A positive-feedback loop between HBx and ALKBH5 promotes hepatocellular carcinogenesis

This study was approved by the Ethical Committee of the First Affiliated Hospital of Kunming Medical University (Clinical Ethic Issue No: kmmu2019483) following guidelines of the Declaration of Helsinki. All the participants signed written informed consents, and they gave consent to have their data published. Twenty pairs of fresh HBV-HCC tissues with matched peri-tumor tissues from Chinese patients who were diagnosed as HBV-related primary HCC and normal liver tissues from patients with repair treatment of traumatic hepatic rupture were collected from the First Affiliated Hospital of Kunming Medical University to test the ALKBH5 mRNA and protein expression level. Seventy-nine pairs of Chinese HBV-HCC tissues and paried peritumor tissues from commercial tissue microarrays supplied by Shanghai Superbiotech (LD-LVC1805, Shanghai, China) were processed into immunohistochemical (IHC) staining together with integrated follow-up and clinical information data.

The human HCC cell lines L02, HepG2, MHCC97H, HepG2.2.15 were from ATCC (Manassas, USA). The primary human hepatocytes (PHH) were purchased from the iCELL company (Shanghai, China) and the HepG2-NTCP cells were previously constructed by our institute. The stable ALKBH5-knockdown (KD) cell lines of HepG2.2.15 were generated by lentivirus infection with puromycin selection according to the manufacturer’s protocol (Santa Cruz). Male athymic BALB/c nude mice (4 weeks old) were used for tumor formation assay. The animal experiments were approved according to the guidelines of the Animal Care and Use Committee of the First Affiliated Hospital of Kunming Medical University (IACUC Issue No: kmmu2019370), and the study was carried out in compliance with the ARRIVE guidelines. HepG2.2.15 cells with ALKBH5-KD or control were prepared in PBS and injected subcutaneously with 1 × 10⁷ cells of the nude mice. After 3 weeks, tumors were excised to measure weight.

HBV infection assay was performed according to previous reports [16]. Briefly, plasmid of wildtype HBV 1.1mer (carrying a greater-than-unit-length HBV genome) or mutation of HBV 1.1mer with HBx abrogation (the same plasmid with a stop codon for amino acid 7 of HBx) were transiently transfected into HepG2 cells and the supernatants were collected by polyethylene glycol (PEG) 8000 and centrifugated to enrich the viral particles. For the following HBV infection of HepG2-NTCP or PHH cells, media were supplemented with final 4% PEG-8000 plus HBV particles and added to the cells with 50–80% confluency, 3 days after HBV infection, cells were collected to conduct related experiments. Silencing RNA oligos for knockdown of HBx or WDR5 were synthesized according to our published literature [6]. The wild-type HBx 3’UTR and the mutated HBx 3’UTR (A1907C mutation, diminishing the site of m⁶A modification), namely HBx 3’UTR WT and HBx 3’UTR MT, respectively, overexpressing plasmids were constructed based on endonuclease digestion and T4 DNA ligation. The primers used in these procedures are listed in Table 1. RNA oligos or plasmids were transfected into cells using Lipofectamine 3000 reagent (Thermo Fisher Scientific, USA) and cells were collected after 48 h to be used in the related experiments.

RT-qPCR, Western blot and IHC were performed according to the previous reports [6, 17]. Antibodies: anti-ALKBH5 and anti-HBx antibodies (Abcam, USA) and anti-β-actin antibody (Cell Signaling Technology, USA). Primers sequences used for RT-qPCR were included in Table 1.

The ChIP assay was performed as previously described [18]. In brief, final 1% concentration formaldehyde fixed chromatin from 10 million HepG2.2.15 cells with different treatment and were sonicated to about 500 bp fragments, antibody-bound protein A/G Dynabeads (Invitrogen) were incubated with sonicated chromatin at 4 °C overnight. After 4 rounds of washing, the precipitated protein-DNA complex was reverse-crosslinked at 65 °C overnight. DNA were purified by Phenol chloroform. The ChIP-qPCR primers used in this study are presented in Table 1. Antibodies used in ChIP were H3K4me3 (ab8580, Abcam).

Cell proliferation was assessed by cell counting kit 8 (CCK8) assay (Beyotime, China). In brief, cells were seeded onto 96-well plates. CCK8 solution (10 μL/well) was added to the cells after incubating for the indicated time points, and the reaction product was quantified according to the manufacturer’s instructions. Transwell assays were performed using 24-well Transwell plates (8-μm pore size; Millipore, Burlington, MA, USA). 1 × 10⁵ cells were seeded in serum-free medium in the upper chamber, whereas medium supplemented with 20% fetal bovine serum was applied to the lower chamber as a chemoattractant. After 48 h of incubation, the migrated cells at the bottom surface of the filter were fixed, stained, and counted.

The MeRIP assay was conducted following the manufacturer’s procedure (Merck Millipore). In brief, total RNA were purified by TRIzol reagent (Japan, TAKARA) and mRNA were enriched and fragmented by the commercial kit (Invitrogen). m⁶A antibody conjugated proteinA/G Dynabeads were incubated with mRNA at 4 °C for 4 h. After washing, the precipitated mRNA was submitted to library construction or RT-qPCR assay. MeRIP-qPCR primers used in this study are presented in Table 1.

RNA-seq and Mass spectrum (MS) data of 159 liver samples (paired tumor and peri-tumor) from Fudan Chinese HBV-HCC data was downloaded through the URL: https://www.biosino.org/node/project/detail/↗ OEP000321. RNA-seq data of 115 HBV-HCC samples from The Cancer Genome Atlas (TCGA) were downloaded from the URL: https://cbioportal-datahub.s3.amazonaws.com/lihc_tcga.tar.gz↗ and the HBV infection and genotype information was based on ViralMine analysis [19]. MeRIP-seq and input RNA were sequenced and mapped to human genome using HISAT2 (http://daehwankimlab.github.io/hisat2↗). The m⁶A peaks were called by exomePeak (https://bioconductor.org/packages/exomePeak↗).

All statistical analysis was performed using GraphPad Prism 8 software (GraphPad Software, La Jolla, CA, USA). Unpaired or paired two-tailed Student’s t-test were applied to assess comparison between two groups; Welch’s ANOVA test was used for comparison among groups; Kaplan-Meier survival curves and log-rank tests were used to estimate survival and differences between groups; The Pearson’s chi-square test was used for correlation analysis with calculation of correlation coefficients (r); The statistical significance was indicated by *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 and not significant (NS) when P > 0.05.

Several studies reported the increased ALKBH5 in HCC based on TCGA dataset [20, 21] while other reports indicated decreased ALKBH5 in HCC [22, 23], suggesting ALKBH5 might function differently in different HCC background (such as HBV versus non-HBV related HCC). To define ALKBH5 expression pattern in HBV-HCC, we collected 20 pairs of tumor and peri-tumor tissues with HBV infection history from Chinese patients. ALKBH5 mRNA and protein levels were significantly increased in HCC tissues compared to peri-tumor (Fig. 1A and B). IHC assay further confirmed that ALKBH5 was mainly stained in the nuclei of tumor cells and there was the highest expression of ALKBH5 in tumor cells compared with moderate expression in the peritumor tissues and the negative staining in normal liver tissues (Fig. 1C). We further use a tissue microarray including another 79 pairs of tumor and peri-tumor from Chinese HBV-HCC patients to assert the ALKBH5 expression pattern. The IHC score of ALKBH5 was predominantly higher in tumors than their paired peri-tumors from the 79 Chinese HBV-HCC patients (Fig. 1D). And the higher IHC score of ALKBH5 in HBV-HCC tissue microarray predicted a poor prognosis (Fig. 1E).

To expand the sample size and dissect ALKBH5 role in HBV-HCC patients, we data mined a recently comprehensive Fudan university’s study of HBV-HCC including 159 Chinese patients’ transcriptome and proteome [24]. The ALKBH5 mRNA and protein level in tumors and peri-tumors from each of 159 Chinese patients were extracted to assess ALKBH5 expression pattern. The average mRNA and protein level of ALKBH5 were significantly higher in tumors than in peri-tumors of the 159 Chinese patients (Fig. 2A). 115/159 of Fudan HBV-HCC samples displayed a higher ALKBH5 mRNA level, 96/159 for higher ALKBH5 protein level, in tumor compared to paired peri-tumor tissues and there were significant differences by paired t test (Fig. 2B). Considering that Fudan HBV-HCC samples come all from Chinese patients and HBV genotype B and C are more popular in China than other genotypes [25], we further data mined TCGA HBV-HCC datasets to explore whether the expression level of ALKBH5 in the HBV background is related to different ethnicity or different HBV genotypes. We found that the ALKBH5 level was higher in Asian groups than White and Black groups (Fig. 2C); However, there’s no significant difference of the ALKBH5 expression level among HBV genotypes B, C, and other phenotypes (Fig. 2D). These TCGA HBV-HCC results suggest that the ALKBH5 expression might be correlated to ethnicity but not the HBV phenotype.

Survival analysis of both Fudan and TCGA HBV-HCC datasets indicated that patients with higher mRNA and protein levels of ALKBH5 had markedly less total survival rate than patients with lower ALKBH5 expression (Fig. 2E and F). Furthermore, significant positive associations of ALKBH5 expression were identified with gender, tumor size, AFP level, TNM stage and HCC proliferation subtype, whereas there was no significant association with age, liver cirrhosis, tumor encapsulation and tumor number (Table 2). Taken together, ALKBH5 was highly expressed in most HBV-HCC tissues and could predict poor prognosis for HBV-HCC patients.

As HBV could dysregulate host genes, we next explored whether HBV causes the increased ALKBH5 expression. Firstly, we checked ALKBH5 protein level in L02, HepG2, HepG2.2.15 and MHCC97H cell lines as the latter two cell lines contain HBV genome and produce HBV virus compared to the first two. Western blot indicated a higher ALKBH5 expression in HBV-infected cell lines (Fig. 3A) which implied the regulation relationship between HBV and ALKBH5. To directly assess the effects of HBV on ALKBH5 expression, we conducted in vitro HBV infection assay on PHH cells and HepG2-NTCP cells. Wild type HBV virion (HBV WT) could increase the ALKBH5 protein level while a mutation of HBx-depletion HBV virion (HBV MT) could not (Fig. 3B and C). HBx is not only vital to boost HBV transcription and replication, but also play critical roles in liver carcinogenesis by directly dysregulating host oncogenes and tumor suppressing genes. We further observed there were positive correlations between HBx and ALKBH5 in the HBV-HCC tissue (Fig. 3D), and we thus check whether overexpression of HBx promotes ALKBH5 expression. Indeed, HBx up-regulates both mRNA and protein level of ALKBH5 in HepG2 cells after HBx-overexpression plasmids transfection (Fig. 3E and F).

Our previous study indicated that HBx hijacks WDR5 to promote target oncogenes’ promoter H3K4me3 modification [6], which suggests that the increased ALKBH5 expression in HBV-HCC might be mediated through the HBx-WDR5-H3K4me3 axis. After re-analyzing the HBx, WDR5 and H3K4me3 ChIP-seq data in our previous study, we found that HBx and WDR5 binds to the promoter of ALKBH5 gene with the H3K4me3 modification as shown in the UCSC genome browser (Fig. 4A). Furthermore, knocking down either HBx or WDR5 by small interfering RNA transfection decreased H3K4me3 modification level of ALKBH5 gene’s promoter as evidenced by ChIP-qPCR (Fig. 4B). We further data mined 159 Fudan HBV-HCC cohort to analyze the correlation between WDR5 and ALKBH5 expression, we found that WDR5 was indeed increased in the HBV-HCC at both mRNA and protein level (Fig. 4C) and the increased WDR5 level predicted a poor prognosis in accordance with previous reports [6] (Fig. 4D, Left panel). Moreover, the expression of WDR5 were positively related to the expression of ALKBH5 in the 159 Fudan HBV-HCC cohort (Fig. 4D, Right panel). The relationship between WDR5 and ALKBH5 was further defined by knockdown of ALKBH5 in HepG2.2.15 cells, and we observed that the down-regulated expression of WDR5 after ALKBH5 knockdown (Fig. 4E). These results support the notion that HBV up-regulates ALKBH5 via the HBx-WDR5-H3K4me3 axis.

As ALKBH5 is highly expressed in HBV-HCC, we wondered the tumorigenic function of ALKBH5 in HBV-induced carcinogenesis. Silencing ALKBH5 inhibited proliferation rates of both HepG2 and HepG2.2.15 cells (Fig. 5A and B). Correspondingly, transwell assay showed that ALKBH5 supported tumor’s migration ability of both HepG2 and HepG2.2.15 cells (Fig. 5C). These silencing results indicated that ALKBH5 functions as the oncogene in HBV-independent manner. The role of ALKBH5 in HCC was further addressed in a nude mouse model. When HepG2.2.15 cells with stable ALKBH5 knockdown were transferred to the subcutaneous of nude mice, the sizes of xenografted tumors and the weights decreased than the control group (Fig. 5D and E). These results illustrated the tumor-promoting activity of HBV might be mediated, at least partially, by the ALKBH5 in HBV-HCC cells.

A recently published study reported that HBV encodes RNA including pre-genomic RNA with m⁶A modification at the 3’UTR and 5′ UTR, and the m⁶A modification of 3’UTR affects the HBV RNA stability in HBV expressing HepG2 cells [15]. Therefore, the increased ALKBH5 in HepG2.2.15 cells of the present study probably involve in epitranscriptomics regulation of HBV RNA. We thus use MeRIP-seq for HepG2.2.15 to detect whether there are m⁶A modification on HBV RNA in our study. We finally detected one major m⁶A peak on HBV RNA, similar with the previous report [15] (Fig. 6A). As HBx is the determinant of both HBV replication and HBV-induced carcinogenesis, we focused on the m⁶A modification on HBx 3’UTR. Firstly, we constructed both wildtype and A1907C mutation of HBx 3’UTR overexpression plasmids, and transfected the plasmids into HepG2 cells. MeRIP-qPCR validated that there was m⁶A modification on HBx 3’UTR WT mRNA compared to the positive control (Fig. 6B). Furthermore, the m⁶A modification level was dramatically decreased at the HBx 3’UTR MT mRNA than the WT (Fig. 6C). We also noticed that HBx 3’UTR MT mRNA processed more stability than the WT by RNA decay assay (Fig. 6D), which indicates that m⁶A modification on HBx 3’UTR promotes its mRNA decay. To further directly determine whether the ALKBH5 affect the m⁶A modification of HBx mRNA, we knocked down ALKBH5 in the HepG2-HBx 3’UTR WT cells, and the MeRIP-qPCR assay indicated that ALKBH5 silencing induced a higher m⁶A modification of HBx mRNA (Fig. 6E), in accordance with the decreased protein level of HBx (Fig. 6F), which suggests that ALKBH5 could regulate HBx expression level by modifying m⁶A level of HBx mRNA. In addition, we also conducted western blot assay for WDR5 and H3K4me3 with or without knockdown of ALKBH5 gene (Fig. 6F), and we found that silencing of ALKBH5 decreased not only HBx protein but also WDR5 and H3K4me3 level, which indicated that ALKBH5 might participate in the epigenetic regulation of HBx-WDR5-H3K4me3 axis [6].

Therefore, we verified that HBx up-regulates ALKBH5 by the histone modification of H3K4me3 while ALKBH5 increases HBx via the mRNA modification of m⁶A, which may form a positive loop between HBx and ALKBH5 (Fig. 6G).

we confirm that all methods were carried out in accordance with relevant guidelines and regulations.

This study involving human data and human tissue was approved by the Ethical Committee of the First Affiliated Hospital of Kunming Medical University (Clinical Ethic Issue No: kmmu2019483) following guidelines of the Declaration of Helsinki. All the patients signed written informed consents, and they gave consent to have their data published. The animal experiments were approved by the Animal Care and Use Committee of the First Affiliated Hospital of Kunming Medical University (IACUC Issue No: kmmu2019370) in compliance with the ARRIVE guidelines.

Not applicable.

The authors declare no conflict of interest.

RNA-seq and Mass spectrum (MS) data of 159 liver samples (paired tumor and peri-tumor) from CHCC-HBV patients are available downloaded through the URL: https://www.biosino.org/node/project/detail/↗ OEP000321. RNA-seq data of 115 HBV-HCC samples from TCGA were downloaded from the URL: https://cbioportal-datahub.s3.amazonaws.com/lihc_tcga.tar.gz↗ and the HBV infection and genotype information were downloaded from the URL: https://ars.els-cdn.com/content/image/1-s2.0-S2589004221003369-mmc2.zip↗. MeRIP-seq and input RNA datasets generated and analysed during the current study are not publicly available due other relevant study is still in progress but are available from the corresponding author on reasonable request.