Cluster of Differentiation 44 Promotes Liver Fibrosis and Serves as a Biomarker in Congestive Hepatopathy

Study Approval

All analyses using human samples were approved by the institutional research ethics committees at Saiseikai Yokohamashi Tobu Hospital and the National Center for Global Health and Medicine (NCGM) Kohnodai Hospital (NCGM-G-002339-02, NCGM-G-001967-04, and NCGM-G-003231-00).

All animal experiments were conducted in accordance with institutional guidelines (Guide for the Care and Use of Laboratory Animals, prepared by the National Academy of Sciences), and the protocol was approved by the Animal Research Committee of the NCGM (20055 and 20056).

Patients

Serum samples (n = 15) and liver biopsy specimens (n = 7) were obtained from patients who had had Fontan procedures. Among them, three sets of serum and liver biopsy samples were obtained from the same 3 patients. These patients were not infected with hepatitis B virus or hepatitis C virus (HCV). Additionally, 14 sets of blood samples and liver biopsy specimens were obtained from patients who were HCV positive and who were going to receive antiviral drugs (n = 14). Blood and liver samples were obtained during a single session at our hospital. For obtaining healthy control samples, blood was donated by healthy volunteers (HVs; n = 15) who were officers at the hospitals and did not present with any abnormalities in the annual health check, including blood examination. For the inactive control for pathology, liver biopsy specimens that showed minimal abnormalities were used (n = 8). These included inactive autoimmune hepatitis and complete remission from mild hepatitis. Liver fibrosis was evaluated according to the meta-analysis of histological data in viral hepatitis (METAVIR) F stage and congestive hepatic fibrosis scoring systems.⁽²³⁾ Clinical data were obtained from the electronic clinical records. Clinical characteristics of the patients were determined based on blood samples and liver specimens, which are summarized in Supporting Tables S1-S3. Liver stiffness measurement (LSM) (FibroScan 502; Echosens, Waltham, MA) and virtual touch quantification (VTQ) (ACUSON S3000; Siemens, Erlangen, Germany) were performed only on patients who were HCV positive.

Animals and Treatments

Male C57BL/6J mice were obtained from Japan SLC (Shizuoka, Japan). Transgenic mice expressing enhanced green fluorescent protein (EGFP) under the transcriptional control of the collagen type I α1 (COL1a1) gene promoter (Coll-GFP mice)⁽²⁴⁾ were kindly provided by Dr. David A. Brenner (University of Calfornia, San Diego). The animals were bred for studies in the animal facility of the Research Center for Hepatitis and Immunology (NCGM) and subjected to pIVCL at 8 to 12 weeks of age, as reported.⁽⁷⁾ Animals were anesthetized with medetomidine, midazolam, and butorphanol tartrate. The suprahepatic inferior vena cava was ligated with 6-0 silk and shrunk to 0.6 mm in diameter. Abdominal closure and atipamezole administration were performed on the animals while keeping them warm. After the operation, some mice were treated with the rat anti-mouse CD44 monoclonal antibody IM7 (BE0039; Bio X Cell, Lebanon, NH) or rat anti-human CD44 monoclonal antibody Hermes-1 (BE0262; Bio X Cell), and the effects were compared to immunoglobulin G (IgG) isotype control administration (BE0089; Bio X Cell). The antibodies were injected intravenously (100 μg/mouse) twice a week over the course of 2 weeks after the operation (administered 4 times in total). Other liver fibrosis models analyzed included mice receiving a choline-deficient diet (CDD) (8 weeks) (Research Diets, Inc., New Brunswick, NJ), common bile duct ligation (BDL) (2 weeks), or intraperitoneal CCl₄ injections at 0.5 mL/kg body weight (1:5 volume [vol]/vol in corn oil, twice a week for 8 weeks) (Sigma-Aldrich, St Louis, MO). For induction of liver tumors, diethylnitrosamine (DEN) (Sigma-Aldrich) was dissolved in phosphate-buffered saline (10 g/L) and administered intraperitoneally to male mice at postnatal day 14 (0.2 mg/mouse) and at 10 weeks of age (2 mg/mouse). For the DEN-only model, the animals were euthanized 1 year after the final DEN administration. For the DEN + CCl₄ model, DEN-administered mice were intraperitoneally injected with CCl₄ at 0.5 mL/kg body weight (1:9 vol/vol in corn oil twice a week over the course of 12 weeks) (Sigma-Aldrich), starting at 12 weeks of age. NASH-mediated liver tumors were induced by feeding a CDD to 5-week-old male animals for 9 months. The animals were euthanized by exsanguination. The liver was immediately removed, and a part of the dissected tissue was frozen in liquid nitrogen, fixed with 10% formalin, or embedded in optical cutting temperature compound (Sakura Finetek Japan, Tokyo, Japan) for histologic analysis.

Other Experimental Procedures

Other experimental procedures are described in the Supporting Experimental Procedures. These include histologic analysis, proteome analysis of exosomes, CD44 measurement, hydroxyproline measurement, quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR), exosome isolation, western blotting, HA measurement, and statistical analysis.

CD44 is Increased in the Serum and Liver of Patients With FALD

Proteome analysis was performed on serum-derived exosomes to identify biomarkers of liver fibrosis in patients with CH. Those from 3 patients after the Fontan procedure were compared with those from 3 HVs (Supporting Table S4; Supporting Fig. S1A). Although several factors were significantly changed, pathway analysis revealed no relation between these factors (Supporting Fig. S1B). Thus, we focused on CD44, the top hit in the analysis. CD44 was significantly increased in patient-derived exosomes, and a fragment of the intracellular domain of CD44 was identified by proteome analysis (Supporting Fig. S1C). In addition, serum levels of soluble CD44 were also increased in patients after the Fontan procedure and in patients who were HCV positive (Fig. 1A). Because the enzyme-linked immunosorbent assay (ELISA) kit allowed for detection of the extracellular domain of CD44, the shedding and release of CD44 was also found to be increased in these patients. In the liver, the expression of CD44 was increased in nonparenchymal cells after the Fontan procedure, mainly at perivenous sites, and in patients who were HCV positive, mainly at periportal sites, compared to specimens obtained from the controls (Fig. 1B,C; Supporting Fig. S2). These results suggest that the increase in serum levels of CD44 reflects the enhanced expression of CD44 in the liver. The number of CD44-positive cells in the livers of patients after the Fontan procedure seemed to be associated with fibrosis scores according to METAVIR F stage and congestive hepatic fibrosis scoring systems (Fig. 1D). In contrast, well-known liver fibrosis markers in the blood (HA, type IV collagen 7s, and Mac-2 binding protein glycosylation isomer [M2BPGi]) did not correlate with the fibrosis score (Supporting Fig. S3). Because we were not able to analyze a sufficient number of paired blood samples and liver biopsy specimens of patients with FALD collected at the same time (Supporting Fig. S4A), we examined the relationship between serum CD44 levels and liver fibrosis using samples from patients who were HCV positive and found that the levels were positively correlated with CD44-positive cells in the livers of patients who were HCV positive (Supporting Fig. S4B). Sirius red staining revealed collagen deposits in the livers of these patients (Supporting Fig. S4C), and the CD44 level correlated with the F stage and sirius red-positive area (Supporting Fig. S4D). In addition, LSM and VTQ values also increased with soluble CD44 levels (Supporting Fig. S4D). These results suggest that CD44 could serve as a biomarker for liver fibrosis in patients.

Anti-CD44 Antibody Treatment Inhibits Liver Fibrosis Induced by Chronic Liver Congestion in Mice

We performed pIVCL on mice to investigate the mechanisms of liver fibrosis in chronic liver congestion. Liver congestion and fibrosis were observed 6 weeks after the procedure, as indicated by hematoxylin and eosin and sirius red staining as well as by hydroxyproline measurements, and progressed chronologically (Fig. 2A). Infiltration of inflammatory cells, increase of CD3-positive T cells and F4/80-positive macrophages, hepatocellular damage, and alanine aminotransferase (ALT) elevation were inconspicuous compared to the CDD-induced NASH model (Fig. 2A; Supporting Fig. S5A), as reported.⁽⁷⁾ An increase in α-smooth muscle actin (α-SMA), a marker of HSC activation, was already detectable 2 weeks after the operation, accompanied by increased CD44 expression in nonparenchymal cells mainly at perivenous sites (Fig. 2A-C). The apparent molecular weight of CD44 was approximately 80 kDa, according to western blot analysis (Fig. 2C), suggesting that the induced CD44 was of the standard form. Similar to the observations made for samples from patients with FALD, levels of CD44 in the plasma and exosomes were increased by pIVCL (Fig. 2D). CD44 was also found to be increased in the liver and plasma of mice reflecting other conditions, including BDL, CDD, and CCl₄-induced liver fibrosis models (Supporting Fig. S5B). Among the liver tumor models, an increase in CD44 expression was observed in nonparenchymal cells of CDD and DEN + CCl₄ models and was accompanied by liver fibrosis. In contrast, CD44 expression was not detectable in the DEN-only model, which does not cause liver fibrosis (Supporting Fig. S5C). These results suggest that CD44 is involved in liver fibrosis, regardless of its etiology, and that CD44 is barely expressed in hepatocytes or hepatoma cells at steady state. As with the expression of CD44, increased accumulation of HA was observed in the liver after pIVCL (Fig. 3A). In addition, plasma HA levels were also slightly increased (Fig. 3B). It has been reported that anti-CD44 antibodies prevent lung collagen accumulation in a bleomycin-induced pulmonary fibrosis mouse model.⁽¹⁷⁾ To explore the role of CD44 in liver fibrosis, we administrated monoclonal antibodies against CD44, clones IM7 and Hermes-1, which recognize different epitopes⁽²⁵⁾ to the pIVCL mice. The blockade of CD44 by IM7 inhibited liver fibrosis (Fig. 3C,D), suggesting that the signaling molecules downstream of CD44 are involved in fibrogenesis during chronic liver congestion. Plasma levels of ALT and CD3 messenger RNA (mRNA) expression in the livers were not increased by pIVCL and were not affected by IM7 and Hermes-1 (Fig. 3E), suggesting that the antifibrotic effect of IM7 is not mediated by the reduction of liver injury or immune responses.

CD44 Expression is Increased in Activated HSCs in the Congestive Liver

As observed in the patients who were HCV positive, plasma CD44 had a positive correlation with liver fibrosis examined by hydroxyproline measurement in the pIVCL mice (Fig. 4A). mRNA expression of CD44 was correlated with that of HSC activation factors (Fig. 4B), whereas no correlation was found with endothelial cell marker CD31 or T-lymphocyte marker CD3 (Supporting Fig. S5C), suggesting that CD44 is involved in the activation of HSCs. Indeed, the CD44-specific signal colocalized with that of α-SMA (Fig. 4C) and collagen-expressing cells (Fig. 4D), suggesting that CD44 expression is increased in activated HSCs in congestive livers. In addition, the GFP intensity of the exosomes isolated from the Coll-GFP mice was increased in response to pIVCL (Fig. 4E). Thus, the release of exosomes from activated HSCs may have been increased in these pIVCL mice.

We also examined the relationship between CD44 and S100A4 in order to investigate the mechanisms of CD44-induced liver fibrosis. S100A4 expression was increased in the livers of pIVCL mice (Fig. 5A), which further correlated with collagen expression (Fig. 5C). The induction of S100A4 in pIVCL mice was reduced by CD44 blockade with IM7 (Fig. 5B,D). The expression of CD44 mRNA was positively correlated with that of S100A4 (Fig. 5E). In the pIVCL liver, CD44 expression overlapped with that of S100A4 (Fig. 5F). These results suggest that CD44 is involved in S100A4 expression to activate HSCs.