Multi-parameter magnetic resonance imaging of zebularine in liver fibrosis treatment and calcineurin/NFAT3 mechanism

Abstract

BACKGROUND

Hepatic stellate cell (HSC) activation is key to liver fibrosis. Targeting DNA methylation shows promise. Zebularine, a methylation inhibitor, may suppress HSC activation via the calcineurin (CaN)/NFAT3 pathway. Magnetic resonance imaging (MRI) is a noninvasive tool for evaluating liver fibrosis evaluation tool, but multiparametric MRI for zebularine’s effects in liver fibrosis mouse models has not been studied.

AIM

To clarify the anti-fibrosis mechanism and MRI-evaluated efficacy of zebularine.

METHODS

In vitro, transforming growth factor (TGF)-β1-stimulated human HSCs (LX-2) were treated with zebularine. α-smooth muscle actin, fibrotic and anti-fibrotic gene levels, and regulator of calcineurin1 (RCAN1) regulation were measured. In vivo, carbon tetrachloride (CCl₄)-induced liver fibrosis in mice was treated with zebularine, and fibrosis was evaluated using various biochemical, histopathological, and MRI methods.

RESULTS

Zebularine upregulated RCAN1.4 protein (P < 0.01) and inhibited the CaN/NFAT3 pathway (P < 0.05). In HSCs, TGF-β1 reduced anti-fibrotic gene massage RNA (mRNA) and increased fibrotic mRNA (P < 0.05), whereas zebularine had the opposite effects (P < 0.01, P < 0.05). CCl₄-treated mice exhibited increases in various fibrosis-related indices, all of which were reversed by zebularine treatment (P < 0.05).

CONCLUSION

Zebularine may reduce LX-2 activation and extracellular matrix deposition via RCAN1.4 and CaN/NFAT3 pathways. Multiparametric MRI can assess its efficacy, suggesting zebularine’s potential as a liver fibrosis treatment.

Key Words: Zebularine; Regulator of calcineurin1.4; Calcineurin/NFAT3 pathway; Hepatic stellate cell; Liver fibrosis; DNA methylation

Core Tip: This study explored zebularine’s role in liver fibrosis. Given that hepatic stellate cell activation drives fibrosis and zebularine’s potential via calcineurin/NFAT3, it aimed to clarify its mechanism and efficacy using multiparametric magnetic resonance imaging (MRI). In vitro and in vivo tests showed zebularine regulated genes, inhibited the pathway, and reduced fibrosis, with MRI validating its efficacy.

INTRODUCTION

Liver fibrosis results from a combination of long-term liver damage and excessive connective tissue deposition[1]. If left untreated, it can progress to irreversible cirrhosis, characterized by portal hypertension and severe impairment of liver function, posing a significant threat to human health[2]. Therefore, liver fibrosis should be treated to prevent cirrhosis. Clinical approaches often involve multi-faceted strategies targeting various aspects of the disease process, such as pathogenesis, hepatoprotection, and anti-fibrotic, anti-inflammatory, and antioxidant properties[3]. However, there are currently no effective therapeutic drugs or biological agents for the specific treatment of liver fibrosis, making the reversal of liver fibrosis and prevention of its progression to cirrhosis a major challenge in clinical research.

Hepatic stellate cells (HSCs), stimulated by various liver fibrosis-promoting factors, transform into myofibroblasts, secreting an abundant extracellular matrix (ECM)[4], a core component of fibrosis. HSCs exhibit overexpression of α-smooth muscle actin (α-SMA) and type I collagen (CoL1a1)[5]. Activation of HSCs and liver fibrosis are regulated by the same signaling factors, such as transforming growth factor-β (TGF-β)[6], with TGF-β being the major fibroblast factor responsible for HSC activation. Thus, targeting HSCs is a fundamental approach for treating liver fibrosis.

DNA methylation is an essential component of epigenetic modification and plays a crucial role in aggravating liver fibrosis. Genes such as hMSH3, MGMT, SPP1, and GSTM3 that promote liver fibrosis are susceptible to upregulation through hypomethylation[7], while Smad7, PTCH1, and PPAR-γ, which inhibit liver fibrosis, are susceptible to episodic silencing by hypermethylation[8]. Preclinical studies have shown that DNA methylation inhibitors, such as 5-aza-deoxycytidine, inhibit HSC activation[9], resulting in improved liver function and regression of fibrosis[10]. By competitively inhibiting DNA methyltransferase (DNMT) activity, 5-aza-deoxycytidine reduces the binding of CpG island 5-cytosine to methyl groups, thereby restoring the expression levels of genes involved in the regulation of fibrosis[11]. However, the drawbacks of 5-azacytidine and 5-aza-deoxycytidine, such as instability, toxicity, and short half-lives, limit their clinical application. Zebularine, a new-generation chemically synthesized DNA methylation inhibitor, effectively maintains the demethylated state of genes with fewer side effects and improved specificity[12]. It has been shown to successfully reverse TGF-β-induced α-SMA expression in a time- and dose-dependent manner[13]. It also restores the responsiveness of nonfibrotic fibroblasts to protective factors through demethylation[14]. However, the molecular mechanism by which zebularine inhibits fibroblast proliferation via demethylation remains unclear.

Calcineurin (CaN) is the only phosphatase activated by calcium and calmodulins. During liver fibrosis, NFAT3 in the cytoplasm undergoes dephosphorylation by CaN and translocates into the nucleus. It promotes the apoptosis of activated HSCs, leading to excessive deposition of ECM, which in turn facilitates the progression of fibrosis[10,15]. The CaN/NFAT3 signaling pathway is an important pathway for the deposition of ECM. The regulator of calcineurin1 (RCAN1), an endogenous inhibitor of CaN, is regulated by DNA methylation. RCAN1.4, the major isoform, inhibits HSC activation by antagonizing the CaN/NFAT3 signaling pathway[10]. Studies have indicated that upregulation of the RCAN1.4 gene results in therapeutic remission of various diseases, such as liver[10] and renal fibrosis[13]. Whether zebularine can upregulate RCAN1.4 and block the CaN/NFAT3 signaling pathway, thereby antagonizing HSC activation, remains to be investigated.

The “gold standard” for evaluating liver fibrosis is liver biopsy. However, biopsies have limitations, such as sampling error, poor reproducibility, bleeding, and inconvenience for dynamic observation and follow-up. Magnetic resonance imaging (MRI) provides several advantages, including a lack of radiation exposure, superior soft-tissue resolution, and capability for multi-parameter and multidirectional imaging[16]. Our preliminary study found a strong correlation between multi-parameter MRI data and the pathological staging of fibrosis and inflammation[17,18].

We hypothesized that the methylation inhibitor zebularine may negatively regulate the CaN/NFAT3 signaling pathway by upregulating RCAN1.4 expression, thereby inhibiting HSC activation and ECM production, ultimately reversing liver fibrosis. Additionally, multiparametric MRI could allow for noninvasive liver fibrosis grading and assessment of antifibrotic therapy efficacy, enabling the evaluation of the efficacy of zebularine treatment in a liver fibrosis model.

In this study, we aimed to provide new insights into the role of zebularine in epigenetic regulation and its potential as a therapeutic agent for liver fibrosis and to validate MRI as a diagnostic tool for fibrosis evaluation.

MATERIALS AND METHODS

The Animal Care and Use Committee of the Second Xiangya Hospital in Central South University (approval number: 20230520). All procedures were performed in full compliance with relevant Chinese laws and institutional ethical guidelines.

Cell culture

The human HSC line LX-2 (Procell, CL-0560, Wuhan, Hubei Province, China) was cultured in a constant temperature incubator (37 °C, 5% carbon dioxide) using Dulbecco’s modified eagle medium (Corning, 10-013-CVRC, United States) containing 10% fetal bovine serum (Procell, 164210-50, Wuhan, Hubei Province, China).

Cell counting kit-8 assay

Cell viability was assessed using the cell counting kit (CCK)-8 assay. LX-2 cells (activated with 10 ng/mL TGF-β1) were seeded into 96-well plates (3000 cells per well, n = 3). Cells were treated with 0, 25, 50, 100, 200, 400, 800, and 1600 mM zebularine for 72 hours. The waste solution was removed, and a medium containing 10% CCK-8 reagent (Solarbio, Beijing, China) was added to each well and incubated for 4 hours. The absorbance (A) at 450 nm was measured using a microplate reader. Cell viability was calculated as follows: Cell viability (%) = (A_{experimental group} - A_{blank group})/(A_control - A_{blank group}) × 100%.

Real-time fluorescent quantitative polymerase chain reaction

Total RNA was extracted and reverse-transcribed into complementary DNA using RNA extraction and reverse transcription kits (AG21023, AG11728; AG Accurate Biotechnology, Co., Ltd., Hunan Province, China), respectively. Quantitative real-time polymerase chain reaction (PCR) was performed using PCR Master Mix to evaluate gene expression levels (AG Accurate Biotechnology Co. Changsha, Hunan Province, China). The primers used for PCR are listed in Table 1. The relative gene expression was calculated using the comparative computed tomography method.

Table 1 Primer sequences for real-time polymerase chain reaction.

Gene (human)	Forward	Reverse
CoL1a1	5’-AGACGAAGACATCCCACCA-3’	5’-GTCGCAGACGCAGATCC-3’
ACTA2 (α-SMA)	5’-GTTACGAGTTGCCTGATGG-3’	5’-AGGTGGTTTCATGGATGC-3’
SEPTIN9	5’-GTGAAGAACTCAGAACCCTCG-3	5’-TCAATGGACAGCTCAGTGC-3’
PPAR-γ	5’-AGATCATTTACACAATGCTGGC-3’	5’-TAAAGTCACCAAAAGGCTTTCG-3’
PTEN	5’-GACCAGAGACAAAAAGGGAGTA-3’	5’-ACAAACTGAGGATTGCAAGTTC-3’
RCAN1.4	5’-TTTAGCTCCCTGATTGCCTGT-3’	5’-AAAGGTGATGTCCTTGTCATACG-3’
PTCH1	5’-TTTTCTGCTGTTTTACAAGCCC-3’	5’-CATGGTAATCTGCGTTTCATGG-3’
β-actin (loading control)	5’-TCCCTGGAGAAGAGCTACGA-3’	5’-TGAAGGTAGTTTCGTGGATGC-3’

RCAN: Regulator of calcineurin; CoL1a1: Type I collagen; α-SMA: α-smooth muscle actin.

Western blot

Proteins were extracted from HSCs using radio immunoprecipitation assay reagent (89900, Thermo Scientific, United States), separated by polyacrylamide sodium dodecyl sulfate gel electrophoresis, and transferred to polyvinylidene difluoride (PVDF) membranes (ISEQ0010, Millipore, MA, United States). The membranes were incubated overnight at 4 °C with the primary antibodies listed in Table 2. The PVDF membranes were subsequently incubated with horseradish peroxidase-coupled secondary antibodies, and protein bands were detected using enhanced chemiluminescence reagent (K-12045-d10, Advansta, United States). Western blotting was performed three times, and the target protein bands were scanned for optical densitometry and normalized by glyceraldehyde-3-phosphate dehydrogenase, β-tubulin, and Lamin B1 signal intensities using Image-J processing software.

Table 2 Antibody characteristics.

Protein	Application	Origin	Dilution
α-SMA	WB and IHC-P	ab5694, Abcam, United Kingdom	1:200
RCAN1.4	WB	ab140131, Abcam, United Kingdom	1:10000
CaN	WB	55147-1-AP, Proteintech, China	1:1000
NFAT3	WB and IHC and IF	bs-6461R, Bioss, China	1:1000
β-tubulin	WB and IHC and IF	10094-1-AP, Proteintech, China	1:5000
GAPDH	WB and IHC and IF	ab9485, Abcam, United Kingdom	1:2500
Lamin B1	WB and IHC and IF	12987-1-AP, Proteintech, China	1:5000
HRP-conjugated affinipure goat anti-rabbit IgG (High + Low)	ELISA, WB	SA00001-2, Proteintech, China	1:5000

WB: Western blot; IF: Immunofluorescence; IHC: Immunohistochemistry; ELISA: Enzyme-linked immunosorbent assay; IgG: Immunoglobulin; HRP: Horseradish peroxidase; RCAN: Regulator of calcineurin; α-SMA: α-smooth muscle actin; CaN: Calcineurin; GAPDH: Glyceraldehyde-3-phosphate dehydrogenase.

Establishment of a mouse liver fibrosis model and zebularine treatment

Female Balb/C mice (6-8 weeks old, 18-22 g) were purchased from the Laboratory Animal Center of the Second Xiangya Hospital, Central South University. Mice were housed in pathogen-free animal facilities and temperature-controlled rooms (20 °C-26 °C) with a 12-hour regular light-dark cycle and ad libitum access to food and water. After 1 week of adaptive breeding, the mice were randomly divided into the control, a carbon tetrachloride (CCl₄)_, and a CCl₄ + zebularine group (36 mice per group) and into 2-, 4-, and 6-week subgroups.

The treatments were as follows: The CCl₄ + zebularine group was administered intraperitoneal injections of zebularine (225 mg/kg/day)[13]. The control group received an equivalent volume of saline containing 3% dimethyl sulfoxide. In the CCl₄-induced hepatic fibrosis model group and CCl₄ + zebularine group, mice were administered a mixture (1:3 v/v) of CCl₄ and olive oil[19] via intraperitoneal injection at a dose of 0.7 mL/kg twice a week. The control group received an equivalent dose of olive oil. At the end of the modeling period, MRI scans were performed on all mice.

MRI in vivo and image processing

A 3.0T clinical MR scanner (uMR790; United Imaging Medical, Shanghai, China) equipped with an 8-channel mouse coil was used for MRI. Initially, unenhanced whole-liver sequences were obtained, including T1-weighted fast spin-echo, T2-weighted single-shot fast spin-echo, diffusion-weighted imaging (DWI) with two b values (b = 0, 800 s/mm), variable flip-angle T1 mapping, and multi-echo spin-echo T2 mapping. Subsequently, a contrast agent (80 μL saline plus 20 μL gadoxetate disodium) was injected via the tail vein, and enhanced T1-weighted imaging and T1 mapping were performed. The detailed parameters are listed in Table 3.

Table 3 Sequence-specific parameters.

Parameter/sequence	T1WI	T2WI	EPI-DWI	T1 MAP	T2 MAP
TR	585	3243	2000	15	3000
TE	11.32	87.38	80	3.87	17.4, 34.8, 52.2, 69.6, 87.0, 104.4
Flip angle (°)	150°	140°	90°	5°, 25°	180°
Voxel size (mm³) (X × Y × Z)	0.16 × 0.16 × 1	0.30 × 0.30 × 1	0.59 × 0.59 × 2	0.13 × 0.13 × 1	0.50 × 0.50 × 2
FOV (mm²)	40 × 40	40 × 40	50 × 80	75 × 83	90 × 100
Slices	20	20	15	20	15
Acquisition time	4:47	5:24	3:22	3:8	5:39

T1WI: T1-weighted imaging; T2WI: T2-weighted imaging; EPI: Echo planar imaging; DWI: Diffusion-weighted imaging; TR: Repetition time; TE: Echo time; FOV: Field of view; T1 MAP: T1 mapping; T2 MAP: T2 mapping.

Postprocessing and image analysis

The captured images were sent to a post-processing workstation (uWS-MR, United Imaging Medical, Shanghai, China). Pixel-level T1 maps were developed using the Levenberg-Marquardt curve-fitting algorithm, while T2 maps were generated by fitting the echo intensity of each pixel using the log-linear least squares method. Two radiologists with 11 years of experience in diagnostic abdominal imaging (Xiao EH and Luo YH) selected six regions of interest in the whole liver cross-section, avoiding large vessels and bile ducts, with an averaged area of 0.75 mm² (Figure 1). The difference in the TI values (ΔT1) was calculated as T1_pre - T1_post. DWI data were processed using custom software developed in MATLAB R2018b (MathWorks, Natick, MA, United States). The apparent diffusion coefficient (ADC) was calculated voxel-wise by applying the following formula: S_b/S₀ = e^-bADC, where S_b and S₀ represent the signal intensities with and without diffusion weighting, respectively, and b is the diffusion-sensitizing gradient factor. The ADC values were derived using b values of 0 and 800 s/mm².

Figure 1 Example of region of interest placement in Balb/C mice. A: Three regions of interest (ROIs) were allocated to the same liver-parenchyma region in the plain stage, and T1 relaxation time was measured; B: Three ROIs were allocated to the same liver-parenchyma region in the hepatobiliary stage, and T1 relaxation time was measured.

Liver function biochemistry

To evaluate liver function, serum samples were extracted from the retroorbital venous plexus of each mouse, and the levels of the enzymes aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were determined.

Pathological and histological examinations

Following MRI at the designated time points, the mice were humanely euthanized and liver tissue was promptly harvested for histological analysis. The liver weights (G) were recorded, the hepatosomatic index (HSI) was calculated as the liver weight-body ratio, and the tissues were labeled, fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin to obtain pathological sections. The embedded tissues were stained with hematoxylin and eosin (HE) and Masson’s trichrome staining kit (G1006; ServiceBio, Wuhan, Hubei Province, China).

Observations were conducted using an Olympus microscope (Olympus Corporation, Tokyo, Japan), and the collagen fiber area fraction was quantified using ImageJ software (Rawak Software, Stuttgart, Germany).

Fibrosis and inflammation were evaluated and staged according to the METAVIR scoring system for fibrosis[20,21] and activity[22].

Statistical analysis

Statistical comparisons between groups were performed using one-way analysis of variance (ANOVA) for normally distributed data with homogeneous variances. For non-normally distributed data, the Kruskal-Wallis test with Bonferroni correction was applied. When data were normally distributed but exhibited heterogeneity of variance, one-way ANOVA with Tamhane’s T2 post hoc test was employed. Spearman’s rank correlation analysis was conducted to assess the relationships between MRI parameters and the pathological and inflammatory stages of liver fibrosis. All statistical analyses were performed using SPSS software (version 27.0; SPSS Inc., Chicago, IL, United States). Statistical significance was defined as P < 0.05. The statistical review of the study was carried out under the supervision of a biomedical statistician.

RESULTS

Zebularine toxicity in activated human LX-2 cells

The CCK8 assay showed that the survival rate of activated LX-2 cells decreased as zebularine concentration increased (Figure 2A). The IC₂₅ and IC₅₀ of zebularine in activated LX-2 cells were 410 mmol/L and 1050 mmol/L, respectively (Figure 2B). A concentration of 400 mmol/L zebularine was used in subsequent experiments.

Figure 2 Toxicity of zebularine in LX-2 cells. A: Histogram showing zebularine concentration-cell viability; B: Nonlinear regression curves [log (zebularine) - relative cell viability]. ^cP < 0.001. ^dP < 0.0001.

Effect of zebularine on RCAN1.4 and CaN/NFAT3 signaling in activated LX-2 cells

Compared to the TGF-β1 group, the protein expression of RCAN1.4 was significantly increased in the TGF-β1 + zebularine group (P < 0.01), whereas the expression of CaN was decreased (P < 0.01, Figure 3A). Additionally, zebularine decreased NFAT3 protein expression in the nucleus of activated LX-2 cells (P < 0.05, Figure 3B) and increased its expression in the cytoplasm (P < 0.01, Figure 3C).

Figure 3 Expression of regulator of calcineurin1.4, calcineurin, and NFAT3 pairs in transforming growth factor-β1-activated LX-2 by zebularine action. A: Expression of regulator of calcineurin1.4 and calcineurin proteins; B: Intranuclear protein expression of NFAT3; C: Cytoplasmic protein expression of NFAT3. ^aP < 0.05. ^bP < 0.01. RCAN: Regulator of calcineurin; CaN: Calcineurin; TGF: Transforming growth factor.

Zebularine inhibits TGF-β1-induced LX-2 activation

In the resting state, LX-2 cells exhibit a polygonal morphology (Figure 4A). Following TGF-β1 treatment, the cells adopted a long spindle shape, and the number of interstitial laser breaks increased (Figure 4B). Compared to cells in the TGF-β1 group, cells in the TGF-β1 + zebularine group were thinner and less elongated, with many cells maintaining a star or polygonal shape (Figure 4C and D).

Figure 4 Microscopic changes of LX-2 cells activated by transforming growth factor-β1 by zebularine action. A: Control group; B: Transforming growth factor-β1 group; C: Zebularine group; D: Transforming growth factor-β1 + zebularine group.

Compared to the TGF-β1 group, the massage RNA (mRNA) expression of α-SMA and CoL1a1 was significantly reduced in the TGF-β1 + zebularine group (P < 0.0001, Figure 5A and B). In contrast, when the concentration was 400 mM, the mRNA levels of anti-fibrotic genes in the TGF-β1 + zebularine group were significantly elevated (P < 0.01, Figure 5C-F), and the protein expression of α-SMA was reduced (P < 0.05, Figure 5G).

Figure 5 Effect of zebularine on anti-fibrotic and fibrotic gene expression in transforming growth factor-β1-activated LX-2. The message RNA levels of the following genes in LX-2 cells were measured before and after transforming growth factor-β1 activation and after zebularine treatment. A: α-smooth muscle actin (α-SMA); B: Type I collagen; C: SEPTIN9; D: PTEN; E: PPAR-γ; F: PTCH1; G: Protein expression of α-SMA in LX-2 cells following zebularine intervention at varying concentrations. ^aP < 0.05. ^bP < 0.01. ^cP < 0.001. ^dP < 0.0001. ¹P vs control. ²P vs transforming growth factor-β1. ³P vs transforming growth factor-β1 + zebularine 70 mmol/L. ⁴P vs zebularine 70 mM. α-SMA: α-smooth muscle actin; TGF: Transforming growth factor; CoL1a1: Type I collagen; GAPDH: Glyceraldehyde-3-phosphate dehydrogenase.

MRI parameters and their diagnostic efficacy for assessing fibrosis

Receiver operating characteristics analyses were also performed. The single-parameter results indicated that the area under the curve (AUC) was highest for the ΔT1 value (0.961). Notably, the AUC for the combined parameters (0.989) exceeded that of any single parameter (Table 4 and Figure 6). Correlation analyses of each parameter with METAVIR fibrosis and inflammation scores were performed (Figure 7 and Table 5). Changes in T1, ∆T1, and T2 values were consistent with changes in liver fibrosis and inflammation, whereas ADC values were negatively correlated with these changes. The correlation of each parameter with fibrosis staging was higher than that with inflammation staging (P < 0.001).

Figure 6 Receiver operating characteristic curves for T1pre, DT1, apparent diffusion coefficient, T2, and the combined diagnostic approach for liver fibrosis. AUC: Area under the receiver operating characteristic curve; ADC: Apparent diffusion coefficient; Combination: Combined diagnosis.

Figure 7 Spearman correlation analysis of magnetic resonance parameters with fibrosis and inflammation. A-D: Fibrosis; E-H: Inflammation. ADC: Apparent diffusion coefficient.

Table 4 Diagnostic performance of magnetic resonance parameters in the assessment of liver fibrosis.

Parameters	T1pre (ms)	△T1 (ms)	ADC (× 10-3, mm2/second)	T2	Combined diagnosis
AUC	0.926	0.961	0.809	0.926	0.989
95%CI	0.902-0.951	0.945-0.976	0.768-0.851	0.899-0.953	0.982-0.996
P value	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
Sensitivity	0.889	0.898	0.810	0.829	0.940
Specificity	0.852	0.907	0.708	0.940	0.977

ADC: Apparent diffusion coefficient; AUC: Area under the curve; CI: Confidence interval.

Table 5 Correlation matrix between magnetic resonance parameters and histological features of liver fibrosis.

Parameters	Fibrosis		Inflammation
	rs	P value	rs	P value
T2	0.861	< 0.001	0.795	< 0.001
△T1	0.853	< 0.001	0.763	< 0.001
T1pre T1	0.779	< 0.001	0.673	< 0.001
ADC	-0.692	< 0.001	-0.577	< 0.001

ADC: Apparent diffusion coefficient.

MRI multiparametric analysis confirmed the antifibrosis effect of zebularine

The liver lobe structure was more clearly observed in the T2-weighted imaging images (Figure 8A).

Figure 8 Changes in magnetic resonance parameters in the control, carbon tetrachloride, and carbon tetrachloride + zebularine groups. Red dotted lines indicate the entire liver border in: A: T2-weighted images; B: Pre-contrast T1 maps; C: Post-contrast T1 maps; D: T2 maps; E: Apparent diffusion coefficient maps. ^aP < 0.05. ^bP < 0.01. ^cP < 0.001. ^dP < 0.0001. ¹P vs control. ²P vs carbon tetrachloride. T2WI: T2-weighted imaging; CCl₄: Carbon tetrachloride; ADC: Apparent diffusion coefficient.

In the CCl₄ group, the T1_pre, ∆T1, and T2 values were significantly increased, whereas ADC values were notably reduced compared to the control group (all P < 0.001). In contrast, compared with the CCl₄ group, the CCl₄ + zebularine treatment group exhibited a decrease in T1_pre, DT1 and T2 values, along with an increase in ADC values, with the change in ADC reaching statistical significance at 2 weeks. At 4 weeks, the T1_pre, DT1, and ADC values were significantly different (P < 0.05), and at 6 weeks, all parameters showed significant differences (P < 0.01). Multi-parameter MRI effectively detected the therapeutic effects of zebularine in the liver fibrosis mouse model at week 6 (Table 6, Figure 8B-E).

Table 6 Comparison of magnetic resonance parameters between the control, carbon tetrachloride, and carbon tetrachloride + zebularine groups.

Parameters	2 weeks				4 weeks				6 weeks
	Control	CCl4	CCl4 + zebularine	P value	Control	CCl4	CCl4 + zebularine	P value	Control	CCl4	CCl4 + zebularine	P value
T1pre (millisecond)	632.554 ± 91.474	783.493 ± 47.454	790.09 ± 57.283	1.000	650.917 ± 14.543	845.448 ± 98.6	694.370 ± 52.447	< 0.001	632.712 ± 27.834	934.073 ± 38.4291	760.6564 ± 61.1215	< 0.001
△T1	445.771 ± 91.877	657.431 ± 51.533	633.599 ± 72.49	1.000	445.996 ± 69.364	724.751 ± 102.433	584.67 ± 55.173	< 0.001	439.015 ± 53.865	816.836 ± 41.547	666.994 ± 64.864	< 0.001
T2 (millisecond)	24.54 ± 1.253	30.124 ± 2.269	29.9 ± 1.592	1.000	25.278 ± 1.473	33.544 ± 2.575	30.698 ± 1.582	0.12	25.777 ± 1.304	37.305 ± 1.547	34.641 ± 1.489	< 0.001
ADC (× 10-3, mm2/second)	1.397 ± 0.190	1.076 ± 0.129	1.271 ± 0.135	< 0.001	1.178 ± 0.099	1.028 ± 0.106	1.454 ± 0.147	0.012	1.388 ± 0.229	0.946 ± 0.043	1.164 ± 0.155	0.005

ADC: Apparent diffusion coefficient; CCl₄: Carbon tetrachloride. The P values for the carbon tetrachloride group, compared to the control group, were < 0.001 at all three time periods, therefore, the P values presented in Table 6 correspond specifically to comparisons between the liver fibrosis and treatment groups.

Animal model

Liver volumes in the CCl₄ group were visibly larger than those observed in the control and treatment groups (Figure 9). The average weight loss in the CCl₄ group was 11% (P < 0.01) but only 3% in the CCl₄ + zebularine group (P > 0.05) relative to that in the control group. Conversely, the CCl₄ + zebularine group exhibited a weight gain of 8% compared with the group treated with CCl₄ at 6 weeks (P < 0.05, Figure 10A). The mean liver mass was increased by 31% in the CCl₄ group (P < 0.0001) and by 16% in the CCl₄ + zebularine group (P < 0.01) compared to the control group. However, the 15% reduction in mean liver weight between the CCl₄ + zebularine and the CCl₄ group at 6 weeks did not reach statistical significance (P > 0.05, Figure 10B). Analysis of the HSI revealed that compared to the control group, the CCl₄ group exhibited significantly higher HSI (P < 0.05). In contrast, The HSI of the CCl₄ + zebularine group was lower than that of the CCl₄ group, with statistically significant differences becoming evident at 4 weeks (P < 0.05). However, the HSI of the CCl₄ + zebularine group remained elevated compared to that of the control group (Figure 10C).

Figure 9 Representative images of mouse livers after carbon tetrachloride and zebularine treatment. A: Carbon tetrachloride group; B: Carbon tetrachloride + zebularine group; C: Control group.

Figure 10  Changes in body weight, liver weight, and hepatosomatic index in mice treated with carbon tetrachloride and zebularine. A: Body weight at 2, 4, and 6 weeks; B: Mouse liver mass-timed histograms for each group; C: Hepatosomatic index-timed histograms for each group (n = 12). ^aP < 0.05. ^bP < 0.01. ^dP < 0.0001. ¹P vs control. ²P vs carbon tetrachloride. CCl₄: Carbon tetrachloride.

Serological examination

Serum ALT and AST levels were significantly elevated in the CCl₄ group compared with the control group at each of the three time points (2 weeks, 4 weeks, and 6 weeks) (all P < 0.001). However, in the CCl₄+zebularine group, serum ALT and AST levels were lower than those in the CCl₄ group (all P < 0.05), although they remained higher than those in the control group. Intraperitoneal injection of CCl₄ caused marked increases in ALT and AST levels, whereas zebularine treatment reduced these levels, although it did not restore them to normal levels (Figure 11).

Figure 11  Effects of carbon tetrachloride and zebularine treatments on serum aspartate aminotransferase and aspartate aminotransferase levels in each experimental group. A: Alanine aminotransferase (U/L) time histogram; B: Aspartate aminotransferase (U/L) time histogram (n = 12). ^aP < 0.05. ^bP < 0.01. ^cP < 0.001. ^dP < 0.0001. ¹P vs control. ²P vs carbon tetrachloride. CCl₄: Carbon tetrachloride; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase.

Histopathologic analysis

Masson’s trichrome staining revealed that the deposited collagen fibers were interconnected and formed pseudonodules in the liver parenchyma, disrupting liver structure and impairing function (Figure 12A). The proportion of collagen fibers was markedly elevated in the CCl₄ group (P < 0.0001). In contrast, the fibrotic percentage in the CCl₄+zebularine group was consistently lower than that in the CCl₄ group at 2 weeks and 6 weeks (P < 0.05, Figure 12B), with a statistically significant difference in METAVIR fibrosis grading between the two groups becoming evident at 4 weeks (P < 0.05; Figure 12C). At 4 weeks, the percentage of fibrosis was consistently lower in the CCl₄ group than in the CCl₄ group (both P > 0.05, Figure 12B).

Figure 12  Masson’s staining showing the effect of carbon tetrachloride and zebularine treatment on collagen fibers in mouse liver tissue. A: Representative images of liver sections stained with Masson’s trichrome; B: Collagen fiber percentage at 2, 4, and 6 weeks; C: METAVIR fibrosis grading at 2, 4, and 6 weeks (n = 12 mice/group). Original magnification: × 200. ^aP < 0.05. ^bP < 0.01. ^dP < 0.0001. ¹P vs control. ²P vs carbon tetrachloride. CCl₄: Carbon tetrachloride.

HE staining showed that hepatocyte cords became increasingly disorganized and inflammatory cell infiltration increased with time in the experimental liver injury models, reflecting progression toward more severe liver damage or fibrosis (Figure 13). Compared with the findings in the CCl₄ group, the degree of liver tissue necrosis in the CCl₄ + zebularine group was relatively mild (P > 0.05). Moreover, the inflammatory scores did not differ between the two groups.

Figure 13  Hematoxylin and eosin staining showing the effect of carbon tetrachloride and zebularine on inflammation in mouse liver tissue. A: Hematoxylin and eosin staining of liver sections from each experimental group; B: Liver inflammation grading (METAVIR) at 2, 4, and 6 weeks (n = 12 mice/group). Original magnification: × 200. ^cP < 0.001. ^dP < 0.0001. P vs control. HE: Hematoxylin and eosin; CCl₄: Carbon tetrachloride.

DISCUSSION

Our study highlights the promising role of zebularine in the treatment of liver fibrosis. Zebularine appears to effectively upregulate RCAN1.4 and inhibit the CaN/NFAT3 signaling pathway, which reduces LX-2 cell activation and ECM deposition in vitro. In vivo, zebularine mitigated liver fibrosis and inflammation while showing consistent results across different MRI parameters (T1_pre, DT1, T2, and ADC). By integrating molecular, cellular, and imaging data, we provide a comprehensive approach to evaluating the efficacy of zebularine.

HSC activation contributes to ECM remodeling and fibrosis[23]. Our results-demonstrating reduced α-SMA and CoL1a1 expression following zebularine treatment support the hypothesis that targeting epigenetic mechanisms can modulate HSC activation and ECM composition. Zebularine mitigates TGF-β1-induced activation injury in LX-2 cells by targeting human HSCs and promoting ECM degradation. Specifically, it functions via the RCAN1.4-CaN/NFAT3 signaling pathway.

CaN is an important factor in ECM accumulation, and the CaN/NFAT3 signaling pathway has been implicated in fibrosis across various cells[24]. In this study, zebularine inhibited the CaN/NFAT3 signaling pathway in activated LX-2 cells, a process closely related to the activation and fibrotic progression of HSCs[25]. Previous studies have shown that CaN inhibitors successfully prevent ECM overaccumulation[26-28], which is consistent with our findings. Furthermore, DNMT is upregulated in activated HSCs, resulting in reduced RCAN1.4 expression[10]. This suggests that low RCAN1.4 expression in activated LX-2 cells is associated with DNA methylation. Our study demonstrated that zebularine increased RCAN1.4 expression in activated LX-2 cells, in addition to upregulating the expression of the anti-fibrotic genes SEPTIN9, PTEN, PPAR-γ, and PTCH1, while inhibiting the expression of the fibrotic-related genes α-SMA and CoL1a1. This suggests that in activated LX-2 cells, zebularine upregulates RCAN1.4 and reverses the progression of TGF-β1-induced RCAN1.4-CaN/NFAT3 signaling in fibrosis. Demethylating agents may serve as a new therapeutic approach to inhibit fibrosis by maintaining the hypomethylated state of RCAN1.4, increasing its expression, and inhibiting the CaN/NFAT3 pathway.

However, some studies have reported that RCAN1.4 overexpression does not reduce fibrosis but instead increases fibronectin and CoL1a1 production in mouse glomerular mesangial cells, promoting renal fibrosis[29]. This paradox arises because RCAN1.4 not only acts on the CaN/NFAT3 pathway but also interacts with other proteins, such as regulating nuclear factor κB[30], acting on extracellular regulated protein kinases[31] and Raf[32], to co-regulate the fibrotic process. We propose that a dynamic equilibrium exists in the body, with RCAN1.4 potentially induced by other compounds that activate the CaN/NFAT3 signaling pathway, leading to a negative feedback loop that regulates the pathway. This suggests the presence of a self-regulatory mechanism in fibrosis that helps achieve effective tissue repair and prevents excessive fibrosis that could impair tissue function.

Although zebularine is approximately one-tenth as potent as 5-azacytidine as a demethylating agent, both compounds share similar inhibitory effects on DNMT when tested with small oligodeoxynucleotide fragments[8]. The IC₅₀ values for zebularine were lower in other cells, likely due to differences in cell type, experimental conditions, and methodologies[33-35]. Zebularine is a noteworthy DNMT inhibitor with lower toxicity, better stability, and a stronger safety profile than 5-azacytidine[8]. In this study, we used a higher concentration of zebularine to increase DNA incorporation and improve metabolic activation efficiency, thereby inhibiting HSC activation.

A robust correlation between MRI parameters (T1pre, DT1, and T2) and the severity of liver fibrosis has been well-documented in the literature[36,37]. In the CCl₄-induced mouse model of liver fibrosis, differences in the parameters were gradually revealed on MRI after prolonged zebularine treatment. Consistent with previous studies, T1 relaxation time reflects protein and ECM accumulation due to liver fibrosis[38,39]. Our results corroborate these findings, showing both an increase in the extent of fibrosis and a corresponding prolongation of the T1 relaxation time. This is attributable to the sensitivity of T1 mapping to low-frequency movements and static processes, making it a reliable indicator of the presence of macromolecules in an organism[40]. T2 mapping, which is used to quantitatively analyze lesions in chronic liver disease, reflects subtle changes in water molecules in tissues[41]. In this study, as liver fibrosis progressed, T2 relaxation time increased, corresponding to worsening liver fibrosis and inflammation. Additionally, as liver fibrosis aggravated, the DWI increased while ADC values decreased, likely due to the proliferation of collagen fibers limiting molecular movement[42]. ADC values, which reflect the diffusion limitation of molecular movement between tissues, are useful for assessing the stage of liver fibrosis in patients[43]. However, ADC values are not a common indicator of cirrhosis[44]. Therefore, the ability of ADC values to diagnose mild liver fibrosis is not as accurate as ΔT1, T1pre, and T2[45], which aligns with our experimental findings.

Gadoxetate disodium is a paramagnetic contrast agent used for T1-weighted imaging that specifically accumulates in hepatocytes and plays a key role in detecting mild to severe liver fibrosis[46]. Liver fibrosis leads to a decrease in the number of normal cells, which in turn affects the balance between the uptake and excretion of gadoxetate disodium[47], resulting in reduced contrast enhancement of the liver parenchyma in the hepatocyte phase. Our results revealed significant differences in T1 values between the zebularine treatment group and the CCl₄ model group beginning at week 4. This suggests that prolonged zebularine administration gradually alleviated fibrosis. Recent MRI scans in a mouse liver fibrosis model showed an increase in ΔT1 value with worsening fibrosis[36], which aligns with our findings. However, in our study, image acquisition was performed immediately after the tail vein injection of the contrast agent, during which the vast majority of the contrast agent remained in the liver. Although opinions differ, a strong correlation between liver fibrosis and ΔT1 has been consistently observed[17]. The statistically significant differences observed at weeks 4 and 6 suggest that longitudinal studies and longer treatment periods may be required to detect the measurable effects of zebularine treatment on MRI parameters related to liver fibrosis.

Animal organ weights and organ-to-body ratios are commonly used in drug safety evaluation[48]. Early hepatic fibrosis leads to ECM and collagen fiber proliferation, which increases liver weight and HSI[36]. This suggests that CCl₄ increases liver mass but decreases body weight in mice, whereas zebularine treatment alleviates disease progression. These findings suggest that zebularine holds promise as a therapeutic agent for liver fibrosis and underscores its role in reducing collagen deposition.

In LX-2 cells stimulated with TGF-β1, zebularine exerted its antifibrotic effects through gene demethylation, particularly by upregulating antifibrotic genes and inhibiting the CaN/NFAT3 signaling pathway via RCAN1.4 gene demethylation. This resulted in the suppression of HSC activation and ECM production, ultimately alleviating the pathological process of liver fibrosis. These findings provide new directions for liver fibrosis treatment. However, the mechanism by which zebularine protects HSCs requires further exploration. This includes investigating the role of hypermethylation of other anti-fibrotic genes that undergo hypermethylation in other signaling pathways, as well as the effects of demethylating drugs on the expression of other key genes and proteins, with a view to comprehensively understanding the mechanism of action of demethylation on liver fibrosis.

This study has some limitations. First, the exact molecular mechanism of zebularine in the regulation of other signaling pathways involved in liver fibrosis remains unclear and requires further investigation. Second, although multiparametric MRI was effective in assessing fibrosis, the results were based on a preclinical mouse model, and translation to human studies is yet to be validated. Additionally, the long-term safety and efficacy of zebularine in the treatment of liver fibrosis require further exploration through larger, more extended clinical studies.

CONCLUSION

Zebularine demonstrated a multifaceted antifibrotic effect, primarily through its effects on the regulation of gene expression and signaling pathways. The regulation of RCAN1.4, inhibition of the CaN/NFAT3 pathway, and reduction in ECM production collectively contributed to its efficacy in mitigating liver fibrosis. These findings highlight the potential of zebularine as a therapeutic agent for liver fibrosis and other fibrotic conditions. Further investigations are required to confirm these results and explore the full range of their effects.