Evaluation of ALKBH2 and ALKBH3 gene regulation in patients with adult T-cell leukemia/lymphoma Authors: Yuji Wada, Tadasuke Naito, Takuya Fukushima, Mineki Saito Published: 2024 ABSTRACT BackgroundHuman T-cell leukemia virus type 1 (HTLV-1) is an oncogenic virus that causes malignant adult T-cell leukemia/lymphoma (ATL). Patients infected with HTLV-1 are considered HTLV-1 carriers, and a small proportion of patients progress to life-threatening ATL after a long asymptomatic phase. No antiviral agent or preventive vaccine specific for HTLV-1 infection is established in current situation. For development of countermeasures to combat HTLV-1 infection and ATL, it is essential to expand our knowledge about their pathogenesis. Recently, AlkB homolog (ALKBH) family have been shown to participate in the oncogenesis of various cancer types.MethodsTo investigate the potential role of ALKBH family members in the pathogenesis of ATL, we analyzed their gene expression dynamics in HTLV-1-infected T-cell lines and peripheral blood mononuclear cell-derived clinical specimens obtained from asymptomatic HTLV-1 carriers and patients with acute-type ATL. Epigenetic analysis was performed to dissect the mechanisms of ALKBH3 gene regulation using cultivated cells and a public dataset.ResultsThe mRNA expression levels of ALKBH2 and ALKBH3 were significantly or suggestively decreased in asymptomatic HTLV-1 carriers, but reverted in acute-type ATL patients, correlating with HTLV-1 basic leucine zipper factor gene expression. Intriguingly, the pre-mRNA expression of ALKBH2 and ALKBH3 was significantly suppressed in patients infected with HTLV-1, but not in healthy controls. Epigenetic analysis was performed to dissect the mechanisms of ALKBH3 gene regulation. In vitro analysis suggested a possible relationship between DNA methylation and ALKBH3 gene expression. Investigation of a public dataset revealed that specific CpG sites exhibited characteristically regulated methylation states in HTLV-1-infected T-cell subsets.ConclusionWe discovered dynamically regulated patterns of ALKBH2 and ALKBH3 gene expression in patients infected with HTLV-1, and specific CpG sites epigenetically regulated by HTLV-1 infection. This study provides novel insights into HTLV-1 infection and contributes to the elucidation of ATL pathogenesis.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12985-024-02590-w. Background Human T-cell leukemia virus type 1 (HTLV-1) is an oncogenic retrovirus belonging to the genus Deltaretrovirus, family Retroviridae. HTLV-1 is widespread worldwide and estimated to infect 5–10 million people [1, 2]. Well-documented endemic areas include southwestern Japan, central and western Africa, Iran, the Caribbean islands, South America, Australia, and Melanesia [3–5]. The World Health Organization currently promote global activities for awareness and prevention of HTLV-1 infection, and epidemiological surveillance was activated in non-endemic areas such as Europe and North America [6–8]. Cell-to-cell contact is essential for HTLV-1 infection, and three major transmission routes are currently known: mother-to-child transmission, mainly via breastfeeding; sexual transmission; and unscreened blood donation [3, 9]. HTLV-1 infection is established by the integration of viral genes into the host genome, mainly targeting CD4-positive T-cells, and infected patients are considered HTLV-1 carriers [3, 9]. The majority of HTLV-1 carriers are asymptomatic; however, less than 10% of cases cause harmful HTLV-1-associated diseases, such as adult T-cell leukemia/lymphoma (ATL) and HTLV-1-associated myelopathy/tropical spastic paraparesis, after decades of latency [3, 9]. The HTLV-1 genome encodes the retroviral structural and accessory proteins. In particular, the Tax and HTLV-1 basic leucine zipper factor (HBZ) proteins have been extensively investigated as oncogenic viral proteins that play critical roles in HTLV-1 pathogenesis [10]. ATL is classified into four subtypes, depending on its clinical manifestations: acute, lymphoma, chronic, and smoldering [3, 9, 11]. Acute and lymphoma are classified as aggressive types owing their high malignancy, with a median survival of less than 1 year [3, 9, 12]. Retrospective research performed in Japan reported that 4-year overall survival rates of acute, lymphoma, chronic, and smoldering types of ATL were 11%, 16%, 36%, and 52%, respectively [13]. Recent research has greatly contributed to the development of ATL prophylaxis, therapeutics, and diagnosis. For example, Kobayashi et al. reported that the population of a certain subtype of HTLV-1-infected T-cells in peripheral blood mononuclear cells (PBMCs) reflects disease progression [14, 15]. Yamagishi et al. and Kataoka et al. further suggested that the accumulation of genetic mutations causes the clonal expansion of ATL cells [16–18]. Elucidation of the pathogenesis of ATL is essential for developing countermeasures against this disease; therefore, further investigations are needed to clarify the genomic, molecular, and cellular mechanisms related to the emergence and propagation of ATL cells. The AlkB homolog (ALKBH) family comprises nine homologous molecules (ALKBH1–8 and FTO), which share alpha-ketoglutarate and Fe2+-dependent dioxygenase [19, 20]. ALKBH family members mainly repair 1-methyladenine (1-meA) and/or 3-methylcytosine (3-meC) lesions, which potentially cause somatic mutations in the host genome by targeting various conformations of DNA and RNA [19–26]. Although precise mechanisms of tumorigenesis remain unclear, some ALKBH family members have been intensively investigated as novel oncogenic and/or malignancy-related factors [27]. Representatively, Chen et al. revealed that ALKHB3 promoted cancer cell proliferation and invasion in vitro and in vivo [23]. To the best of our knowledge, no study has evaluated ALKBH family members as pathogenic factors of ATL. Thus, we explored the possible relationship between ALKBH family members and ATL using HTLV-1-infected T-cell lines and patient-derived PBMCs. According to the results of gene expression screening in vitro, ALKBH family genes exhibited a trend of low expression in HTLV-1-infected T-cell lines. Analysis of clinical specimens revealed that the mRNA expressions of ALKBH2 and ALKBH3 were regulated depending on the disease stage of ATL, correlating with HBZ mRNA expression. It was also discovered discrepancies between the pre-mRNA and mRNA expression of ALKBH2 and ALKBH3. To further advance HTLV-1 and ATL research, characteristically modified epigenomes were identified in patients infected with HTLV-1. Our findings provide novel insight into genetically and epigenetically regulated host factors dependent on HTLV-1 infection, and would contribute to elucidate pathogenesis of ATL. Methods Cells Three human leukemic T-cell lines (Jurkat, CEM, and Molt4) and five HTLV-1-infected human T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1) were maintained in RPMI 1640 medium (Sigma-Aldrich, St. Louis, MO, USA) containing 10% fetal calf serum (FCS), 100 unit/mL penicillin, and 100 μg/mL streptomycin (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) at 37 °C in 5% CO2. Jurkat, CEM and Molt4 cells were obtained from Riken BioResource Research Center Cell Bank (Tsukuba, Japan). MT1, MT2, MT4 and HUT102 cells were kindly provided by Prof. Masahiro Fujii at Niigata University [28–31]. TLOm1 cells were kindly provided by Prof. Masao Matsuoka at Kumamoto University [32]. No mycoplasma contamination was observed during the cell culture process. Peripheral blood was collected from 16 healthy volunteers (healthy control; HC), 11 asymptomatic HTLV-1 carriers (ACs), and 10 acute-type ATL patients with precise informed consent and written permission as previously described [33]. All patients were clinically diagnosed according to the Shimoyama criteria [11]. PBMCs were isolated using Histopaque-1077 (Sigma-Aldrich) immediately after blood collection. After washing with RPMI 1640 medium containing 10% FCS, the PBMCs were stored in liquid nitrogen until use. RNA extraction and qRT-PCR assay RNA was extracted from cultured cell lines using NucleoSpin RNA Plus (Takara Bio Inc., Shiga, Japan) and from frozen PBMCs using the AllPrep DNA/RNA/Protein Mini Kit (QIAGEN, Hilden, Germany). The RNA was then used for cDNA synthesis, using the PrimeScript® RT Reagent Kit (Takara Bio Inc.) with random primers. All procedures were performed in accordance with the manufacturer’s protocol. Host and viral RNA expression was quantified by qRT-PCR using SYBR Green dye. Target genomes and β-actin mRNA, used as an internal control, were amplified using TB Green Premix EX Taq II (Tli RNase Plus, Takara Bio Inc.) in an Mx3000P QPCR system (Agilent Technologies, Santa Clara, CA, USA) following the manufacturer’s protocol. The primers used in this study were synthesized following previous reports or made in-house and are listed in the S1 Fig and S1 Table [34–38]. All experiments were performed in duplicate. The obtained gene expression values of target genes were normalized by that of β-actin mRNA, and equalized gene expression values (2^−ΔΔCt) were calculated comparing to control group (Jurkat, HC, and AC in respective experiment), as following ΔΔCt method [39]. Finally, the calculated 2^−ΔΔCt values were standardized by the mean value of 2^−ΔΔCt in the control group, and relative expression values were calculated (control group = 1). The melt curve was manually checked in all assays, and data were omitted from the analysis when nonspecific amplification was expected. Western blotting Total proteins were extracted from 1 × 106 cells of the cultured cell lines using 100 µl of lysis buffer (50 mM Tris–HCl [pH 7.5], 200 mM NaCl, 1mM EDTA, 10% Glycerol, 1% NP-40) with fresh cOmplete Protease Inhibitor Cocktail (Sigma-Aldrich). Subsequently, an equal amount of sample buffer (100 mM Tris–HCl [pH 6.8], 4% SDS, 20% Glycerol, 0.2% bromophenol blue, 20% 2-mercaptoethanol) was added and boiled for 3 min at 95 °C. Proteins were separated by electrophoresis on a 10% acrylamide gel which was created by TGX Stain-Free Fast Cast Acrylamide Kit (Bio-Rad, Hercules, CA), and then transferred to a 0.2µm PVDF membrane by Trans-Blot Turbo Transfer (Bio-Rad) and Trans-Blot Turbo Transfer Pack (Bio-Rad) in accordance with the manufacturer’s protocol. The membranes were blocked using Blocking One (NACALAI TESQUE, INC, Kyoto, Japan) for 1 h at room temperature with shaking. After washing with PBS containing 0.05% Tween 20 (PBS-T), the blocked membrane was incubated with rabbit anti-ALKBH2 monoclonal antibody (clone: EPR6177(2), Abcam, Cambridge, UK), rabbit anti-ALKBH3 polyclonal antibody (clone: 12292-1-AP, Proteintech Group, Inc., Rosemont, IL) or rabbit anti-beta actin polyclonal antibody (clone: 20536-1-AP, Proteintech Group) in antibody dilution buffer (PBS-T containing 5% blocking solution) for 1 h at room temperature with shaking. The stained membranes were washed with PBS-T and incubated with Anti-rabbit IgG, HRP-linked Antibody (clone: 7074, Cell Signaling Technology, Inc., Danvers, MA) for 1 h at room temperature with shaking. After washing the incubated antibody, the proteins were colored by Western BLoT Ultra Sensitive HRP Substrate (Takara Bio Inc.) and detected by a LAS-3000 imaging system (Fujifilm, Tokyo, Japan). The band intensity was calculated using the Image J software. Exploration of HTLV-1 PVL HTLV-1 PVL in PBMCs was obtained from a data set previously published by our research group [40]. Briefly, copy numbers of HTLV-1 tax were quantified using genomic DNA extracted from PBMCs, and normalize by that of β-actin. Then, HTLV-1 tax copy numbers per 1 × 104 PBMCs were calculated and shown as HTLV-1 PVL. All samples were analyzed in triplicate. Methylation analysis of genomic DNA The methylation state of CpG sites around the TSS of ALKBH3 was analyzed using the bisulfate method, according to a previously established protocol [41]. Total DNA was extracted from cultured T-cells using NucleoSpin Tissue (Takara Bio Inc.), and the extracted DNA was used for bisulfate conversion using the EZ DNA Methylation-Gold Kit (Zymo Research, Irvine, CA, USA). PCR was performed using the TaKaRa Taq Hot Start Version (Takara Bio Inc.), and the primers are listed in the S1 Table [41]. The amplified target genomes were gel-extracted and purified using NucleoSpin Gel and PCR Clean-up (Takara Bio Inc.). Subsequently, TA cloning was performed using the Mighty TA-cloning kit and E. coli HST08 Premium Competent Cells (Takara Bio Inc.). The cloned plasmids were purified using NucleoSpin Plasmid EasyPure (Takara Bio Inc.), and sequence analysis was performed using a commercial DNA sequencing service (Fasmac Co., Ltd., Kanagawa, Japan). The obtained results were analyzed using GENETYX ATGC software (Genetyx Corporation, Tokyo, Japan) and QUMA, a quantification tool for methylation analysis. All procedures were performed in accordance with the manufacturer’s protocol. Demethylation assay in vitro Decitabine (DAC, Sigma-Aldrich) was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 50 μM, and aliquots were stored at −30 °C. MT1 cells (5 × 104) were seeded into 24-well plates to establish a suspension culture (Sumitomo Bakelite Co., Ltd., Tokyo, Japan) containing 0.01–10 nM DAC and cultivated for 2 days at 37 °C in 5% CO2 (500 μL of culture medium per well). Total DNA and RNA were extracted from cells after washing with PBS. Subsequently, the gene expression levels and methylation states of CpG sites were analyzed as described above. As a control, an equal amount of DMSO containing 1 nM of DAC was co-cultivated with MT1 cells under the same conditions. The cell toxicity of up to 100 nM of DAC and an equal amount of DMSO was evaluated by microscopical observation and qRT-PCR quantification, with β-actin mRNA used as a housekeeping gene. Investigation of genome-wide DNA methylation status in patients infected with HTLV-1 using a public dataset We utilized the Gene Expression Omnibus to explore the DNA methylation states of CpG sites around the TSS of ALKBH3 in HTLV-1-infected patients (GSE136189). Statistical analysis Statistical significance was calculated using an unpaired Student's t-test, and one-way ANOVA followed by Tukey’s multiple comparisons test. Correlation analysis was performed using the Pearson correlation coefficient by only exploiting the data that were detectable in both analyzed parameters. GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA) was used for all calculations. Experiments with cultured cell lines were performed using at least three biological replicates. Study approval This study was approved by the Research Ethics Committee of Kawasaki Medical School (approval number: 3447-01). The study was fully explained to all participants orally and in writing. Written informed consent was obtained from all participants. All procedures using human subjects were performed in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan, and Declaration of Helsinki. Results Screening of ALKBH family gene expression in cultured T-cell lines First, we measured the mRNA expression of ALKBH family genes in HTLV-1-infected and uninfected T-cell lines to investigate whether their gene expression was affected by HTLV-1 infection. The mRNA expression of the ALKBH family genes was similar in the three uninfected T-cell lines (Jurkat, CEM, and Molt4), except for that of ALKBH4, which was not detected in any of the assessed T-cell lines (Fig. 1a). The other five HTLV-1-infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1) exhibited relatively low expression levels compared to the uninfected ones. To clarify the impact of HTLV-1 infection on the gene expression of ALKBH family genes, the mean values of their gene expression were analyzed after dividing into two groups, which were uninfected (including Jurkat, CEM, and Molt4) and HTLV-1 (including MT1, MT2, MT4, HUT102, and TLOm1) groups. The mean values of ALKBH1, ALKBH2, ALKBH3, ALKBH5, ALKBH6, ALKBH7, ALKBH8, and FTO mRNA expression were 2.47, 2.29, 5.92, 1.69, 1.89, 2.64, 2.05, and 7.00 times lower in HTLV-1 group, respectively (S2 Fig. a). Statistical significance was shown in the analyzed ALKBH family genes, except for ALKBH8. Remarkably, the mRNA expression of ALKBH3 was undetectable in MT1 cells and differed in a statistically significant manner in multiple pairs of HTLV-1-infected and uninfected T-cell lines. ALKBH3 mRNA expression in CEM cells was 11.91 times higher than that in HUT102 cells; in Molt4 cells, its expression was 10.84 and 12.71 times higher than that in MT4 and HUT102 cells. It was also observed that protein expressions of ALKBH3 were undetectable or disturbed in MT1, MT4 and HUT102 cells; however, no statistical significance was not observed between the uninfected and HTLV-1 groups (Fig. 1b and S2 Fig. b). ALKBH3 was recently recognized as an oncogenic molecule present in various types of cancers [41–43]; therefore, we investigated ALKBH3 further.Fig. 1Screening of ALKBH family genes by comparing mRNA expression levels between HTLV-1-infected and uninfected T-cell lines. a The mRNA expression of ALKBH family genes was measured by qRT-PCR in uninfected (Jurkat, CEM, and Molt4) and HTLV-1-infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1). Gene expression was calculated using the ΔΔCt method. The equalized data are shown as relative expression, standardized by the mean value of Jurkat cells (Jurkat = 1) on the y-axis. The data represent the mean ± SD of four independent replicates. Significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test and is indicated by an asterisk. Statistical significance was observed in the analysis of ALKBH3 (CEM vs. HUT102: *p = 0.0467, Molt4 vs. MT4: *p = 0.0313, Molt4 vs. HUT120: *p = 0.0280). b ALKBH2, ALKBH3 and β-actin protein expressions in cultured T-cell lines were analyzed by western blotting. The figure is a representative of the independent experiments. Original images of western blotting were shown in S2 Fig. a. The proteins detection and imaging were performed by a LAS-3000 imaging system. The equalized data are shown as relative band intensity (Jurkat = 1) above the detected-bands, respectively. The graphs represent the mean ± SD of the relative band intensity in uninfected and HTLV-1 groups. Significance was analyzed by unpaired Student's t-test (ALKBH2: p = 0.1503, ALKBH3: p = 0.4797). Uninfected, group of uninfected T-cell lines; HTLV-1, group of HTLV-1–infected T-cell lines Screening of ALKBH family genes by comparing mRNA expression levels between HTLV-1-infected and uninfected T-cell lines. a The mRNA expression of ALKBH family genes was measured by qRT-PCR in uninfected (Jurkat, CEM, and Molt4) and HTLV-1-infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1). Gene expression was calculated using the ΔΔCt method. The equalized data are shown as relative expression, standardized by the mean value of Jurkat cells (Jurkat = 1) on the y-axis. The data represent the mean ± SD of four independent replicates. Significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test and is indicated by an asterisk. Statistical significance was observed in the analysis of ALKBH3 (CEM vs. HUT102: *p = 0.0467, Molt4 vs. MT4: *p = 0.0313, Molt4 vs. HUT120: *p = 0.0280). b ALKBH2, ALKBH3 and β-actin protein expressions in cultured T-cell lines were analyzed by western blotting. The figure is a representative of the independent experiments. Original images of western blotting were shown in S2 Fig. a. The proteins detection and imaging were performed by a LAS-3000 imaging system. The equalized data are shown as relative band intensity (Jurkat = 1) above the detected-bands, respectively. The graphs represent the mean ± SD of the relative band intensity in uninfected and HTLV-1 groups. Significance was analyzed by unpaired Student's t-test (ALKBH2: p = 0.1503, ALKBH3: p = 0.4797). Uninfected, group of uninfected T-cell lines; HTLV-1, group of HTLV-1–infected T-cell lines Quantification of ALKBH2 and ALKBH3 gene expression in PBMCs derived from patients infected with HTLV-1 To assess the gene expression of ALKBH3, we used total RNA extracted from PBMCs collected from 16 HCs, 11 ACs, and 10 patients with acute-type ATL. The ALKBH3 coding region overlapped with an antisense non-coding RNA, named ALKBH3-AS1. To ensure accurate evaluation, we individually quantified the gene expression of ALKBH3 mRNA, ALKBH3 pre-mRNA, and ALKBH3-AS1 to distinguish among the detected gene species (S1 Fig and S1 Table). The gene expression of ALKBH3-AS1 could not be confirmed in our investigation; therefore, we considered that they would not affect the gene expressions of other genes and omitted it from our analysis. The ALKBH3 mRNA expression in the AC group was 1.77 and 1.69 times lower than that in the HC and ATL groups, respectively, but the differences were not statistically significant (Fig. 2a). Conversely, compared to that in the HC group, the gene expression of ALKBH3 pre-mRNA was 11.24 and 3.59 times lower in the AC and ATL groups, respectively; these differences were statistically significant. The detection rates of the analyzed genes also supported the observation that ALKBH3 mRNA was highly expressed in all specimens, whereas ALKBH3 pre-mRNA was not detected in the AC and ATL groups (Table 1).Fig. 2Analysis of ALKBH2 and ALKBH3 gene expression using PBMC specimens derived from patients infected with HTLV-1. a, b The gene expression levels of ALKBH2 and ALKBH3 were calculated using the ΔΔCt method. The data are shown as relative expressions, standardized by the mean value of the HC group (HC = 1). All experiments were performed in duplicate. Significance was analyzed by one-way ANOVA followed by Tukey's multiple comparisons test and is indicated by an asterisk. Statistical significance was observed for ALKBH3 pre-mRNA (HC vs. AC: **p = 0.0013, HC vs. ATL: *p = 0.0152), ALKBH2 mRNA (HC vs. AC: *p = 0.0127), and ALKBH2 pre-mRNA (HC vs. AC: **p = 0.0091, HC vs ATL: *p = 0.0173). Scatter plots represent the median values. c, d Correlation between the mRNA and pre-mRNA expression of ALKBH2 (left panel) and ALKBH3 (right panel) in the HC group are shown. Correlation analysis was performed using the Pearson correlation coefficient, and data are shown using a simple linear regression model. The p-values and Pearson r2 values are shown on the panels (ALKBH2: ****p < 0.0001 and R = 0.6858, ALKBH3: ***p = 0.0004 and R = 0.6027). HC, healthy control; AC, asymptomatic HTLV-1 carrier; ATL, acute-type adult T-cell leukemiaTable 1Detection efficiency of ALKBH3 transcriptsmRNApre-mRNAHCACATLHCACATLNumber of detected/analyzed16/1611/1110/1016/168/117/10Detection rate (%)10010010010072.770.0 Analysis of ALKBH2 and ALKBH3 gene expression using PBMC specimens derived from patients infected with HTLV-1. a, b The gene expression levels of ALKBH2 and ALKBH3 were calculated using the ΔΔCt method. The data are shown as relative expressions, standardized by the mean value of the HC group (HC = 1). All experiments were performed in duplicate. Significance was analyzed by one-way ANOVA followed by Tukey's multiple comparisons test and is indicated by an asterisk. Statistical significance was observed for ALKBH3 pre-mRNA (HC vs. AC: **p = 0.0013, HC vs. ATL: *p = 0.0152), ALKBH2 mRNA (HC vs. AC: *p = 0.0127), and ALKBH2 pre-mRNA (HC vs. AC: **p = 0.0091, HC vs ATL: *p = 0.0173). Scatter plots represent the median values. c, d Correlation between the mRNA and pre-mRNA expression of ALKBH2 (left panel) and ALKBH3 (right panel) in the HC group are shown. Correlation analysis was performed using the Pearson correlation coefficient, and data are shown using a simple linear regression model. The p-values and Pearson r2 values are shown on the panels (ALKBH2: ****p < 0.0001 and R = 0.6858, ALKBH3: ***p = 0.0004 and R = 0.6027). HC, healthy control; AC, asymptomatic HTLV-1 carrier; ATL, acute-type adult T-cell leukemia Table 1Detection efficiency of ALKBH3 transcriptsmRNApre-mRNAHCACATLHCACATLNumber of detected/analyzed16/1611/1110/1016/168/117/10Detection rate (%)10010010010072.770.0 Detection efficiency of ALKBH3 transcripts We additionally analyzed mRNA and pre-mRNA expressions of ALKBH2 for comparing to that of ALKBH3, because ALKBH2 is functionally and phylogenetically related to ALKBH3 and a decrease trend of their protein expressions was observed in HTLV-1-infected T-cell lines (Fig. 1b and S2 Fig. b) [20]. The mRNA expression of ALKBH2 in the AC group was 2.55 and 2.25 times lower than that in the HC and ATL groups, respectively, and the difference between the HC and AC groups was statistically significant (Fig. 2b). The pre-mRNA expression of ALKBH2 in the AC and ATL groups was significantly decreased compared to that in the HC group (by factors of 24.18 and 110.46, respectively), and a decrease in the detection rate was also observed, consistent with the results for ALKBH3 (Fig. 2b and Table 2).Table 2Detection efficiency of ALKBH2 transcriptsmRNApre-mRNAHCACATLHCACATLNumber of detected/analyzed16/1610/119/1015/167/115/10Detection rate (%)10090.990.093.863.650.0 Table 2Detection efficiency of ALKBH2 transcriptsmRNApre-mRNAHCACATLHCACATLNumber of detected/analyzed16/1610/119/1015/167/115/10Detection rate (%)10090.990.093.863.650.0 Detection efficiency of ALKBH2 transcripts Correlation analysis revealed that the mRNA and pre-mRNA expression of both ALKBH2 (R = 0.6858) and ALKBH3 (R = 0.6027) were significantly correlated in the HC group (Fig. 2c, d). Unfortunately, due to a number of undetected pre-mRNAs, it was difficult to perform a similar correlation analysis of the AC and ATL groups. However, in the AC and ATL groups, sustainably detected ALKBH2 and ALKBH3 mRNA exhibited different expression trends from pre-mRNA. Interestingly, the mRNA expression of ALKBH2 and ALKBH3 was strongly correlated in the AC and ATL groups (R = 0.7782), but not in the HC group (R = 0.0768, S3 Fig. a-b). It would be possible that some kind of gene regulation event might be occurred in patients dependent on HTLV-1 infection. These observations revealed significant and suggestive decreases of the mRNA expression of ALKBH2 and ALKBH3 in the AC group. It was also discovered significant decrease in the pre-mRNA expression of ALKBH2 and ALKBH3 in AC and ATL groups, suggesting transcriptional suppression in patients infected with HTLV-1. This phenomenon contained another aspect that was unbalanced mRNA and pre-mRNA expression of ALKBH2 and ALKBH3. Associations between HTLV-1 viral factors and mRNA expression of ALKBH2 and ALKBH3 HBZ is an HTLV-1 viral protein constantly detected in HTLV-1-infected cells. Consistent with general knowledge, HBZ mRNA was only detected in the AC and ATL groups, and HBZ mRNA expression was significantly higher in the ATL group (by a factor of 38.36; Fig. 3a and Table 3). We also attempted to evaluate Tax mRNA expression, but it could not be detected in any of the analyzed specimens (Table 3). The practicality of the assay methods to detect HBZ and Tax mRNA was confirmed in the cultured T-cell lines (S2 Fig. c). To evaluate the propagation of HTLV-1-infected cells, we analyzed a dataset of HTLV-1 proviral load (PVL) that was previously reported by our research group [40]. Consistent with HBZ mRNA expression, the HTLV-1 PVL in the ATL group was significantly higher than that in the AC group, by a factor of 7.14 (Fig. 3b).Fig. 3Evaluation of HBZ mRNA expression and effect of HTLV-1 infection on ALKBH2 and ALKBH3 mRNA expression. a The HBZ mRNA value was shown as relative expression standardized by the mean value of the AC group (AC = 1). All experiments were performed in duplicate. Significance was analyzed by unpaired Student's t-test and is indicated by an asterisk (*p = 0.0229). b HTLV-1 PVL is shown as the copy numbers of HTLV-1 per 1 × 104 PBMCs. Significance was analyzed by unpaired Student's t-test and is indicated by an asterisk (****p < 0.0001). Scatter plots represent the median values. c, d Correlation between HBZ mRNA expression and the mRNA expressions of ALKBH2 or ALKBH3 in the AC and ATL groups are shown. The p-values and Pearson r2 values are shown on the panels. Statistical significance was observed for HBZ vs. ALKBH2 mRNA (AC and ATL: **p = 0.0047 and R = 0.4456, AC: ***p = 0.0002 and R = 0.9542, ATL: **p = 0.0036 and R = 0.7237) and HBZ vs. ALKBH3 mRNA (AC and ATL: **p = 0.0035 and R = 0.4447, AC: ****p < 0.0001 and R = 0.9524, ATL: **p = 0.0084 and R = 0.6526). e, f Correlation between HTLV-1 PVL and the mRNA expressions ALKBH2 mRNA or ALKBH3 in the AC and ATL groups are shown. The p-values and Pearson r2 values are shown on the panels (HTLV-1 PVL vs. ALKBH2 mRNA: p = 0.1099 and R = 0.1434, HTLV-1 PVL vs. ALKBH3 mRNA: p = 0.5183 and R = 0.0223). Scatter plots represent the median values. Correlation analysis was performed using the Pearson correlation coefficient, and data are shown using a simple linear regression model. AC, asymptomatic HTLV-1 carrier; ATL, acute-type adult T-cell leukemia; PVL, proviral loadTable 3Detection efficiency of Tax and HBZ mRNATaxHBZHCACATLHCACATLNumber of detected/analyzed0/160/110/100/168/119/10Detection rate (%)000072.790.0 Evaluation of HBZ mRNA expression and effect of HTLV-1 infection on ALKBH2 and ALKBH3 mRNA expression. a The HBZ mRNA value was shown as relative expression standardized by the mean value of the AC group (AC = 1). All experiments were performed in duplicate. Significance was analyzed by unpaired Student's t-test and is indicated by an asterisk (*p = 0.0229). b HTLV-1 PVL is shown as the copy numbers of HTLV-1 per 1 × 104 PBMCs. Significance was analyzed by unpaired Student's t-test and is indicated by an asterisk (****p < 0.0001). Scatter plots represent the median values. c, d Correlation between HBZ mRNA expression and the mRNA expressions of ALKBH2 or ALKBH3 in the AC and ATL groups are shown. The p-values and Pearson r2 values are shown on the panels. Statistical significance was observed for HBZ vs. ALKBH2 mRNA (AC and ATL: **p = 0.0047 and R = 0.4456, AC: ***p = 0.0002 and R = 0.9542, ATL: **p = 0.0036 and R = 0.7237) and HBZ vs. ALKBH3 mRNA (AC and ATL: **p = 0.0035 and R = 0.4447, AC: ****p < 0.0001 and R = 0.9524, ATL: **p = 0.0084 and R = 0.6526). e, f Correlation between HTLV-1 PVL and the mRNA expressions ALKBH2 mRNA or ALKBH3 in the AC and ATL groups are shown. The p-values and Pearson r2 values are shown on the panels (HTLV-1 PVL vs. ALKBH2 mRNA: p = 0.1099 and R = 0.1434, HTLV-1 PVL vs. ALKBH3 mRNA: p = 0.5183 and R = 0.0223). Scatter plots represent the median values. Correlation analysis was performed using the Pearson correlation coefficient, and data are shown using a simple linear regression model. AC, asymptomatic HTLV-1 carrier; ATL, acute-type adult T-cell leukemia; PVL, proviral load Table 3Detection efficiency of Tax and HBZ mRNATaxHBZHCACATLHCACATLNumber of detected/analyzed0/160/110/100/168/119/10Detection rate (%)000072.790.0 Detection efficiency of Tax and HBZ mRNA We performed a correlation analysis between HTLV-1 viral factors and mRNA expression of ALKBH2 and ALKBH3. In each analysis, incomplete data that contained an undetected result for either parameter were omitted (Tables 1, 2, 3). The mRNA expression levels of both ALKBH2 (R = 0.4456) and ALKBH3 (R = 0.4447) were moderately, but significantly, correlated with those of HBZ (Fig. 3c, d). In contrast, the mRNA expression of ALKBH2 (R = 0.1434) and ALKBH3 (R = 0.0223) was not correlated with the HTLV-1 PVL (Fig. 3e–f). These observations suggested that HTLV-1 infection, which is possibly associated with the expression of HBZ mRNA, is a key factor in the regulation of ALKBH2 and ALKBH3 mRNA. DNA methylation analysis near the transcription start site (TSS) of ALKBH3 in vitro To assess the complicated gene regulation mechanisms observed in PBMCs derived from patients infected with HTLV-1, we analyzed the transcriptional suppression of ALKBH3. Esteve-Puig et al. previously revealed that the gene expression of ALKBH3 was inhibited by DNA hypermethylation of a CpG island over the TSS of ALKBH3 in Hodgkin lymphoma cell lines [41]. Based on this report, DNA methylation was analyzed by targeting CpG sites from position −126 to +183 bp relative to the TSS of the ALKBH3 gene, where the CpG island was located, in HTLV-1-infected and uninfected T-cell lines. Methylation states were evaluated by bisulfate sequencing, targeting 10 individual clones. A similar methylation trend was observed among the CpG sites in each T-cell line. Thus, the methylation states of the analyzed region were evaluated as a cluster, and the results are shown as the total percentage of methylated CpG sites (Table 4, S4 Fig. a). Among the analyzed cell lines, MT1 cells exhibited a particularly high methylation state (91.7%) whose ALKBH3 mRNA expression was diminished (Fig. 1a and Table 4). To confirm the effect of hypermethylation on the gene expression of ALKBH3, we performed demethylation assays using MT1 cells and 5-aza-2′-deoxycytidine, also known as DAC (Decitabine), a well-known demethylation agent used in HTLV-1-infected T-cell lines [44, 45]. It was confirmed that 1 nM DAC sufficiently demethylated MT1 cells, decreasing the methylation state from 90.0–91.4% to 38.1–77.9% in the analyzed genomic region (Fig. 4a). The demethylating effect of DAC was further verified by evaluating the gene expression of Tax and HBZ. Taniguchi et al. reported that the mRNA expression of Tax is inhibited by hypermethylation, whereas that of HBZ is not regulated in an epigenetic manner [46]. Consistent with a previous report, the mRNA expression of Tax was restored by DAC in MT1 cells, whereas that of HBZ was not affected (Fig. 4b). In this setting, it was confirmed that the mRNA and pre-mRNA expression of ALKBH3, but not that of ALKBH2, was restored in the presence of DAC (Fig. 4c). Therefore, we concluded that hypermethylation of the CpG island suppressed ALKBH3 gene expression via an epigenetic mechanism, suggesting that transcriptional inhibition occurred in vitro.Table 4Methylation states around ALKBH3 TSS region (%)Position from TSSJurkatCEMMolt4MT1MT2MT4HUT102TLOm1−126 ~ +183047.4091.712.576.946.938.4−269 ~ −178053.7091.773.381.758.388.3TSS, transcription start siteFig. 4DNA demethylation assay for MT1 cells. a DNA methylation states around the TSS of ALKBH3 were analyzed by the bisulfate method. The figures are representative of three independent assays, and each assay analyzed sequences of 10 individual clones. The CpG sites are indicated by vertical lines, and the TSS is indicated by a bending arrow. Unmethylated and methylated CpG sites are represented by white and black circles, respectively. b, c The gene expression of HBZ mRNA, Tax mRNA, ALKBH2 mRNA, ALKBH3 mRNA, and ALKBH3 pre-mRNA are shown in terms of the 2−ΔCt values on the y-axis. The data represent the mean ± SD of four independent replicates. TSS, transcription start site; DMSO, dimethyl sulfoxide; DAC, decitabine Table 4Methylation states around ALKBH3 TSS region (%)Position from TSSJurkatCEMMolt4MT1MT2MT4HUT102TLOm1−126 ~ +183047.4091.712.576.946.938.4−269 ~ −178053.7091.773.381.758.388.3TSS, transcription start site Methylation states around ALKBH3 TSS region (%) −269 ~ −178 TSS, transcription start site DNA demethylation assay for MT1 cells. a DNA methylation states around the TSS of ALKBH3 were analyzed by the bisulfate method. The figures are representative of three independent assays, and each assay analyzed sequences of 10 individual clones. The CpG sites are indicated by vertical lines, and the TSS is indicated by a bending arrow. Unmethylated and methylated CpG sites are represented by white and black circles, respectively. b, c The gene expression of HBZ mRNA, Tax mRNA, ALKBH2 mRNA, ALKBH3 mRNA, and ALKBH3 pre-mRNA are shown in terms of the 2−ΔCt values on the y-axis. The data represent the mean ± SD of four independent replicates. TSS, transcription start site; DMSO, dimethyl sulfoxide; DAC, decitabine Investigation of DNA methylation states around the TSS of ALKBH3 in patients infected with HTLV-1 using a public dataset To investigate whether HTLV-1 infection causes the epigenetic gene regulation of ALKBH3, we explored a public dataset which contained comprehensive DNA methylation profiles in peripheral T-cell subsets derived from patients infected with HTLV-1 [44]. The T-cell subsets were isolated on the basis of CADM1 and CD7 expression following a well-established method, and classified into three populations according to HTLV-1 infection status and disease progression. Briefly, HTLV-1-infected cells are more concentrated in the CADM1-positive population, and CD7 expression decreases from dimly positive to negative with disease progression (P: CD3+CADM1−CD7+; D: CD3+CADM1+CD7dim; N: CD3+CADM1+CD7−) [14, 15]. First, we analyzed the methylation states of the CpG island around the TSS of ALKBH3; however, no characteristic differences were observed among T-cell subsets (Fig. 5a). Subsequently, the analyzed genomic regions were extended to the upstream and downstream regions of the CpG island, and four CpG sites were detected that were significantly methylated in the D and N populations, compared to those in the P population (Fig. 5b, c). Hypermethylation of the analyzed CpG sites was defined in the original article as an increase in the β-value of over 0.2 points in N populations, compared to that in P populations [44]. The CpG sites at positions −269, −247, and +661 bp relative to the TSS of ALKBH3 satisfied this criterion, exhibiting 0.221-, 0.205-, and 0.243-point increases in the mean values, respectively [44]. Notably, N populations from acute-type ATL patients (Acu-N), representing a highly malignant cell population, showed similar levels of methylation as the P population at the CpG sites at positions −269 and −247 bp relative to the TSS of ALKBH3, despite the high methylation states observed in the D and N populations. These observations suggest that these two CpG sites were hypermethylated in the N population but unmethylated in the Acu-N population. It was also confirmed that the methylation states were mostly consistent among the T-cell subsets but not among ATL subtypes, except for Acu-N (S5 Fig).Fig. 5Investigation of DNA methylation states around the TSS of ALKBH3 in patients infected with HTLV-1. a–c CpG methylation states of T-cell subsets derived from HTLV-1-infected patients were investigated using a public dataset (GSE136189). Each CpG site was designated by a unique CpG cluster ID identifying the unique locus and by the number of nucleotide bases from the TSS of ALKBH3. Data are shown by β-value, which ranged from 0 to 1, denoting complete unmethylation and methylation, respectively. Scatter plots represent the median values. Significance was analyzed by one-way ANOVA followed by Tukey's multiple comparisons test and is indicated by an asterisk. Statistical significance was observed in the analysis of cg14908875 (P vs. D: *p = 0.0272, P vs. N: ****p < 0.0001, P vs. Acu-N: ***p = 0.0004), cg21125143 (P vs. D: **p = 0.0069, P vs. N: ****p < 0.0001), cg26579032 (P vs. D: **p = 0.0038, P vs. N: ****p < 0.0001, N vs. Acu-N: *p = 0.0468), and cg26985203 (P vs. D: ****p < 0.0001, P vs. N: ****p < 0.0001, P vs. Acu-N: ****p < 0.0001, D vs. N: *p = 0.0390, D vs. Acu-N: *p = 0.0154). P, CD3+CADM1−CD7+ population; D, CD3+CADM1+CD7dim population; N, CD3+CADM1+CD7− population from chronic- and smoldering-type ATL; Acu-N, CD3+CADM1+CD7− population from acute-type ATL Investigation of DNA methylation states around the TSS of ALKBH3 in patients infected with HTLV-1. a–c CpG methylation states of T-cell subsets derived from HTLV-1-infected patients were investigated using a public dataset (GSE136189). Each CpG site was designated by a unique CpG cluster ID identifying the unique locus and by the number of nucleotide bases from the TSS of ALKBH3. Data are shown by β-value, which ranged from 0 to 1, denoting complete unmethylation and methylation, respectively. Scatter plots represent the median values. Significance was analyzed by one-way ANOVA followed by Tukey's multiple comparisons test and is indicated by an asterisk. Statistical significance was observed in the analysis of cg14908875 (P vs. D: *p = 0.0272, P vs. N: ****p < 0.0001, P vs. Acu-N: ***p = 0.0004), cg21125143 (P vs. D: **p = 0.0069, P vs. N: ****p < 0.0001), cg26579032 (P vs. D: **p = 0.0038, P vs. N: ****p < 0.0001, N vs. Acu-N: *p = 0.0468), and cg26985203 (P vs. D: ****p < 0.0001, P vs. N: ****p < 0.0001, P vs. Acu-N: ****p < 0.0001, D vs. N: *p = 0.0390, D vs. Acu-N: *p = 0.0154). P, CD3+CADM1−CD7+ population; D, CD3+CADM1+CD7dim population; N, CD3+CADM1+CD7− population from chronic- and smoldering-type ATL; Acu-N, CD3+CADM1+CD7− population from acute-type ATL Finally, we analyzed DNA methylation states, targeting positions −319 to −247 bp relative to the TSS of ALKBH3, using HTLV-1-infected and uninfected T-cell lines. High methylation states were commonly observed at position −319 bp relative to the TSS of ALKBH3 (S4 Fig. b). This result was consistent with the observation that the methylation state of the P population remained high, despite being significantly decreased compared to that of the other populations (mean β-value ± SD: 0.77 ± 0.06; Fig. 5b). Because the T-cell lines exhibited common methylation trends at the CpG sites between positions −269 and –178 bp relative to the TSS of ALKBH3, as a cluster, as described above (Table 4 and S4 Fig. b). In these clusters, HTLV-1-infected T-cell lines exhibited moderate to high methylation states, which were higher than those observed in the analyzed region between positions −126 and +183 bp relative to the TSS of ALKBH3 (Table 4). In contrast, the DNA methylation states of uninfected T-cell lines showed similar trends in the analyzed regions. Consistent with the results shown in Fig. 5b, all HTLV-1-infected T-cell lines exhibited higher methylation states than the uninfected T-cell lines in the analyzed region between positions −269 and −178 bp relative to the TSS of ALKBH3. These observations suggest that the CpG sites between positions −269 and −178 bp relative to the TSS of ALKBH3 had the potential to play some roles in the epigenetic gene regulation of ALKBH3. In conclusion, this series of epigenetic analyses of clinical specimens and T-cell lines suggested that the methylation state of CpG sites in a specific region or locus around the TSS of ALKBH3 was characteristically regulated by HTLV-1 infection. Discussion HTLV-1 is an oncogenic virus that causes highly malignant ATL after a long latency period. It is still difficult to predict disease progression and provide effective therapy precisely. It is also not sufficient to access clinical specimens obtained from patients infected with HTLV-1, because global recognition of HTLV-1 infection is still in developing stage [2]. To control the burden of HTLV-1, recent research has aimed to develop our understanding of host immune responses and viral oncogenic effects and tackled to clarify the pathological process of ATL [10, 47, 48]. Here, we revealed dynamic patterns of ALKBH2 and ALKBH3 gene regulation that occur in healthy carriers versus acute-type of ATL patients. In particular, a characteristic discrepancy between the mRNA and pre-mRNA expression of ALKBH2 and ALKBH3 was observed, providing novel insight into ATL pathogenesis. It was also discovered that DNA methylation states around the TSS of ALKBH3 were specifically regulated by HTLV-1 infection. ALKBH family members commonly function as demethylases. For the purpose of genome maintenance, they remove naturally occurring nucleotide modifications by targeting various conformations of DNA and RNA [19, 20]. Recently, their functions have been intensely investigated in oncology because some ALKBH family members have been implicated in oncogenesis and/or disease progression in various types of cancers [27]. For example, overexpression of ALKBH2 activates cancer cell proliferation in bladder cancer and colorectal cancer, and ALKBH3 is considered a diagnostic marker and promising drug target for prostate carcinoma [36, 49–51]. In our study, the mRNA expression of ALKBH2 was significantly decreased only in the AC group, and that of ALKBH3 showed a similar trend, even though statistical significance was not observed. Both ALKBH2 and ALKBH3 commonly play roles in the restoration of 1-meA and 3-meC, which are known to cause somatic mutations in the host genome if they were unrepaired [19–24]. Mounting evidence suggests that accumulated somatic mutations are a key factor in the onset of ATL; therefore, it is possible that the sustained suppression of ALKBH2 and ALKBH3 insidiously supports disease progression in ACs [16–18]. In particular, ALKBH2 is a prominent candidate for future studies because its mRNA expression was significantly suppressed in the AC group; therefore, further analysis would be valuable such as exploring ALKBH2-related molecular pathways in HTLV-1-infected cells. It would be also possible that the ALKBH2 mRNA expression could serve as prominent candidate of genetic marker to predict disease progression of patients infected with HTLV-1. Interestingly, the mRNA expression of ALKBH2 was not correlated with PVL in patients infected with HTLV-1, even though PVL is a reliable indicator of ATL [52]. This observation indicated that ALKBH2 mRNA has the potential to evaluate ATL progression independently from PVL. To develop ALKBH2 mRNA as practical gene marker for ATL diagnosis, an additional large-scale clinical study is needed to target not only patients with acute type of ATL but also patients with lymphoma, chronic, and smoldering types of ATL. We unexpectedly discovered that the pre-mRNA expression of ALKBH2 and ALKBH3 was significantly inhibited in the AC and ATL groups. Regarding the observation in the ATL group, to our knowledge, there has been no report that HTLV-1 infection selectively inhibits host genomic pre-mRNA expression with no effect on mRNA expression. Although we could not investigate the relationship between these types of gene regulation in ATL pathogenesis, there would be some mechanisms to maintain mRNAs under the pre-mRNA exhausted situation in patients infected with HTLV-1, such as getting resistance against RNA decay [53, 54]. We evaluated the expression of ALKBH3-AS1 to accurately determine the mRNA and pre-mRNA expression of ALKBH3. However, gene expression could not be detected using a primer set designed to amplify multiple exons. An alternative assay using another primer set designed to amplify a single exon demonstrated a gene expression trend similar to that of ALKBH3 pre-mRNA. Based on these observations, we concluded that ALKBH3-AS1 was not expressed in the analyzed T-cell lines and was unlikely to affect the quantification of the mRNA and pre-mRNA gene expressions of ALKBH3. We also revealed that HBZ mRNA expression was significantly correlated with the mRNA expression of ALKBH2 and ALKBH3, suggesting that HBZ would be a key factor in inducing these gene regulation patterns in patients infected with HTLV-1. The AC group showed significantly suppressed pre-mRNA expression of ALKBH2 and ALKBH3, even though PVL analysis confirmed that the population of HTLV-1-infected cells was limited in the PBMCs analyzed (Figs. 2 and 3). This observation suggests a possibility that HBZ-expressing T-cells communicate with other immune cells and regulate their gene expression. Intercellular communication between immune cells and malignant cells has been vigorously investigated as they are known to construct a ‘tumor microenvironment’ that accelerates oncogenesis in various type of cancers [55, 56]. It is well known that HBZ, in both form of RNA and protein, regulates a variety of gene expressions in the HBZ-expressing cells and modulates their differentiation into specific phenotypes [57–59]. In addition to intracellular function of HBZ, further analyses focusing on the intercellular effects of HBZ-expressing cells in vivo would benefit the comprehension of HTLV-1 pathogenesis in humans. The DNA methylation analysis conducted in this study provided valuable insights into gene regulation of ALKBH3. We discovered novel CpG sites at positions −269 and −247 bp relative to the TSS of ALKBH3 that were significantly hypermethylated in the N population, but not in the Acu-N population. It is well documented that CD7 expression is downregulated in response to the progression of HTLV-1 infection, and an expanded CD7-negative population indicates the onset of ATL [14, 15]. Summarizing this series of events, the appearance of each population can be placed on a timeline: the first stage of HTLV-1 infection is characterized by the presence of the D population. The progression of infection is characterized by the transition of the D population to the N population, and the onset of acute-type ATL is characterized by the Acu-N population. This suggests that these two CpG sites exhibit methylation in the D and N populations and then revert to a demethylated state in the Acu-N population. This consideration is consistent with the obtained results that the mRNA expressions of ALKBH2 and ALKBH3 were decreased in the ACs but reverted in the patients with acute-type ATL. Although we could not investigate the mechanisms that maintain mRNAs under the pre-mRNA exhausted situation in patients infected with HTLV-1, it is possible that a group of enzymes regulating DNA methylation states also affect mRNA methylation states and provide stability as a side effect [60]. On the other hand, Acu-N–dependent demethylation was not observed at other analyzed CpG sites, suggesting that this is a locus-specific event. These observations suggest that the methylation states of these CpG sites could be affected by the progression of HTLV-1 infection and have potential contributing to the onset of acute-type ATL. In our knowledge, this is first report to discover the hyper- and hypomethylation events at CpG sites in host genome depending on HTLV-1 infection and/or disease progress of ATL. Conclusions In this study, we revealed dynamically regulated patterns of ALKBH2 and ALKBH3 gene expression in patients infected with HTLV-1 and discovered specific CpG sites around the TSS of ALKBH3, the methylation states of which were specifically regulated by HTLV-1 infection. This is a first report to investigate relationship between ALKBH family gene and HTLV-1 infection, and our study indicated ALKBH2 and ALKBH3 as possible pathogenic factor of ATL. Our findings also suggest that evaluation of pre-mRNA, which have not been investigated in clinical patients previously, is an interesting target for future research exploring oncogenic mechanism of ALKBH family or other host genes. This research would contribute to elucidate pathogenesis of ATL, and also expand understanding of ALKBH2 and ALKBH3 in the research field of oncology. Supplementary Information Additional file1 S1 Table. Primer setsAdditional file 2 S1 Fig. Primer targets in qRT-PCR assay. Primer targets of (a) ALKBH2, (b) ALKBH3 in qRT-PCR assay were graphically represented. S2 Fig. Analysis of the expressions of ALKBH family genes, and HTLV-1 Tax and HBZ mRNA in cultured T-cell lines. (a) The mRNA expressions of ALKBH family genes were measured by qRT-PCR. Uninfected (Jurkat, CEM, and Molt4) and HTLV-1-infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1) were grouped, and their mean values of gene expression were used for the analysis. The mRNA expression of ALKBH3 in MT1 cells was omitted from the analysis because it was undetectable. Gene expression was calculated using the ΔΔCt method. The equalized data are shown as relative expression, standardized by the mean value of uninfected group (uninfected = 1) on the y-axis. The data represent the mean ± SD of the gene expression levels in uninfected and HTLV-1 groups. Significance was analyzed by unpaired Student's t-test and is indicated by asterisks (ALKBH1: *p = 0.0103, ALKBH2: **p = 0.0085, ALKBH3: ***p = 0.0004, ALKBH5: *p = 0.0442, ALKBH6: **p = 0.058, ALKBH7: *p = 0.137, ALKBH8: p = 0.0738, and FTO: ****p < 0.0001). Abbreviations: Uninfected, group of uninfected T-cell lines; HTLV-1, group of HTLV-1–infected T-cell lines. (b) Original images of ALKBH2, ALKBH3 and β-actin protein were shown, respectively. The figure is a representative of the independent experiments. The proteins detection and imaging were performed by a LAS-3000 imaging system. (c) The mRNA expression of Tax and HBZ mRNA were measured by qRT-PCR in uninfected (Jurkat, CEM, and Molt4) and HTLV-1–infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1). Gene expression was calculated using the ΔΔCt method. The equalized data are shown as relative expression, standardized by the mean value of MT2 cells (MT2 = 1) on the y-axis. The data represent the mean ± SD of four independent experiments. Significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test and is indicated by an asterisk. Statistical significance was observed in the analysis of HBZ mRNA (MT1 vs. MT4: *p = 0.0330, MT1 vs. HUT102: *p = 0.0363, MT1 vs TLOm1: *p = 0.0494). S3 Fig. Correlation analysis between mRNA expression of ALKBH2 and ALKBH3. (a-b) Correlation between mRNA expression of ALKBH2 and ALKBH3 in the HC group (left panel) and the AC and ATL groups (right panel) are shown. Correlation analysis was performed using the Pearson correlation coefficient, and data are shown using a simple linear regression model. The p-values and Pearson r2 values are shown on the panels (HC group: p = 0.0768 and R = 0.2068, AC and ATL groups: ****p < 0.0001 and R = 0.7782, respectively). Abbreviations: HC, healthy control; AC, asymptomatic HTLV-1 carrier; ATL, acute-type adult T-cell leukemia. S4 Fig. DNA methylation states around the TSS of ALKBH3 in T-cell lines. CpG methylation states between positions (a) −126 and +183, and (b) −501 and −178 relative to the TSS of ALKBH3 were analyzed by bisulfate method against HTLV-1-infected or uninfected T-cell lines. Each T-cell line was sequenced 10 individual clones. The CpG sites are indicated by vertical lines, and the TSS is indicated by a bending arrow. Upward arrows indicated the CpG sites analyzed in Fig. 5a-b. Unmethylated and methylated CpG sites are represented by white and black circles, respectively. Abbreviations: TSS, transcription start site. S5 Fig. DNA methylation states around the TSS of ALKBH3 in patients infected with HTLV-1. CpG methylation states of T-cell subsets derived from patients infected with HTLV-1 were investigated using a public dataset (GSE136189). Each CpG site was designated by a unique CpG cluster ID identifying the unique locus and by the number of nucleotide bases from the TSS of ALKBH3. Data are shown by β-value, which ranged from 0 to 1, denoting complete unmethylation and methylation, respectively. Scatter plots represent the median values. Abbreviations: HC, healthy control; AC, asymptomatic carrier; S, smoldering-type of ATL; C, chronic-type of ATL; A, acute-type of ATL; P, CD3+CADM1−CD7+ population; D, CD3+CADM1+CD7dim population; N, CD3+CADM1+CD7− population; TSS, transcription start site. Additional file1 S1 Table. Primer sets Additional file 2 S1 Fig. Primer targets in qRT-PCR assay. Primer targets of (a) ALKBH2, (b) ALKBH3 in qRT-PCR assay were graphically represented. S2 Fig. Analysis of the expressions of ALKBH family genes, and HTLV-1 Tax and HBZ mRNA in cultured T-cell lines. (a) The mRNA expressions of ALKBH family genes were measured by qRT-PCR. Uninfected (Jurkat, CEM, and Molt4) and HTLV-1-infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1) were grouped, and their mean values of gene expression were used for the analysis. The mRNA expression of ALKBH3 in MT1 cells was omitted from the analysis because it was undetectable. Gene expression was calculated using the ΔΔCt method. The equalized data are shown as relative expression, standardized by the mean value of uninfected group (uninfected = 1) on the y-axis. The data represent the mean ± SD of the gene expression levels in uninfected and HTLV-1 groups. Significance was analyzed by unpaired Student's t-test and is indicated by asterisks (ALKBH1: *p = 0.0103, ALKBH2: **p = 0.0085, ALKBH3: ***p = 0.0004, ALKBH5: *p = 0.0442, ALKBH6: **p = 0.058, ALKBH7: *p = 0.137, ALKBH8: p = 0.0738, and FTO: ****p < 0.0001). Abbreviations: Uninfected, group of uninfected T-cell lines; HTLV-1, group of HTLV-1–infected T-cell lines. (b) Original images of ALKBH2, ALKBH3 and β-actin protein were shown, respectively. The figure is a representative of the independent experiments. The proteins detection and imaging were performed by a LAS-3000 imaging system. (c) The mRNA expression of Tax and HBZ mRNA were measured by qRT-PCR in uninfected (Jurkat, CEM, and Molt4) and HTLV-1–infected T-cell lines (MT1, MT2, MT4, HUT102, and TLOm1). Gene expression was calculated using the ΔΔCt method. The equalized data are shown as relative expression, standardized by the mean value of MT2 cells (MT2 = 1) on the y-axis. The data represent the mean ± SD of four independent experiments. Significance was analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test and is indicated by an asterisk. Statistical significance was observed in the analysis of HBZ mRNA (MT1 vs. MT4: *p = 0.0330, MT1 vs. HUT102: *p = 0.0363, MT1 vs TLOm1: *p = 0.0494). S3 Fig. Correlation analysis between mRNA expression of ALKBH2 and ALKBH3. (a-b) Correlation between mRNA expression of ALKBH2 and ALKBH3 in the HC group (left panel) and the AC and ATL groups (right panel) are shown. Correlation analysis was performed using the Pearson correlation coefficient, and data are shown using a simple linear regression model. The p-values and Pearson r2 values are shown on the panels (HC group: p = 0.0768 and R = 0.2068, AC and ATL groups: ****p < 0.0001 and R = 0.7782, respectively). Abbreviations: HC, healthy control; AC, asymptomatic HTLV-1 carrier; ATL, acute-type adult T-cell leukemia. S4 Fig. DNA methylation states around the TSS of ALKBH3 in T-cell lines. CpG methylation states between positions (a) −126 and +183, and (b) −501 and −178 relative to the TSS of ALKBH3 were analyzed by bisulfate method against HTLV-1-infected or uninfected T-cell lines. Each T-cell line was sequenced 10 individual clones. The CpG sites are indicated by vertical lines, and the TSS is indicated by a bending arrow. Upward arrows indicated the CpG sites analyzed in Fig. 5a-b. Unmethylated and methylated CpG sites are represented by white and black circles, respectively. Abbreviations: TSS, transcription start site. S5 Fig. DNA methylation states around the TSS of ALKBH3 in patients infected with HTLV-1. CpG methylation states of T-cell subsets derived from patients infected with HTLV-1 were investigated using a public dataset (GSE136189). Each CpG site was designated by a unique CpG cluster ID identifying the unique locus and by the number of nucleotide bases from the TSS of ALKBH3. Data are shown by β-value, which ranged from 0 to 1, denoting complete unmethylation and methylation, respectively. Scatter plots represent the median values. Abbreviations: HC, healthy control; AC, asymptomatic carrier; S, smoldering-type of ATL; C, chronic-type of ATL; A, acute-type of ATL; P, CD3+CADM1−CD7+ population; D, CD3+CADM1+CD7dim population; N, CD3+CADM1+CD7− population; TSS, transcription start site.