases, using previously reported replication stress-related gene sets. This data demonstrated the high enrichment score of the pathways of the replication stress signature in HBV-KMT2B integration cases, consistent with our iPSC-model findings.

Immunoblot analysis validated enhanced protein expression of ATR, Chk1, ATM, and Chk2 in KMT2B-Int HPCs (Figure 17C). As the replication stress led to the induction of DNA damage, we assessed DNA damage in HPCs by γH2AX (a representative marker of DNA damage). Immunoblot analysis revealed that γH2AX was increased in KMT2B-Int HPCs compared with KMT2B-HT HPCs (Figure 17D). To confirm the γH2AX expression in HPCs more specifically, we performed immunostaining of γH2AX and quantified the intensity of γH2AX. HNF4α-negative cells were mitomycin (MMC)-treated feeder mouse embryonic fibroblasts (MEFs) with high DNA damage as a positive control. HNF4α-positive KMT2B-Int HPCs presented significantly high DNA damage compared with HNF4α-positive KMT2B-HT HPCs (Figure 17E). GSEA analyses also indicated significant enrichment of replication stress gene signature in KMT2B-Int HPCs compared with KMT2B-WT HPCs (Figure 18A). Protein expression levels of ATM, Chk1, ATR, Chk2, and γH2AX were also elevated in KMT2B-Int HPCs compared with KMT2B-WT HPCs (Figure 18B and C).

Figure 18.

KMT2B-Int HPCs induced higher levels of replication stress and DNA damage compared with KMT2B-WT HPCs. (A) GSEA using microarray data of HPCs showed significant enrichment of the DNA repair gene set and the replication stress gene set in KMT2B-Int HPCs. (compared to KMT2B-WT HPCs) (B, C) Representative images of proteins related to DNA damage in HPCs using capillary electrophoresis system (SimpleWestern). KMT2B-Int HPCs represented higher intensity of ATR, Chk1, ATM, Chk2, and γH2AX compared with KMT2B-WT HPCs.

To validate these results obtained from our iPSC model, we used Gene Expression Omnibus (GEO) dataset (GSE65485), which included the RNA-seq data of HBV-KMT2B integrated HCCs.¹¹ We analyzed RNA-seq data of 3 samples with HBV integration into exons of KMT2B locus as well as our sample, 6 HBV-negative HCC samples, and 4 non-tumorous samples. In GSVA analysis, all 3 HBV-KMT2B integrated HCCs showed high enrichment scores of the pathway related to replication stress, consistent with our iPSC model findings, whereas HBV-negative HCCs showed heterogeneous enrichment scores (Figure 17F). Reflecting normal tissue characteristics, non-tumorous liver samples showed distinctly low enrichment scores. Furthermore, we analyzed the EGA dataset (EGAD00001004484) including RNA-seq data for the AAV2-KMT2B-integrated HCC, and we performed GSVA analysis of them and non-tumorous samples. The results demonstrated that the AAV2-KMT2B-integrated HCC belongs to the cluster that exhibits a high enrichment of replication stress signatures (Figure 19) as well as CCNA2/E1-integrated HCCs, which were previously shown high replication stress signature.²¹ These data strongly suggest that HCC caused by viral integration into the KMT2B locus is associated with replication stress caused by a viral species-independent mechanism.

Figure 19.

The AAV2-KMT2B integrated HCC sample showed an enhanced signature of replication stress. Heatmap of enrichment score of the replication stress-related pathways in AAV2-integrated HCCs. RNA-seq of AAV2-integrated HCCs (from EGAD00001004484) and non-tumorous samples from GSE65485, as control, were analyzed by GSVA. AAV2 integrated loci were also shown in the gray box. CHC1185T, the AAV2-KMT2B integrated HCC, exhibited a high enrichment of the replication stress signatures.

In summary, our data demonstrated that hepatic cells with HBV-KMT2B integration are in an early oncogenic process through aberrant cell proliferation, replication stress, and DNA damage, which arise from abnormal KMT2B function. This disease model implies that the oncogenic axis derived from abnormal KMT2B function could be a potential therapeutic target of HBV-KMT2B integrated HCC.

Discussion

Our study identified the detailed role of HBV-KMT2B integration in the tumorigenesis of HBV-related HCC. Through the analysis of a tumor sample, we determined complete sequences of HBV integration on the KMT2B locus and reproduced it using genome-intact iPSCs. From the analyses of these iPSC-derived hepatic cells, HBV-KMT2B integration induced aberrant cell proliferation and upregulation of expression and H3K4me3 of cell cycle-related genes, which were similar phenotype to the original HCC sample. The long-read RNA-seq revealed truncated and HBV-KMT2B fusion transcripts. Among these transcripts, only the truncated KMT2B (#20), detected in both HCC samples and KMT2B-Int HPCs, was able to promote cell proliferation in both HPCs and iPS-Heps. Furthermore, HBV-KMT2B integration, as an oncogenic process, caused replication stress and DNA damage (Graphical abstract). Thus, this is the first evidence that clearly shows oncogenic mechanisms of HBV-KMT2B integration based on both clinical and basic research.

In this study, we used iPSCs and their hepatic derivations to reveal hepatic lineage-specific change due to HBV-KMT2B integration (Figure 2A). Although mature hepatocytes, such as primary human hepatocytes and iPSC-derived hepatocytes, did not maintain their proliferative potential in vitro, our iPSC-derived HPCs maintain such phenotypes, enabling us to assess the direct relationship between HBV-KMT2B integration and cell proliferation in hepatic lineages. Moreover, using genome-intact iPSCs for gene editing, we obtained HBV-KMT2B-integrated cells without any other oncogenic mutations. This allowed us to observe the specific phenotype of HBV-DNA integration. As HBV-DNA integration in HCC is considered an early event during tumorigenesis,²² preexisting oncogenic mutations in cancer cell lines affect and conceal the mechanism of HBV-KMT2B integration. Although it is a limitation that our culture system consists solely of hepatic cells, preventing analysis of cell-cell interaction, such as the reported association between KMT2B and the inflammatory tumor microenvironment,^23,24 our iPSC-disease model demonstrated the specific function of HBV-KMT2B integration in initiating oncogenesis in hepatic cells with an originally normal genome.

HBV integration induces production of viral protein in host cells. HBx protein is known to have oncogenic potential, and C-terminal-depleted HBx also induce cell proliferation of HCC cell lines.²⁵ HBx protein in HCC cell lines interacted with WDR5, resulting in the global alternation of H3K4me3 and increased proliferation of such cells.²⁶ However, HBx-ΔC overexpression in our HPCs significantly decreased the number of colonies (Figure 13E). In previous reports, the HBx/KMT2B fusion transcripts were detected and showed unique gene regulation compared with full-length HBx in HepG2 cell lines.¹² Although HBx/KMT2B fusion was not detected in our Iso-seq analysis (Figure 13A and B), fusion transcripts did not affect the proliferation of HPCs (Figure 13E). Specific studies for the function of fusion proteins will be needed.

This study shows the first evidence that H3K4me3 due to aberrant KMT2B function, which is specific for cell cycle-related genes, induces proliferation, replication stress, and DNA damage in normal cells without other oncogenic mutation. As the molecular mechanisms in HBV-KMT2B integrated HPCs, only the truncated KMT2B transcript, which encodes the protein from the PHD4 domain to the C-terminal, reproduced the upregulation in colony numbers in the colony formation assay. In contrast, KMT2B transcript started from exon 1 was not detected at all in Iso-seq, probably because of low activity of KMT2B itself promoter compared with HBV promoter in hepatic lineages. Moreover, AAV2-genomic insertions in the KMT2B locus in HCC are detected between exon 3 and exon 6 within the KMT2B locus as well as that of HBV.¹³ Our data showed AAV2-KMT2B-integrated HCC belongs to the cluster that exhibits a high enrichment of replication stress signatures as well as HBV-KMT2B-integrated HCCs (Figures 17 and 19). This strongly suggested that the truncated KMT2B (due to viral DNA integration) is a common and viral species-independent mechanism underlying the oncogenic process. The known KMT2B mutations in human diseases (in mainly dystonia²⁷ and cancers in other organs⁶) are predicted protein-truncating variants and missense mutations. However, N-terminal truncated KMT2B has not been reported, thus our findings were the first study about the gain-of-function of N-terminal truncated KMT2B. Whereas it was impossible to detect the N-terminal fragment of truncated KMT2B using the available antibody due to a lack of antibody-recognition site in N-terminal truncated KMT2B, the importance of the PHD4 and SET domains in enhancing proliferative capacity of hepatic cells was confirmed (Figure 16). Thus, truncated transcripts including these domains, like #20, contribute to aberrant proliferation of hepatic cells. Further studies will be needed to clarify the precise molecular interaction of truncated KMT2B.

We demonstrated that HBV-KMT2B-integrated hepatic cells initiated the oncogenic process. However, because KMT2B-Int HPCs did not demonstrate the acquisition of independence from humoral factors, which is a hallmark of in vitro malignancy, there was no evidence of complete malignant transformation of HPCs with HBV-KMT2B integration. Our data implied that KMT2B integration gives the cells a high growth ability, making them expand more than the surrounding non-integrated hepatocytes. During this process, high replication stress will lead to the accumulation of mutations in HCC driver and other genes as a second hit and after, advancing the carcinogenesis process. Future study would be needed to clarify such mechanisms or mutations for subsequent multi-step oncogenic processes.

In the current study, cyclin upregulations, enhanced E2F and ATR pathways, and DNA replication stress signatures were shown in KMT2B-Int HPCs. Although HBV-KMT2B-integrated iPSCs and their derivative HPCs were derived from a single clinical HCC case, high replication stress signatures associated with HBV and AAV2 integration on KMT2B locus were confirmed in other HCC datasets (Figures 17 and 19), consistent with finding in KMT2B-Int HPCs. CCNA2/E1-integrated HCC belonged to the HCC subtype showing integrated cyclin activation and a rearrangement signature of replication stress. This type of HCC tends to have a high frequency of RB1 mutation and is exclusive to CTNNB1 and TERT promoter mutations.²¹ Furthermore, HBV-KMT2B integration in HCC tends to be mutually exclusive with other HCC major mutations (such as TERT, CTNNB1, and TP53).¹¹ Taken together, these findings suggest that DNA replication stress-induced hepatocarcinogenesis based on abnormal cell cycle regulation and aberrant KMT2B function was a significant mechanism of oncogenesis triggered by viral integration. This replication stress will potentially be a treatment target for this HBV- or AAV2-integrated subgroup of HCC.

In conclusion, our study demonstrates the oncogenic pathophysiological mechanisms of HBV-KMT2B integration. The detailed analysis of HBV-KMT2B-integrated HCC provided precise sequences of integration and various transcription variants. Our iPSC model with the recapitulation of the HBV-KMT2B integration revealed the oncogenic mechanism of HBV-KMT2B integration: the elevated H3K4me3 of cell cycle-related genes induced aberrant cell proliferation, replication stress, and DNA damage, suggesting that HBV-KMT2B integration drives oncogenic process for early stage of HCC development. Our data also showed that abnormal KMT2B function could be a potential therapeutic target of HBV-related HCC.

Materials and Methods

Human Clinical Samples and Ethical Statement

Patients with HCC who received locally curative hepatectomy at Tokyo Medical and Dental University/Institute of Science Tokyo were included in the present study. Written informed consent was obtained from all patients, and the Human Ethics Review Committee of Tokyo Medical and Dental University/ Institute of Science Tokyo approved this study, which was conducted in accordance with the Declaration of Helsinki.

Cell Culture

Human iPS cell line P11025 was established from male fibroblasts reprogrammed by the overexpression of Oct3/4, Sox2, Klf4, and c-Myc transduced by episomal vectors (purchased from Takara Bio). All iPSC lines (KMT2B-HT, KO, Int) were established from P11025. iPSCs were maintained and cultured by the Cellartis DEF-CS 500 culture system (Takara Bio) according to the manufacturer’s protocol in 37 °C incubator with 5% CO₂ in air.

Generation of KMT2B Mutated and HBV-KMT2B Integrated iPSCs

Electroporation was performed to establish KMT2B-KO iPSCs and KMT2B-HT iPSCs through homologous recombination of the gene cassette into P11025 iPS cells. 1 × 10⁶ iPSCs were transfected using Lonza 4D-Nucleofector (Lonza) with 2.5 μg of a plasmid expressing Cas9 and gRNA and 2.5 μg of the targeting vector containing a puromycin resistance gene cassette. After electroporation, the cells were resuspended in iPS cell culture medium and then incubated at 37 °C. The culture medium was changed every 24 hours after transfection. Puromycin (0.5 μg/mL) was added to the culture medium from 3 to 6 days after the transfection to select the mutated cells. The clones were picked with a microscope and cultured. The genotype of the clones was determined by PCR. PCR products of the mutant fragment were selected, and the DNA sequences were determined by Sanger sequencing. Clones with starting codon deleted in KMT2B exon 1 were selected as KMT2B-KO iPSCs, and the loss of protein expression of KMT2B in KMT2B-KO iPSCs was confirmed by immunoblotting. To exclude the possibility that an off-target mutation by the CRISPR/Cas9 genome-editing method affected the phenotype of such clones, 2 different gRNA sequences were used for genome editing, and the phenotype of such cells was analyzed using more than 2 clones derived from genome-edited clones using each gRNA sequence. For establishing KMT2B-Int iPSCs, homologous recombination of the gene cassette into P11025 iPS cells was performed by electroporation, as in the establishment of KMT2B-KO iPSCs. The targeting vector contained puromycin resistance and TagRFP gene cassettes. The genotype of the clones was determined by PCR after puromycin (0.5 μg/mL) selection. Two different gRNA sequences were used for genome editing, and the phenotype of such cells was analyzed using more than 2 clones derived from genome-edited clones using each gRNA sequence. Then, transduction of the Cre recombinase vector into the edited iPSCs was performed by lipofection. At 80% confluency of the dish, the iPSCs were transfected using 12 μL of viafect (Promega) with 2 μg of Cre vector (Santa Cruz). Three days after the transfection, for flow cytometric analysis, transfected iPSCs were dissociated in TrypLE Select Enzyme (ThermoFisher Scientific), and washed twice with phosphate-buffered saline (PBS), then stained with propidium iodide (PI, ThermoFisher) to identify dead cells. RFP negative cells were sorted on a single cell per well, using BD Aria2 (Becton Dickinson).

To confirm gene editing in iPSCs, DNA genotyping of the clones was determined by PCR. DNA was extracted from the clones using a DNeasy Blood & Tissue Kit (Qiagen). The region around exon 1 of KMT2B was amplified by PCR. The amplified bands were extracted, and their sequences were determined by Sanger sequencing. KMT2B-HT iPSCs #1 had an intact allele on one side and another allele with a sequence incorporating a selection marker containing SV40 and puromycin-resistant gene (PuroR) between the arms of the vector on the other side. KMT2B-HT iPSCs #2 also has an intact allele and another allele with a sequence incorporating the 3' arm of the vector followed by a selection marker. The sequence of KMT2B-KO iPSCs #1 is a sequence that incorporates a selection marker containing SV40 and PuroR between the arms on one side, and on the other side a sequence is missing 15 bases, including the Exon1 of the KMT2B start codon. KMT2B-KO iPSCs #2 has a sequence that incorporates a part of 3' arm of the vector followed by a selection marker in both alleles (Figure 3B). For KMT2B-Int iPSCs, DNA genotyping of the first genome-editing clones (before transfected Cre recombinase) was determined by PCR. The second genome-editing clones (after transfected Cre recombinase) were also determined by PCR. The gRNA sequences for genome editing are listed in Table 2.

Table 2.

gRNA Sequences

gRNA1 for HBV-KMT2B integrated iPSCs	TTCTTTGCAACCCCCTAACC
gRNA2 for HBV-KMT2B integrated iPSCs	GGACACTTTCCAGCATTGCG
gRNA1 for KMT2B-HT/KO iPSCs	GGCGGGGGCCGCGGCGGACG
gRNA2 for KMT2B-HT/KO iPSCs	CAACGGGGCCGAAAGAGTGC

HBV, Hepatitis B virus; iPSCs, induced pluripotent stem cells; KMT2B-KO, KMT2B homozygous-knockout.

Identification of Complete Sequences of HBV-KMT2B Integration in HCC Samples

To detect HBV integration in HCC samples, deep sequencing analyses were conducted as previously reported.³ Briefly, DNA was extracted from samples by DNeasy Mini Kit (Qiagen), and target amplicon libraries were generated by the Ion AmpliSeq Cancer Panel v2 (Life Technologies). Multiplexed barcoded libraries were amplified by emulsion PCR ion sphere particles (ISPs) using the Ion OneTouch 200 Template Kit v2 (Life Technologies) according to the manufacturer’s instructions. After the template ISPs were recovered from the emulsion, the positive template ISPs were biotinylated during the emulsion process and enriched with Dynabeads MyOne Streptavidin C1 beads (Life Technologies). Sequencing was performed on a Personal Genome Machine Sequencer (Ion PGM, Life Technologies) using the Ion PGM 200 Sequencing Kit (Life Technologies) according to the manufacturer’s instructions. In addition to 28 HBV-related HCC cases in the previous study,³ 17 additional HBV-related HCC cases were analyzed. The analysis identified one available sample with HBV-KMT2B integration that can be further analyzed in addition to DNA analysis. The full sequence of HBV genome integration in the HCC specimen was obtained by applying PCR to genomic DNA. The sequences were amplified by PCR. The sequences of PCR products were determined by Sanger sequencing.

Hepatic Differentiation of iPSCs and Establishment of iPSC-derived HPCs

Human iPSCs were differentiated into the hepatic lineage using Cellartis Definitive Endoderm Differentiation Kit and Cellartis Hepatocyte Differentiation Kit (Takara Bio) according to the manufacturer’s protocol. During hepatic differentiation, iPSCs were differentiated into hepatoblasts (hepatoblast-like cells) at day 14 and hepatocytes (hepatocyte-like cells: iPS-Heps) after day 21.

To establish HPCs, iPSCs were differentiated into hepatoblasts, and then the CD13^high/CD133⁺ fraction, which contains hepatic progenitor cells, was sorted onto feeder MEFs using a fluorescence-activated cell sorter (FACS). Human HPCs were passaged and expanded in HPC medium [1:1 mixture of Dulbecco's Modified Eagle Medium (DMEM)/F12 and hepatic colony-forming units in culture (H-CFU-C) medium supplemented with 40 ng/mL recombinant human hepatocyte growth factor (HGF), 20 ng/mL recombinant human epidermal growth factor (EGF), 0.25 μM A83-01, 10 μM Y-27632, and 10% fetal bovine serum [FBS]). H-CFU-C medium consisted of DMEM/F12 supplemented with 2 × non-essential amino acids, 2 × insulin-transferrin-selenium, 5 mM HEPES, 2 × 10^-7 M dexamethasone, and 2 × antibiotic-antimycotic mixed solution. The culture medium was changed every 3 days.

Generation of iPS-LOs (Liver Organoids)

Firstly, iPSCs were differentiated into hepatocytes using another protocol reported by Ang et al²⁸ with some modifications. In brief, iPSCs were seeded on a Matrigel-coated dish at about 3 × 10⁴ cells/cm² (Day 0). For differentiation into definitive endoderm (Day 1-2), iPSCs were cultured with Iscove’s Modified Dulbecco’s Medium (IMDM)/F12 (1:1) supplemented with 1× insulin-selin-transferrin (ITS), 1× lipid concentrate (LC), 1× penicillin/streptomycin (PS), 0.1% polyvinyl alcohol (PVA), 100 ng/mL activin A, 2 ×M CHIR99021 (Day 1), 250 nM LDN193189 (Day 2), and 50 nM PI-103. For differentiation into the hepatic lineage (Day 3-6), the cells were cultured with IMDM/F12 (1:1) supplemented with 10% knock-out serum replacement (KOSR), 1× LC, 1×PS, 0.1% PVA, basic FGF 10 ng/mL (Day 3), 10 ng/mL activin A (Day 4 -6), 30 ng/mL BMP-4, 1 μM A83-01 (Day3), 2 μM all trans-retinoic acid (Day 3), 1 μM forskolin (Day 4-6), 1 μM C59 (Day 4-5), 1 μM CHIR99021 (Day 6). To differentiate into hepatocytes, for the next 6 days (Day 7-12), the cells were cultured with IMDM/F12 (1:1) supplemented with 1 × ITS, 1 × Glutamax, 1 × non-essential amino acids (NEAA), 1 × LC, 1 × PS, 10 ng/mL BMP-4, 10 ng/mL oncostatin M,10 μM forskolin, 10 μM DAPT, 10 μM dexamethasone, 200 μg/mL ascorbic acid-2 phosphate (AAP), and 1 μM SB505124 (Day 7-8). For further differentiation, the cells were cultured with IMDM/F12 (1:1) supplemented with 10% KOSR, 1 × ITS, 1 × Glutamax, 1 × NEAA, 1 × LC, 1 × PS, 10 μM forskolin, 10 μM DAPT, 10 μM dexamethasone, 200 μg/mL AAP (Day 13-18). After 18 days, iPSCs were differentiated into iPS-Heps.

Secondly, human iPS-Heps differentiated in the above-mentioned 2D culture were dissociated by TrypLE select, washed twice with PBS, mixed with Matrigel, and seeded. The dissociated 1 × 10⁴ cells were seeded per a well of 24-well plate and cultivated in the human fetal hepatocyte organoid medium whose protocol is reported by Hu et al¹⁸ with some modifications. The human fetal hepatocyte organoid medium consisted of AdDMEM/F12 supplemented with 10 mM HEPES, 1 × Glutamax, 1 × PS, 2% B27, 1.25 mM N-acetylcysteine, 10 mM nicotinamide, 15% R-spondin-conditioned medium (home-made), 10 ng/mL gastrin, 2 μM A83-01, 10 μM Y-27632, 50 ng/mL EGF, 50 ng/mL HGF, 100 ng/mL FGF-7, 100 ng/mL FGF-10, 20 ng/mL TGFa, and 3 μM CHIR99021. The medium was changed every 2 to 3 days. After 10 days of culture, 3 wells were used for RNA extraction, 3 wells for CellTiter-Glo 3D cell viability assay (Promega), and the remaining wells for ChIP.

CellTiter-Glo 3D cell viability assay was performed according to the manufacturer’s protocol. In brief, 500 μL of CellTiter-Glo 3D reagent was added to each well of 24-well culture plate containing 500 μL of culture medium. After dissociating the Matrigel by pipetting, the solution containing the cells was transferred to a 15-mL Falcon tube. The solution was mixed at 22 °C and 1000 rpm for 5 minutes using a thermomixer and incubated for 25 minutes, and 9 mL of PBS was added to the mixture. Then, 100 μL of CellTiter-Glo 3D reagent was added onto each well of a 96-well plate, and 100 μL of the solution in a 15mL Falcon tube was mixed on the plate. The absorbance was measured using a Glomax Discover Microplate Reader (Promega).

qPCR

Total RNA was extracted using RNeasy mini kit (Qiagen) according to the manufacturer’s protocols. First-strand cDNA was synthesized using a PrimeScript IV 1st strand cDNA Synthesis Mix (Takara Bio) and used as a template for PCR amplification. qPCR analysis was performed using the Universal Probe Library system (Roche Diagnostics) or using a QuantiTect SYBR Green PCR Kit (Qiagen) in a reaction volume of 20 μL. All data were normalized against GAPDH. For the quantification of GAPDH mRNA, the expression levels were quantified using TaqMan GAPDH Control Reagents (human, Thermo Fisher Scientific). The primers used for qPCR are listed in Tables 3 and 4.

Table 3.

Primer Sequences Used for qPCR

Gene	UPL probe #	Forward primer	Reverse primer
OCT3/4	3	GAAACCCACACTGCAGATCA	CGGTTACAGAACCACACTCG
NANOG	69	GAGATGCCTCACACGGAGAC	AAGTGGGTTGTTTGCCTTTG
MLL1/KMT2A	9	GACAGTGTGCGTTATGTTTGACT	TGGCCAATATATAGTAAACGACCA
MLL4/KMT2B(5')	63	AGAGCAGTGACGGGGAATC	GAACTGGGGGCCACATCT
MLL4/KMT2B(3')	62	AGATGGTCCTCCCCAGGT	CTCAGAGCTCGAAGCCTCAC
AFP	66	ATGGCCATCACCAGAAAAAT	CATAAGTGTCCGATAATAATGTCAGC
ALB	27	AATGTTGCCAAGCTGCTGA	CTTCCCTTCATCCCGAAGTT
HNF4α	68	GAGATCCATGGTGTTCAAGGA	GTGCCGAGGGACAATGTAGT
EpCAM	3	CCATGTGCTGGTGTGTGAA	TGTGTTTTAGTTCAATGATGATCCA

qPCR, Quantitative polymerase chain reaction.

Table 4.

Primer Sequences Used for qPCR (SYBERGreen)

Gene	Forward primer	Reverse primer
CDK1	GGAAACCAGGAAGCCTAGCATC	GGATGATTCAGTGCCATTTTGCC
CCNA2	CTCTACACAGTCACGGGACAAAG	CTGTGGTGCTTTGAGGTAGGTC

qPCR, Quantitative polymerase chain reaction.

Microarray analysis

For the microarray analysis, total RNA was extracted using RNeasy mini kit from iPSCs and HPCs. The total RNA of HPCs was extracted after removing MEF using a Feeder Removal Microbeads mouse (Miltenyi Biotec). A detailed description of the methods was presented previously.²⁹ Briefly, we used 3D-Gene Human Oligo chip 25k (24,457 distinct genes, Toray Industries). Total RNA was labeled with Cy3 or Cy5 using the Amino Allyl MessageAMP IIaRNA Amplification Kit (Applied Biosystems). Genes with Cy3/Cy5 normalized ratios of >2.0 or <0.5 were defined as upregulated or downregulated, respectively. Pathways that showed significant gene expression changes were analyzed using GenMAPP ver2.1 (MAPP finder, http://www.genmapp.org/) software. The full microarray data was deposited to GEO of NCBI GEO. Access to the microarray data will be opened after the acceptance of this manuscript.

GSEA was performed using the Broad Institute GSEA tool software (http://broadinstitute.org/gsea/index.jsp) with standard settings, based on genes that were upregulated or downregulated at least 2 folds. Cluster analysis and visualization were performed with the heatmap package. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed with the clusterProfiler package.³⁰ Reactome pathway enrichment analysis was performed with the ReactomePA package. Dot plots of results were generated by the ggplot2 package. GSVA was performed with the GSVA package.³¹ Calculated enrichment scores of the pathways of replication stress signature that were generated by previous articles were visualized by heatmap. These data analyses and visualization were performed using R studio.

RNA-seq and Iso-seq

Sequencing libraries were prepared using the TruSeq Stranded mRNA Library Prep Kit (Illumina) following the manufacturer’s instructions. The libraries were sequenced on the Ilmina HiSeq 2500 sequencing platform using HiSeq SBS Kit v4 reagents with 2 × 10¹ cycles. The adapter sequences and low-quality regions were trimmed by Cutadapt and Trimmomatic, respectively. After pre-processing, the reads were mapped to human (GRCh38.p12) and HBV genomes by using TopHat/Bowtie. The mapped reads to HBV-KMT2B fusion genes were extracted by Samtools, then, were assembled using Cufflinks. Visualization of RNA seq coverage on KMT2B was made by IGV.

The PacBio Iso-Seq Long-Read sequencing was prepared with the standard protocol in “Procedure & Checklist Iso-Seq Express Template Preparation for Sequel and Sequel II Systems” (Pacific Biosciences). The library was prepared using the SMRTbell express template prep kit 2.0. Then, the result of the library was single molecule real-time (SMRT) sequenced using a Sequel II binding plate 2.0, sequel II sequencing plate 2.0 with a 20 h movie time on a PacBio Sequel II system using a SMRT cell 8M. Next, the Iso-Seq application was run in SMRT Link v9.0 (Pacific Bioscience) using the continuous long reads sub-read dataset and default parameters. Reads identified as full-length non-chimeric (FLNC) were considered for de novo clustering to generate unique isoforms. Individual Iso-seq reads of KMT2B and HBV-KMT2B fusion were visualized by IGV.

We obtained the RNA-seq data (GSE65485) from GEO to validate the upregulation of the replication stress signature. GSVA was conducted using the data of fragments per kilobase of transcript per million mapped reads (FPKM) on R studio. Furthermore, RNA-seq data of AAV2-integrated HCCs were obtained from the European Genome-Phenome Archive (EGA). The TPM of downloaded samples from (EGAD00001004484) and non-tumorous samples from GSE65485 were evaluated by salmon, then conducted GSVA using GSVA package on R-studio. The calculated enrichment scores of the pathways of replication stress signature were visualized by heatmap.

CUT&Tag analysis

For CUT&Tag analysis, HPCs were prepared by removing MEF using a Feeder Removal Microbeads mouse (Miltenyi Biotec). They were sent to Active Motif for CUT&Tag. Briefly, cells were incubated overnight with Concanavalin A beads and 1.3 μL of the primary anti-H3K4me3 antibody per reaction (Active Motif, catalog number 39159). After incubation with the secondary antibody (1:100), cells were washed and tagmentation was performed at 37 °C using protein-A-Tn5. Tagmentation was halted by the addition of EDTA, SDS, and proteinase K at 55 °C, after which DNA extraction and ethanol purification was performed, followed by PCR amplification and barcoding (see Active Motif CUT&Tag kit, catalog number 53160 for recommended conditions and indexes). Following SPRI bead cleanup (Beckman Coulter), the resulting DNA libraries were quantified and sequenced on Illumina’s NextSeq 550 (8 million reads, 38 paired end).

Reads were aligned using the BWA algorithm (mem mode; default settings). Duplicate reads were removed and only reads that mapped uniquely (mapping quality ≥1) and as matched pairs were used for further analysis. Alignments were extended in silico at their 3′-ends to a length of 200 bp and assigned to 32-nt bins along the genome. The resulting histograms (genomic “signal maps”) were stored in bigWig files. Peaks were identified using the MACS 2.1.0 algorithm at a cutoff of P-value 1e-7, without control file, and with the –nomodel option. Peaks that were on the ENCODE blacklist of known false ChIP-Seq peaks were removed. Signal maps and peak locations were used as input data to Active Motifs proprietary analysis program, which creates Excel tables containing detailed information on sample comparison, peak metrics, peak locations, and gene annotations. For differential analysis, reads were counted in all merged peak regions (using Subread), and each condition was compared using DESeq2. Average plot and heatmaps were generated by Seqplots package and enrichment analysis were performed by the ChIP-Enrich package,³² and the results were visualized using dot plot by the ggplot2 package.

Immunoblotting

Cultured cells or liver tissue samples were homogenized and lysed in RIPA buffer (50 mM, pH 8.0, Tris-HCl, 150 mM NaCl, 0.1% NP-40, 1% sodium deoxycholate, 1% sodium dodecyl sulfate) containing a Halt proteinase and phosphatase inhibitor cocktail (ThermoFisher) and benzonase nuclease (Millipore). After sodium dodecyl sulfate/polyacrylamide gel electrophoresis of cell homogenates (20 μg protein) and transfer to PVDF membranes, the blots were incubated with primary antibodies (listed in Table 5), followed by incubation with peroxidase-labeled secondary antibodies (GE Healthcare) and visualization using the ECL Western Blotting Analysis System (GE Healthcare). A detailed protocol for immunoblotting was previously described.²⁹

Table 5.

Antibodies Used for Immunostaining, Immunoblots, and FACS

	Dilution	Source	Catalog number
Primary antibodies for immunostaining
AFP	1/300	Spring	E-2954
ALB	1/1000	DAKO	A0001
HNF4α	1/300	Santa Cruz	sc-6556
Secondary antibodies for immunostaining
AlexaFluor-488 donkey anti-rabbit IgG	1/1000	Thermo Fisher	A21206
AlexaFluor-568 donkey anti-goat IgG	1/1000	Thermo Fisher	A11057
Primary antibodies for immunoblots
MLL2/KMT2B(D6X2E)	1/1000	Cell Signaling	63735
β-actin	1/5000	Sigma	A5441
Caspase-3	1/1000	Cell Signaling	9662
CDK1	1/2000	Abcam	ab18
Cyclin A2 (E1D9T)	1/1000	Cell Signaling	91500
H3K4me3-ChIP Grade	1/10000	Abcam	ab8580
Rabbit IgG,polyclonal-Isotype Control(ChIP Grade)	1/1000	Abcam	ab171870
γH2AX(p-S33)	1/1000	Cell Signaling	2577
ATR	1/1000	Cell Signaling	2790
ATM	1/1000	Abcam	ab32420
Chk1	1/1000	Abcam	ab40866
Chk2	1/1000	Cell Signaling	2662
Secondary antibodies for immunoblots
Anti-mouse IgG (HRP)	1/2000	GE Healthcare	NA931V
Anti-rabbit IgG (HRP)	1/5000	GE Healthcare	NA934V
Antibodies for flow cytometry analysis
TRA-1-60 (FITC)		BD Biosciences	560173
SSEA-4 (PE)		R & D Systems	FAB1435P
CD13 (BV421)		BioLegend	301715
CD133 (APC)		Miltenyi Biotec	130-113-670
CD44 (FITC)		eBioscience	11-0441-82

AFP, Alpha fetoprotein; ALB, albumin; CDK, cyclin-dependent kinase; FACS, fluorescence-activated cell sorter.

Immunoblotting Using a Capillary Electrophoresis Automated System (Simple Western)

Capillary electrophoresis automated system on Jess Simple Western system (Bio-techne) was used. Sample preparation, antibody application (antibody dilution; 1:50), and running conditions were performed according to the manufacturer’s instructions. In experiments, we applied 3 μg protein per a lane. The 12-230 kDa Wes Separation Module (Bio-techne) was used when Chk1, Chk2, and γH2AX were analyzed, whereas the 66-440 kDa Wes Separation Module was used when ATM and ATR were analyzed. Production of β-actin were analyzed using RePlex assay (Bio-techne) in every experiment.

Immunostaining

Cultured cells were fixed with 50% acetone/50% methanol and permeabilized with 0.5% Triton X-100/PBS. After blocking by 10% Blocking One (Nakalai)/PBS, cells were incubated in primary antibodies overnight at 4 °C. The cells were washed with PBS twice and were incubated with diluted secondary antibodies for 1 hour at room temperature. Then, the cells were washed with PBS twice and their nuclei were stained with DAPI. For each analysis, the addition of an appropriate immune serum was used as a negative control. The microscopic photographs were taken using a BZ-X710 microscope (Keyence). The primary and second antibodies and their dilution are listed in Table 5.

Quantification of the Intensity of γH2AX Immunofluorescence in HPCs

To quantify the immunofluorescence intensity of γH2AX in KMT2B-Int HPCs and KMT2B-HT HPCs, colonies of both HPCs were stained using anti-γH2AX and anti-HNF4α antibodies as described above. All images were acquired under the same setting of a confocal microscope (FV3000, Olympus). Images of γH2AX were converted into grayscale and the intensity of nuclei in each cell was obtained by using ImageJ. HNF4α positive cells were HPCs, and negative cells were MMC-treated feeder MEFs, which were positive control as DNA damaged cells. At least 100 cells were measured, and 3 independent staining and quantification experiments were conducted.

FACS

For flow cytometric analysis, cells were dissociated in 0.05% trypsin/0.5 mM EDTA and washed twice with PBS. The cells were incubated with an antibody and washed with 3% FBS/PBS. The antibodies used in this study are listed in Table 5. The sample incubation in 3% FBS/PBS alone was used as a negative control. The cells were then washed in 3% FBS/PBS and stained with PI to identify dead cells. BD Canto II (Becton Dickinson), BD Aria II (Becton Dickinson), or MoFlo XDP (Beckman Coulter) was used for the flow cytometric analysis. For cell sorting, cells were sorted using a BD Aria II or MoFlo XDP.

Cell Proliferation Assay

Human iPSCs were cultured in a 96-well culture plate at a density of 2 × 10³ cells per well (Day −1). On Days 0, 1, 2, and 3, Cell Titer 96 Aqueous One Solution (20 μL/well, Promega) was added to the plate, and the cells were again incubated for 1 hour at 37 °C and 5% CO₂. In each well, the absorbance at 490 nm was measured using a Glomax Discover Microplate Reader (Promega). Background absorbance was determined by wells containing medium alone and subtracted from that of the sample wells. The cell proliferation ratio of each cell line was calculated by the ratio of absorbance on Day 1, 2, and 3 to that on Day 0.

For HPCs, MEFs were seeded in 12-well plates (Day −1). After HPCs were purified by removing MEF using Feeder Removal Microbeads mouse (Miltenyi Biotec), HPCs were seeded there onto MEF at 3 × 10³ cells / well. Cell growth was measured from Day 0 to Day 9 (Day 0, 3, 6, and 9). Background absorbance was determined by wells containing MEF alone and subtracted from that of the sample wells. The cell proliferation ratio of each cell line was calculated by the subtraction of absorbance on Day 3, 6, and 9 to that on Day 0.

Colony Formation Assay

MEFs were seeded in 12-well plates (Day −1). For flow cytometric analysis, HPCs were dissociated in 0.05% trypsin/0.5 mM EDTA and washed twice with PBS, then treated with PI to identify dead cells. 3 × 10³ live cells were sorted on MEFs of every well, using BD Aria II (Becton Dickinson). After 10 days of cultivation, HPCs formed colonies. Then, the cells were washed with PBS and fixed with 4% paraformaldehyde/PBS, and permeabilized using 0.5% Triton X-100/PBS. The cells were washed with PBS, and their nuclei were stained with 4’,6-diamidino-2-phenylindole dihydrochloride (DAPI). Images were taken using a BZ-X710 microscope (Keyence). The number of colonies over 500 μm was counted.

ChIP-qPCR

High-sensitivity ChIP Kit (ab185913, Abcam) was used for ChIP according to the manufacturer's instructions. Briefly, 5 × 10⁵ cells or 10 mg for HCC specimens at room temperature were crosslinked in 9 mL cross-link solution (cell culture medium with 1% paraformaldehyde) for 10 minutes with on a rocking platform (50-100 rpm), and the reaction was terminated by the addition of 1 mL 1.25 M glycine for 5 minutes, followed by extracting the nuclear fraction using the lysis buffer. Crosslinked chromatin was sonicated to an average fragment size of 200–300 bp in ChIP buffer using a QSONICA (WAKEN). Sonicated chromatin was centrifuged for 10 minutes at 12000 rpm, and its supernatant was used on the antibody binding wells. Antibodies used for immunoprecipitation included anti-H3K4me3 and anti-rabbit IgG, which are listed in Table 5. The relative enrichment of promoters was determined by qPCR using the primers listed also in Table 6.

Table 6.

Primer Sequences Used for ChIP-qPCR

Gene	Forward primer	Reverse primer
CDK1	GAGGGCGAGTATTGAGGAAC	CTAAAGCCAGCCCAGTGTGG
CCNA2	CCGAAGACGAGACGGGTAAA	CATGCCTTCTCGCCACTCTT

ChIP-qPCR, Chromatin immunoprecipitation quantitative polymerase chain reaction.

Vector Construction

For generating KMT2B-KO iPSCs and KMT2B-HT iPSCs, the plasmid, pBS-polyA-LoxP-PuroR, which carries a poly(A) signal - SV40 promoter puromycin resistant gene cassette was used at the insertion site. To insert this cassette near the KMT2B start codon (located at exon 1) of host cells, knockout arms were designed up- and down-stream of exon1 and then cloned into pBS-polyA-LoxP-PuroR, as pBS-5′arm-polyA-LoxP-PuroR-3′arm. A plasmid expressing Cas9 (pSp-Cas9(BB)-2A-GFP(PX458)) was obtained from Addgene. Two different gRNAs were designed to target near the KMT2B start codon (located at exon 1). The target sequences for KMT2B were determined using the CRISPRdirect website (https://crispr.dbcls.jp/), and appropriate gRNA oligos with an adaptor sequence were synthesized and annealed. gRNAs were inserted into the gRNA scaffold site of pSp-Cas9(BB)-2A-GFP, and the plasmids carrying both Cas9 and gRNA were used for gene editing.

For the establishment of HBV-KMT2B integrated iPSCs, the plasmid, pUCFa-LoxP-polyA-TagRFP-PuroR, which carries a poly(A) signal – SV40 promoter puromycin resistant TagRFP gene cassette was used at the insertion site. To insert this cassette to the HBV genome integration site of KMT2B exon3 of host cells, the sequences were amplified by PCR as previously described and then cloned into the plasmid, as pUCFa-5′arm-HBV-KMT2B-LoxP-polyA-TagRFP-PuroR-3′arm. As with the establishment of KMT2B-KO iPSC lines, pSp-Cas9(BB)-2A-GFP was used. Two different gRNAs were designed to target the intron3 of KMT2B. The target sequences for KMT2B were determined and gRNAs were inserted into the plasmid as described in KMT2B-KO iPSC lines.

The lentiviral vector, CSII-EF-MCS-IRES2-Venus was used for overexpression assay for HBx-DC, #F2, #20, #33, #20-DPHD4, #20-DC, KMT2B-N-HBV as CSII-EF-HBx-DC-IRES2-Venus, CSII-EF-#F2-IRES2-Venus, CSII-EF-#20-IRES2-Venus, CSII-EF-#33-IRES2-Venus, CSII-EF-#20-DPHD4-IRES2-Venus, CSII-EF-#20-DC-IRES2-Venus, and CSII-EF-KMT2B-N-HBV-IRES2-Venus, respectively (Figures 13 and 16). The objective fragments were inserted into the MCS of the vector. Inserted sequences were artificially synthesized with partial synonymous substitution (FASMAC).

Overexpression Experiments in KMT2B-HT HPCs

Lentivirus was produced in HEK293T cells as previously described. In brief, HEK293T cells were seeded on a poly-L-lysine-coated plate and transfected at 80% confluency with expression and packaging vectors. Transfections were performed with lipofectamine 2000 according to the manufacturer’s protocol. The next day of transfection, the medium was changed to DMEM containing 10% FBS. Lentiviruses were harvested at 3 days after transfection and filtered with 0.45 μm and concentrated by Amicon ultra 30000 NMWL (Merck Millipore). To estimate the virus titer, NIH3T3 cells were infected, and titers were calculated by GFP-positive percentage of infected NIH3T3 cells. Concentrated virus solutions were stored at −80 ˚C until usage. For transduction, KMT2B-HT HPCs were seeded onto MEFs and infected with a lentiviral vector at a multiplicity of infection (MOI) 10. After 24 h infection, the culture medium, HPC medium, was changed. At 10 days after infection, cells were dissociated using 0.05% trypsin/0.5 mM EDTA, and GFP-positive cells were sorted onto MEFs at 3 × 10³ cells/well by using FACS. After 10 days of cultivation, the number of GFP-positive colonies over 500 μm was counted using a BZ-X710 microscope (Keyence).

Overexpression Experiments in iPS-Heps

KMT2B-HT iPSCs were differentiated into the hepatic lineage using Cellartis Definitive Endoderm Differentiation Kit and Cellartis Hepatocyte Differentiation Kit (Takara Bio) according to the manufacturer’s protocol. For transduction, on day 22, the gel covering the plate was dissociated by collagenase, and the cells were infected with a lentiviral vector at a MOI 10. After 24 hours of infection, the culture medium, hepatocyte maintenance medium, was changed. At 5 days after infection, cells were dissociated using TrypLE Select Enzyme (ThermoFisher), and GFP-positive cells were sorted onto MEFs at 1 × 10³ cells/well by using FACS. After 10 days of cultivation, they were fixed with 4% paraformaldehyde/PBS and stained with crystal violet. Stained areas were measured using a BZ-X710 microscope (Keyence).

For crystal violet staining, medium was aspirated from the plate. Cultured cells were fixed with 4% paraformaldehyde at room temperature for 15 minutes. The cells were washed with PBS twice and incubated with 0.5% crystal violet solution (Nacalai Tesque) at room temperature for 30 minutes. Then, the cells were washed with PBS twice and analyzed crystal violet stained area.

KMT2B Inhibition Experiment

MEFs were seeded in 24-well plates (Day −1). For flow cytometric analysis, KMT2B-HT and Int HPCs were dissociated in 0.05% trypsin/0.5 mM EDTA and washed twice with PBS, then stained with PI (ThermoFisher) to identify dead cells. 3 × 10³ live cells were sorted onto MEF of every well, using MoFlo XDP (Beckman Coulter) (Day 0). The culture medium with a KMT2B inhibitor MI-2 (Selleck Chemicals) was changed on Day 1 and Day 4. Control cells were treated with dimethyl sulfoxide (DMSO) only. On Day 6, colony numbers were analyzed.

CDK1 Inhibition Experiment

For flow cytometric analysis, KMT2B-HT and Int HPCs were dissociated in 0.05% trypsin/0.5 mM EDTA and washed twice with PBS, then stained with PI (ThermoFisher) to identify dead cells. 3 × 10³ live cells were sorted onto MEF of every well, using MoFlo XDP (Beckman Coulter, Day 0). The culture medium with a CDK1 inhibitor, Ro-3306 (Selleck Chemical), was changed on Day 1, Day 4, and Day 7. Control cells were treated with DMSO only. On Day 8, colony numbers were analyzed.

Quantification and Statistical Analysis

All error bars in the figures indicate the standard deviation (SD). The quantified data represent the findings of 3 or more independent experiments, except for the experiments using tumor samples (2 independent experiments were performed, because of sample usage limitation). Statistical analyses were performed using the Prism 9 software program (GraphPad). After assessing the normal distribution of the data using Shapiro-Wilk’s normality test, significant differences in comparisons of 2 groups were determined by 2-sided Student’s t-test, and comparisons of multiple groups were determined by 1-way analysis of variance (ANOVA). For data that were not normally distribution, the Mann-Whitney test was used for analysis. Analyses, using R, were performed on R studio, with the packages including heatmap, Clusterprofiler, ChIP-enrich, ggplot2, and GSVA.