Jin-Yong Zhuo , , # , Di Lu , , # , Zu-Yun Lin , , Bei-Ni Cen , Xu-Yong Wei , , Hi-Yng Xie ,Shu-Sen Zheng , , c , Xio Xu , , ?

a Department of Hepatobiliary and Pancreatic Surgery, the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310 0 03, China

b NHC Key Laboratory of Combined Multi-organ Transplantation, Key Laboratory of the Diagnosis and Treatment of Organ Transplantation, CAMS, Hangzhou 310 0 03, China

c Department of Hepatobiliary and Pancreatic Surgery, Shulan (Hangzhou) Hospital, Hangzhou 310 0 03, China

Keywords:

ABSTRACT

Introduction

Liver cirrhosis results from chronic liver damage, including hepatitis B virus (HBV) infection, hepatitis C virus (HCV) infection,alcohol abuse, and non-alcoholic steatohepatitis (NASH) [1 , 2].In the development of liver cirrhosis, extracellular matrix (ECM) and abnormal inflammation play an important role [3].The accumulation of ECM subsequently results in hepatic insufficiency, portal hypertension, and increases the risk for hepatocellular carcinoma(HCC) [4].

Conventional laboratory tests such as alanine aminotransferase(ALT) are used to assess liver inflammation activity.Nevertheless, they have poor capability to predict progression of cirrhosis alone [5].Imaging examination is recommended for the early diagnosis of liver cirrhosis.Considering the expensive economic costs and the uncertainty of early screening, high frequency of imagine examination is infeasible for each individual, especially in rural areas.Liver biopsy is the golden criterion for diagnosis of cirrhosis.However, its clinical practice is limited by the invasiveness, risk of complications and lack of patient acceptance [6 , 7].

Gene Expression Omnibus (GEO) database has been widespread applied in hepatitis, liver fibrosis, and liver cirrhosis studies [8-10].To identify distinguished biomarkers in liver cirrhosis, this study analyzed gene expression profile from GEO database (GSE15654)with 216 cirrhotic patients [11].We aimed to find a sensitivebiomarker that can indicate the occurrence and progression of liver cirrhosis and to investigate the mechanism involved.

Methods

GEO database analysis

We searched GEO database by keywords “l(fā)iver cirrhosis”and found out GSE15654 with 216 early stage (Child-Pugh class A) liver cirrhotic patients.Expression matrix profile and microarray platform information of GSE15654 were downloaded by R package“GEOquery”.According to whether Child-Pugh class progression(from class A to B or C) during the whole follow-up period, 216 patients were divided into Child stable group (n= 150) and Child progression group (n= 66).Differential genes were identified using R, version 3.4.0 ( https://www.R-project.org/ ).The expression of a specific gene was divided into the high mRNA expression group and low mRNA expression group according to the median value of mRNA.The survival of Child-Pugh progression, defined as the time of class A progressing to class B or C, and the overall survival were analyzed based on the follow-up time in the GEO dataset.

Patients and samples preparation

A total of 181 plasma specimens were collected between September 2015 and May 2017 at the First Affiliated Hospital,Zhejiang University School of Medicine, including 20 healthy control (HC), 77 chronic hepatitis B (CHB) and 84 liver cirrhosis(LC).Plasma samples were stored at ?80 °C.Seven liver cirrhotic tissue samples were collected.This study was approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine.

Cell culture

Human normal liver Cell line (LO2) and human hepatic stellate cell line (LX-2) were purchased from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China), and were cultured at 5% CO 2 and 37 °C in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine plasma(Gibco, Waltham, USA) and 1% penicillin/streptomycin.

Enzyme linked immunosorbent assay (ELISA)

C -C motif chemokine ligand 16 (CCL16) level in plasma and supernatant was determined using ELISA kits (ab100532, Abcam,Cambridge, USA) according to the manufacturer’s protocol.The absorbance intensity was read at 450 nm by microplate reader(BioTek, Winooski, USA).

Western blotting

Cells after different treatments were harvested.The proteins were separated with SDS-PAGE and transferred to polyvinylidene difluoride membranes.The membranes were incubated at 4 °C overnight with the primary antibody with anti-CCL16 (1:50 0 0,ab134917, Abcam), anti-HIF-1α(1:50 0, NB10 0-105, Novus, Littleton, USA), anti-Collagen I (1:10 0 0, ab34710, Abcam), anti-Collagen IV (1:10 0 0, ab6586, Abcam), and anti-GAPDH (1:50 0 0, EM1101,HuaAn, Hangzhou, China).The bands were visualized by chemiluminescence using Image Lab (Bio-Rad, Hercules, USA).

Immunohistochemistry

Immunohistochemistry (IHC) was stained with anti-CCL16(1:50, MAB328, R&D, Minneapolis, USA) and anti-HIF-1α(1:50,NB100-105, Novus, Littleton, USA) antibodies.Immunostaining of CCL16 and HIF-1αwas semi-quantitatively scored as previously described [12].Shortly, the intensity of staining was scored as absent(0), weakly positive (1), moderately positive (2), or strongly positive (3).The area was scored as no positive area (0), less than 10%of positive area (1), 11%-50% of positive area (2), 51%-75% of positive area (3), and more than 75% of positive area (4).

Liver cirrhosis model

Four-week-old Balb/c mice were tail vein injected with CCL16 adeno-associated virus (AAV-CCL16) or AAV only (AAV-control)(3.75 ×1011 vg per mouse,n= 5).After 6 weeks, all mice were administrated with 25% carbon tetrachloride (CCl 4 ) twice a week for another 6 weeks.Then mice were sacrificed and liver tissues were harvested.Animal studies were performed according to the guidelines approved by the First Affiliated Hospital, Zhejiang University School of Medicine.

Statistical analysis

All statistical analyses were performed with GraphPad Prism 6.0(GraphPad Software, La Jolla, California, USA) or R ( www.r-project.org ).Two-way Student’st-test was used for parametric variables,Mann-WhitneyUtest was used for non-parametric variables, and Chi-square test was used for categorical variables.Receiver operating characteristic (ROC) curve was used to determine the predictive value.Kaplan-Meier method was used for survival analysis.APvalue less than 0.05 was considered statistically significant.

Results

GEO dataset identified CCL16 as a potential biomarker of liver cirrhosis

We first analyzed the GEO dataset GSE15654 [11].Child-Pugh class is able to distinguish advanced-stage cirrhosis from earlystage cirrhosis [13].Two hundred and sixteen individuals were divided into the Child stable group (n= 150) and Child progression group (n= 66) based on whether Child-Pugh class A progressing to B or C after more than two-decade follow-up.Then we obtained 158 differentially expressed genes, includingCCL16( Fig.1 A), and based on relative expression of matrix,CCL16expression in the Child progression group was lower than that in the Child stable group (P<0.0 0 01, Fig.1 B).

Two hundred and sixteen cirrhotic patients were divided into the lowCCL16group (n= 107) and the highCCL16group (n= 109)according toCCL16relative expression.The data showed that the rate of Child-Pugh progression in the lowCCL16group was higher than that in the highCCL16group (P= 0.034, Fig.1 C), and the overall survival of the lowCCL16group was poorer than that of the highCCL16group (P= 0.025, Fig.1 D).

Plasma CCL16 distinguished the different stage of liver diseases

To validate whether CCL16 was associated with cirrhotic development, the plasma level of CCL16 in 181 samples (20 HC, 77 CHB and 84 LC) were evaluated by ELISA.The plasma CCL16 median concentrations in the HC, CHB and LC groups were 10.29, 6.57 and 4.47 ng/mL, respectively.The plasma CCL16 in the LC group was significantly lower than that in the HC group (P<0.001).And there was a statistical difference in plasma CCL16 level between the CHB and HC groups (P= 0.010, Fig.2 A).

We found that the ROC curve showed a significant discriminatory ability between the CHB and HC groups (AUC = 0.68,P= 0.013,Fig.2 B).And the ROC curve showed a significant discriminatoryability between the LC and HC groups (AUC = 0.79,P<0.001,Fig.2 C).The ROC curves of CCL16, alanine aminotransferase (ALT),aspartate aminotransferase (AST), prothrombin time (PT) and albumin (ALB) to distinguish the CHB and LC were 0.580, 0.583,0.4 84, 0.4 83 and 0.560, respectively.Multivariate logistic analysis showed that CCL16 and ALT were the two independent factors for the LC group.Furthermore, a model combining with CCL16 and ALT showed preferable capability (AUC = 0.679,P<0.0 0 01, Fig.2 D)to distinguish the CHB and LC groups comparing to CCL16 alone(AUC = 0.580,P= 0.081).

Fig.1.GEO datasets analysis identified CCL16.A: Heatmap of differentially expression genes according to Child-Pugh class, Child stable group ( n = 150) and Child progression group ( n = 66); B: Relative expression of CCL16 between in Child stable and Child progression groups; C: Cumulative survival of Child-Pugh progression; D: Overall survival in the low ( n = 107) and high ( n = 109) CCL16 groups.CCL16: C -C motif chemokine ligand 16.

In the CHB group, CCL16 concentration was inversely related to ALT, AST, PT and model for end-stage liver disease (MELD) score( Fig.3 A).The high CCL16 subgroup and the low CCL16 subgroup showed significant difference in liver function indexes, including ALT, AST, PT, MELD and indirect bilirubin ( Table 1 ).However, in the LC group CCL16 concentration was positively related to ALB,but inversely related to ALT, PT and MELD score ( Fig.3 B).The high CCL16 and low CCL16 subgroups showed significant difference in hyaluronic acid (HA), ALT, AST, PT, MELD scores, total bilirubin and direct bilirubin, most importantly, in ALB ( Table 2 ).

Table 1Clinical characteristic in the chronic hepatitis B group.

CCL16 was regulated by lipopolysaccharide (LPS) stimulation and hypoxia inducible factor 1 alpha (HIF-1 α)

To examine whether CCL16 was regulated by inflammation factor, we stimulated the cell line LO2 with LPS with differentconcentration (0, 0.5, 1, and 2 μg/mL) for 48 h.CCL16 expression decreased in the LPS concentration-dependent pattern ( Fig.4 A).

Fig.2.Plasma CCL16 level in clinical patient.A: Different plasma CCL16 level in HC ( n = 20), CHB ( n = 77) and LC ( n = 84); B: ROC curve between the CHB and the HC group; C: ROC curve between the LC and the HC group; D: ROC curves compare the capability to distinguish the CHB and the LC group by alone or combination of CCL16 and ALT.HC: healthy control; CHB: chronical hepatitis B; LC: liver cirrhosis; ROC: receiver operating characteristic; CCL16: C -C motif chemokine ligand 16; ALT: alanine aminotransferase.

Table 2Clinical characteristic in the liver cirrhosis group.

Furthermore, to test whether CCL16 was regulated by hypoxia,LO2 was treated with hypoxia environment (1% O 2 ).CCL16 expression was decreased in LO2 with the exposure time of hypoxia.Overexpression of HIF-1αin the normoxia had no influence on the expression of CCL16, however CCL16 was obviously decreased and HIF-1αwas overexpression in the hypoxia ( Fig.4 B).Our IHC in the 7 liver cirrhotic tissues showed that CCL16 expression was negatively related to HIF-1α(P= 0.018, Fig.4 C, D).

CCL16 inhibited collagen production from hepatic stellate cells

Collagen is deposited in the progression of cirrhotic liver, due to activation of hepatic stellate cells (HSC) [14].Overexpression of CCL16 in LX2 resulted in downregulation of collagen-related markers, Col1 and Col4 ( Fig.5 A).CCL16 overexpressed alone in LO2 had no influence to the expression of Col1 and Col4 compared with negative control ( Fig.5 C).In order to evaluate the regulation of CCL16 from hepatocytes to hepatic stellate cells, we established aninvitroculture system, mixing LO2 and LX2 with different cell number ratios (LO2/LX2, 6:1 and 2:1).In the mixed culture, the level of CCL16 sharply elevated in the supernatant( Fig.5 B).And Col1 and Col4 were significantly downregulated in 6:1 ratio ( Fig.5 C,D).

Fig.3.Plasma CCL16 was associated with liver function in the CHB ( n = 77) and LC groups ( n = 84).A: Plasma CCL16 was adversely related to ALT, AST, prothrombin time and MELD in the CHB group; B: Plasma CCL16 was positively related to albumin, adversely related to ALT, prothrombin time and MELD in the LC group.CCL16: C -C motif chemokine ligand 16; CHB: chronical hepatitis B; LC: liver cirrhosis; ALT: alanine aminotransferase; AST: aspartate aminotransferase; MELD: model for end-stage liver disease score.

Fig.4.CCL16 was regulated by LPS and HIF-1 α.A: LO2 was treated with 0, 0.5, 1 and 2 μg/mL LPS for 48 h; B: LO2 with HIF-1 α overexpression or control were exposed to hypoxia or normoxia for 24 h; C: Sections of liver cirrhosis tissues were stained with CCL16 and HIF-1 α, respectively (original magnification ×100 and ×400); D: Visualization of relative expression of CCL16, according to HIF-1 αexpression in liver cirrhosis tissues; E: Schematic diagram of CCL16 regulating collagen production.Hypoxia and inflammation inhibited CCL16 expression and secretion.Decrease of CCL16 secreted from hepatocytes promoted the Col1 and Col4 production in hepatic stellate cells,furthermore, increased the extracellular matrix.CCL16: C -C motif chemokine ligand 16; LPS: lipopolysaccharide; HIF-1 α: hypoxia inducible factor 1 alpha.

CCL16 inhibited liver cirrhosis in CCl 4 -induced animal model

We established a hepatic cirrhosis model by intraperitoneal injection of CCl 4 for 6 weeks in mice after tail vein injection of AAV-CCL16 or AAV-control.Sirius Red and Masson staining showed that the degree of fibrosis was significantly decreased in the AAV-CCL16 group compared to the AAV-controls( Fig.5 E).

Discussion

Liver cirrhosis results from an accumulation of ECM in all background of chronic liver injuries [3 , 15 , 16].HBV infection is one of the most common chronic damages to promote inflammation activity and cirrhosis progression.In China, HBV accounts for a large proportion of liver disease and is the most vital risk factor for cirrhosis and liver cancer [17 , 18].And the incidence of non-HBV associated cirrhosis, including NASH, is rising in China as well [19].Despite that several important biomarkers have been identified to relate to liver cirrhosis in East-Asian population, such as taurocholic acid [20], disorganized gut microbiome [21]and Lachnospiraceae in faeces [22]; more markers need to be identified to promote the sensitivity and specificity for the detection and prognostic prediction of cirrhosis.The present study showed that CCL16 was associated with hepatic dysfunction both in patients with chronic hepatitis B and liver cirrhosis.CCL16 had the discriminatory ability for liver cirrhosis.CCL16 could predict the prognosis of patients with cirrhosis, and inhibit the progression of liver cirrhosis.

Fig.5.CCL16 reduced collagen production in HSC and inhibited cirrhosis development in CCl 4 -induced animal model.A: Western blotting detected Col1 and Col4 levels in LX2 expressing CCL16 or negative control; B: ELISA detected CCL16 levels in the supernatant in LO2 mix-cultured with or without LX2; C: Western blotting detected Col1 and Col4 levels in LX2 mixed without or with LO2 in varying proportions (2:1 or 6:1); D: Quantification of Col1 and Col4 levels in LX2 mixed without or with LO2; E:Representative images of HE staining, Sirius Red staining, and Masson staining showing fibrosis degree in mouse liver sections, n = 5 per group, magnification was 100×for HE, 400× for Sirius Red staining and Masson staining.Values are mean ±SD.?: P< 0.05 vs.Control group.n = 3 per group.HSC: hepatic stellate cell; HE: hematoxylin-eosin staining; SD: standard deviation; ELISA: enzyme linked immunosorbent assay; CCL16: C -C motif chemokine ligand 16.

Our study analyzed GEO dataset (GSE15654), and found that CCL16 had significantly distinguished expression in the individuals with different Child-Pugh score status.Unfortunately, we could not access more objective indexes to evaluate cirrhotic classification,such as MELD, which may limit the power of the study to a certain extent.The data showed that low CCL16 expression was significantly associated with poor prognosis in cirrhotic patients.CCL16 is also known as liver-expressed chemokine (LEC), and was reported to specifically express in normal liver and present in the healthy plasma with high concentrations [23 , 24].Therefore, we assumed that CCL16 would work as a biomarker reflecting liver function.Our data showed that the plasma level of CCL16 was negatively related to ALT, AST, PT and MELD score in patients with chronic HBV hepatitis.The plasma level of CCL16 in liver cirrhosis was significantly declined compared with healthy control.CCL16 was positively related to albumin, and negatively related to hyaluronic acid, ALT, PT and MELD in cirrhotic patients.CCL16 combined with ALT showed more robust capability to distinguish the CHB and LC groups than CCL16 alone.These data indicated that CCL16 was significantly associated with hepatic dysfunction both in the CHB and LC groups.

Chemokines and chemokine-receptors have been described to play a vital role in liver cirrhosis [25 , 26].CCL16 is a member of CC chemokine subfamily, and is located on chromosome 17q in the C -C chemokines cluster [27].CCL16 is uniquely upregulated by IL-10, which plays an anti-inflammation role [24].Therefore, we assumed that CCL16 might be an anti-inflammation chemokine.Once liver was injured, damaged hepatocyte decreased CCL16 expression.

To test whether CCL16 would be regulated by inflammation stimulation, LO2 cell line was treated with LPS.CCL16 expression gradually decreased after the treatment with different concentration of LPS (0, 0.5, 1, and 2 μg/mL), which suggests that the inflammation may reduce the expression of CCL16.Furthermore,cirrhotic liver is surrounded by the hypoxia environment where the hypoxia-marked protein HIF-1αelevates concomitantly [28 , 29].In order to investigate whether HIF-1αregulated CCL16 expression, LO2 overexpressed with HIF-1αplasmid or control were exposed to normoxia and hypoxia environment.Our data showed that CCL16 expression was significantly decreased in hypoxia environment with HIF-1αoverexpression.Altogether, CCL16 might be regulated by inflammation stimulation and hypoxia.

Liver cirrhosis is the surplus accumulation of extracellular matrix containing collagen.Activated hepatic stellate cells is one of the major collagen-producing cells in the chronic liver diseases [1].Therefore, it is needed to explore whether CCL16 is able to impact hepatic stellate cells.Our data showed that overexpression of CCL16 in LX-2 inhibited the expression of Col1 and Col4.Our mice model also confirmed that CCL16 inhibited the pathological features of cirrhosis.Therefore, we draw the schematic diagram of CCL16 regulating collagen production to illustrate our results( Fig.4 E).

CCL16 induces a pathway that interacts with chemokine receptors on the cell membrane CCR1, CCR2, CCR5 and CCR8 [22 , 26].CCR1 or CCR5 could activate HSC and promote collagen production at liver cirrhosis [30 , 31].We assumed that CCL16 binding to CCR1 and CCR5 on the surface of HSC might block the signature transmit at the quiescent state.When liver was injured, damaged hepatocytes decreased CCL16 expression, meanwhile, CCR1 and CCR5 signature were activated.At the point, activated HSC produced and secreted collagen, which further promoted fibrogenesis.However,it needs further validation in the future.

In conclusion, this study demonstrated that CCL16 could be a potential plasma marker to predict the prognosis of patients with liver cirrhosis.Plasma CCL16 is decreased in chronic hepatitis B and cirrhotic patients, and is significantly associated with hepatic dysfunction.Combination of CCL16 and ALT has preferable capability to distinguish liver cirrhosis from chronic hepatitis B.CCL16 inhibits the development of liver cirrhosis through inactivating hepatic stellate cells.

Acknowledgments

We thank Dr.Bao-Hong Wang, Dr.Jian-Guo Wang, Dr.Run-Zhou Zhuang, Dr.Fang Yang, Dr.Mo-Dan Yang, and Mr.Zheng-Xing Lian for their excellent and considerable supports to this research.

CRediT authorship contribution statement

Jian-Yong Zhuo:Data curation, Formal analysis, Resources,Writing-original draft.Di Lu:Data curation, Writing-original draft.Zu-Yuan Lin:Data curation, Software.Bei-Ni Cen:Project administration.Xu-Yong Wei:Investigation, Supervision.Hai-Yang Xie:Supervision, Validation.Shu-Sen Zheng:Investigation, Writingreview & editing.Xiao Xu:Conceptualization, Funding acquisition,Writing-review & editing.

Funding

This work was supported by grants from the National Science and Technology Major Project of China [ 2017ZX10203205 ], the National Natural Science Funds for Distinguished Young Scholar of China [ 81625003 ], Projects of Medical and Health Technology Program in Zhejiang Province [ WKJ-ZJ-1514 ], China Postdoctoral Science Foundation [ 2017M612014 ]and Zhejiang Medical and Technological Program [ 2018263185 ].

Ethical approval

This study was approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University School of Medicine.

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.