Hepatocyte CD146 and soluble CD146 promote hepatic stellate cell activation via the integrin αvβ1–p38 MAPK signaling axis to drive liver fibrosis

Abstract

During chronic liver injury, hepatocytes promote liver fibrosis by interacting with hepatic stellate cells (HSCs). However, the key intracellular factors that regulate intercellular communication in hepatocytes remain poorly understood. Although cluster of differentiation 146 (CD146) is predominantly recognized as an endothelial cell marker, our study found that CD146 is substantially upregulated in hepatocytes under fibrotic stress and may play an independent role in transcellular signaling. This study aimed to elucidate the functions and underlying mechanisms of hepatocyte CD146 in liver fibrosis. We evaluated CD146 expression in liver tissues and serum soluble CD146 (sCD146) levels in patients with liver fibrosis and mouse models. Hepatocyte-specific CD146-overexpressing mice (AAV8-CD146) were generated and fed either a normal chow diet (NCD) or a high-fat diet (HFD); hepatocyte-specific CD146 knockout mice (CD146^Alb−Cre) were subjected to carbon tetrachloride (CCl₄)-induced fibrosis. Comprehensive molecular, biochemical, and histological analyses were performed to systematically assess the role of CD146 in the initiation and progression of liver fibrosis. A co-culture system of hepatocytes and HSCs was also established to investigate the paracrine regulatory effects of CD146 on HSCs. The levels of sCD146 were positively correlated with the progression and severity of chronic liver disease. A diagnostic model incorporating sCD146 and other serological markers showed improved diagnostic performance for decompensated cirrhosis. Compared with healthy controls, hepatocyte CD146 expression was markedly upregulated in liver tissues from patients with cirrhosis and fibrotic mice. Under HFD conditions, hepatocyte-specific CD146 overexpression exacerbated hepatic inflammation and fibrosis in mice, whereas hepatocyte-specific CD146 deletion significantly attenuated CCl₄-induced liver fibrosis. Transforming growth factor-β1 (TGF-β1) induced CD146 upregulation in hepatocytes and increased the release of sCD146. Released sCD146 bound to integrin αvβ1 on the surface of HSCs, activated the p38 mitogen-activated protein kinase (p38 MAPK) pathway, and promoted HSC activation and type I collagen production. These findings identify hepatocyte-derived CD146/sCD146 as a TGF-β1-inducible mediator of hepatocyte–HSC communication that promotes liver fibrosis through the CD146/sCD146–integrin αvβ1–p38 MAPK signaling axis. They also support serum sCD146 as a potential non-invasive biomarker for advanced chronic liver disease.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00018-026-06324-9.

Introduction

Liver fibrosis is a prevalent pathological consequence of chronic liver diseases, including viral hepatitis, alcohol-associated liver disease, metabolic dysfunction-associated steatotic liver disease (MASLD), and cholestatic liver disorders. It poses a serious threat to public health [1, 2]. Liver disease accounts for approximately two million deaths annually worldwide, with mortality largely attributable to complications of cirrhosis and hepatocellular carcinoma (HCC) [3–5]. Although diagnostic approaches and therapeutic interventions for liver fibrosis have improved, specific antifibrotic therapies and sensitive early diagnostic biomarkers remain lacking, limiting timely detection and effective intervention [6]. Therefore, elucidating the molecular mechanisms underlying liver fibrosis, identifying specific diagnostic biomarkers, and developing targeted therapeutic strategies are of great clinical significance.

An imbalance between the synthesis and degradation of extracellular matrix (ECM) components constitutes a hallmark of liver fibrosis, with hepatic stellate cell (HSC) activation serving as the principal driving force. Upon activation, quiescent HSCs differentiate into myofibroblast-like cells that exhibit proliferative and migratory capacities and produce large amounts of collagen, thereby contributing to abnormal ECM deposition [7–11]. Interactions between HSCs and other hepatic cells, including hepatocytes and macrophages, constitute key regulatory mechanisms for HSC activation and phenotypic transition [12–15]. Damaged hepatocytes can indirectly promote HSC activation by recruiting macrophages and stimulating their secretion of profibrotic factors, such as transforming growth factor-β1 (TGF-β1), platelet-derived growth factor (PDGF), and tumor necrosis factor-α (TNF-α) [16–18], or directly drive the phenotypic conversion of HSCs by releasing profibrotic mediators and reactive oxygen species [13, 19, 20]. Although multicellular interaction networks are critical in liver fibrosis, hepatocyte-derived profibrotic molecules that directly mediate HSC activation remain incompletely characterized.

Cluster of differentiation 146 (CD146), also known as MCAM, is a cell adhesion molecule with elevated expression in various fibrosis-related diseases. Beyond endothelial cells, CD146 is markedly enriched in fibroblasts and smooth muscle cells [21–25]. A growing body of evidence indicates that CD146 exerts regulatory signaling functions in non-endothelial cells. In systemic sclerosis and cardiac fibrosis, CD146 expression is elevated in activated fibroblasts, where it promotes ECM deposition by triggering the Wnt/β-catenin pathway [26]. In HCC, CD146 is upregulated in tumor cells and facilitates migration, invasion, and epithelial–mesenchymal transition (EMT), and is associated with poor clinical outcomes [27]. Membrane-bound CD146 can undergo proteolytic shedding to generate a soluble form, sCD146, that enters the circulation. Accordingly, increased serum sCD146 levels have been observed in several pathological conditions, including atherosclerosis and liver cirrhosis [28, 29]. As a proteolytic product of membrane-bound CD146, sCD146 may reflect tissue pathological states and retain ligand-binding capacity, thereby participating in intercellular signaling [30].

Although CD146/sCD146 has attracted attention in various fibrosis-related diseases, its role and mechanism of action in liver fibrosis remain poorly understood. In this study, we found that hepatocyte CD146 was upregulated under fibrotic stress and associated with increased release of sCD146. Our data further suggest that sCD146 binds to integrin αvβ1 on HSCs, activates p38 mitogen-activated protein kinase (p38 MAPK) signaling, and promotes type I collagen expression and ECM deposition. These findings support a potential role for CD146/sCD146 in hepatocyte–HSC communication and suggest its possible value in the diagnosis and mechanistic investigation of liver fibrosis.

Materials and methods

A detailed description of the methods is provided in the Supplementary Materials.

Human serum samples

Serum samples were collected from healthy controls undergoing routine physical examinations (n = 54) and from patients with chronic liver disease (n = 246) at Beijing Ditan Hospital, Capital Medical University (Beijing, China), between October 2022 and June 2023. Inclusion and exclusion criteria for the clinical cohort are described in the Supplementary Materials.

Human liver tissue samples

The liver tissue samples used in this study were obtained from 35 patients who underwent liver biopsy for chronic liver disease evaluation and from four histologically normal controls at Beijing Ditan Hospital, Capital Medical University. Although these four control individuals had serum biochemical abnormalities that were suggestive of liver injury, histological examination of their liver biopsy specimens revealed no structural abnormalities. All samples were initially assessed by clinicians and subsequently confirmed by histopathological examination. Liver tissue sections were independently evaluated in a blinded manner by two experienced hepatopathologists, and fibrosis staging was performed according to the Scheuer scoring system as follows: S0, no fibrosis; S1, portal fibrosis without septa; S2, portal fibrosis with a few portal–portal septa; S3, numerous septa without pseudolobule formation; and S4, extensive septa with pseudolobule formation.

Experimental animal models

Details of the animal models are provided in the Supplementary Materials.

Statistical analysis

Quantitative data are presented as mean ± SD or mean ± SEM and as median (interquartile range [IQR]) for normally and non-normally distributed variables, respectively. Categorical variables are presented as counts and percentages, n (%). Normality was assessed using the Shapiro–Wilk test, and homogeneity of variance was assessed using Levene’s test. For continuous variables, between-group comparisons were performed using the Student’s t-test or the Mann–Whitney U test, as appropriate, whereas comparisons among multiple groups were performed using one-way ANOVA or the Kruskal–Wallis test. Categorical variables were compared using the chi-square or Fisher’s exact test, as appropriate. Receiver operating characteristic (ROC) curves were used to assess the diagnostic performance, and the area under the ROC curve (AUC) was calculated. Image quantification was performed using Fiji/ImageJ, version 1.53t. Data analysis and visualization were performed using SPSS version 25.0, GraphPad Prism version 9.5.0, and R version 4.4.1. A two-tailed P value < 0.05 was considered statistically significant.

Results

Result 1. Improved diagnostic performance of the serum sCD146-based combined diagnostic model for decompensated cirrhosis

Serum sCD146 levels were measured in 54 healthy individuals and 246 patients with chronic liver disease, including those with chronic hepatitis, liver fibrosis, compensated cirrhosis, and decompensated cirrhosis (Table 1, Table S1). Serum sCD146 levels were significantly higher in patients with chronic liver disease than in healthy controls (P < 0.01; Fig. 1A). Analysis stratified by disease progression revealed that sCD146 levels increased stepwise with disease progression and worsening Child–Pugh class (Figs. 1B and 1D) and showed a positive correlation with the fibrosis-4 index (FIB-4; ρ = 0.670, P < 0.01; Fig. 1C). In contrast, sCD146 levels did not differ significantly between viral and non-viral etiologies in the overall chronic liver disease cohort or within cirrhosis subgroups (Table S2).

Table 1.

Baseline characteristics of healthy controls and patients at different stages of chronic liver disease

Characteristics	Total (n = 300)	Healthy controls (n = 54)	Chronic hepatitis (n = 89)	Liver fibrosis (n = 42)	Compensated cirrhosis (n = 67)	Decompensated cirrhosis (n = 48)	P value
Age (years)	43.0 (33.0, 54.0)	35.0 (31.8, 39.3)	38.0 (31.0, 46.5)	44.0 (32.0, 52.8)	54.0 (43.0, 60.0)	54.0 (45.0, 64.0)	< 0.01
WBC (109/L)	5.4 (4.3, 6.5)	5.9 (5.1, 6.6)	5.8 (4.9, 6.8)	5.7 (4.3, 6.6)	4.7 (3.3, 5.8)	4.6 (2.7, 6.7)	< 0.01
PLT (109/L)	208.0 (110.0, 247.3)	245.0 (220.0, 290.7)	239.0 (217.0, 291.0)	208.0 (150.5, 234.5)	107.5 (67.0, 163.3)	79.0 (45.0, 109.0)	< 0.01
HGB (g/L)	135.0 (122.0, 145.6)	132.5 (127.0, 136.8)	145.0 (133.0, 160.0)	145.0 (131.5, 156.0)	134.5 (118.8, 151.5)	107.0 (85.0, 121.0)	< 0.01
ALT (U/L)	24.3 (15.7, 47.5)	13.6 (10.1, 18.1)	25.7 (19.9, 42.0)	80.7 (39.5, 178.4)	32.0 (18.3, 49.2)	18.5 (13.7, 32.7)	< 0.01
AST (U/L)	24.5 (17.7, 41.0)	16.3 (13.5, 19.2)	20.7 (18.0, 26.5)	55.9 (31.1, 132.9)	28.7 (21.6, 44.1)	34.5 (21.4, 60.7)	< 0.01
TBIL (μmol/L)	14.7 (11.4, 22.0)	12.3 (9.7, 16.7)	12.4 (10.7, 14.8)	14.1 (12.3, 21.2)	20.9 (13.4, 33.1)	30.6 (18.0, 53.8)	< 0.01
DBIL (μmol/L)	4.8 (3.4, 8.4)	3.5 (2.5, 4.9)	3.7 (3.1, 5.0)	4.8 (3.6, 8.0)	7.8 (4.5, 15.2)	13.7 (7.3, 37.3)	< 0.01
TP (g/L)	74.4 (69.4, 77.2)	74.6 (71.3, 77.2)	75.6 (73.2, 77.5)	75.9 (71.8, 78.7)	72.7 (68.0, 76.4)	64.9 (58.0, 74.2)	< 0.01
ALB (g/L)	44.8 (38.9, 47.1)	46.3 (45.3, 48.0)	46.5 (44.9, 48.7)	44.4 (39.9, 47.2)	40.9 (35.2, 45.0)	31.2 (26.6, 34.9)	< 0.01
GLO (g/L)	29.5 (26.9, 32.9)	28.5 (26.1, 29.6)	28.7 (26.8, 30.2)	30.5 (27.5, 34.6)	30.1 (26.9, 34.7)	32.8 (27.3, 39.9)	< 0.01
ChE (U/L)	7456.5 (5023.3, 9197.8)	7521.5 (6816.0, 8904.5)	9219.0 (8355.8, 10492.8)	8041.5 (5992.3, 9252.5)	5440.0 (3951.8, 7657.5)	2693.0 (1879.0, 3503.0)	< 0.01
Scr (μmol/L)	60.9 (51.8, 71.5)	51.6 (47.8, 54.4)	60.4 (51.4, 71.5)	61.1 (52.5, 70.6)	67.4 (60.0, 78.5)	61.1 (52.4, 72.6)	< 0.01
FIB-4	0.99 (0.63, 3.22)	0.64 (0.46, 0.81)	0.61 (0.49, 0.86)	1.51 (1.00, 2.64)	3.17 (1.80, 5.97)	6.70 (3.02, 9.70)	< 0.01
APRI	0.36 (0.20, 1.01)	0.16 (0.13, 0.20)	0.22 (0.17, 0.30)	0.68 (0.40, 2.19)	0.90 (0.48, 1.70)	1.12 (0.67, 2.02)	< 0.01
sCD146 (ng/mL)	433.48 (369.41, 552.44)	370.07 (315.31, 418.86)	405.88 (349.14, 463.08)	428.43 (394.61, 527.18)	549.14 (444.78, 731.88)	771.62 (531.97, 991.62)	< 0.01

Data are presented as median (interquartile range, IQR). P values were calculated using the Kruskal–Wallis H test. The Total (n = 300) column includes healthy controls and all patient groups

Abbreviations: WBC, white blood cell count; PLT, platelet count; HGB, hemoglobin; ALT, alanine aminotransferase; AST, aspartate aminotransferase; TBIL, total bilirubin; DBIL, direct bilirubin; TP, total protein; ALB, albumin; GLO, globulin; ChE, cholinesterase; Scr, serum creatinine; FIB-4, Fibrosis-4 index; APRI, AST-to-platelet ratio index; sCD146, soluble CD146

Fig. 1.

Serum sCD146 levels and hepatic CD146 expression in chronic liver disease stages. (A) ELISA analysis of serum sCD146 levels in NC (n = 54) and patients with chronic liver disease (n = 246). (B) Serum sCD146 levels across chronic liver disease stages: NC (n = 54), chronic hepatitis (n = 89), liver fibrosis (n = 42), compensated cirrhosis (n = 67), and decompensated cirrhosis (n = 48). (C) Correlation analysis between serum sCD146 levels and FIB-4 (n = 300). (D) Serum sCD146 levels according to Child–Pugh class: NC (n = 54), Child–Pugh class A (n = 67), Child–Pugh class B (n = 32), and Child–Pugh class C (n = 16). (E–H) ROC curve analyses of serum sCD146 levels for distinguishing different chronic liver disease stages. (I) Nomogram integrating serum sCD146 levels and clinical indicators for the diagnosis of decompensated cirrhosis. (J) ROC curve analyses of FIB-4, APRI, S index, GPR, serum sCD146, and the nomogram for distinguishing decompensated cirrhosis from compensated cirrhosis. (K) Volcano plot of cirrhotic and normal liver RNA-seq data. (L–M) H&E staining, Masson’s trichrome staining, and CD146 IHC of liver tissues from NC (n = 4) and patients with chronic liver disease stratified by Scheuer fibrosis stage (S0: n = 8; S1–2: n = 10; S3–4: n = 17). Masson’s trichrome staining for collagen deposition across fibrosis stages. Scale bar: 100 μm. (N–O) Dual IF staining for CD146 and albumin in liver tissue sections from NC (n = 4) and patients with cirrhosis (n = 5), with DAPI nuclear counterstaining. Scale bar: 100 μm. Note: CD146, cluster of differentiation 146; sCD146, soluble CD146; ELISA, enzyme-linked immunosorbent assay; FIB-4, fibrosis-4 index; ROC, receiver operating characteristic; AUC, area under the ROC curve; APRI, aspartate aminotransferase-to-platelet ratio index; GPR, gamma-glutamyl transpeptidase-to-platelet ratio; RNA-seq, RNA sequencing; H&E, hematoxylin and eosin; NC, normal control; IHC, immunohistochemistry; IF, immunofluorescence; DAPI, 4′,6-diamidino-2-phenylindole; SEM, standard error of the mean. The CD146-positive IHC staining area and CD146/albumin double-positive fluorescence area were quantified using Fiji/ImageJ. For quantification of human tissue staining, five randomly selected non-overlapping fields per sample were analyzed under the same magnification and threshold settings, and the mean value was used for statistical analysis. sCD146 values are presented as medians for non-normally distributed data, whereas IHC and IF quantification data are presented as mean ± SEM for normally distributed data. For continuous quantitative data with normal distribution and equal variance, comparisons between two and multiple groups were performed using the Student’s t-test and one-way ANOVA, respectively. For non-normally distributed data, the Mann–Whitney U or Kruskal–Wallis test was used. Correlation analyses were performed using the Spearman’s rank correlation. ns, not significant; *P < 0.05; **P < 0.01

The ROC analysis showed increasing diagnostic performance of sCD146 across different stages of chronic liver disease, with the highest AUC observed for decompensated cirrhosis (0.842), corresponding to a sensitivity of 75.0% and specificity of 86.5% (Fig. 1E–H, Table S3).

In patients with cirrhosis, we developed a noninvasive diagnostic model for decompensated cirrhosis using multivariable logistic regression. The final model included serum sCD146, platelet count (PLT), albumin grade (ALB grade), and prothrombin time (PT) and was presented as a nomogram (Fig. 1I). We compared the diagnostic performance of sCD146, the combined model, and established noninvasive indices, including the FIB-4, aspartate aminotransferase-to-platelet ratio index (APRI), S index, and gamma-glutamyl transpeptidase-to-platelet ratio (GPR), for decompensated cirrhosis. The combined diagnostic model yielded the highest AUC (0.847) among the evaluated indices and showed improved diagnostic performance for decompensated cirrhosis (Fig. 1J, Table S4). After adjustment for etiology, sCD146 remained independently associated with decompensated cirrhosis, whereas etiology itself was not significantly associated with decompensation in the model (Table S5). After further stratification by etiology, the combined model achieved AUCs of 0.818 and 0.848 for the viral and non-viral subgroups, respectively. The interaction test was not significant (P = 0.772), indicating comparable diagnostic performance across etiologic subgroups (Table S6). Thus, serum sCD146 levels were associated with the severity of chronic liver disease and showed diagnostic potential for identifying decompensated cirrhosis. Compared with the conventional indices FIB-4, APRI, S index, and GPR, the combined diagnostic model incorporating serum sCD146 and routine clinical laboratory indicators showed better diagnostic performance in cirrhosis patients.

Result 2. Upregulated CD146 expression in liver tissues of patients with liver fibrosis

Given that serum sCD146 may arise from the proteolytic shedding of membrane-bound CD146, we next assessed CD146 expression in liver tissues. Analysis of the RNA sequencing (RNA-seq) dataset GSE164799 showed significantly higher CD146 mRNA levels in liver tissues from patients with cirrhosis than in those from normal controls (P < 0.05; Fig. 1K). Subsequently, CD146 expression was examined in liver tissues from four histologically normal controls and 35 patients with liver fibrosis at different Scheuer fibrosis stages (S0, S1–2, and S3–4; Table 2). Immunohistochemistry (IHC) showed that CD146 protein expression increased significantly with advancing fibrosis stage (Fig. 1L–M). Immunofluorescence (IF) staining of liver sections from normal controls and patients with cirrhosis for CD146 and the hepatocyte marker albumin demonstrated increased colocalization of CD146 with albumin in cirrhotic liver tissues, which was further supported by quantitative analysis (Fig. 1N–O).

Table 2.

Clinical characteristics and laboratory parameters across different liver fibrosis stages

Characteristics	NC (n = 4)	S0 (n = 8)	S1–2 (n = 10)	S3–4 (n = 17)
Male, n (%)	3 (75.0)	4 (50.0)	7 (70.0)	16 (94.1)
Age (years)	32.0 ± 6.7	36.5 ± 11.5	38.6 ± 8.6	54.6 ± 8.7
ALT (U/L)	29.4 (25.9, 36.2)	21.0 (14.7, 28.2)	38.9 (23.4, 57.0)	28.2 (20.8, 39.7)
AST (U/L)	17.5 (16.8, 18.6)	17.1 (15.9, 24.5)	35.1 (18.9, 55.9)	34.0 (23.6, 50.1)
ALP (U/L)	41.3 (35.0, 46.4)	66.2 (52.9, 91.1)	70.3 (62.1, 75.9)	91.4 (56.4, 94.2)
GGT (U/L)	19.5 (13.6, 24.1)	12.5 (11.0, 18.7)	25.9 (21.2, 62.5)	36.9 (15.5, 118.5)
TBIL (μmol/L)	14.1 (9.4, 15.0)	13.5 (11.8, 15.9)	13.5 (10.7, 15.9)	16.7 (11.8, 21.2)
DBIL (μmol/L)	4.1 (2.9, 5.6)	4.2 (3.6, 5.2)	5.0 (3.9, 5.5)	5.9 (4.3, 7.5)
TP (g/L)	74.9 (72.1, 77.4)	70.6 (67.8, 79.0)	73.8 (72.1, 80.3)	68.0 (65.6, 71.1)
ALB (g/L)	49.3 (44.2, 49.8)	46.9 (40.1, 48.4)	44.0 (41.6, 46.4)	41.0 (35.1, 43.5)
PLT (109/L)	311.5 (253.3, 367.5)	204.0 (146.0, 247.5)	167.0 (158.2, 172.0)	149.0 (93.0, 177.0)
FIB-4	0.30 (0.30, 0.40)	0.80 (0.60, 0.90)	1.70 (0.80, 2.60)	2.80 (1.70, 3.70)

Data are presented as mean ± SD for normally distributed variables, median (IQR) for non-normally distributed variables, or n (%) for categorical variables. Fibrosis stage (S0–S4) was defined according to the Scheuer scoring system

Abbreviations: NC, normal control; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ALP, alkaline phosphatase; GGT, gamma-glutamyl transferase; TBIL, total bilirubin; DBIL, direct bilirubin; TP, total protein; ALB, albumin; PLT, platelet count; FIB-4, Fibrosis-4 index

Result 3. CD146 upregulation in liver fibrosis models and hepatocytes

To investigate the association between CD146 and liver fibrosis, we established two murine models of liver fibrosis and assessed CD146 levels in liver tissues. In the carbon tetrachloride (CCl₄)-induced fibrosis model, IHC and quantitative analysis revealed a significant increase in CD146 expression in liver tissues of fibrotic mice compared with controls (Fig. 2A–D). Reverse transcription quantitative polymerase chain reaction (RT-qPCR) and Western blot showed significantly increased CD146 mRNA and protein levels (Fig. 2E–I). A similar increase in hepatic CD146 expression was observed in the methionine- and choline-deficient (MCD) diet–induced fibrosis model (Fig. 2J–L), suggesting that CD146 upregulation occurs in both toxic injury- and diet-induced fibrosis models. To identify the cell type in which CD146 was upregulated during murine liver fibrosis, we performed IF staining for CD146 and albumin. Analysis showed a significant increase in the colocalization of CD146 with albumin in liver tissues from fibrotic mice compared with controls (Fig. 2M–N).

Fig. 2.

CD146 expression in liver fibrosis models and hepatocytes. (A–D) H&E staining, Masson’s trichrome staining, Sirius Red staining, and CD146 IHC of liver tissues from Oil- and CCl₄-treated mice (n = 5 mice per group). (E–G) Western blot analysis and densitometric quantification of COL1A1, α-SMA, and CD146 protein levels in liver tissues from Oil- and CCl₄-treated mice (n = 5 mice per group). (H–I) RT-qPCR analysis of Col1a1, Acta2, and Cd146 mRNA levels in liver tissues from Oil- and CCl₄-treated mice (n = 5 mice per group). (J–K) Western blot analysis and densitometric quantification of CD146 protein levels in liver tissues from NCD- and MCD-fed mice (n = 5 mice per group). (L) RT-qPCR analysis of Cd146 mRNA levels in liver tissues from NCD- and MCD-fed mice (n = 5 mice per group). (M–N) Dual IF staining for CD146 and albumin in liver tissues from Oil- and CCl₄-treated mice, with DAPI nuclear counterstaining (n = 5 mice per group). Scale bar: 50 μm. (O–Q) Western blot analysis and densitometric quantification of CD146 protein levels, and RT-qPCR analysis of CD146 mRNA levels, in Huh7 cells after 24 h of TGF-β1 treatment (n = 3 independent biological replicates). (R–S) Western blot analysis of CD146 protein levels, and RT-qPCR analysis of CD146 mRNA levels, in LX2 cells after 24 h of TGF-β1 treatment (n = 3 independent biological replicates). Note: H&E, hematoxylin and eosin; IHC, immunohistochemistry; CCl₄, carbon tetrachloride; COL1A1, collagen type I alpha 1; α-SMA, alpha-smooth muscle actin; RT-qPCR, reverse transcription quantitative polymerase chain reaction; NCD, normal chow diet; MCD, methionine- and choline-deficient diet; IF, immunofluorescence; DAPI, 4′,6-diamidino-2-phenylindole; TGF-β1, transforming growth factor-β1; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; SEM, standard error of the mean. For Western blot quantification, target protein levels were normalized to GAPDH and expressed relative to the corresponding control group. Histological IHC and IF quantification were performed using Fiji/ImageJ. Masson- and Sirius Red-positive areas, CD146-positive IHC-stained areas, and CD146/albumin double-positive fluorescent areas were quantified as indicated in the corresponding panels. For quantification of tissue staining, five randomly selected non-overlapping fields per mouse were analyzed under the same magnification and threshold settings, and the mean value was used as one biological replicate for statistical analysis. For the in vitro experiments, n represents independent biological replicates generated from independent cell cultures and treatments. Technical replicates were averaged before statistical analysis and were not treated as independent biological replicates. Data are presented as mean ± SEM for normally distributed data. Between-group comparisons were performed using a two-tailed unpaired Student’s t-test. ns, not significant; *P < 0.05; **P < 0.01

In vitro, the profibrotic factor TGF-β1 significantly upregulated CD146 expression in Huh7 cells at both the mRNA and protein levels, whereas no significant difference was observed in the human HSC line LX2 (Fig. 2O–S). These findings suggest that TGF-β1 may induce CD146 expression in hepatocyte-derived cells during liver fibrosis.

Taken together, these findings indicate that CD146 is upregulated in liver tissues from mouse models of liver fibrosis, with increased localization in albumin-positive hepatocytes under fibrotic conditions and upregulation in Huh7 cells following TGF-β1 treatment.

Result 4. Hepatocyte-specific deletion of CD146 attenuates CCl₄-induced liver injury and fibrosis

To evaluate the role of hepatocyte CD146 in liver fibrosis, we generated hepatocyte-specific CD146 knockout mice (CD146^Alb−Cre) and control mice (CD146^flox/flox), and established a CCl₄-induced liver fibrosis model (Fig. 3A). Following CCl₄ treatment, no significant between-group differences were observed in body weight or liver weight; however, serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were significantly lower in the CD146^Alb−Cre group than in control mice (P < 0.05; Fig. 3B–D). Compared with controls, CD146^Alb−Cre livers showed better-preserved liver surface architecture with reduced fibrous adhesions (Fig. 3E). Scanning electron microscopy showed reduced collagen fibers and ECM deposition in the CD146-deficient group (Fig. 3F). Masson and Sirius Red staining, together with quantitative analysis of fibrotic areas, showed significantly reduced collagen accumulation and septa formation in the knockout group (Fig. 3G–I). IF and flow cytometric analyses showed significantly reduced infiltration of F4/80⁺ macrophages, Ly6C^hi monocytes, and Ly6G⁺CD11b⁺ neutrophils in liver tissues from CD146 knockout mice compared with controls (Fig. 3J–M). Collectively, these findings indicate that hepatocyte-specific deletion of CD146 attenuates CCl₄-induced liver inflammation and fibrosis.

Fig. 3.

Hepatocyte-specific deletion of CD146 in a CCl₄-induced mouse model of liver fibrosis. (A) Schematic of the CCl₄-induced liver fibrosis model. (B–C) Body weight and liver weight of CD146^flox/flox mice (Ctrl) and CD146^Alb−Cre mice (CD146-KO) after CCl₄ treatment (n = 5 mice per group). (D) Serum ALT and AST levels in Ctrl and CD146-KO mice after CCl₄ treatment (n = 5 mice per group). (E) Gross liver morphology of Ctrl and CD146-KO mice after CCl₄ treatment. (F) Scanning electron microscopy of liver tissues from Ctrl and CD146-KO mice after CCl₄ treatment (n = 5 mice per group). Scale bar: 100 μm. (G–J) H&E staining, Masson’s trichrome staining, Sirius Red staining, F4/80 IF staining, and quantitative analysis of liver fibrosis and macrophage infiltration in Ctrl and CD146-KO mice after CCl₄ treatment (n = 5 mice per group). Scale bar: 100 μm. (K–M) Flow cytometry analysis of hepatic immune cell subsets, including Ly6C^hi monocytes and Ly6G⁺CD11b⁺ neutrophils, in Ctrl and CD146-KO mice after CCl₄ treatment (n = 5 mice per group). Note: CD146, cluster of differentiation 146; CCl₄, carbon tetrachloride; ALT, alanine aminotransferase; AST, aspartate aminotransferase; H&E, hematoxylin and eosin; IF, immunofluorescence; F4/80, macrophage marker F4/80; Ly6C, lymphocyte antigen 6 complex, locus C; Ly6G, lymphocyte antigen 6 complex, locus G; CD11b, cluster of differentiation 11b; SEM, standard error of the mean. Histological and IF quantifications were performed using Fiji/ImageJ. Masson- and Sirius Red-positive areas and F4/80-positive fluorescent areas were quantified as indicated in the corresponding panels. For quantification of tissue staining, five randomly selected non-overlapping fields per mouse were analyzed under the same magnification and threshold settings, and the mean value was used as one biological replicate for statistical analysis; n represents biologically independent mice per group. Data are presented as mean ± SEM for normally distributed data. Statistical comparisons between two groups were performed using a two-tailed unpaired Student’s t-test. ns, not significant; *P < 0.05; **P < 0.01

Result 5. Hepatocyte-specific CD146 overexpression exacerbates high-fat diet-induced hepatic inflammation and fibrosis in mice

To further assess the effects of adeno-associated virus serotype 8 (AAV8)-mediated hepatocyte-specific CD146 overexpression on liver inflammation and fibrosis, we established a mouse model by tail-vein injection of AAV8-CD146 or AAV8-GFP as a control. RT-qPCR and Western blot analyses showed that CD146 mRNA and protein levels were significantly increased in the livers of mice injected with AAV8-CD146 (Fig. 4A–C). Under normal chow diet (NCD) conditions, no significant between-group differences were observed in body weight, liver weight, serum ALT and AST levels, or histological indicators of hepatic inflammation and fibrosis (Fig. 4D–J). Under high-fat diet (HFD) conditions, CD146-overexpressing mice exhibited significantly higher body and liver weights than controls (P < 0.01), and serum ALT and AST levels were significantly increased (P < 0.01) (Fig. 4K–N). Masson staining and F4/80 IHC, supported by quantitative analysis of positive staining areas, showed significantly increased collagen deposition and macrophage infiltration in the livers of CD146-overexpressing mice (Fig. 4O–Q). Overall, hepatocyte-specific CD146 overexpression aggravated hepatic inflammation and fibrosis under HFD conditions.

Fig. 4.

CD146 overexpression in mice under NCD and HFD conditions. (A–B) Western blot analysis and densitometric quantification of hepatic CD146 protein levels in mice injected with AAV8-GFP (Ctrl) or AAV8-CD146 (CD146-OE) (n = 6 mice per group). (C) RT-qPCR analysis of Cd146 mRNA levels in liver tissues (n = 6 mice per group). (D) Schematic of the CD146-OE model under NCD conditions. (E–F) Body weight and liver weight of Ctrl and CD146-OE mice under NCD conditions (n = 6 mice per group). (G) Serum ALT and AST levels in Ctrl and CD146-OE mice under NCD conditions (n = 6 mice per group). (H–J) H&E staining, Masson’s trichrome staining, F4/80 IHC, and quantitative analysis of liver fibrosis and macrophage infiltration in Ctrl and CD146-OE mice under NCD conditions (n = 6 mice per group). Scale bar: 100 μm. (K) Schematic of the CD146-OE model under HFD conditions. (L–M) Body weight and liver weight of Ctrl and CD146-OE mice under HFD conditions (n = 6 mice per group). (N) Serum ALT and AST levels in Ctrl and CD146-OE mice under HFD conditions (n = 6 mice per group). (O–Q) H&E staining, Masson’s trichrome staining, F4/80 IHC, and quantitative analysis of liver fibrosis and macrophage infiltration in Ctrl and CD146-OE mice under HFD conditions (n = 6 mice per group). Scale bar: 100 μm. Note: CD146, cluster of differentiation 146; NCD, normal chow diet; HFD, high-fat diet; Ctrl, control; OE, overexpression; AAV8, adeno-associated virus serotype 8; GFP, green fluorescent protein; RT-qPCR, reverse transcription quantitative polymerase chain reaction; ALT, alanine aminotransferase; AST, aspartate aminotransferase; H&E, hematoxylin and eosin; IHC, immunohistochemistry; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; SEM, standard error of the mean. For Western blot quantification, target protein levels were normalized to those of GAPDH and expressed relative to the corresponding Ctrl group. Histological and immunohistochemical quantification was performed using Fiji/ImageJ. Masson-positive and F4/80-positive immunohistochemically stained areas were quantified as indicated in the corresponding panels. For quantification of tissue staining, five randomly selected non-overlapping fields per mouse were analyzed under the same magnification and threshold settings, and the mean value was used as one biological replicate for statistical analysis; n represents biologically independent mice per group. Data are presented as mean ± SEM for normally distributed data. Statistical comparisons were performed using a two-tailed unpaired Student’s t-test. ns, not significant; *P < 0.05; **P < 0.01

Result 6. CD146 regulates the expression of inflammation- and fibrosis-related factors in the liver

To elucidate the mechanisms by which hepatocyte CD146 contributes to liver fibrosis, we analyzed the expression of inflammation- and fibrosis-related genes in liver tissues from hepatocyte-specific CD146 knockout and overexpression mouse models. In the CCl₄-induced liver fibrosis model, the expression of several inflammatory genes (e.g., Il1β, Il6, Tnfα) and fibrosis markers (e.g., Tgfβ1, Col1a1, Acta2) was significantly lower in CD146^Alb−Cre mice than in CD146^flox/flox mice. At the protein level, Western blot showed that IL-1β, α-SMA, and COL1A1 were significantly reduced (Fig. 5A–E). In CD146-overexpressing mice, the expression of inflammation-related genes (e.g., Il6, Il1β, Tnfα, Mcp1) and fibrosis-related genes (e.g., Tgfβ1, Col1a1, Col3a1) was significantly increased under both NCD and HFD conditions. In the HFD group, Western blot showed that COL3A1 protein levels were significantly elevated, whereas COL1A1 levels showed an upward trend that did not reach statistical significance (Fig. 5F–K).

Fig. 5.

Hepatocyte CD146 regulates inflammation- and fibrosis-related factor expression in the liver. (A–B) RT-qPCR analysis of inflammatory genes (Il1β, Il6, and Tnfα) and fibrotic genes (Tgfβ1, Col1a1, and Acta2) in CD146^flox/flox mice (Ctrl) and CD146^Alb−Cre mice (CD146-KO) after CCl₄ treatment (n = 5 mice per group). (C–E) Western blot analysis and densitometric quantification of TNF-α, IL-1β, α-SMA, and COL1A1 protein levels in liver tissues from CCl₄-treated Ctrl and CD146-KO mice (n = 5 mice per group). (F–G) RT-qPCR analysis of inflammatory genes (Il6, Il1β, Tnfα, Mcp1, Il10, and Mgl2) and fibrotic genes (Tgfβ1, Col1a1, Col3a1, and Acta2) in liver tissues from NCD-fed mice injected with AAV8-GFP (Ctrl) or AAV8-CD146 (CD146-OE) (n = 6 mice per group). (H–I) RT-qPCR analysis of inflammatory genes (Il6, Il1β, Tnfα, Mcp1, Adgre1, and Mgl2) and fibrotic genes (Tgfβ1, Col1a1, Col3a1, and Acta2) in liver tissues from HFD-fed Ctrl and CD146-OE mice (n = 6 mice per group). (J–K) Western blot analysis and densitometric quantification of COL1A1 and COL3A1 protein levels in liver tissues from HFD-fed Ctrl and CD146-OE mice (n = 6 mice per group). Note: CD146, cluster of differentiation 146; RT-qPCR, reverse transcription quantitative polymerase chain reaction; Ctrl, control; KO, knockout; CCl₄, carbon tetrachloride; NCD, normal chow diet; HFD, high-fat diet; OE, overexpression; AAV8, adeno-associated virus serotype 8; GFP, green fluorescent protein; TNF-α, tumor necrosis factor-α; IL-1β, interleukin-1β; α-SMA, alpha-smooth muscle actin; COL1A1, collagen type I alpha 1; COL3A1, collagen type III alpha 1; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; SEM, standard error of the mean. In panels A–E, Ctrl refers to CD146^flox/flox mice and CD146-KO refers to CD146^Alb−Cre mice. In panels F–K, Ctrl refers to the AAV8-GFP group and CD146-OE refers to the AAV8-CD146 group. For Western blot quantification, target protein levels were normalized to those of GAPDH and expressed relative to the corresponding Ctrl group; n represents biologically independent mice per group. Data are presented as mean ± SEM for normally distributed data. Statistical comparisons were performed using a two-tailed unpaired Student’s t-test. ns, not significant; *P < 0.05; **P < 0.01

Collectively, these findings indicate that hepatocyte CD146 may contribute to the progression of liver fibrosis by modulating the expression of inflammation- and fibrosis-related factors. CD146 deletion was associated with reduced expression of these factors, whereas CD146 overexpression was associated with their increased expression.

Result 7. CD146 activates the p38 MAPK signaling pathway and promotes hepatic inflammation and fibrosis in mice

To explore the potential downstream mechanisms by which CD146 promotes liver fibrosis, we performed RNA-seq on liver tissues from CD146-overexpressing mice under NCD and HFD conditions to identify differentially expressed genes (DEGs).

Under NCD conditions, several inflammation- and fibrosis-associated genes, including Ccl5 and Ctsg, were significantly upregulated in the CD146-overexpressing mice (Fig. S1A–B). Under HFD conditions, a larger number of fibrosis-associated genes were upregulated, including Acta2 and Timp3, as well as Tgfβ1, Col1a1, and Col3a1, which were validated by RT-qPCR (Fig. 6A and B). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses revealed that the upregulated genes were significantly enriched in biological processes, such as inflammatory responses and ECM remodeling, as well as in relevant signaling pathways (Fig. S1C–D, 6C–D). Notably, among the fibrosis-related KEGG pathways, the MAPK signaling pathway was significantly enriched (Fig. S1E–F, 6E–F).

Fig. 6.

CD146 promotes hepatic inflammatory and fibrotic responses in association with p38 MAPK pathway activation. (A) Volcano plot of DEGs in liver tissues between the AAV8-CD146 (CD146-OE) and AAV8-GFP (Ctrl) groups under HFD conditions (n = 3 mice per group). Significant DEGs were identified based on the thresholds of |log2FC|> 1 and adjusted P value < 0.05. (B) Heatmap of DEGs in liver tissues from CD146-OE and Ctrl mice under HFD conditions (n = 3 mice per group). Upregulated genes are shown in pink, and downregulated genes in green. (C) GO enrichment analysis of DEGs, with the top 20 enriched GO terms shown (n = 3 mice per group). (D) KEGG pathway enrichment analysis of DEGs, with the top 20 enriched pathways shown (n = 3 mice per group). (E) Heatmap of fibrosis-related DEGs in liver tissues from CD146-OE and Ctrl mice under HFD conditions (n = 3 mice per group). (F) KEGG pathway enrichment analysis of fibrosis-related DEGs, with the top 30 enriched pathways shown (n = 3 mice per group). (G–H) Western blot analysis and densitometric quantification of p-p38 and p38 protein levels in liver tissues from Ctrl and CD146-OE mice under NCD conditions (n = 6 mice per group). (I–J) Western blot analysis and densitometric quantification of p-p38 and p38 protein levels in liver tissues from Ctrl and CD146-OE mice under HFD conditions (n = 6 mice per group). (K–L) Western blot analysis and densitometric quantification of p-MKK3, MKK3, p-p38, and p38 protein levels in liver tissues from CCl₄-treated CD146^flox/flox mice (Ctrl) and CD146^Alb−Cre mice (CD146-KO) (n = 5 mice per group). Note: CD146, cluster of differentiation 146; MAPK, mitogen-activated protein kinase; DEGs, differentially expressed genes; Ctrl, control; OE, overexpression; AAV8, adeno-associated virus serotype 8; GFP, green fluorescent protein; HFD, high-fat diet; FC, fold change; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; NCD, normal chow diet; CCl₄, carbon tetrachloride; KO, knockout; MKK3, mitogen-activated protein kinase kinase 3; SEM, standard error of the mean. For Western blot quantification, phosphorylated protein levels were normalized to the corresponding total protein levels. Specifically, p-p38 was normalized to total p38 and p-MKK3 was normalized to total MKK3; n represents biologically independent mice per group. Data are presented as mean ± SEM for normally distributed data. Statistical comparisons between groups were performed using a two-tailed unpaired Student’s t-test. ns, not significant; *P < 0.05; **P < 0.01

Western blot analysis showed that CD146-overexpressing mice exhibited increased activation of the p38 MAPK signaling pathway, with significantly elevated p-p38 levels normalized to total p38, under both NCD and HFD conditions (Fig. 6G–J). Conversely, in CD146 knockout mice subjected to CCl₄-induced fibrosis, analysis showed reduced p-MKK3 and p-p38 levels after normalization to their corresponding total protein levels (Fig. 6K–L).

Taken together, these findings suggest that CD146 promotes liver fibrosis, at least in part, through the activation of the p38 MAPK signaling pathway, which may contribute to hepatic inflammatory responses and ECM deposition.

Result 8. Hepatocyte CD146 promotes hepatocyte–HSC crosstalk through paracrine release of sCD146

In vitro experiments were performed to investigate the effects of CD146 overexpression in hepatocyte-derived cells on HSC activation. Huh7 cells were transfected with a human CD146 overexpression plasmid or the corresponding empty vector and the supernatants were collected. Proteomic analysis revealed that, compared with controls, sCD146 levels were significantly increased in the supernatant of CD146-overexpressing Huh7 cells (fold change > 1.5, P < 0.05), identifying sCD146 as one of the most prominently altered soluble proteins (Fig. 7G–I).

Fig. 7.

CD146 overexpression in hepatocyte-derived cells promotes hepatocyte–HSC crosstalk through paracrine release of sCD146. (A) Schematic of the Huh7–LX2 co-culture system. (B–D) Western blot analysis and densitometric quantification of CD146 protein levels, and RT-qPCR analysis of CD146 mRNA levels, in Huh7 cells transfected with CD146 overexpression plasmid (Ad-CD146) or the corresponding empty vector (Ad-Ctrl) (n = 3 independent biological replicates). (E–F) Western blot analysis and densitometric quantification of COL1A1 protein levels in LX2 cells treated with CM from Ad-CD146- or Ad-Ctrl-infected Huh7 cells, with TGF-β1-treated LX2 cells as a positive control (n = 3 independent biological replicates). (G–I) Proteomic analysis of CM collected from Huh7 cells after infection with Ad-CD146 or Ad-Ctrl, including volcano plot, radar plot, and heatmap visualization (n = 3). (J–K) Western blot analysis and densitometric quantification of COL1A1 protein levels in LX2 cells treated with rhsCD146 (n = 3 independent biological replicates). (L–P) Western blot analysis and densitometric quantification of CD146, COL1A1, and α-SMA protein levels, and RT-qPCR analysis of CD146, ACTA2, COL1A1, COL1A2, and COL3A1 mRNA levels, in LX2 cells infected with Ad-CD146 or Ad-Ctrl (n = 3 independent biological replicates). Note: CD146, cluster of differentiation 146; HSC, hepatic stellate cell; sCD146, soluble CD146; Ctrl, control; CM, conditioned medium; TGF-β1, transforming growth factor-β1; COL1A1, collagen type I alpha 1; rhsCD146, recombinant human soluble CD146; α-SMA, alpha-smooth muscle actin; RT-qPCR, reverse transcription quantitative polymerase chain reaction; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; SEM, standard error of the mean. For Western blot quantification, target protein levels were normalized to GAPDH and expressed relative to the corresponding control group. For the in vitro experiments, n represents independent biological replicates generated from independent cell cultures and treatments. Technical replicates were averaged before statistical analysis and were not treated as independent biological replicates. Data are presented as mean ± SEM for normally distributed data. Statistical comparisons were performed using a two-tailed unpaired Student’s t-test or one-way ANOVA, as appropriate. ns, not significant; *P < 0.05; **P < 0.01

Conditioned medium (CM) collected from Huh7 cells was used to treat LX2 cells (Fig. 7A). Compared with controls, CM derived from CD146-overexpressing Huh7 cells significantly increased COL1A1 expression in LX2 cells, as shown by RT-qPCR and Western blot analysis, suggesting HSC activation (Fig. 7B–F). Similarly, exogenous recombinant human sCD146 protein (rhsCD146) increased COL1A1 and TIMP1 expression in LX2 cells (Figs. 7J–K, S2A–F). These findings suggest that CD146 overexpression in hepatocytes promotes HSC activation and collagen expression, at least partly through increased paracrine release of sCD146.

To determine whether CD146 overexpression exerted a direct profibrotic effect on HSCs, LX2 cells were transfected with a human CD146 overexpression plasmid or the corresponding empty vector control. No significant changes were observed in the mRNA levels of fibrosis-related genes (COL1A1, COL1A2, COL3A1, and ACTA2) or in the protein levels of COL1A1 and α-SMA (Fig. 7L–P), suggesting that CD146 overexpression did not directly induce a profibrotic phenotype in LX2 cells under these conditions.

Overall, these data suggest that CD146 overexpression in hepatocyte-derived cells increases sCD146 secretion, which may promote HSC activation and collagen synthesis through a paracrine mechanism.

Result 9. sCD146 activates HSCs via the integrin αvβ1–p38 MAPK signaling axis

Based on our finding that hepatocyte CD146 promotes liver fibrosis through the p38 MAPK signaling pathway in mice, we further investigated whether sCD146 regulates HSC activation via the same pathway. Incubation of LX2 cells with CM from CD146-overexpressing Huh7 cells or direct stimulation with rhsCD146 increased p-MKK3 and p-p38 protein levels (Fig. 8A–E). These results suggest that sCD146 may activate the p38 MAPK pathway and promote HSC activation and collagen expression.

Fig. 8.

sCD146 activates HSCs through an integrin αvβ1-dependent p38 MAPK signaling pathway. (A–B) Western blot analysis and densitometric quantification of p38 MAPK pathway-related protein levels in LX2 cells treated with CM from Huh7 cells infected with Ad-CD146 or Ad-Ctrl (n = 3 independent biological replicates). (C–E) Western blot analysis and densitometric quantification of p-MKK3, MKK3, p-p38, and p38 protein levels in LX2 cells treated with rhsCD146 (n = 3 independent biological replicates). (F–H) Western blot analysis and densitometric quantification of COL1A1 and p38 MAPK pathway-related protein levels in LX2 cells treated with rhsCD146 alone or in combination with an integrin αvβ1 inhibitor (n = 3 independent biological replicates). Note: sCD146, soluble CD146; HSC, hepatic stellate cell; MAPK, mitogen-activated protein kinase; CM, conditioned medium; Ctrl, control; rhsCD146, recombinant human soluble CD146; MKK3, mitogen-activated protein kinase kinase 3; COL1A1, collagen type I alpha 1; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; SEM, standard error of the mean. For Western blot quantification, COL1A1 protein levels were normalized to GAPDH and expressed relative to the corresponding control group. Phosphorylated protein levels were normalized to corresponding total protein levels. Specifically, p-p38 was normalized to total p38 and p-MKK3 was normalized to total MKK3. For in vitro experiments, n represents independent biological replicates generated from independent cell cultures and treatments. Technical replicates were averaged before statistical analysis and were not treated as independent biological replicates. Data are presented as mean ± SEM for normally distributed data. Statistical comparisons were performed using a two-tailed unpaired Student’s t-test or one-way ANOVA, as appropriate. ns, not significant; *P < 0.05; **P < 0.01

Previous studies have reported that sCD146 binds to cell-surface integrin αvβ1 to regulate vascular function [30]. To assess the role of this signaling axis in HSCs, we performed functional blockade of integrin αvβ1 in LX2 cells. Compared with rhsCD146 treatment alone, co-treatment with rhsCD146 and an integrin αvβ1 inhibitor significantly reduced p-MKK3, p-p38, and COL1A1 levels (Fig. 8F–H), suggesting that integrin αvβ1 may be involved in rhsCD146-induced p38 MAPK signaling activation and HSC activation.

Together, these findings suggest that hepatocyte-derived sCD146 activates HSCs and promotes collagen expression, at least in part, via the integrin αvβ1–p38 MAPK signaling axis.

Discussion

This study provides evidence supporting a profibrotic role for hepatocyte-derived CD146 in liver fibrosis and suggests that sCD146 mediates intercellular communication between hepatocytes and HSCs. TGF-β1 induced CD146 upregulation in hepatocytes and promoted its release as sCD146, thereby engaging the CD146/sCD146 signaling axis. Upon binding to integrin αvβ1 on the surface of HSCs, sCD146 activated the p38 MAPK signaling pathway, thereby promoting HSC activation and collagen production. In contrast to the traditional view of CD146 as an endothelial marker, our findings showed that hepatocyte CD146 expression was elevated under fibrotic stress. CD146 deletion alleviated inflammation and fibrosis, whereas its overexpression exacerbated these pathological processes. These observations broaden the understanding of CD146 function beyond endothelial cells, suggest its novel role in hepatocyte–HSC interactions, and provide new insights into noninvasive diagnosis and targeted therapeutic strategies for liver fibrosis.

Among profibrotic signaling pathways, TGF-β1 is widely recognized as a central driver because it not only directly induces HSCs to synthesize and secrete collagen but also contributes to myofibroblast activation and phenotypic remodeling [31–34]. Although the pivotal role of TGF-β1 in liver fibrosis is well established, the precise molecular components of its downstream amplification mechanisms remain unclear. Previous studies have suggested that paracrine signals secreted by hepatocytes may enhance activation of the TGF-β or other pathway [19, 35, 36]. Our results suggest that TGF-β1 markedly induces CD146 expression in hepatocytes and promotes its release as sCD146. Released sCD146 acts on HSCs via a paracrine mechanism, suggesting the CD146/sCD146 signaling axis may complement the traditional view of TGF-β1 acting directly on HSCs and may serve as a potential amplification mechanism within the TGF-β1 profibrotic pathway. The intercellular signaling network in the liver exhibits remarkable complexity under both physiological and pathological conditions, with crosstalk between hepatic cell populations and HSCs serving as a major regulator of HSC activation and phenotypic transitions [37, 38]. In recent years, several studies have identified important paracrine pathways that are mediated by hepatocytes and macrophages. Lü et al. reported that hepatocyte sex determining region Y (SRY) amplifies PDGF signaling by upregulating platelet-derived growth factor receptor-α (PDGFR-α) and high mobility group box 1 (HMGB1), thereby contributing to the greater severity of fibrosis in males [39]. Wang et al. found that hepatocyte ninjurin 2 (NINJ2) promotes PDGF-BB secretion and activates HSCs [40]. Li et al. demonstrated that macrophage-specific fibroblast growth factor 12 (FGF12) drives fibrosis via the monocyte chemoattractant protein-1 (MCP-1)–CCR2 axis [41]. Collectively, these studies highlight how diverse molecules act through intercellular communication to influence HSCs and thereby accelerate fibrotic progression [42]. Building on this foundation, our study further showed that hepatocyte CD146 promotes HSC activation and collagen deposition through increased release of sCD146, which acts in a paracrine manner to mediate hepatocyte–HSC communication and contribute to fibrotic progression. These findings provide new molecular evidence for understanding the complex mechanisms of intercellular interactions in liver fibrosis.

Although elevated sCD146 levels have been reported in diseases such as diabetes, atherosclerosis, and multiple sclerosis [28, 43, 44], the precise pathological functions of sCD146 remain unclear. Our data support a previously unrecognized profibrotic role of sCD146 in liver fibrosis, whereby sCD146 binds to integrin αvβ1 on the surface of HSCs and activates the p38 MAPK pathway, thereby promoting HSC activation and type I collagen expression. Therefore, sCD146 may function as a profibrotic signaling mediator in hepatocyte–HSC interactions.

In clinical samples, we observed that serum sCD146 levels progressively increased with the progression and severity of liver disease, ranging from chronic hepatitis to decompensated cirrhosis. When combined with routine laboratory indicators, the constructed diagnostic model showed improved diagnostic performance for decompensated cirrhosis and outperformed commonly used noninvasive indices, such as FIB-4, APRI, and the S index. Although existing serum diagnostic markers (e.g., hyaluronic acid, N-terminal propeptide of type III procollagen, and matrix metalloproteinases) provide some reference value for assessing advanced liver fibrosis, they remain limited by poor organ specificity and high testing costs [45]. In this context, the combined diagnostic model may help optimize the use of existing noninvasive indicators, although its relatively better performance was observed only in this cohort, particularly for identifying decompensated cirrhosis.

Although direct clinical intervention targeting sCD146 remains challenging, its receptor, integrin αvβ1, and the downstream p38 MAPK pathway may represent pharmacologically targetable nodes for antifibrotic intervention using small-molecule inhibitors or blocking antibodies [46, 47]. Regarding mechanistic validation, we observed that hepatic inflammation and fibrosis were aggravated in CD146-overexpressing mice under HFD conditions, whereas CD146 deletion in the CCl₄-induced model reduced p38 MAPK pathway activity and attenuated tissue pathology. These complementary in vivo findings further support the profibrotic role of CD146 and its potential as an antifibrotic therapeutic target.

This study has several limitations. First, although CD146 is highly expressed in liver sinusoidal endothelial cells and may also be relevant in other hepatic cell populations, lineage-specific interventions were not performed in these non-hepatocyte populations [48]. Second, the specific proteases responsible for CD146 shedding and the cleavage process by which membrane-bound CD146 is converted into sCD146 were not defined. Third, circulating sCD146 levels may be influenced by clinical and biological heterogeneity. In addition, the clinical model was derived from a single-center cohort and requires validation in larger multicenter cohorts and independent external validation cohorts. Fourth, the in vivo findings were based mainly on diet-induced and CCl₄-induced mouse models of liver fibrosis, which may not fully capture the etiological heterogeneity of human chronic liver diseases. Further studies using lineage-specific approaches, etiology-specific models, and clinically stratified validation cohorts are warranted.

In conclusion, this study supports the existence of a novel profibrotic signaling axis in which TGF-β1 induces CD146 upregulation in hepatocytes and enhances the release of sCD146, which binds to integrin αvβ1 on the surface of HSCs, activates the p38 MAPK pathway, thereby promoting HSC activation and ECM deposition. These findings suggest that the CD146/sCD146 axis may represent a potential target for future antifibrotic strategies, while circulating sCD146 may serve as a candidate diagnostic biomarker for decompensated cirrhosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Qi Wang, Wen Xie, and Yue Li conceived and mentored this study. Bingqing Yang, Chenxue Hou, Rongcai Ye, Jingya Yin, and Yixue Shu designed and performed experiments, acquired and analyzed data, and wrote the draft of the manuscript. Xiaoyan Cheng, Lei Sun, Kun Yang, and Quan Sun gave technical support and conceptual advice.

Funding

This study was supported by the National Natural Science Foundation of China (82170591) and Beijing High-level Public Health Technical Personnel Construction Project (Discipline backbone-02-29).

Data availability

The raw sequencing data have been submitted to the NCBI Sequence Read Archive (SRA) under submission ID SUB11418856 and will be made publicly available upon publication. In the meantime, the data can be obtained from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The collection and use of human serum and liver tissue samples complied with international ethical guidelines for biomedical research involving human participants and were approved by the Ethics Committee of Beijing Ditan Hospital, Capital Medical University (Ethics Approval No. DTEC-KY2021-046-04). Additionally, all animal experiments adhered to institutional and national guidelines for the care and use of laboratory animals, with approval from the Animal Care and Use Committees of the Health Science Center of Capital Medical University and the Institute of Zoology, Chinese Academy of Sciences (Ethics Approval No. AEEI-2023-339).

Consent for publication

All authors have read and approved the final manuscript and consent to its publication.