Abstract
Objective
Metabolic dysfunction-associated steatohepatitis (MASH) is characterized by inflammation, fibrosis, and accumulation of fatty acids in the liver. MASH disease progression has been associated with reduced thyroid hormone (TH) signaling in the liver, including reduced expression of deiodinase type I (DIO1) and TH receptor beta (THRB). However, the underlying mechanisms mediating these effects remain elusive. Here, we hypothesized that epigenetic mechanisms may be involved in modulating hepatic TH action.
Methods
Liver samples from patients with and without MASH were analyzed by qRT-PCR and correlated with clinical parameters. Luciferase reporter assays and overexpression of miRNA in HepG2 cells were used to validate the functional binding of miRNA to predicted targets. DNA methylation was analyzed by bisulfite pyrosequencing.
Results
miR-34a-5p was upregulated in MASH patients and correlated positively with the clinical parameters of MASH. Using in silico and in vitro analysis, we demonstrate that miR-34a-5p is capable of targeting several modulators of local hepatic TH action, as evidenced by the functional binding of miR-34a-5p to the seed sequence in the THRB and DIO1 genes. Consequently, overexpression of miR-34a-5p in HepG2 cells reduced the expression of THRA, THRB, DIO1, and SLC10A1, thus potentially mediating an acquired hepatic resistance to TH in MASH. As an additional regulatory mechanism, DNA methylation of THRB intron 1 was increased in MASH and negatively correlated with THRB expression.
Conclusion
miR-34a-5p constitutes a possible epigenetic master regulator of hepatic TH action, which together with THRB-specific DNA methylation could explain a possible developing TH resistance in the liver during MASH progression on the molecular level.
Introduction
Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as non-alcoholic fatty liver disease (NAFLD), is the hepatic manifestation of the metabolic syndrome and is characterized by the accumulation of fatty acids in hepatocytes (1). To date, MASLD affects 30% of the global population (2) and is one of the main indicators for liver transplantation (3). The more severe stage, metabolic dysfunction-associated steatohepatitis (MASH), is defined by additional inflammation and fibrosis, with an increased risk of permanent liver damage, cirrhosis, and hepatocellular carcinoma (4).
The hepatic glucose and lipid metabolism, which is dysregulated in MASLD, is strongly regulated by thyroid hormone (TH) action (5, 6). Previously, we have shown that the MASLD activity score (MAS) is associated with reduced TH receptor beta (THRB) expression in the liver of obese individuals (7). Additionally, hypothyroidism is a known risk factor for MASLD progression, severity, and development of fibrosis (8). The connection between THRB and MASLD is underlined by the successful phase 3 clinical trial of resmetirom, a liver-specific THRB agonist, which emphasizes the positive effect of hepatic TH signaling activation on MASLD (9). The activity of TH in the liver is not only dependent on the availability of THRB but also on the import of TH into the liver by specific transporters such as solute carrier family 10 member 1 (SLC10A1) or family 16 member 2 (SLC16A2). Moreover, the subsequent intracellular activation of thyroxine (T4) into the active form 3,3′,5-triiodothyronine (T3) by deiodinase type 1 (DIO1) is a second layer controlling local TH action. However, little is known about the regulation of these genes during liver disease.
Epigenetic mechanisms dynamically modify gene expression and protein synthesis through histone modification, DNA and RNA methylation, and non-coding RNAs. Cytosines are commonly methylated on their fifth carbon atom in the context of CpG dinucleotides, modifying the binding affinity of DNA-binding proteins. Therefore, changes in DNA methylation lead to alterations in gene transcription, and differently methylated cytosines (DMCs) are known risk factors for metabolic diseases (10). While DNA methylation works on the transcriptional level, micro-RNAs (miRNAs) influence translation. miRNAs are small, ~22 nucleotide-long, non-coding RNAs that bind to target mRNAs through the short complementary seed sequence and regulate protein production by initiating ribosomal drop off, preventing mRNA circularization and causing mRNA degradation (11). The role of epigenetics in metabolic diseases has been well established (12); however, the alteration of TH signaling in MASLD by epigenetics has not been investigated yet.
Materials and methods
Human cohort
Liver biopsies of patients with obesity were obtained from segment III during bariatric surgery at the University Hospital Eppendorf (Hamburg, Germany) (7). The study was approved by the local ethics committee (PV4889), and all participants signed an informed consent. The MAS was determined by two pathologists according to current standards. Clinical parameters were determined as described previously (7). Cohort statistics and clinical parameters are represented in Table 1.
Clinical parameters of the cohort. Values in bold indicate statistical significance. Data are presented as n or as mean±S.D.
| Parameters | Non-MASH | MASH | P-value | n |
|---|---|---|---|---|
| Sex | 68 | |||
| M/F (% M) | 6/35 (14.6%) | 9/18 (33.3%) | ||
| Age, years | 40.07 ± 12.53 | 49.07 ± 13.11 | 0.0059 | 68 |
| BMI, kg/m2 | 51.97 ± 11.75 | 54.37 ± 8.582 | 0.3649 | 68 |
| TSH, mU/L | 2.015 ± 1.125 | 3.62 ± 8.467 | 0.5669 | 63 |
| T3, ng/mL | 0.8872 ± 0.2492 | 1.048 ± 0.1480 | 0.0057 | 59 |
| T4, µg/dL | 5.183 ± 1.124 | 5.872 ± 1.438 | 0.0448 | 59 |
| Triglycerides, mmol/L | 176.8 ± 98.18 | 255.4 ± 140.5 | 0.0120 | 66 |
| Total cholesterol, mg/dL | 182.3 ± 35.79 | 195.8 ± 50.84 | 0.4832 | 66 |
| LDL, mg/dL | 102.7 ± 36.11 | 99.67 ± 36.00 | 0.7598 | 60 |
| HbA1c, % | 5.880 ± 1.665 | 7.222 ± 1.934 | 0.0005 | 68 |
| Blood glucose, mg/dL | 110.1 ± 41.86 | 154.5 ± 72.36 | 0.0003 | 68 |
| Diabetes diagnosis | 68 | |||
| Yes/No (% Yes) | 9/32 (21.9%) | 16/11 (59.2%) | ||
| MASLD activity score, n | 68 | |||
| 0 | 19 | |||
| 1 | 5 | |||
| 2 | 10 | |||
| 3 | 7 | |||
| 4 | 9 | |||
| 5 | 10 | |||
| 6 | 8 | |||
| Fibrosis diagnosis, n | 68 | |||
| Yes/No (% Yes) | 9/32 (21.9%) | 21/6 (77.7%) | ||
| AST, U/L | 19.78 ± 9.627 | 37.56±19.86 | <0.0001 | 68 |
| ALT, U/L | 24.95 ± 16.18 | 41.74±18.98 | <0.0001 | 68 |
M, males; F, females; BMI, body mass index; TSH, thyroid-stimulating hormone; LDL, low-density-lipoprotein; MASH, Metabolic dysregulation-associated steatohepatitis; AST, Aspartate aminotransferase; ALT, Alanine aminotransferase.
Cell culture
HepG2 cells (American Type Culture Collection (ATCC HB-8065)) were maintained in DMEM medium (Gibco) with 1.5 g/L glucose, 10% fetal bovine serum (FBS), and 1% Pen/Strep at 37°C and 5% CO2. Seeding and transfection of HepG2 cells were performed simultaneously onto a six-well plate (Sarstedt, Nümbrecht, Germany). The transfection mix for one well contained 10 nM miRVana™ (Life Technologies) miR-34a-5p mimic or miR-34a-5p inhibitor or negative controls nc#1 or nc#1 inhibitor, with RNAiMAX (Thermo Fisher Scientific). Cells were incubated for 48 h before harvest. The experiment was performed with technical and biological triplicates. To test for the effect of miR-34a-5p on T3 response, HepG2 cells were kept in charcoal-stripped FBS medium for 48 h before transfection with either one – miR-34a-5p mimic, nc#1, miR-34a-5p inhibitor, or nc#1 inhibitor. Each condition was treated with 0 nM, 10 nM, or 50 nM T3 directly after transfection, and after 24 h, the cells were harvested after a total incubation time of 48 h. For this experiment, technical duplicates and biological triplicates were performed. Cell passage numbers between 15 and 20 were used for experiments.
RNA extraction, cDNA synthesis, and RT-qPCR
RNA was extracted with the miRNeasy Mini Kit (Qiagen) with on-column DNase treatment. About 1000 ng RNA were transcribed into DNA using the high-capacity cDNA reverse transcription kit (Thermo Fisher Scientific). For miRNA detection, RT-qPCR with TaqMan Fast Advanced Master Mix was performed using the assays hsa-miR-34a-5p (478048_mir) and housekeeper hsa-miR-24-3p (477992_mir). Target genes were analyzed with FastStart Universal SYBR Green (Roche). The primer sequences were: CASC3, forward-5′ACC TCG GAA AGG GCT CTT CTT and reverse-5′CGA CCC TCA TCC TTC CAT AGC; THRA, forward-5′AGG TCA CCA GAT GGA AAG CG and reverse-5′AGT GAT AAC CAG TTG CCT TGT C; THRB, forward-5′TGG GAC AAA CCG AAG CAC TG and reverse-5′TGG CTC TTC CTA TGT AGG CAG; DIO1, forward-5′GTC GTG GGT AAA GTG CTT CTG and reverse-5′GTT CCG CTT GAC TCT GTC TGG; SLC10A1, forward-5′AAG GAC AAG GTG CCC TAT AAA GG and reverse-5′TTG AGG ACG ATC CCT ATG GTG; SLC16A2, forward-5′CCA CGC CTA CGG TAG AGA C and reverse-5′CAG AGT TAT GGA TGC CGA AGA TG; CYP7A1, forward-5′AGCATTGACCCGATGGATGG and reverse-5′TCCGTGAGGGAATTCAAGGC. The relative miRNA expression was calculated with the ΔΔCt method and mRNA expression after Pfaffl (13). Gene expression was normalized to CASC3 and miR-34a-5p to miR-24-3p.
Luciferase reporter assay
HEK293T cells (ATCC) were maintained in DMEM (Gibco, 4.5g/L glucose, 10% FBS, 1% Pen/Strep) at 37°C and 5% CO2. Cell passages 10–15 were used for experiments. The cells were seeded on a 96-well plate (Sarstedt) and transfected with 100 ng of the pmirGLO vector (Promega) containing 3′UTR of THRA, THRB, DIO1, or SLC16A2 approximately ±200 bp from the seed sequence, and 10 nM miR-34a-5p mimic miRVana™ or scrambled nc#1 for control and incubated for 48 h. An additional control measurement was also conducted with the seed sequence mutated in the THRB 3′UTR from ACTGCC to ACTTCC. Each experiment was conducted with three technical and five biological replicates. Luciferase activity was measured with the dual-luciferase reporter assay (Promega) using the ClarioStar plate reader.
RNA-interacting protein immunoprecipitation (RIP)-sequencing
For RIP-seq, HepG2 cells were maintained in DMEM medium (Gibco) with 1.5 g/L glucose, 10% FBS, and 1% Pen/Strep at 37°C and 5% CO2. 1×106 cells were seeded in each well of a six-well plate and simultaneously transfected with 10 nM miRVana™ miR-34a-5p mimic or nc#1 for 24 h before harvest. Three wells were combined for one experiment, and the protocol from Meier et al. (14) was followed with the following modifications: lysate and beads were incubated at 4°C overnight and protein degradation was carried out in NT2 buffer. Five micrograms of both argonaute 2 (AGO2) antibodies (Abnova, clone 2E12-1C9 LOT M9221-S2, and Sigma Aldrich, clone 11A9 LOT3894983) were used. The experiment was repeated three times to generate a biological triplicate. RNA isolation was performed with miRNeasy (Qiagen) using the protocol for RNA isolation from cells. Input control samples were rRNA depleted using the RiboCop rRNA Depletion Kit from Lexogen. Library prep, RNA sequencing, and analysis of data were performed by Novogene (München, Germany). Detected genes with a log2-fold change of at least 1 were considered bound to the AGO2.
DNA methylation
DNA isolation, bisulfite conversion, assay validation, and pyrosequencing were performed as described previously (15). For primer design, to detect DNA methylation at chromosome 3 position 24477865 (C-24477865), the reference genome hg19 was used. The following primers were designed using the PyroMark Assay Design 2.0 software (Qiagen): forward-5′biotin GGT TAG GTT TAG AGG AAA GTT AAA AAG TA, reverse-5′ TAC ACC ACT CTA CAT TCC TCA TAA TCT C, and sequencing-5′CTC AAA AAA AAA TAA CCC AAT T.
Target gene prediction
Target genes for miR-34a-5p were predicted as described previously (16).
Statistical analysis
Statistical analysis was performed with GraphPad Prism 10 using the unpaired two-tailed t-test, Mann–Whitney, multiple t-tests, and one- or two-way ANOVA when appropriate. Correlations were calculated using Pearson correlation, and for partial correlation, Jamovi 1.8.1 was used. The correlation matrix was generated with Python using the packages pandas, numpy, seaborn, and matplotlip.pyplot. Methylation results were analyzed with the Mann–Whitney test, and correlations were calculated using the Spearman correlation test. Results under P < 0.05 were considered significant; and multiple testing was corrected using the Benjamini–Hochberg method.
Results
miR-34a-5p is upregulated in patients with MASH
To test the hypothesis that the dysregulation of TH metabolism in MASH could be facilitated by epigenetic mechanisms, we used data from our previously published miRNA 4.0 array in the livers of obese humans (16) to identify regulated miRNAs with the potential to bind THRB by in silico analysis. Among those, miR-34a-5p was identified as the most promising candidate due to its significant correlation with the MAS (P = 0.027), confirming previous findings of increased miR-34a-5p in association with MASLD (17). Intriguingly, analysis for 8mer, 7mer, and 6mer seed sequences subsequently identified multiple TH metabolism-related genes (THRA, DIO1, SLC10A1, and SLC16A2) as potential miR-34a-5p targets (Fig. 1A), indicating that miR-34a-5p could act as an epigenetic master regulator to suppress TH function in the liver. Increased expression of miR-34a-5p in MASH was validated by individual qPCR in the same liver biopsies as used for the miRNA 4.0 array, and additional patients with morbid obesity (BMI ≥ 35, total n = 68) were classified according to their MAS into non-MASH (≤3 MAS, n = 37) and MASH (≥4 MAS, n = 29) (Table 1). miR-34a-5p was overexpressed by a fold-change of 1.7 (P = 0.0006) in liver biopsies derived from MASH patients compared to non-MASH (Fig. 1B). Correlation analysis of miR-34a-5p expression with mRNA of possible target genes and clinical parameters showed negative correlations between miR-34a-5p and SLC10A1 (r = −0.26, P = 0.01) and SLC16A2 (r = −0.42, P < 0.001) and positive correlations with HbA1c (r = 0.56, P = 0.007), glucose (r = 0.59, P = 0.001), liver fat content (r = 0.57, P ≤ 0.001), and MAS (r = 0.48, P = 0.006), after correction for age and gender (Fig. 1C).
miR-34a-5p in a human cohort with obesity. (A) miR-34a-5p sequence is shown compared to predicted binding sites. (B) Relative expression of miR-34a-5p in a human cohort with and without MASH (MAS ≥ 4, n = 27 and MAS ≤ 3, n = 41), measured by RT-qPCR, shown as mean ± s.d. (C) Pearson correlation of miR-34a-5p with predicted target gene expression and clinical parameters. *<0.05, **<0.01, ***<0.001, ****<0.0001.
Citation: European Thyroid Journal 13, 5; 10.1530/ETJ-24-0080
Overexpression of miR-34a-5p in HepG2 cells represses genes of TH metabolism
To study the effect of miR-34a-5p on the expression of genes controlling TH metabolism, we transfected HepG2 cells with a miR-34a-5p mimic or inhibitor and quantified mRNA of the predicted target genes using qPCR. Overexpression of miR-34a-5p (Fig. 2A) resulted in significant downregulation of THRA (P = 0.0227), THRB (P = 0.0014), DIO1 (P = 0.0014), and SLC10A1 (P = 0.0017). Interestingly, SLC16A2 (P = 0.0014) was upregulated (Fig. 2B), suggesting an additional indirect mechanism potentially involving circadian factors as shown previously (18). In contrast, inhibition of miR-34a-5p in HepG2 cells (Fig. 2C) resulted in significant upregulation of THRB (P = 0.0187), whereas the other predicted target genes remained unchanged (Fig. 2D). To verify the effect of miR-34a-5p overexpression on TH action, we transfected HepG2 cells with mimic and inhibitor, stimulated with T3 at different concentrations, and quantified the downstream target genes DIO1 and cytochrome P450 family 7 subfamily A member 1 (CYP7A1). Both DIO1 and CYP7A1 showed the expected positive regulation upon T3 stimulation but exhibited indeed significantly reduced expression after mimic transfection and simultaneous stimulation with 10 nM T3 (CYP7A1P = 0.0170, DIO1P = 0.0003) and 50 nM T3 (CYP7A1P = 0.0113, DIO1P = 0.0003) compared to nc#1-treated cells (Fig. 2E and F). Transfection with a miRNA-34a-5p inhibitor did not result in significant changes in DIO1 or CYP7A1 expression (data not shown).
Expression of miR-34a-5p in HepG2 cells. (A) Expression of miR-34a-5p after treatment with miR-34a-5p mimic. (B) Expression of predicted target genes after mimic treatment. (C) Expression of miR-34a-5p after treatment with miR-34a-5p inhibitor. (D) Expression of target genes after inhibitor treatment. (E) Expression of miR-34a-5p after treatment with miR-34a-5p mimic and T3 stimulation at different concentrations. (F) Expression of T3-responsive genes after miR-34a-5p mimic transfection with T3 stimulation. Data shown as mean ± s.e.m. *<0.05, **<0.01, ***<0.001, ****<0.0001.
Citation: European Thyroid Journal 13, 5; 10.1530/ETJ-24-0080
Luciferase reporter assays and RIP-seq validate miR-34a-5p and predicted target binding
For direct validation of target binding of miR-34a-5p to the TH-related genes, we performed luciferase reporter assays using the pmirGLO vector system. While THRA and SLC10A1 expressions were not changed, luciferase expression was significantly reduced by 28% in plasmids containing the THRB-binding site compared to nc#1 (P = 0.0005). The specificity of this effect was confirmed by using a mutated THRB seed sequence, which did not result in a significant reduction of luciferase expression (Fig. 3A). Co-transfection of a plasmid containing the DIO1 seed sequence and miR-34a-5p mimic significantly reduced the luciferase expression by 21% (P = 0.0031) (Fig. 3B). This confirms that miR-34a-5p is capable of binding the sequences and subsequent regulation of THRB and DIO1 expression.
Target binding of miR-34a-5p. Luciferase reporter assay of (A) THRB 3′UTR and (B) DIO1 3′UTR with miR-34a-5p treatment, depicted as mean ± s.d., n = 5, **<0.01, ***<0.001. (C) KEGG pathway analysis of RIP-seq data for miR-34a-5p.
Citation: European Thyroid Journal 13, 5; 10.1530/ETJ-24-0080
In addition, RIP-seq was performed to identify mRNAs that are directly targeted by miR-34a-5p. Since binding between miRNA and mRNA is facilitated by the Argonaute 2 (AGO2) protein, we transfected HepG2 cells with miR-34a-5p mimic and control and used AGO-specific antibodies to identify mRNAs bound to this complex. Only gene products that were enriched at least two-fold in the miR-34a-5p condition were considered. Here, we identified a total of 544 gene products to be present at the Argonaute miRNA complex, of which 265 were predicted in silico as miR-34a-5p target genes (Supplementary Table 1, see the section on supplementary materials given at the end of this article). This included THRB, which was two-fold increased in cells treated with miR-34a-5p mimic as compared to the input control, thus confirming that overexpression of miR-34a-5p causes an increased localization of THRB mRNA to the degradation complex. A pathway analysis of all bound genes resulted in significant enrichment of metabolic pathways as well as genes, like carbohydrate-responsive element-binding protein (ChREBP) or adiponectin receptor 1 (ADIPOR1), involved in MASLD (Fig. 3C). In conclusion, we could experimentally validate the targeting of THRB and DIO1 by miR-34a-5p in human liver cells and identified multiple other genes that are targeted by miR-34a-5p to the AGO2 and are associated with metabolic pathways and MASLD.
DNA methylation of THRB is increased in liver of MASH patients
Because epigenetic mechanisms often act synergistically (19), we next investigated DNA methylation of the THRB and DIO1 genes in the liver biopsies of patients without and with MASH. While the DNA methylation in DIO1 was not altered, we identified a DMC in intron 1 of the THRB gene. The region was selected for analysis due to the relevance of DNA methylation in intron 1 (20) and the high number of predicted transcription factor-binding sites according to JASPAR 2022 (21). Pyrosequencing of this cytosine at chromosome 3 position 24477865 (C-24477865) of THRB showed significantly increased DNA methylation by 9.7% (P = 0.002) in the liver of MASH patients (Fig. 4A). After adjustments for the confounding factors age and gender, C-24477865 methylation correlated negatively with THRB expression (r = −0.51, P < 0.001) (Fig. 4B) and positively with the disease severity markers MAS (r = 0.47, P = 0.004), HbA1c (r = 0.44, P = 0.02), aspartate aminotransferase (AST) (r = 0.56, P ≤ 0.001), alanine aminotransferase (ALT) (r = 0.49, P = 0.003), and liver fat (r = 0.49, P = 0.02) (Fig. 4C-G). Thus, the THRB gene expression seems to be repressed by both miR-34a-5p and DNA methylation in the liver of individuals with MASH.
DNA methylation of THRB in humans. (A) Methylation differences in THRB for patients with and without MASH were analyzed with the Mann–Whitney test. Spearman correlation of methylation with (B) THRB, (C) HbA1c, (D) aspartate aminotransferase (AST), (E) alanine aminotransferase (ALT), (F) liver fat, and (G) MAS. *<0.05.
Citation: European Thyroid Journal 13, 5; 10.1530/ETJ-24-0080
Discussion
In this study, we tested the hypothesis that local TH action in MASH could be regulated by epigenetic mechanisms. We indeed found an increased expression of miR-34a-5p in patients with MASH in agreement with previous studies (17, 22), which, together with altered THRB DNA methylation, can reduce local THRB expression and could constitute a novel mechanism explaining the declining responsiveness of the liver to TH on the molecular level during MASH progression.
Previous studies have already identified miR-34a-5p as a potential contributor to MASLD pathogenesis in mouse and human liver. The main focus in these studies, however, was on the miR-34a-5p target genes peroxisome proliferator-activated receptor alpha (23), sirtuin-1 (23, 24), and hepatocyte nuclear factor 4 alpha (22), as they are well-known contributors to MASLD development and progression by regulating lipid metabolism, mitochondrial function, and inflammation (22, 25). Our data now add the concerted local control of TH metabolism by transporters, deiodinases, and receptors as additional targets for this miRNA. This control is of great pathological relevance as TH regulates several aspects of liver health including mitochondrial function (26), and hypothyroidism is a well-documented risk factor for MASLD development and progression (4). Altering local TH action would in turn affect a plethora of possible downstream target genes involved in fatty acid metabolism, including fatty acid synthase, malic enzyme (ME), acetyl-CoA carboxylase alpha, and TH-responsive Spot14 homolog (Thrsp), which are part of the de novo lipogenesis (6), and hepatic lipase activity as part of beta oxidation (6). Therefore, an epigenetic reduction of local TH action would lead to impaired lipid metabolism, causing further lipid accumulation in the liver.
As we identified THRB and DIO1 as direct miR-34a-5p targets by luciferase reporter assays and TH transporters SLC10A1 and SLC16A2 as further potential targets, this miRNA has the potential to regulate local hepatic TH signaling in a concerted manner on multiple levels: TH transport, TH activation, and TH receptor availability. Together with an increased methylation of THRB in patients with MASH, which can arrest the receptor expression in a pathologically reduced state, our finding provides a fascinating interplay of epigenetic regulation of TH action in MASH, which cannot only explain the pathological progression of the disease but could potentially be targeted for therapeutic benefits. While altered expression of TH transporters might be of limited relevance for local TH action, given that they are passive transporters, changes in the expression of DIO1 are relevant for disease progression. The expression of DIO1 tends to be reduced by age in humans (7), thus potentially making them more susceptible to metabolic disease, and mouse studies have shown that the expression of DIO1 is upregulated as a possible compensation mechanism in early MASLD and reduced in later stages (27), which could be facilitated by the increased expression of miR-34a-5p in MASH. As DIO1 activates T4 to T3, a reduction of DIO1 expression could lead to a reduction of local T3 and an increase in fT4, as seen in patients with a DIO1 activity reducing polymorphism (28), ultimately leading to a local hepatic hypothyroid-like state, which would further aggravate the disease. On the other hand, studies reducing hepatic Dio1 in mice have not shown any change in the hepatic T3 content, suggesting that changes in DIO1 alone may not be sufficient to alter hepatic TH action (27, 29). Nevertheless, a recent study using RNA sequencing in a large cohort of individuals with MASLD identified reduced THRB expression and THRB regulon activity during disease progression, supporting the concept of a developing TH resistance (30). Importantly, given that DIO1 itself is a positive TH target gene, regulated by hepatic THRB (31), this might start a vicious cycle toward escalating hepatic hypothyroidism. The negative consequences are illustrated by a liver-specific knockdown of Dio1 in mice fed with a Western diet with fructose, which showed an increase in liver triglycerides and cholesterol as well as increased fatty acid synthesis and a reduction of beta oxidation (27).
A reduced availability of T3 can be rescued by levothyroxine, as shown in previous studies (32) or by the use of resmetirom, a phase 3 clinical trial liver-specific THRB agonist (33). However, if THRB expression is additionally downregulated by the identified epigenetic mechanisms, an increased ligand availability might not be sufficient to overcome local hypothyroidism, as a sufficient amount of receptors would not be available to start the action. The identification of new MASLD therapies is still of clinical relevance as some patients do not respond to resmetirom - an effect that could very well be explained by very low THRB expression in the hepatocytes, possibly mechanistically caused by the increased expression of miR-34a-5p as well as the increased DNA methylation of the THRB gene. Here, an additional treatment with an miR-34a-5p antagonist might be required prior to resmetirom therapy to increase receptor availability. Inhibition of miR-34a-5p alone might also be beneficial for MASLD treatment, as miR-34a-5p knockout mice show a better phenotype under high-fat diet challenge (34). However, a longitudinal study and further investigation of additional samples, including published sequencing and DNA methylation data, especially in the resmetirom non-responder groups, would be required to validate our findings. Moreover, the precise consequences of a gradual simultaneous decrease in hepatic DIO1 and THRB for the cellular actions of TH in the different liver cell types need to be evaluated, ideally on the single-cell level.
In conclusion, we provide evidence that epigenetic mechanisms could contribute to the regulation of TH signaling in MASLD, with potentially devastating effects on disease progression. Our data thus support miR-34a-5p as a potential therapeutic target in MASLD, possibly by improving local TH signaling.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/ETJ-24-0080.
Declaration of interest
J Mittag is an editorial board member of the European Thyroid Journal. He was not involved in the editorial or peer review process for this paper. The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the study reported.
Funding
Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 424957847 – TRR 296; KI1887-2/2; KI1887-3; GRK-1957 and by the German Centre for Diabetes Research (DZD).
Institutional Review Board Statement
The studies on human biopsies were conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Ärztekammer Hamburg, Germany (PV4889).
Patient consent
Informed consent was obtained from all subjects involved in the study.
Author contribution statement
A-MN, JM, and HK designed the study and wrote the manuscript. A-MN, JHB, and NT performed experiments. CK provided intellectual input. JM and HK acquired funding, and HK supervised the project.
References
- 1↑
Chalasani N, Younossi Z, Lavine JE, Charlton M, Cusi K, Rinella M, Harrison SA, Brunt EM, & Sanyal AJ. The diagnosis and management of nonalcoholic fatty liver disease: practice guidance from the American Association for the Study of Liver Diseases. Hepatology 2017 67 328–357. (https://doi.org/10.1002/hep.29367/suppinfo)
- 2↑
Younossi ZM, Golabi P, Paik JM, Henry A, Dongen Van C, & Henry L. The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review. Hepatology 2023 77 1335–1347. (https://doi.org/10.1097/HEP.0000000000000004)
- 3↑
Younossi ZM, Stepanova M, Al R, Eberly KE, Shah D, Nguyen V, Ong J, Henry L, & Alqahtani SA. The changing epidemiology of adult liver transplantation in the United States in 2013–2022: the dominance of metabolic dysfunction–associated steatotic liver disease and alcohol-associated liver disease. Hepatology Communications 2024 8 (https://doi.org/10.1097/HC9.0000000000000352)
- 4↑
Marjot T, Moolla A, Cobbold JF, Hodson L, & Tomlinson JW. Nonalcoholic fatty liver disease in adults: current concepts in etiology, outcomes, and management. Endocrine Reviews 2020 41 66–117. (https://doi.org/10.1210/endrev/bnz009)
- 5↑
Seifert J, Chen Y, Schöning W, Mai K, Tacke F, Spranger J, Köhrle J, & Wirth EK. Hepatic energy metabolism under the local control of the thyroid hormone system. International Journal of Molecular Sciences 2023 24 4861. (https://doi.org/10.3390/ijms24054861)
- 6↑
Sinha RA, Singh BK, & Yen PM. Direct effects of thyroid hormones on hepatic lipid metabolism. Nature Reviews Endocrinology 2018 14 259–269. (https://doi.org/10.1038/nrendo.2018.10)
- 7↑
Krause C, Grohs M, El Gammal A, Wolter S, Lehnert H, Mann O, Mittag J, & Kirchner H. Reduced expression of thyroid hormone receptor beta in human nonalcoholic steatohepatitis. Endocrine Connections 2018 7 1448–1456. (https://doi.org/10.1530/ec-18-0499)
- 8↑
He W, Huang C, Wang L, Su W, Wang S, Huang P, Zhang X, Huang Y, Zhao Y, Lin M, et al.The correlation between triiodothyronine and the severity of liver fibrosis. BMC Endocrine Disorders 2022 22 313. (https://doi.org/10.1186/s12902-022-01228-8)
- 9↑
Harrison SA, Taub R, Neff GW, Lucas KJ, Labriola D, Moussa SE, Alkhouri N, & Bashir MR. Resmetirom for nonalcoholic fatty liver disease: a randomized, double-blind, placebo-controlled phase 3 trial. Nature Medicine 2023 29 2919–2928 (https://doi.org/10.1038/s41591-023-02603-1)
- 10↑
Zhu H, Wang G, & Qian J. Transcription factors as readers and effectors of DNA methylation. Nature Reviews Genetics 2016 17 551–565. (https://doi.org/10.1038/nrg.2016.83)
- 11↑
Aranda A. MicroRNAs and thyroid hormone action. Molecular and Cellular Endocrinology 2021 525 111175. (https://doi.org/10.1016/j.mce.2021.111175)
- 12↑
Wu YL, Lin ZJ, Li CC, Lin X, Shan SK, Guo B, Zheng MH, Li F, Yuan LQ, & Li ZH. Epigenetic regulation in metabolic diseases: mechanisms and advances in clinical study. Signal Transduction and Targeted Therapy 2023 8 98 (https://doi.org/10.1038/s41392-023-01333-7)
- 13↑
Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Research 2001 29 e45. (https://doi.org/10.1093/nar/29.9.e45)
- 14↑
Meier J, Hovestadt V, Zapatka M, Pscherer A, Lichter P, & Seiffert M. Genome-wide identification of translationally inhibited and degraded miR-155 targets using RNA-interacting protein-IP. RNA Biology 2013 10 1017–1029. (https://doi.org/10.4161/rna.24553)
- 15↑
Krause C, Sievert H, Geißler C, Grohs M, Gammal El AT, Wolter S, Ohlei O, Kilpert F, Krämer UM, Kasten M, et al.Critical evaluation of the DNA-methylation markers ABCG1 and SREBF1 for type 2 diabetes stratification. Epigenomics 2019 11 885–897. (https://doi.org/10.2217/epi-2018-0159)
- 16↑
Krause C, Britsemmer JH, Bernecker M, Molenaar A, Taege N, Geißler C, Kaehler M, Iben K, Judycka A, Wagner J, et al.Liver microRNA transcriptome reveals miR-182 as link between type 2 diabetes and fatty liver disease in obesity. eLife 2023 12 (https://doi.org/10.7554/eLife.92075.1)
- 17↑
Liu XL, Pan Q, Zhang RN, Shen F, Yan SY, Sun C, Xu ZJ, Chen YW, & Fan JG. Disease-specific miR-34a as diagnostic marker of nonalcoholic steatohepatitis in a Chinese population. World Journal of Gastroenterology 2016 22 9844–9852. (https://doi.org/10.3748/wjg.v22.i44.9844)
- 18↑
Assis de LVM, Harder L, Lacerda JT, Parsons R, Kaehler M, Cascorbi I, Nagel I, Rawashdeh O, Mittag J, & Oster H. Rewiring of liver diurnal transcriptome rhythms by triiodothyronine. eLife 2022 11 1–35. (https://doi.org/10.7554/ELIFE.79405)
- 19↑
Wang S, Wu W, & Claret FX. Mutual regulation of microRNAs and DNA methylation in human cancers. Epigenetics 2017 12 187–197. (https://doi.org/10.1080/15592294.2016.1273308)
- 20↑
Anastasiadi D, Esteve-Codina A, & Piferrer F. Consistent inverse correlation between DNA methylation of the first intron and gene expression across tissues and species. Epigenetics and Chromatin 2018 11 37 (https://doi.org/10.1186/s13072-018-0205-1)
- 21↑
Castro-Mondragon JA, Riudavets-Puig R, Rauluseviciute I, Lemma RB, Turchi L, Blanc-Mathieu R, Lucas J, Boddie P, Khan A, Manosalva Pérez N, et al.JASPAR 2022: the 9th release of the open-access database of transcription factor binding profiles. Nucleic Acids Research 2022 50 D165–D173. (https://doi.org/10.1093/nar/gkab1113)
- 22↑
López-Sánchez GN, Dóminguez-Pérez M, Uribe M, Chávez-Tapia NC, & Nuño-Lámbarri N. Non-alcoholic fatty liver disease and microRNAs expression, how it affects the development and progression of the disease. Annals of Hepatology 2021 21 100212 (https://doi.org/10.1016/j.aohep.2020.04.012)
- 23↑
Ding J, Li M, Wan X, Jin X, Chen S, Yu C, & Li Y. Effect of MIR-34a in regulating steatosis by targeting PPARα expression in nonalcoholic fatty liver disease. Scientific Reports 2015 5 13729 (https://doi.org/10.1038/srep13729)
- 24↑
Wen Y, Huang H, Huang B, & Liao X. HSA-miR-34a-5p regulates the SIRT1/TP53 axis in prostate cancer. American Journal of Translational Research 2022 14 4493–4504.
- 25↑
Anggreini P, Kuncoro H, Sumiwi SA, & Levita J. Role of the AMPK/SIRT1 pathway in nonalcoholic fatty liver disease (Review). Molecular Medicine Reports 2023 27 (https://doi.org/10.3892/mmr.2022.12922)
- 26↑
Ramanathan R, Patwa SA, Ali AH, & Ibdah JA. Thyroid hormone and mitochondrial dysfunction: therapeutic implications for metabolic dysfunction-associated steatotic liver disease (MASLD). Cells 2023 12 (https://doi.org/10.3390/cells12242806)
- 27↑
Bruinstroop E, Zhou J, Tripathi M, Yau WW, Boelen A, Singh BK, & Yen PM. Early induction of hepatic deiodinase type 1 inhibits hepatosteatosis during NAFLD progression. Molecular Metabolism 2021 53 101266. (https://doi.org/10.1016/j.molmet.2021.101266)
- 28↑
de Jong FJ, Peeters RP, den Heijer T, van der Deure WM, Hofman A, Uitterlinden AG, Visser TJ, & Breteler MMB. The association of polymorphisms in the type 1 and 2 deiodinase genes with circulating thyroid hormone parameters and atrophy of the medial temporal lobe. Journal of Clinical Endocrinology and Metabolism 2007 92 636–640. (https://doi.org/10.1210/jc.2006-1331)
- 29↑
Streckfuß F, Hamann I, Schomburg L, Michaelis M, Sapin R, Klein MO, Köhrle J, & Schweizer U. Hepatic deiodinase activity is dispensable for the maintenance of normal circulating thyroid hormone levels in mice. Biochemical and Biophysical Research Communications 2005 337 739–745. (https://doi.org/10.1016/j.bbrc.2005.09.102)
- 30↑
Kendall TJ, Jimenez-Ramos M, Turner F, Ramachandran P, Minnier J, McColgan MD, Alam M, Ellis H, Dunbar DR, Kohnen G, et al.An integrated gene-to-outcome multimodal database for metabolic dysfunction-associated steatotic liver disease. Nature Medicine 2023 29 2939–2953. (https://doi.org/10.1038/s41591-023-02602-2)
- 31↑
Bruinstroop E, van der Spek AH, & Boelen A. Role of hepatic deiodinases in thyroid hormone homeostasis and liver metabolism, inflammation, and fibrosis. European Thyroid Journal 2023 12 (https://doi.org/10.1530/ETJ-22-0211)
- 32↑
Bruinstroop E, Dalan R, Cao Y, Bee YM, Chandran K, Cho LW, Soh SB, Teo EK, Toh SA, Leow MKS, et al.Low-dose levothyroxine reduces intrahepatic lipid content in patients with type 2 diabetes mellitus and NAFLD. Journal of Clinical Endocrinology and Metabolism 2018 103 2698–2706. (https://doi.org/10.1210/jc.2018-00475)
- 33↑
Harrison SA, Bedossa P, Guy CD, Schattenberg JM, Loomba R, Taub R, Labriola D, Moussa SE, Neff GW, Rinella ME, et al.A Phase 3, randomized, controlled trial of resmetirom in NASH with liver fibrosis. New England Journal of Medicine 2024 390 497–509. (https://doi.org/10.1056/NEJMoa2309000)
- 34↑
Xu Y, Zhu Y, Hu S, Pan X, Bawa FC, Wang HH, Wang DQH, Yin L, & Zhang Y. Hepatocyte miR-34a is a key regulator in the development and progression of non-alcoholic fatty liver disease. Molecular Metabolism 2021 51 101244. (https://doi.org/10.1016/j.molmet.2021.101244)
