Abstract

   Metabolic dysfunction-associated steatotic liver disease (MASLD) is a
   major predisposing factor of metabolic dysfunction-associated
   steatohepatitis, hepatic fibrosis and hepatocellular carcinoma.
   Abnormal RNA m6A modification leads to the disturbance of lipid
   metabolism, insulin resistance, and chronic inflammation, all of which
   are critical for the onset and progression of MASLD. It has been shown
   that betaine supplementation has an ameliorative effect on symptoms
   associated with MASLD. The purpose of this study is to explore the role
   of RNA m6A modification in the alleviation of MASLD by betaine, and to
   find out the possible targets of betaine. Methylated RNA
   immunoprecipitation sequencing (MeRIP-seq) and RNA sequencing (RNA-seq)
   were respectively employed to contrast m6A methylation profiles of
   livers from high-fat diet mice with betaine supplementation vs. those
   without betaine supplementation (HB vs. HFD). Functional enrichment
   analysis showed that up-methylated genes are mainly related to mTOR
   signaling pathway, Rap1 signaling pathway and MAPK signaling pathway.
   In addition, differentially expressed mRNAs were enriched in pathways
   such as promoting lipid catabolism and reducing lipid accumulation. By
   combining analyses of MeRIP-seq and RNA-seq, we identified 27 genes
   exhibiting significant differences in m6A abundance and mRNA levels.
   Notably, a new candidate gene, Trub2, was screened out and identified.
   Inhibition of Trub2 increased intracellular TG levels in AML12 cells,
   and this promotion was reduced by betaine, suggesting that betaine
   reduces intracellular TG levels by increasing Trub2 expression. Our
   findings in this study provide a new target and a new approach for the
   treatment of MASLD.

Supplementary Information

   The online version contains supplementary material available at
   10.1038/s41598-025-91573-0.

   Keywords: Betaine, MASLD, N6-methyladenosine (m6A), MeRIP-seq, Trub2

   Subject terms: Computational biology and bioinformatics, Non-alcoholic
   fatty liver disease

Introduction

   MASLD is a chronic liver disease characterized by lipid accumulation
   and hepatocyte steatosis^[44]1. It is the most important factor in a
   range of serious metabolic diseases, including nonalcoholic
   steatohepatitis (NASH), liver fibrosis, cirrhosis, and even
   hepatocellular carcinoma (HCC)^[45]2,[46]3. The incidence of MASLD has
   increased year by year, and it has become the largest chronic liver
   disease in the world, affecting about 25%~30% of the global
   population^[47]4. Similarly, the reported prevalence of MASLD has
   reached 29.2% in China, which has emerged as a major public health
   issue^[48]5. However, no specific drug has been found to be effective
   in the treatment of MASLD. Therefore, there is an urgent need to
   further explore the mechanism of action of MASLD and develop new
   therapeutic targets.

   N6-methyladenosine (m6A) methylation, the most prevalent internal
   modification in eukaryotic mRNA, has been reported to be functional in
   many physiological events^[49]6,[50]7. Proteins such as
   methyltransferases (writer), demethylases (eraser), and reader proteins
   (readers) work together to regulate the dynamic reversible modification
   of m6A^[51]8. m6A methyltransferase is responsible for catalyzing the
   formation of the m6A modification, by some methyltransferase-like 3
   (METTL3)/ methyltransferase-like 14 (METTL14) heterologous dimers and
   various other methyltransferase complexes^[52]9.Demethylase catalyzes
   the removal of methyl groups on m6A, which are composed of the fat mass
   and obesity associated (FTO) and alkylation repair homolog protein 5
   (ALKBH5)^[53]10. In patients with non-alcoholic fatty liver disease and
   in mouse models of MASLD, overexpression of FTO promotes hepatic
   steatosis and lipid accumulation^[54]11.m6A methylated reading
   proteins, including YT521-B homology (YTH) domain-containing proteins
   (YTHDF1-3, YTHDC1-2) and Insulin-like growth factor 2 mRNA-binding
   proteins 1/2/3 (IGF2BP1/2/3), play a crucial role in the alternative
   splicing, stability, degradation and translation of mRNA by recognizing
   m6A modifications on RNA^[55]10. Hu et al. found that IGF2BP3 enhanced
   the mRNA stability of protein kinase, membrane-associated
   tyrosine/threonine 1 (PKMYT1) through its m6A modification site, thus
   promoting the development of gastric cancer^[56]12. Mu et al. found
   that IGF2BP1 mediates oxidative stress-induced granule-cell dysfunction
   by regulating murine double minute 2 (MDM2) mRNA stability in an
   m6A-dependent manner^[57]13. Given the critical role RNA m6A
   methylation plays in dysregulated lipid metabolism, it is consequently
   considered a highly promising and substantial therapeutic target.

   Betaine (N, N, N-trimethylglycine), a trimethyl derivative of glycine,
   was first found in common plants in the 19th century^[58]14. As a
   natural methyl donor, betaine serves as an indispensable component in
   the metabolic regulation of organisms, playing a critical role in both
   glycemic-lipid metabolism and hepatic function regulation^[59]15. A
   significant amount of data from animal models of liver disease
   indicates that administration of betaine can halt and even reverse
   progression of the disruption of liver function^[60]16. In recent
   years, many pathways have been shown to be closely related to the
   improvement of MASLD by betaine, including inhibiting inflammatory
   responses, reducing oxidative stress in the liver, promoting fatty acid
   oxidation, improving insulin resistance, and regulating epigenetic
   modifications^[61]17. At the same time, clinical trials have shown that
   patients with NASH who receive purified betaine for up to one year can
   reduce their liver steatosis index^[62]18. However, the exact mechanism
   by which betaine participates in post-transcriptional regulation
   through m6A methylation and exerts its beneficial effects remains to be
   elucidated.

   This research delineates a comprehensive map of the role played by
   betaine-induced m6A RNA methylation in attenuating hepatic steatosis,
   and it identifies Trub2 as a target gene subject to m6A modification.
   This discovery furnishes a novel therapeutic target for investigating
   the mechanism by which betaine modulates lipid metabolism in the liver.
   This pioneering study offers a comprehensive depiction of the
   transcriptome-wide m6A modification pattern exerted by betaine in
   mitigating fatty liver degeneration, forging a connection between m6A
   and mRNA. It thereby provides a foundational basis for exploring how
   betaine alleviates MASLD through modulation of m6A modifications.

Method

Experimental animals and sample collection

   All animal experiments were approved by the Committee on the Ethics of
   Animal Experiments of Guangxi University (No. GXU2019-063). A total of
   12 eight-week-old male C57BL/6 mice with comparable body weights were
   randomly assigned to two groups: the high fat diet (HFD) group, and the
   high fat diet with betaine (HB) group, each consisting of six mice.
   Mice were housed in a facility with 12-h light/12-h dark cycle at
   23 ± 3 °C and 30–70% humidity. In this study, HFD mice were given a
   high-fat diet (45% kcal fat; Jiangsu Medicience Ltd., Jiangsu, China),
   while the HB group was on the same high-fat diet but with added 2% w/v
   betaine in drinking water. Mice in the HFD group developed MASLD, while
   those in the HB group showed significant improvement in MASLD^[63]19.
   These findings have been thoroughly validated. After 17 weeks of
   dietary intervention, the mice inhaled isoflurane in a closed room,
   quickly reached an effective concentration (5%) and were
   euthanized^[64]20. All research involving animal experimentation was
   conducted in accordance with the National Institutes of Health
   guidelines on animal care and use. All methods are reported in
   accordance with ARRIVE guidelines.

RNA extraction

   Total RNA was extracted from liver tissue using TRIzol reagent and the
   extracted lysed RNA was transferred to 1.5 ml RNase-free EP tubes. 100
   µl of chloroform was added, and the mixture was shaken vigorously in a
   shaker for 15 s, left at room temperature, centrifuged. The upper water
   phase was taken and placed in a new EP tube, each tube was added with
   equal volume isopropyl alcohol, shaken well, standing, centrifuged.
   Then, the RNA was washed with anhydrous ethanol diluted with DEPC
   water, centrifuged, and the supernatant was discarded. The RNA
   concentration and purity were detected by using an enzyme-labeling
   instrument. The remaining RNA was stored at −80°C for subsequent
   library construction and high-throughput sequencing^[65]19.

RNA dot blot

   Cut the nitrocellulose (NC) membranes (Biosharp) and soak them in 20X
   Saline-Sodium Citrate Buffer (20X SSC, Biosharp) for 1 h. Dilute the
   RNA and mix it 1:1 with the denaturing solution (20X SSC: 37% deionized
   formaldehyde = 3:2, v/v), then heat it at 95 °C for 5 min to denature.
   Spot RNA samples (100 ng, 200 ng, and 400 ng) onto the NC membranes.
   After UV cross-linking, block the membrane with 1% Bovine Serum Albumin
   (BSA, Solarbio) for 1 h at room temperature. Incubate the membrane with
   anti-m6A antibody (1:2000, Beyotime) overnight at 4 °C, followed by
   incubation with the secondary antibody for 1 h. Detect the signal using
   a chemiluminescence technique after washing with PBST. As a loading
   control, stain the membranes in 0.3 mol/L sodium acetate (pH 5.2)
   containing 0.02% methylene blue.

Library construction, MeRIP-seq and RNA-seq

   The quality of total mRNA and m6A-enriched RNA was detected by 2.5%
   agarose gel electrophoresis. m6A antibody-immunoprecipitated RNA
   underwent quality control using Qubit^®3.0 (Thermo Fisher Scientific)
   and Agilent 2200 TapeStation. The qualified RNA was then used for
   library construction according to the NEBNext^® Ultra™ RNA Library Prep
   Kit protocol for Illumina platforms (NEB). Post-library preparation,
   products were evaluated with Agilent 2200 TapeStation and Qubit^®3.0
   before sequencing on an Illumina HiSeq X Ten platform with paired-end
   standard sequencing. High-throughput sequencing was performed on
   qualified mRNA and m6A-modified RNA. According to the protocol of VAHTS
   mRNA-seq and library preparation kit, the original sequencing reads
   were generated on the HiSeq X Ten sequencer.

Bioinformatics analysis

   For mRNA and m6A sequencing datasets, high-quality reads were mapped to
   RefSeq mRNA reference sequences (mm10) using the FANSe2 algorithm. The
   reads mapped to variable splicing variants of a gene were merged. The
   abundance of mRNA and m6A in each sample was normalized per kilobase
   per million reads (RPKM)^[66]21. The differentially expressed genes
   (DEGs) were selected with the standard of absolute |Log2 fold change| ≥
   1 and P < 0.05 via R package edgeR. Kyoto Encyclopedia of Genes and
   Genomes (KEGG^[67]22,[68]23) enrichment, Gene Ontology (GO) analysis
   and Reactome Gene Sets analysis were performed using the Metascape
   online tools ([69]http://metascape.org/gp/index.html#/main/step1).

Construction of mouse Trub2 overexpression plasmid

   The result of enzyme digestion showed a band length of about 996 bp,
   which was consistent with the expected length, indicating that the CDS
   region fragment of mouse Trub2 gene ([70]NM_145520.5) was successfully
   constructed. The Trub2 expression plasmid was constructed, and the
   primer sequences used for amplification were respectively
   AAGCTGGCTAGTTAAGCTTGCCACCATGGGTTTC (positive) and
   TCCGAGCTCGGTACAAGCTTTTTACTGCC (reverse). The plasmid was transfected
   transiently with Liposome 3000 reagent^[71]24.

Cell culture and transfection

   AML12 cells were maintained in Dulbecco’s modified Eagle’s Ham’s F-12
   supplemented with 10% fetal bovine serum, ITS Liquid Media Supplement,
   and dexamethasone (40 ng ml^−1) under an atmosphere of 5% CO[2] at
   37℃^[72]25and in an environment supplemented with 10% fetal bovine
   serum and 5% carbon dioxide. When the fusion rate reached 80%, the
   cells were seeded into the 24-well plate^[73]26. Cells were cultivated
   until they reached a density of approximately 80% confluence, at which
   point they were subjected to a 24-hour exposure to a mixture of oleic
   and palmitic acids (OA/PA) with the aim of inducing fat deposition.
   DMEM medium enriched with 200 µM oleic acid (OA) and 100 µM palmitic
   acid (PA) was employed to replicate the high-fat diet milieu in a
   cellular context. 2 mM betaine (Beijing Solarbio Science & Technology
   Co., Ltd. Beijing, China) was used to treat AML12 cells supplemented
   with OA/PA medium.

   Upon reaching 60% confluence, cells were switched to serum-free DMEM
   (Gibco) and transfected with plasmids or siRNA using Hieff TransTM
   Liposomal Transfection Regent. The pcDNA3.1 or Trub2 plasmid were
   transfected into AML12 cells. AML12 cells were transfected with
   siTrub2, or the negative control siRNA (siNC). The previously confirmed
   siRNA targeting sequences used in this study. siRNAs sequences are
   found in Table [74]S1.

Real time quantitative PCR

   According to the manufacturer’s protocol, reverse transcription was
   performed to convert total RNA into cDNA, followed by quantitative PCR
   (qPCR) analysis on the Roche LightCycler^® 480 Real-Time PCR system.
   So, all primers were synthesized by Qingke Biotech Co., Ltd (Beijing,
   China). β-actin gene was used as an internal reference gene. The
   2^−ΔΔCt method was used to analyze the relative mRNA expression levels
   of genes. For this experiment, four biological repeats and technical
   repetitions were performed. Table [75]S2 contains the necessary primer
   sequences.

Western blotting

   Cultured AML12 cells were lysed with RIPA lysate. After centrifugation,
   the protein concentration was measured and diluted to 1 µg/µl, and 2X
   Laemmli loading buffer was added to load the buffer. The sample was
   boiled for 10 min, quickly frozen and stored at −20℃. The 20 µg protein
   was isolated with 10% SDS-PAGE gel and transferred to a PVDF membrane
   (Bio-Rad). The membranes were then cleaned with TBS-Tween and sealed
   with a sealer made of skim milk powder (5% skim milk powder) for 1 h.
   After blocking , the membranes were washed with TBST. The membranes
   were divided into two groups and incubated with the antibody against
   alpha-tubulin (Proteintech, China, 1:10,000 dilution), the antibody
   against  β-actin (Proteintech, China, 1:10,000 dilution) and the
   antibody against Trub2 (Proteintech, China, 1:2000 dilution).

Statistical analysis

   The experimental results are expressed as mean ± SD^[76]27.SPSS 20.0
   software (SPSS Inc., Chicago, IL, USA) was used to conduct t-tests to
   analyze the significance of intergroup differences, and one-way
   analysis of variance (ANOVA) was performed on two or more groups.
   Significance is represented by an asterisk: *P < 0.05, **P < 0.01 and
   **** P < 0.0001.

Results

Differential m6A methylation analysis between HFD and HB groups

   In order to investigate the disparity in m6A modification between the
   HFD and HB groups, we initially identified genes exhibiting
   significantly distinct m6A peaks between the two cohorts. We found
   2,035 common genes in the HB group compared to the HFD group, with
   significant differences in their m6A peaks. A total of 934 genes
   exhibited up-regulated m6A methylation levels, while 1,101 genes showed
   down-regulated m6A methylation (Fig. [77]1A).

Fig. 1.

   [78]Fig. 1
   [79]Open in a new tab

   The landscape of m6A methylation in the HFD and HB groups. (A) Volcano
   plot of differentially expressed m6A methylation mRNAs. (B) The pie
   chart shows the percentage of m6A peaks in the five non-overlapping
   segments of the mRNA transcript. (C) GO enrichment analysis showed the
   top 20 functional analyses of enriched differentially methylated mRNAs
   (P < 0.05). (D) KEGG pathway analysis showed the top 20 pathways of
   Upmethylated mRNAs (P < 0.05). (E) KEGG pathway analysis revealed the
   top 20 pathways of Downmethylated mRNAs (P < 0.05). HFD, High fat diet;
   HB, High fat diet with betaine; GO, Gene Ontology; KEGG, Encyclopedia
   of Genes and Genomes.

Distribution of m6A peaks on gene functional elements

   To identify the primary m6A localization within transcripts, we
   partitioned the m6A peak into five distinct and non-overlapping
   segments: 5’-untranslated regions (5’ UTR), exon, intron, stop codon,
   and 3’-untranslated regions (3’ UTR). Our results indicate that the
   differentially methylated m6A peak is predominantly concentrated in the
   vicinity of the stop codon and 3’ UTR (38.7%), where the stop codon
   accounts for 20.5% and the 3’ UTR accounts for 18.2%, sometimes also
   near the exon (38.2%) (Fig. [80]1B), which is consistent with previous
   experimental results^[81]28.

Analysis of differential methylation enrichment pathway between HFD and HB
groups

   To deepen our comprehension of the involvement of m6A modifications in
   the pathogenesis of MASLD, we carried out GO and KEGG^[82]22,[83]23
   pathway enrichment analyses on genes exhibiting significant
   differential m6A methylation patterns. GO analysis indicated that
   differentially m6A methylated mRNAs were significantly enriched in
   biological processes such as positive regulation of mitogen-activated
   protein kinase (MAPK) cascade, positive regulation of Notch receptor
   target gene transcription and dosage compensation. In terms of
   molecular function, it is enriched in phospholipid binding and
   phosphatidylinositol phosphate phosphatase activity. In terms of cell
   components, target of rapamycin (TOR) complex and Activating
   transcription factor 4 - the cAMP response element binding protein 1
   (ATF4-CREB1) transcription factor complex are enriched. In particular,
   a large number of differential m6A methylation genes are involved in
   the positive regulation of the mitogen-activated protein kinase (MAPK)
   cascade, phospholipid binding, and aromatic compound catabolic process
   (Fig. [84]1C). Some of the pathways mentioned above are associated with
   MASLD progression, such as phospholipid binding^[85]29, and
   phosphatidylinositol phosphate phosphatase activity^[86]30.

   KEGG pathway analysis disclosed that mRNAs subject to upregulated m6A
   methylation predominantly implicated the mechanistic target of
   rapamycin (mTOR) signaling cascade, the Ras-related protein 1 (Rap1)
   signaling pathway, and MAPK signaling pathway (Fig. [87]1D). mRNAs
   experiencing downregulated m6A methylation were predominantly
   associated with the wingless/integrated (Wnt) signaling pathway,
   neurotrophic factor signaling pathways, and transforming growth
   factor-beta (TGF-β) signaling cascades (Fig. [88]1E). mTOR signaling
   pathway is involved in inducing autophagy to reduce hepatic steatosis
   and promote fatty acid oxidation^[89]31. Studies have shown that CD147
   reprograms fatty acid metabolism in hepatocellular carcinoma cells
   through Akt/mTOR/SREBP1c and P38/PPARα pathways^[90]32. Deletion of
   Telomere-Rap1 aggravated cardiac aging by imparing fatty acid oxidation
   via PPARα signaling^[91]33. In addition, Qin et al. found that
   isoquercetin improved hepatic lipid accumulation in HFD-induced
   nonalcoholic fatty liver rat models by activating AMP-activated protein
   kinase (AMPK) pathway and inhibiting TGF-β signaling^[92]34, which is
   very consistent with our findings. Taken together, these data suggest
   that the underlying mechanism of m6A methylation in betaine in
   alleviating the progression of MASLD is closely related to the disorder
   of glucose and lipid metabolism and the activation of liver
   inflammatory response, but further evidence is needed to support this.

Differential transcriptome and functional analysis between HFD and HB groups

   To investigate the relationship between gene expression and phenotype,
   transcriptome differential analysis was conducted for the HFD and HB
   groups. A total of 348 DEGs were analyzed, with 173 up-regulated genes
   and 175 down-regulated genes, respectively (Fig. [93]2A). In the GO
   functional enrichment analysis, DEGs were significantly enriched in
   biological processes such as lipid biosynthesis, small molecule
   biosynthesis, lipid metabolism, and small molecule metabolism. Among
   them, the processes of cholesterol biosynthetic process, secondary
   alcohol biosynthetic process and sterol biosynthetic process are highly
   significant and reliable (Fig. [94]2B). In the KEGG^[95]22,[96]23
   enrichment analysis, DEGs were found to be mainly involved in pathways
   related to metabolism, human diseases, environmental information
   processing, and cellular processes. These included 36 genes associated
   with signal transduction mechanisms, 31 genes implicated in lipid
   metabolism, 22 genes relevant to endocrine and metabolic disorders, and
   13 genes involved in carbohydrate metabolism pathways (Fig. [97]2C).
   Reactome Gene Sets showed that in mRNAs with upregulated expression,
   DEGs were mainly enriched in reactions such as cholesterol
   biosynthesis^[98]35, metabolism of lipids, metabolism of
   steroids^[99]36and fatty acyl-coenzyme A (acyl-CoA)
   biosynthesis^[100]37. DEGs of downregulated mRNAs were mainly enriched
   in responses such as Generic Transcriptional pathway,
   Estrogen-dependent gene expression^[101]38, and ESR-mediated
   signaling^[102]39 (Fig. [103]2D, E). In summary, compared to the HFD
   group, the HB group displays an upregulated expression level for a
   substantial number of reactions implicated in liver lipid metabolism,
   particularly those promoting accelerated lipid catabolism and
   mitigating hepatic lipid accumulation.

Fig. 2.

   [104]Fig. 2
   [105]Open in a new tab

   Transcriptional differences between HFD and HB groups. (A) Volcano map
   of differentially expressed mRNAs. (B) GO enrichment analysis showed
   the top 10 functional analyses of enriched differential mRNAs
   (P < 0.05). (C) The number of differential genes enriched by different
   types of pathways on KEGG (P < 0.05). (D) Reactome Gene Sets showed the
   top 15 pathways of upregulated mRNAs (P < 0.05). (E) Reactome Gene Sets
   showed the top 15 pathways of downregulated mRNAs (P < 0.05).

Combined analysis of enrichment pathways of differentially methylated genes
and differentially expressed genes

   To elucidate the relationship between differentially methylated genes
   and differentially expressed genes, we initially discerned the
   expression upregulation pathways in which they both participate. In the
   context of GO Biological Processes, we ascertained that 670 distinct
   processes were engaged by differentially methylated mRNAs, while 244
   processes were implicated in differentially expressed mRNAs. In
   particular, there were 31 biological processes (7.28%) between the two
   groups, among which the mRNA methylation of negative regulation of
   phosphorus metabolic process was significantly different (Fig. [106]3A,
   B). These results suggest that the methylation status of mRNA
   associated with negative regulation of phosphorus metabolism was
   significantly altered after betaine administration. The
   KEGG^[107]22,[108]23 pathway analysis unveiled that 23 pathways were
   linked to differentially expressed methylated mRNAs, and 17 pathways
   were associated with differentially expressed mRNAs. Among these, 1
   pathway (5.26%) was common to both sets, the alcoholic liver disease
   pathway (Fig. [109]3C, F). Per the Wiki pathway resource, 17 pathways
   were connected to differentially expressed methylated mRNAs, and 14
   pathways were relevant to differentially expressed mRNAs, with 2
   pathways (14.81%) being mutual to both, namely, lipogenesis gene
   networks and metabolic pathways of vitamins and cofactors
   (Fig. [110]3D, F). Lastly, Reactome Gene Sets demonstrated that 110
   pathways were involved in differentially expressed methylated mRNAs,
   and 48 pathways pertained to differentially expressed mRNAs. Of these,
   5 pathways (6.76%) were common to both categories, all pertaining to
   chromatin organization (Fig. [111]3E, F).

Fig. 3.

   [112]Fig. 3
   [113]Open in a new tab

   Venn diagram of overlapped genes of DE methylated mRNAs and DE mRNAs
   and their pathways. (A) The Venn diagram showed the overlap among GO
   biological processes of DE methylated mRNAs and DE mRNAs. (B) Bubble
   plots showed pathways for overlapping genes in GO biological processes.
   (C) The Venn diagram showed the overlap of pathways in KEGG among DE
   methylated mRNAs and DE mRNAs. (D) The Venn diagram showed the overlap
   of pathways in Wiki among DE methylated mRNAs and DE mRNAs. (E) The
   Venn diagram showed the overlap of pathways in Reactome Gene Sets among
   DE methylated mRNAs and DE mRNAs. (F) Bubble plots showed pathways in
   which genes overlap in KEGG, Wiki, Reactome Gene Sets. DE,
   differentially expressed.

   In summary, relative to the HFD group, the HB group exhibits elevated
   m6A levels and mRNA expression in lipid metabolism and negative
   regulation of phosphorus metabolism^[114]40. This suggests that betaine
   may regulate the stability of mRNA by mobilizing these pathways, affect
   the metabolic process, and thus inhibit lipid accumulation, thereby
   inhibiting the formation of obesity and reducing fatty liver.

Gene mRNA-level regulation by m6A modification

   To identify key genes alleviating MASLD, we analyzed differentially
   methylated and expressed genes in HFD and HB groups. After filtering
   out genes with insignificant differences in m6A abundance and mRNA
   expression, 27 genes showed significant differences, suggesting they
   may be regulated by m6A modification. Among them, 11 transcripts showed
   m6A hypermethylation and up-regulated expression (Hyper-up), 13
   transcripts displayed m6A hypermethylation and down-regulated
   expression (Hyper-down), no transcripts exhibited m6A hypomethylation
   and up-regulated expression (Hypo-up), and 3 transcripts had m6A
   hypomethylation and down-regulated expression (Hypo-down)
   (Fig. [115]4A). To assess the reliability of the RNA-seq data, five
   genes (Appl1, Ntrk2, Cbx4, Trub2 and Tlr4) were randomly selected from
   27 DEGs for validation by RT-PCR. Consistently, the mRNA levels of
   Appl1, Ntrk2, and Trub2 were increased, while Cbx4 and Tlr4 mRNA levels
   were decreased in the HB group compared to the HFD group
   (Fig. [116]4B).

Fig. 4.

   [117]Fig. 4
   [118]Open in a new tab

   (A) The four-quadrant plots showed significant changes in mRNA
   expression and m6A methylation in gene distribution. (B) The mRNA level
   of Appl1, Ntrk2, Trub2, Cbx4 and Tlr4 (n = 3 or 4). (C) Dot blot
   detecting the global m6A modification of mRNAs. Methylene blue (MB)
   staining was used as a loading control. Gel images were cropped, and
   the original gels were presented in Supplementary Figure S2. (D)
   MeRIP-seq detected the m6A abundance of Trub2 mRNA in the HFD group and
   HB group. The yellow highlight showed that the intensity of m6A peak
   was significantly higher in the HB group than in the HFD group. (E) The
   protein level of Trub2 in the liver of CK, HFD, HB mice (n = 3). Gel
   images were cropped, and the original gels were presented in
   Supplementary Figure S3. (F) Quantitative results of Trub2 protein
   levels. (G-H) Western blot analysis was employed to assess and quantify
   the expression efficiency of Trub2 in OA/PA-treated AML12 cells after
   its overexpression (n = 4). Gel images were cropped, and the original
   gels were presented in Supplementary Figure S4. (I) Triglyceride levels
   in AML12 cells after transfected with Trub2 after OA/PA treatment
   (n = 3). (J) Triglyceride content in AML12 cells after transfected with
   siNC and siTrub2 (n = 6). The data were expressed as the mean ± SD.
   *P < 0.05, **P < 0.01, ****P < 0.0001. CK, Control mice; HFD, High fat
   diet; HB, High fat diet with betaine.

   To understand the changes in m6A modification caused by betaine
   treatment, RNA m6A dot blot was performed to measure the levels of m6A
   modification. The results revealed that the m6A levels were increased
   in liver tissues of mice treated with betaine (Fig. [119]4C).
   Interestingly, some genes involved in mRNA processing showed
   significant elevations in m6A methylation and mRNA expression. For
   example, Trub2, which is associated with mRNA processing, had
   significantly higher m6A methylation and mRNA levels in the HB group
   than in the HFD group (Fig. [120]4D). Further examination showed that
   the protein level of Trub2 was significantly increased in the HB group
   compared with that in the HFD group (Fig. [121]4E-F). We selected
   Trub2, a Hyper-up gene not previously reported to be involved in MASLD
   progression, as a candidate gene for further analysis.

Betaine reduces intracellular TG levels by regulating Trub2 expression

   To further verify the role of Trub2 in betaine attenuating MASLD, Trub2
   was overexpressed in AML12 cells, and its expression level was detected
   by RT-PCR (Figure [122]S1A). The protein levels of Trub2 were
   significantly elevated in AML12 cells treated with OA/PA after
   overexpression of Trub2 (Fig. [123]4G-H). Notably, intracellular TG was
   significantly reduced after Trub2 overexpression in AML12 cells
   (Fig. [124]4I). This result reveals that Trub2 is a novel target for
   regulating hepatic lipid metabolism. In addition, inhibition of Trub2
   increased intracellular TG levels in AML12 cells, and this increase was
   attenuated by betaine (Fig. [125]4J, Figure [126]S1B). These results
   demonstrate that betaine inhibits intracellular TG accumulation by
   increasing Trub2 expression.

   In summary, betaine increases Trub2 mRNA levels by enhancing its m6A
   methylation level and the stability of the mRNA, and this results in a
   reduction of intracellular TG level, thereby exerting a favorable
   influence on the improvement of MASLD (The whole article is summarized
   in Fig. [127]5).

Fig. 5.

   [128]Fig. 5
   [129]Open in a new tab

   Overview of the experimental design, analysis and validation. The graph
   showed that by combining RNA-seq and MeRIP-seq, we found many candidate
   genes related to lipid metabolism, including Trub2, and the major
   enrichment pathways such as mTOR signaling pathway, Rap1 signaling
   pathway and MAPK signaling pathway. These potential candidate genes and
   pathways are critical to understanding the beneficial effects of
   betaine production, and further research is needed to determine the
   precise processes by which they alleviate metabolic
   dysfunction-associated steatotic liver disease. In conclusion, this
   study provides a new resource for determining the remission effect of
   betaine on metabolic dysfunction-associated steatotic liver disease.

Discussion

   Betaine is a crucial natural methyl donor that plays a significant role
   in metabolism and liver function. Recent research has highlighted its
   active role in fat metabolism. The main objectives of this project are
   to explore betaine’s potential in alleviating hepatic steatosis and to
   map m6A RNA methylation. This study provides insights into betaine’s
   biological function in liver fat metabolism and its regulatory
   mechanisms, offering a solid scientific basis for its application in
   treating MASLD.

   Our study supports the notion that differences in m6A modifications
   between HFD and HB groups significantly affect the expression of lipid
   metabolism-related genes. GO functional enrichment analysis showed that
   m6A-modified genes were enriched in processes like the MAPK cascade,
   Notch receptor target transcription, and dose compensation
   (Fig. [130]1C). Notably, pathways involving positive regulation of MAPK
   cascades and phosphatidylinositol phosphate phosphatase activity
   contained many differentially methylated genes, highlighting m6A
   methylation’s role in regulating lipid metabolism and deposition. This
   enhances our understanding of fat metabolism and provides insights into
   the mechanisms of m6A methylation.

   Moreover, KEGG pathway enrichment analysis revealed that downmethylated
   mRNAs were linked to several critical pathways, including the Wnt
   signaling pathway, animal autophagy, neurotrophic factor signaling,
   renal cell carcinoma, TGF-β pathway, mRNA surveillance, and cAMP
   signaling (Fig. [131]1E). The Wnt signaling pathway is mostly inactive
   in mature healthy livers, but it is consistently overactivated during
   cell renewal or regeneration, as well as in liver fibrosis, cirrhosis,
   and liver cancer^[132]41. In our study, the Wnt pathway in the betaine
   group showed downregulated mRNA m6A methylation, suggesting that
   betaine might mitigate fatty liver and prevent progression to liver
   cirrhosis and cancer by reducing m6A methylation of Wnt pathway-related
   mRNA-a novel finding. While downmethylated mRNA holds significant
   research value for studying liver fat deposition, we surmise that these
   pathways do not play a dominant role in our context. Therefore, we
   investigated mRNA with upregulated methylation for further insights.

   We found that upmethylated mRNAs are strongly associated with a variety
   of biological pathways, including biotin metabolism, lysine
   degradation, Rap1 signaling pathway, mTOR signaling pathway, and
   adhesion junction (Fig. [133]1B). The J Pearce’s study illustrates the
   relationship between biotin deficiency and liver metabolism with fatty
   liver and renal syndrome^[134]42. Non-histone lysine crotonization is
   involved in the regulation of browning of white fat^[135]43. In
   DiZhang’s study, thymoquone was successfully shown to improve glucose
   homeostasis through AMPK/mTOR/ULK1-dependent pathways, alleviate
   hepatic fatty acid salt disease, and reduce hepatic lipid accumulation
   associated with autophagy induction in mice fed a high-fat
   diet^[136]44. In comparison to the downmethylated mRNA pathway, we
   found that the upmethylated mRNA biological pathway in this experiment
   may exhibit a more prominent performance. By KEGG pathway enrichment
   analysis, we further explored the biological pathway affecting mRNA
   methylation level, thus providing a foundation for us to specifically
   explore how m6A affects liver fat deposition or metabolism.

   In this study, the HFD and HB groups showed significant differences at
   the mRNA level. We concentrated on discerning the disparities in
   expression between the two cohorts, with the aim of elucidating the
   underlying biological mechanisms and implications (Fig. [137]2A). From
   a molecular standpoint, GO functional enrichment analysis highlighted
   that the upregulated mRNA was largely concentrated within processes
   involving lipid biosynthesis, small molecule biosynthesis, and both
   lipid and small molecule metabolism (Fig. [138]2B-E). This led to an
   acceleration of liver lipid metabolism and a reduction in TG levels in
   liver cells. Our study also shows that m6A methylation modification
   does play a very crucial role in lipid metabolism, and further proves
   that m6A may regulate gene expression through methylation modification,
   thereby affecting fat deposition metabolism.

   In further research, we have uncovered those co-expressed genes
   enriched in several key pathways associated with negative regulation of
   phosphorus metabolism, alcoholic fatty liver disease, lipid metabolism
   genes, as well as vitamin and cofactor metabolism, show significant
   perturbations at both the m6A and mRNA levels in the betaine-treated
   group (Fig. [139]3A-F). We speculated that these pathways may indicate
   that under the influence of betaine, lipogenesis is significantly
   suppressed. Consequently, betaine holds potential therapeutic value in
   terms of promoting lipid metabolism and preventing the development of
   MASLD. This study provides robust theoretical backing for the
   application of betaine in the treatment of fatty liver conditions.

   27 candidate genes affected by betaine therapy in MASLD models have
   been screened out. Interestingly, the Appl1-rs4640525 variant has been
   proved to be associated with MASLD, with the C allele being more
   prevalent in individuals with MASLD compared to the controls^[140]45.
   However, the mechanisms by which Appl1 regulates MASLD have not been
   fully reported. It has been hypothesized that SNPs may influence
   Appl1 expression levels, thereby altering adiponectin activity. Such
   alterations in adiponectin activity could subsequently lead to reduced
   fatty acid oxidation, potentially contributing to the development of
   fatty liver^[141]46. In addition, betaine treatment reduced serum TC
   and TG levels and the expression of Tlr4in HFD-induced MASLD
   rats^[142]47. These results suggest the beneficial efficacy of betaine
   and the reliability of the candidate genes.

   Multiple studies have demonstrated the critical role of m6A
   modification in the regulation of mRNA stability, splicing, transport,
   localization, and translation. For instance, the m6A demethylase ALKBH5
   facilitates m6A demethylation of glucose transporters type 4 (GLUT4)
   mRNA and enhances its stability in a YTHDF2-dependent manner^[143]48.
   Furthermore, m6A methylation is associated with mRNA degradation, which
   is mainly manifested as reduced stability of the transcript^[144]49.
   During the investigation of differential m6A peaks, we unexpectedly
   discovered that the Trub2 gene transcript level was markedly
   overexpressed in the livers of betaine-treated mice, challenging our
   previous understanding of the Trub2 gene and presenting a fresh avenue
   for exploring the related mechanisms of MASLD. Historically, the Trub2
   gene has been primarily associated with pseudouridine synthase
   activity, contributing to the positive regulation of mRNA pseudouridine
   synthesis and mitochondrial translation, and functioning either
   upstream or within mRNA and tRNA processing events. At the cellular
   level, overexpression of Trub2 significantly decreased the triglyceride
   content of OA/PA treated cells (Fig. [145]4I). From a molecular biology
   standpoint, it seems plausible that the enhanced methylation level of
   m6A might promote the functional expression of Trub2, thereby boosting
   mitochondrial function and lipid metabolism. This finding reiterates
   the significance of m6A in modulating mRNA stability via methylation
   and underscores its pivotal role in regulating the expression of
   lipid-associated genes.

   In summary, this project has further solidified the role of betaine in
   mitigating MASLD and elucidated the mechanism by which betaine
   alleviates MASLD through the lens of m6A methylation, successfully
   constructing an m6A methylation landscape. More notably, we have also
   identified a novel candidate gene, Trub2, which furnishes a fresh
   target for studying MASLD.

   Remarkably, betaine, being a naturally occurring amino acid derivative
   ubiquitous in nature, exhibits distinctive biochemical properties and
   low cost, rendering it a potential therapeutic agent for treating
   MASLD. This suggests its potential to become a transformative
   intervention for MASLD patients and presents exciting prospects for
   future development. As applied sciences continue to evolve, we are
   justified in believing that betaine will assume an increasingly pivotal
   role in the management of MASLD, offering a ray of hope to the
   multitude of affected patients.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [146]Supplementary Material 1^ (34.3MB, tif)
   [147]Supplementary Material 2^ (34.3MB, tif)
   [148]Supplementary Material 3^ (34.3MB, tif)
   [149]Supplementary Material 4^ (130.4KB, jpg)

Acknowledgements