Hepatic cell proliferation plays a pivotal role in the prognosis of alcoholic hepatitis

Research Article

Hepatic cell proliferation plays a pivotal role in the prognosis of alcoholic hepatitis Nicolas Lanthier1,3, Laura Rubbia-Brandt2, Nathalie Lin-Marq2, Sophie Clément2, Jean-Louis Frossard1, Nicolas Goossens1, Antoine Hadengue1, Laurent Spahr1,⇑ 1

Gastroenterology and Hepatology, University Hospitals and Faculty of Medicine, Geneva, Switzerland; 2Clinical Pathology, University Hospitals and Faculty of Medicine, Geneva, Switzerland; 3Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique, Université catholique de Louvain, Brussels, Belgium

Background & Aims: The role of liver progenitor cell (LPC) expansion, known as a marker of disease severity, as well as the impact of macrophage activation on liver regeneration remains unclear in humans. We aimed to characterize the LPC and macrophage compartments in alcoholic hepatitis (AH), as well as gene expression patterns to identify predictors of a good prognosis in this setting. Methods: Immunohistochemical studies for macrophages, proliferative hepatocytes, total and proliferative LPC, as well as whole liver microarray gene expression were performed on baseline liver biopsies of 58 AH patients early after admission. Abstinent cirrhotic patients were used as controls. Patients were qualified as ‘‘improvers’’ or ‘‘non-improvers’’ based on the change in MELD score three months after baseline. Results: Compared to controls, AH patients demonstrated a significant expansion of macrophages, invasion of LPC and a higher number of proliferating hepatocytes and LPC. In AH patients, total LPC expansion (total Keratin7+ cells) was associated with liver disease severity. The group of improvers (n = 34) was characterized at baseline by a higher number of proliferating hepatocytes, proliferative LPC (double Keratin7+Ki67+ cells) and liver macrophages as compared to non-improvers (n = 24), despite similar

Keywords: Alcoholic hepatitis; Regeneration; Liver progenitor cell; Macrophage; Liver biopsy; SPINK1. Received 22 July 2014; received in revised form 10 March 2015; accepted 1 April 2015; available online 11 April 2015 ⇑ Corresponding author. Address: Gastroenterology and Hepatology, University Hospitals of Geneva, 4, Rue Gabrielle Perret-Gentil, CH-1211 Geneva 14, Switzerland. Tel.: +41 223729340; fax: +41 223729366. E-mail address: [email protected] (L. Spahr). Abbreviations: AH, alcoholic hepatitis; ALD, alcoholic liver disease; ASH, alcoholic steatohepatitis; ACLF, acute on chronic liver failure; LPC, liver progenitor cell; K7, cytokeratin 7; KRT7, cytokeratin 7 (gene expression); SCT, stem cell transplantation; TNFa, tumor necrosis factor-a; sTNFR1, soluble form of receptor-1 for TNFa; IL-6, interleukin-6; AFP, alfa-foetoprotein; HGF, hepatocyte growth factor; TGFb, transforming growth factor beta; IL-22, interleukin-22; TWEAK, TNF-like weak inducer of apoptosis; SDF-1, stromal-cell-derived-factor-1; IL-10, interleukin-10; CDE, choline-deficient diet supplemented by ethionin; DR, ductular reaction; iPC, intermediate liver progenitor cells; IH, intermediate hepatocytes; CCNB1, cyclin B1; Cdk1, cyclin-dependent kinase 1; CDC20, cell division cycle 20; RRM2, ribonucleotide reductase M2; Fn14, fibroblast growth factor inducible 14; SPINK1, serine peptidase inhibitor Kazal type 1.

clinical and biological variables. Upregulated genes in improvers were associated with cell cycle mitosis together with a major expression of SPINK1. Conclusions: Higher liver macrophage expansion, increased proliferative hepatocyte but also LPC number, as well as an upregulation of cell proliferation-related genes are associated with a favourable outcome. These new findings open novel therapeutic targets in AH. Ó 2015 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Introduction Alcoholic hepatitis (AH) is a frequent and severe complication of alcoholic liver disease (ALD) that occurs in approximately 20–30% of heavy drinkers, and is associated with a high morbidity and mortality [1,2]. In Western Europe and in the US, this acute inflammatory liver condition typically occurs in patients with chronic, excessive alcohol intake, and clinically manifests as decompensated cirrhosis. AH is recognized as a clinical syndrome which includes recent jaundice and manifestations of liver failure in a patient with excessive alcohol consumption. In this context, when one suspects histological alcoholic steatohepatitis (ASH) in a severe form based on a prognostic model, such as the Maddrey’s score or the MELD score, a liver biopsy is performed to demonstrate the presence of steatosis, and ballooning degeneration of hepatocytes in association with polymorphonuclear neutrophils parenchymal infiltration [1]. Failure to do so may expose a patient with decompensated liver disease secondary to another cause to the adverse effects of steroids (the treatment indicated in severe ASH [1]) without clinical benefit. Whether liver histology brings prognostic information in the setting of AH remains ill-defined. A liver biopsy performed after decompensation in patients with alcoholic liver disease showed no single histological feature that was predictive of mortality [3,4]. In addition to severity of cholestasis and degree of fibrosis assessed early after hospital admission (<1 week) [4,5], presence of hepatic encephalopathy [6] and elevated circulatory white blood cells [7] that have been associated with a poor outcome in AH, impaired liver regeneration also appears to play a pivotal role in patients with cirrhosis of various etiologies [8]. Exposure to alcohol has an additional negative

Journal of Hepatology 2015 vol. 63 j 609–621

Research Article effect on liver cell proliferation [9,10]. Accordingly, hepatocyte proliferative activity declines with the severity of cirrhosis [8], and a superimposed acute liver injury, a situation named «acute on chronic liver failure (ACLF)» [11], is associated with expansion of liver progenitor cell (LPC) compartment, presumably in an attempt to restore hepatocellular function [12]. Indeed, the LPC potential transformation into hepatocytes has been described in animal tracing studies [13,14]. In contrast, other animal studies support the exclusive role of pre-existing hepatocyte replication for liver parenchymal regeneration [15,16]. The plasticity of hepatocytes and their possible dedifferentiation into LPC is also described [17]. Nevertheless, in humans, while LPC expansion is known as a marker of disease severity [18], the putative role of these cells in liver function improvement and better clinical outcome remains at present unclear [19]. The proliferative response of this intrahepatic stem cell population [20] and its interactions with other cell types of the hepatic microenvironment, such as macrophages [21] is a focus of major interest and a potential target for new treatment strategies including cytokine [22,23], macrophage [24] or stem cell [25] therapy. Therefore, the aim of this study was to analyse the LPC proliferation and macrophage activation in patients with AH, in relation to clinical treatment. Changes in global hepatic gene expression profile were determined in order to evaluate potential pathways associated with a good prognosis.

Patients and methods Patients The clinical and biological material used for this study was derived from a recent trial on stem cell transplantation (SCT) in decompensated ALD [19]. Shortly after hospital admission, 58 patients with AH were randomized to a standard of care management (alcohol withdrawal, vitamin supplementation and steroids in the presence of histological ASH with Maddrey’s score P32) alone or combined with autologous bone marrow SCT injected into the hepatic artery (Supplementary Table 1). A close follow-up during a three month period did not demonstrate any benefits in terms of biological parameters and histological alterations following SCT [19]. The primary endpoint to define a good prognosis or improvers was a decrease of at least three points of MELD as compared to baseline, which we consider in our experience as clinically significant in the management of these severely ill patients. Thus, based on this criterion, 34 patients with MELD amelioration of three points or more at three months as compared to MELD baseline value were classified as ‘‘improvers’’ and 24 patients with MELD deterioration or MELD amelioration lower than three points at three months or patients who died during this three month follow-up (n = 6) were classified as ‘‘non-improvers’’, irrespective of treatment allocation. These two groups, which constituted our study population, shared similar baseline characteristics in terms of biological values, histological lesions, and treatments including SCT (Supplementary Table 2). Mean MELD scores at three months were 10.8 in improvers vs. 18.8 in non-improvers.

operation for partial hepatectomy or after euthanasia for other conditions, as previously described [26–28]. Experiments on animals were performed with approval of the University Animal Welfare Committee of the Université catholique de Louvain. Immunohistochemistry Liver biopsy specimens from baseline AH patients and from control patients during pretransplant evaluation were immediately fixed into formalin 10%, and then paraffin-embedded. Immuno-stainings were performed using the automated Ventana Discovery XT system (Ventana Medical Systems, Tucson, AZ, USA) and Ventana reagents were used for all procedures. Briefly, serial 3 lm thick to 5 lm thick sections from the tissue block were incubated 30 mins at 37 °C with mouse monoclonal human antibody MIB1 against the proliferation marker Ki67 (Dako, M7240, 1:100 dilution) and anti-Keratin7 (K7) (Dako, M7018, 1:100 dilution), anti-CD68 (Dako, M0876, 1:100 dilution), anti-cyclin D1 (Ventana 790-4508, ready to use rabbit monoclonal antibody SP4-R), anti-Cdk1 (Abcam ab133327 1:100 dilution), and anti-SPINK1 (Abcam, ab58227, 1:500 dilution) primary antibodies. Antigen retrieval was performed by heating slides in CC1 buffer (CC1; Tris-based buffer pH 8.4) for 36 mins for CD68, K7, Ki67, cyclin D1 and Cdk1. Detection of anti-CD68, anti-Ki67, anti-cyclin D1 and anti-Cdk1 primary antibodies was carried out using the amplified DabMap detection kit based on the conversion of diaminobenzidine to a brown dye with horseradish peroxidase (HRP/SA complex using secondary biotinylated antibodies). Detection of anti-K7 primary antibodies was carried out using the amplified RedMap detection kit based on the conversion of Naphtol/FastRed to a red dye with steptavidine Alkalin phosphatase (AP/SA complex using secondary biotinylated antibodies). The Research IHC DoubleStain XT procedure was used for all co-staining’s. Briefly, after heating slides 36 mins in CC1 buffer, anti-Ki67 (or anti-Cyclin D1, or anti-Cdk1) was first incubated and Ki67 was detected using DAB staining. Then HRP activity was heat-inactivated 8 mins at 85 °C and anti-K7 antibodies were added and K7 was detected by RED staining. As negative control, specific isotype antibodies were used instead of primary antibodies when protocols were setting up. Proliferation was assessed on the whole biopsy specimen by manual counting under high magnification (400). Results are given as the mean number of Ki67 positive cells per high power field. Ki67 positive cells were classified either as K7 negative or positive to assess either hepatocyte or LPC proliferation. K7 positive cells were further classified by their histological appearance, localization in the liver lobule and staining intensity of K7. Cells from the ductular reaction (DR) are cuboid adjacent K7 intense positive cells forming ductules. Intermediate progenitor cells (iPC) are K7 positive cells (often isolated) located in the liver lobule and smaller than hepatocytes. Intermediate hepatocytes (IH) are large cuboid hepatocyte-like cells with a less intense K7 staining. Total K7 and CD68 (a marker for liver macrophages) positive cell area were quantified using tissue section photographs at magnification (100) of the entire liver biopsy and the Leica QWin computer assisted program [29]. Results are given as the percentage of total positive cell area (K7 or CD68) on the total biopsy surface at 100 magnification. In situ hybridization The PCR template coding partially for human SPINK1 was generated from human pancreas RNA (AM7954, AMBION) using the reverse transcription kit from Takara (RR037A). Detailed protocol is given as Supplementary materials and methods. SPINK1 in situ hybridization with sense or antisense probes specific for human SPINK1 was performed on serial sections from paraffin-embedded human liver. Human pancreas tissue sections served as positive controls for SPINK1 in situ hybridization.

Liver tissue studies

Transcription analysis

All patients with AH had a transjugular liver biopsy performed at baseline shortly (mean three days) after hospital admission for diagnostic purposes. A second liver biopsy was performed after four weeks of follow-up. Patients in the control group were abstinent alcoholic cirrhotic patients in which repeated blood ethanol tests were negative (Supplementary Table 3). In these patients, liver tissue was obtained during transjugular biopsy at time of evaluation for liver transplantation or during surgery. In all cases, the liver tissue specimens were separated in two pieces, one for histological studies and the other fragment immediately snap frozen in liquid nitrogen and stored at 80 °C for further gene expression analysis. For mice studies, liver tissue was obtained under anesthesia either during surgical

Transcriptome profiling was performed with Affymetrix arrays containing 49,395 transcripts. Total RNA was extracted from snap frozen liver biopsies obtained at baseline (n = 30), as well as in snap frozen liver samples in abstinent cirrhotic patients (n = 12) by using TissueLyser machine (Qiagen) and the AllPrepDNA/RNA Mini Kit (Qiagen, Hombrechtikon, Switzerland). RNA quality was assessed by capillary electrophoresis on Agilent 2100 Bioanalyzer (Agilent Technologies, Basel, Switzerland). 100 ng was amplified and labelled using the 30 IVT Express kit (Affymetrix). Hybridization on GeneChip PrimeView Human Gene Expression arrays (Affymetrix) was carried out according to the manufacturer’s instructions.

610

Journal of Hepatology 2015 vol. 63 j 609–621

JOURNAL OF HEPATOLOGY The data were robust multi array normalized [30]. Pathway analysis of the genes, which were identified as differentially expressed by microarray experiment, was undertaken by using the MetaCore software (http:// www.genego.com). The gene expression data can be found in ArrayExpress (https://www.ebi.ac. uk/arrayexpress/) accession n°E-MTAB-2664.

A

Nanostring

B

Alcoholic hepatitis % CD68

CD68 IHC

8

***

6 4 2 0

***

% K7

K7 IHC

15 10 5 0 7 6 5 4 3 2 *** 1 0

***

***

**

Controls

Ki

Ki

67

D

R

67 iP Ki C 67 IH Ki 67 H ep

Number of cells/field

C K7-Ki67 IHC

For the confirmation of the Affymetrix transcriptional analysis, the same samples – 12 abstinent alcoholic cirrhosis (controls) and 27 cases – were treated with the nCounter system (NanoString). NanoString synthesized the code set for the analysis of 115 genes of interest and seven normalization genes (Supplementary Table 4). 300 ng of total RNA was hybridized with multiplexed NanoString probes and samples were processed according to the published procedure [31]. Background correction was done by subtracting from the raw counts the mean + 2 SD of counts obtained with negative controls (alien probes lacking spiked transcript for background calculation). Values <1 were fixed to 1 to avoid negative values after log transformation. Positive controls (alien sequences of RNA spiked at various concentrations to assess the overall assay performance) were used as quality assessment: the ratio between the highest and the lowest positive controls average among samples was below three. Then counts for target genes were normalized with the geometric mean of the four reference genes (ACTB, B2M, EEF1A1 and GUSB) selected as the most stable using the geNorm algorithm [32].

Controls

Alcoholic hepatitis (AH)

Measurement of cytokines and blood markers of regeneration A multiplex suspension array technique based on the sandwich immunoassay method (Bioplex, Biorad, Zürich, Switzerland) was performed for multiple measurements on a single plasma sample, as previously described [19]. The following cytokines were measured at different time points: inflammatory cytokines tumor necrosis factor-a (TNFa), soluble form of receptor-1 for TNFa (sTNFR1) and interleukin-6 (IL-6); proliferative cytokine alpha-fetoprotein (AFP), hepatocyte growth factor (HGF), transforming growth factor b (TGFb), interleukin-22 (IL-22), TNF-like weak inducer of apoptosis (TWEAK), stromal-cell-derived-factor-1 (SDF-1) and interleukin-10 (IL-10). Animal studies Different mouse models of injury were analysed: C57BL/6J wild-type mice were fed either a modified Lieber-DeCarli diet for 20 days [26] or a choline-deficient diet supplemented by ethionin (CDE) for 10 days [28]. Chow-fed mice were subjected to a two-third partial hepatectomy and livers were analysed 48 h after surgery [27].

Fig. 1. Macrophages and K7 positive cells are activated in alcoholic hepatitis compared to abstinent alcoholic cirrhosis. Liver sections obtained from AH patients (n = 58) and control patients (n = 15) were stained with antibodies directed against macrophages (CD68 in brown) to identify Kupffer cells (A), or directed against liver progenitor cells (K7 in pink) (B) by immunohistochemistry (IHC). Morphometric quantification of the area occupied by positive cells is presented. Double K7 (in pink) and Ki67 (in brown) (C) was used to identify proliferating hepatocytes and proliferating liver progenitor cells (LPC). Double K7/ Ki67 positive cells were classified and manually counted as cells from the ductular reaction (DR), intermediate progenitor cells (iPC) or intermediate hepatocyte-like cells (IH). K7 negative Ki67 positive cells with a typical hepatocyte morphology (large cuboid cells) were classified as proliferative hepatocytes (Hep) (C). Inserts show typical 400 magnification on each cell subtype. All data are mean (± S.D.), **p <0.01, ***p <0.001 between groups. (This figure appears in colour on the web.)

Continuous variables are presented as means ± SD and were compared using Student t test (two-tails) or paired if appropriate. Categorical variables were compared using the chi square test. Correlations between variables were evaluated using Pearson’s r. To assess the differences in gene expression values (Affymetrix and Nanostring) between the different groups (AH at baseline vs. controls, improvers vs. non-improvers at baseline), we did a 5-way ANOVA with contrast in Partek Genomics Suite (http://www.partek.com). We applied a conservative significance threshold of 5% p value.

to controls (Fig. 1B). Taken together, these results emphasize the association between Kupffer cell and LPC expansion in AH, as previously reported in experimental chronic liver injury [33]. Double immunostaining for K7 and Ki67 was used to quantify proliferating LPC (double K7+/Ki67+ cells) or proliferating hepatocytes (large Ki67+, K7 cells). Proliferating LPC were manually counted in the entire liver biopsy (mean 22.6 range 6–56 high power fields per case) differentiating cells from the DR, iPC and IH by morphology and localization in the liver lobule. Patients with AH demonstrate a higher number of proliferative hepatocytes and K7 positive cells regardless of cell subtype, compared with controls (Fig. 1C).

Results

Different hepatic gene expression profile according to the presence of AH

Statistical analysis

Expansion of macrophage and LPC compartments in AH patients vs. controls Morphometric analysis of liver sections immunostained for CD68 demonstrated a significant expansion of CD68 positive cells in the entire group of patients with AH as compared to abstinent cirrhotic patients serving as controls (Fig. 1A). In addition, patients with AH were characterized by a significantly higher total K7 liver positivity calculated by morphometric analysis compared

To analyse whether those histological changes were associated with particular inflammatory pathways or specific cell function, transcriptome expression profile was assessed in 30 patients with AH and 12 controls. This analysis revealed 753 genes >2-fold differentially expressed with a p value lower than 0.05 between those two groups (Fig. 2). These results were confirmed by NanoString technology (Table 1). The most relevant genes differentially expressed in patients with AH as compared to controls were related with metabolism, fibrogenesis, cell cycle processes, LPC

Journal of Hepatology 2015 vol. 63 j 609–621

611

Research Article Controls

Alcoholic hepatitis OSGIN1 AXIN1 NUMB CCNA2 CDC20 NCAPG CDKN3 CDK1 PRC1 CCNB1 RRM2 RRM2 TOP2A CCNE1 ABCB4 SPINK1 S100P CD68 LPL PDZK1IP1 CD24 KRT7 NQO1 LCN2 MUC20 SAA1 /// SAA2 SOD2 SPP1 CA12 TRNP1 AKR1B15 AKR1B10 CD109 APOA4 COL3A1 COL1A2 COL5A1 COL1A1 COL1A1 COL1A1 PDGFA ACTN1 TGFBI IFIT2 H6PD H6PD H6PD CKAP4 TRAF3 MLXIP KIAA0100 MTHFR FZD4 KIAA1147 EIF4EBP2 RXRA NDST1 AGPAT3 AGPAT3 HIPK2 MLEC H6PD HIPK2 TMCO3 KLF13 RPS6KA2 HKDC1 TNFSF14 TNFSF14 CRIPAK DGKZ IGFBP2 CXCL10 CCL2 IL8 CXCL3 IGHG3 /// IGHM CXCL14 IGFBP3 IGFBP3 IRS2 AGTR1 NFIL3 ELL2 FOXO1 DUSP1 DUSP1 SGK1 HIF1A HIF1A OAT CXCR7 FOS EGR1 EGR2 CYR61 JUN SOCS3 CSRNP1 GADD45B ATF3 LDLR IFI16 IL6 SLC2A14 /// SLC2A3 NR4A3 RGS2 NR4A2 NR4A1 ADAMTS4 ADAMTS1 APOLD1 PTPRC CXCR4 CXCR4 IL7R CD69 CCL19 ID4 ITGBL1 VCAM1 MYH11 CTGF CXCL12 HGF HSPA2 GAS1 F3 SOCS2 PRKCH CYP26A1 CYP7A1 LPA HEPACAM ADH4 ADH6 BHMT RDH16 CYP2C8 SC5DL F9 CYP2B6 CYP2C9 FETUB GSTA2 GSTA1 CYP3A7 CYP3A7 CYP3A7 PCK1 PCK1 PCK1 PPARGC1A G6PC CYP2E1 CYP1A1 CYP1A2 CD5L

Fig. 2. Heatmap showing hierarchical clustering analysis based on genes differentially expressed in alcoholic liver disease between control abstinent alcoholic liver cirrhosis (n = 12) and alcoholic hepatitis (n = 30). The intensity of the color (red: upregulation; green: downregulation) indicates differential expression. (This figure appears in colour on the web.)

(KRT7, SOX9, CD24) and other functions. However, classical immune response pathways derived from animal studies were unchanged (TNF, IL-6, cJUN) or significantly decreased (SOCS3, FOS) at the transcriptome level in our patients (Table 1 and Fig. 2). Relationship between K7 positive cell expansion and the MELD score We explored the relationship between K7 positive cellular compartment and the severity of liver failure in AH patients. The immunohistochemical staining for K7 correlated with liver failure as assessed by the MELD score and the INR (Fig. 3A and C) and at four week follow-up biopsy (Fig. 3B and D), thereby 612

confirming previously published results about LPC activation and expansion in severe liver damage [18,34–38]. Histological baseline parameters according to outcome We compared baseline histological features in the groups of improvers and non-improvers to evaluate the role of LPC expansion as a prognostic determinant at three months. Except for a higher proportion of females in the improvers vs. non-improvers, groups were similar at baseline (Supplementary Table 2). The expression of total K7 was similar in the two groups (Fig. 4A). To analyse the relationship between parenchymal hepatocyte or LPC proliferation and evolution of the MELD score at three months, we manually counted proliferative K7 positive cells (K7+/Ki67+ cells) in different subtypes of LPC (DR, iPC or IH) as well as proliferative hepatocytes (Ki67+/K7 parenchymal cells) in the entire liver biopsy specimen. Interestingly, improvers were characterized at baseline by a higher number of K7+/Ki67+ cells of the DR and intermediate hepatocytes and Ki67+ hepatocytes as compared to non-improvers (Fig. 4B). Mitosis was also present in some K7+ cells (Supplementary Fig. 1). Considering the apparent important role of both macrophages and LPC in the proliferative and regenerative response to injury [33], we compared liver macrophage expansion in improvers vs. non-improvers. Improvers displayed an increased density of liver macrophages (4.4% of surface area) as compared with non-improvers (3.3% of surface area) (Fig. 4C). Collectively, those data demonstrate at baseline liver biopsy a higher liver macrophage expansion together with a higher number of proliferative hepatocytes and K7 positive cells in improvers as compared to non-improvers. Transcription analysis at baseline according to outcome Provided that expansion or proliferation of particular liver cell compartments observed at baseline influences the short-term prognosis, we wanted to determine whether these histological features were associated with particular inflammatory and/or regenerative pathways. Transcriptome analysis on liver tissue revealed that the pathways most upregulated in improvers vs. non-improvers were related to cell cycle (Fig. 5 and Table 2). NanoString technology confirmed this significant link between liver function improvement and upregulation of the cell cycle process with a significant upregulation of mRNA coding for proteins like cyclin B1 (CCNB1), cyclin-dependent kinase 1 (Cdk1), cell division cycle 20 (CDC20), ribonucleotide reductase M2 (RRM2). Immunohistochemistry analysis of cyclin D1 and Cdk1 revealed enhanced proliferation in both hepatocytes and K7 positive cells (Supplementary Fig. 1). Other significant pathways that were upregulated in improvers compared with non-improvers were associated with fibrosis, inflammation, macrophages and hepatic microenvironment including Wnt signalling and fibroblast growth factor inducible 14 (Fn14 or TNFRSF12A), the receptor for TWEAK, a well-known potential activator of LPC proliferation. These results, confirmed by NanoString data, are presented in Table 2. Finally, in patients who qualified as improvers, the transcriptome analysis demonstrated a seven fold increase of the liver serine peptidase inhibitor Kazal type I (SPINK1). Although unexpected, upregulation of SPINK1 was confirmed by NanoString with a 82.92-fold increase (p = 0.0076) (Table 2).

Journal of Hepatology 2015 vol. 63 j 609–621

JOURNAL OF HEPATOLOGY Table 1. Different expression of genes in patients with alcoholic hepatitis compared with controls.

Accession number Metabolism NM_000670

NM_145740 NM_000767

NM_000773

NM_000770 /// NM_001198853 /// NM_001198854 /// NM_001198855 NM_000771

NM_005577 NM_000482 NM_020299

NM_000208 /// NM_001079817 NM_002591

NM_000151

NM_004285

Inflammation NM_001040059 /// NM_001251 NM_000582 /// NM_001040058 /// NM_001040060 NM_003300 /// NM_145725 /// NM_145726 NM_001295 NM_003807 /// NM_172014

NM_016639

NM_000594 NM_003955

Gene name

Affymetrix fold expression

p value

Nanostring fold expression

p value

ADH4 alcohol dehydrogenase 4 (class II), pi polypeptide GSTA1 glutathione S-transferase alpha 1 CYP2B6 cytochrome P450, family 2, subfamily B, polypeptide 6 CYP2E1 cytochrome P450, family 2, subfamily E, polypeptide 1 CYP2C8 cytochrome P450, family 2, subfamily C, polypeptide 8

-2.88

0.0013

-4.40

0.0004

-4.53

0.0004

-12.4

0.0292

-3.11

0.0001

-8.28

0.0006

-2.56

0.0240

-3.26

0.0473

-4.20

0.0001

-4.77

0.0002

-2.79

0.0013

-2.31

0.0012

-2.67

0.0120

-38.45

0.0075

8.24

0.0001

89.86

0.0018

3.85

0.0001

3.76

0.0005

1.95

0.002

1.66

0.0001

-2.21

0.0005

-2.19

0.0026

-2.24

0.0004

-3.52

0.0000

4.05

0.0000

13.62

0.0000

1.61

0.0007

9.21

0.0074

2.58

0.0027

32.48

0.0012

3.02

0.0000

8.05

0.0264

1.03

0.7673

1.52

0.5392

2.42

0.0004

17.79

0.0227

1.4

0.0759

11.71

0.0265

1.07

0.0894

1.10

0.7310

-5.31

0.0000

-27.37

0.0034

CYP2C9 cytochrome P450, family 2, subfamily C, polypeptide 9 LPA lipoprotein, Lp(a) APOA4 apolipoprotein A-IV AKR1B10 aldo-keto reductase family 1, member B10 (aldose reductase) INSR insulin receptor PCK1 phosphoenolpyruvate carboxykinase 1 (soluble) G6PC glucose-6-phosphatase, catalytic subunit H6PD hexose-6-phosphate dehydrogenase (glucose 1-dehydrogenase) CD68 CD68 molecule SPP1 secreted phosphoprotein 1 (osteopontin) TRAF3 TNF receptor-associated factor 3 CCR1 chemokine (C-C motif) receptor 1 TNFSF14 tumor necrosis factor (ligand) superfamily, member 14 TNFRSF12A tumor necrosis factor receptor superfamily, member 12A TNF tumor necrosis factor SOCS3 suppressor of cytokine signaling 3

(continued on next page)

Line missing Journal of Hepatology 2015 vol. 63 j 609–621

613

Research Article Table 1 (continued)

Accession number Inflammation NM_005252

Gene name

FOS FBJ murine osteosarcoma viral oncogene homolog NM_004887 CXCL14 chemokine (C-X-C motif) ligand 14 NM_001100812 /// CXCL16 NM_022059 chemokine (C-X-C motif) ligand 16 ²NM_000636 /// SOD2 NM_001024465 /// superoxide dismutase 2, NM_001024466 mitochondrial Extracellular matrix/cellular adhesion NM_001102 /// ACTN1 NM_001130004 /// actinin, alpha 1 NM_001130005 NM_001002857 /// ANXA2 NM_001002858 /// annexin A2 NM_001136015 /// NM_004039 NM_000088 COL1A1 collagen, type I, alpha 1 NM_000089 COL1A2 collagen, type I, alpha 2 NM_000093 COL5A1 collagen, type V, alpha 1 NM_004932 CDH6 cadherin 6, type 2, K-cadherin (fetal kidney) NM_001565 CXCL10 chemokine (C-X-C motif) ligand 10 NM_005764 PDZK1 PDZK1 interacting protein 1 NM_000358 TGFBI transforming growth factor, betainduced, 68 kDa NM_001136528 /// SERPIN2 NM_001136530 /// serpin peptidase inhibitor, clade E NM_006216 (nexin, plasminogen activator inhibitor type 1), member 2 Liver stem/progenitor cells NM_005556 KRT7 keratin 7 NM_000224 /// NM_199187 KRT18 keratin 18 NM_002276 KRT19 Keratin 19 NM_002354 EPCAM epithelial cell adhesion molecule NM_013230 CD24 CD24 molecule NM_000346 SOX9 SRY (sex determining region Y)-box 9

Affymetrix fold expression

p value

Nanostring fold expression

p value

-13.05

0.0000

-363.79

0.0000

-2.75

0.0023

-9.34

0.0371

1.24

0.0019

11.31

0.0158

2.07

0.0016

1.76

0.0030

1.43

0.0015

2.70

0.0547

1.91

0.0001

17.88

0.0024

3.19

0.0003

61.87

0.0001

3.42

0.0003

7.98

0.0062

2.21

0.0032

15.66

0.0140

1.11

0.5321

8.95

0.0419

1.74

0.0276

14.72

0.0179

3.34

0.0008

65.13

0.0007

1.75

0.0008

1.64

0.0067

1.59

0.0179

10.67

0.0144

1.75

0.0301

31.95

0.0032

1.25

0.0716

1.56

0.0012

1.10

0.6899

8.88

0.0362

1.15

0.6794

4.76

0.2006

2.63

0.0001

10.16

0.0014

1.42

0.0194

3.90

0.0407

Line missing

614

Journal of Hepatology 2015 vol. 63 j 609–621

JOURNAL OF HEPATOLOGY Table 1 (continued)

Accession number Cell cycle NM_001170406 /// NM_001170407 /// NM_001786 /// NM_033379 NM_001255 NM_031966 NM_004701 NM_005764 NM_001034 /// NM_001165931 NM_015675

Wnt signalling NM_003502 /// NM_181050 NM_012193 NM_003881

NM_004626

Notch signalling NM_024408 Others NM_003122

Gene name

Affymetrix fold expression

p value

Nanostring fold expression

p value

CDK1 cyclin-dependent kinase 1

1.63

0.0131

2.34

0.1579

CDC20 cell division cycle 20 homolog CCNB1 Cyclin B1 CCNB2 Cyclin B2 PDZK1 PDZK1 interacting protein 1 RRM2 ribonucleotide reductase M2 GADD45B growth arrest and DNA-damageinducible, beta

1.38

0.0310

2.80

0.0923

1.61

0.0048

5.54

0.0142

1.37

0.0310

1.24

0.5547

3.34

0.0008

65.13

0.0007

2.6

0.0058

11.45

0.0097

-5.96

0.0000

-9.61

0.0006

AXIN1 Axin 1 FZD4 frizzled homolog 4 WISP2 WNT1 inducible signaling pathway protein 2 WNT11 wingless-type MMTV integration site family, member 11

1.15

0.0486

7.16

0.0175

2.77

0.0000

321.75

0.0000

1.14

0.6010

11.06

0.0365

1.27

0.1378

2.96

0.0964

NOTCH2 Notch 2

1.22

0.0007

5.10

0.0011

0.028

19.52

0.0291

SPINK1 3.68 serine peptidase inhibitor, Kazal type 1

Serum values at baseline according to outcome Improvers were characterized at baseline by a higher level of serum HGF, a higher serum level of sTNFR1 and a lower level of SDF-1, when compared to non-improvers (Fig. 4D). Other cytokines remained unchanged between the two groups (IL-10, IL-6, TGFb) (Fig. 4D), TWEAK, IL-22, TNFa, AFP (data not shown). SPINK1: a possible link between inflammation and proliferation in ALD In order to explore the major, yet unexpected increase in SPINK1 expression in liver tissue of improvers, we searched for the protein on liver sections. Despite its presence on exocrine pancreas tissue serving as positive control (data not shown), we were unable to show the presence of SPINK1 protein by immunohistochemistry in the liver. However, SPINK1 mRNA was detected on liver and pancreatic sections using in situ hybridization (Fig. 6A and D). Serial sections analysis showed that SPINK1-mRNA is detected in parenchymal cells (hepatocytes or K7 positive cells with proliferating activity) and co-localized with

areas of inflammatory CD68 macrophage infiltration (Fig. 6A). Furthermore, SPINK1 mRNA appeared to correlate with macrophage marker CD68 and cell cycle markers like Cdk1 and cyclin B1 (Fig. 6B). A 20-fold increase in liver SPINK3 expression, the mouse homolog of human SPINK1 [39], was detected by PCR following partial hepatectomy in mice (Fig. 6C), together with hepatocyte proliferation and macrophage activation (data not shown). SPINK3 was also upregulated in the CDE model in mice, an experimental condition known to induce LPC expansion by blocking hepatocyte proliferation (Fig. 6C). There was no significant increase in SPINK3 in ethanol-fed mice (Fig. 6C).

Discussion Patients with AH manifest as an acute or sub-acute deterioration of liver function in the setting of cirrhosis, the so-called ACLF [11], which is a life threatening condition with high mortality [1,2]. Although steroids improve short-term survival in severe, biopsy proven ASH, a significant proportion of patients remain

Journal of Hepatology 2015 vol. 63 j 609–621

615

Research Article Table 2. Different expression of genes in improvers vs. non-improvers in alcoholic hepatitis.

Accession number Cell cycle NM_001170406 /// NM_001170407 /// NM_001786 /// NM_033379 NM_001255 NM_031966 NM_053056 NM_002592.2 NM_001034 /// NM_001165931 NM_001130851 /// NM_005192 Fibrosis NM_000088

Gene name

Affymetrix fold expression

p value

Nanostring fold expression

p value

CDK1 cyclin-dependent kinase 1

2.07

0.0018

5.47

0.0221

CDC20 cell division cycle 20 homolog CCNB1 Cyclin B1 CCND1 Cyclin D1 PCNA proliferating cell nuclear antigen RRM2 ribonucleotide reductase M2 CDKN3 cyclin-dependent kinase inhibitor 3

1.58

0.0080

7.85

0.0076

1.82

0.0019

30.69

0.0003

1.43

0.02

/

/

1.31

0.0463

/

/

3.6

0.0014

8.85

0.0369

1.69

0.0076

4.93

0.0223

2.41

0.0023

5.05

0.0909

2.42

0.0082

2.04

0.3589

1.55

0.0193

1.70

0.0106

2.18

0.0171

2.02

0.5412

1.47

0.0342

3.30

0.1685

1.27

0.0198

21.08

0.0055

1.24

0.0227

9.90

0.0342

1.73

0.0136

25.07

0.0142

1.40

0.1298

27.60

0.0234

1.93

0.0156

36.01

0.0026

-3.28

0.0058

-24.51

0.0186

-3.43

0.0002

-15.40

0.0002

-2.15

0.0050

-2.20

0.0055

COL1A1 collagen, type I, alpha 1 NM_000089 COL1A2 collagen, type I, alpha 2 NM_000358 TGFBI transforming growth factor, betainduced, 68kDa Cell adhesion, macrophages and chemotaxis NM_004887 CXCL14 chemokine (C-X-C motif) ligand 14 NM_001040059 /// CD68 NM_001251 CD68 molecule NM_003842 /// TNFRSF10B NM_147187 /// tumor necrosis factor receptor NR_027140 superfamily, member 10B NM_001066 TNFRSF1B tumor necrosis factor receptor superfamily, member 1B NM_016639 TNFRSF12A tumor necrosis factor receptor superfamily, member 12A NM_003807 /// TNFSF14 NM_172014 tumor necrosis factor (ligand) superfamily, member 14 NM_001136528 /// SERPIN2 NM_001136530 /// serpin peptidase inhibitor, clade E NM_006216 (nexin, plasminogen activator inhibitor type 1), member 2 Ethanol metabolism NM_145740 GSTA1 glutathione S-transferase alpha 1 NM_000767 CYP2B6 cytochrome P450, family 2, subfamily B, polypeptide 6 NM_000771 CYP2C9 cytochrome P450, family 2, subfamily C, polypeptide 9

Line missing

616

Journal of Hepatology 2015 vol. 63 j 609–621

JOURNAL OF HEPATOLOGY Table 2 (continued)

Accession number Wnt signalling NM_003502 /// NM_181050 NM_012193 NM_003882 /// NM_080838 NM_004626

Others NM_003122 NM_002354

Affymetrix fold expression

p value

Nanostring fold expression

p value

AXIN1 Axin 1 FZD4 Frizzled homolog 4 (Drosophilia) WISP1 WNT1 inducible signaling pathway protein 1 WNT11 wingless-type MMTV integration site family, member 11

1.16

0.0496

32.86

0.0010

2.77

0.0000

321.75

0.0000

1.23

0.0102

4.16

0.0686

1.85

0.025

5.18

0.0354

SPINK1 serine peptidase inhibitor, Kazal type 1 EPCAM epithelial cell adhesion molecule

6.84

0.0053

82.92

0.0076

-2.54

0.0185

-7.73

0.1513

B

Baseline

35

30

30

25

25

20 15 10

20 15 10

R = 0.33 p = 0.011

5 0 0

10

20

30

R = 0.61 p <0.001

5 0 40

K7 (%)

C

0

10

3.5

3.0

3.0

2.5

2.5

2.0 1.5

20

30

K7 (%)

D

3.5

INR

INR

Week 4

35

MELD

MELD

A

Gene name

2.0 1.5 1.0

1.0

R = 0.49 p <0.001

0.5

R = 0.48 p <0.001

0.5 0.0

0.0 0

10

20

K7 (%)

30

40

0

10

20

30

K7 (%)

Fig. 3. Correlation between cytokeratin 7 (K7) expression surface on the entire liver biopsy with MELD at baseline or week 0 (A) and week 4 (B) of followup. The same correlation was found between liver cytokeratin 7 expression and INR both at week 0 (C) and week 4 (D) of follow-up.

exposed to liver related complications due to incomplete liver function recovery. This state of extreme vulnerability is in part associated with insufficient liver regeneration [40] and repair processes, resulting from several factors including the presence of parenchymal inflammation and cell injury, recent exposition to alcohol [41] and replicative senescence of hepatocytes in the setting of advanced cirrhosis [8]. For these reasons, pro-regenerative strategies including the use of growth factors alone [22,23] or in combination with bone marrow SCT [19] have been tested and associated with a significant [22] or no [19,23] improvement in clinical and biological endpoints. In addition, efforts have been made to better predict the prognosis of AH, either by using biological [42], imaging [43] or histological [4]

parameters. Accordingly, histological severity of cholestasis [4,5] and percentage of hepatocyte loss [44] have been associated with a poor outcome. However, the intensity of lobular inflammatory infiltration has been associated with improved short-term survival [4]. Here, we first determined histological and gene expression changes in AH patients vs. alcoholic cirrhosis. These results are in line with previous studies comparing AH and control liver showing an upregulation in some inflammatory, metabolic and profibrotic genes in AH [45]. In addition, we present here histological data on CD68 expansion in AH together with K7 and K7+Ki67+ infiltration in this situation compared with abstinent cirrhosis. Secondly, we determine for the first time a different short-term outcome according to baseline histological features and hepatic gene expression profile in AH. According to our analysis, improvers demonstrate a significant expansion of liver macrophages and a high number not only of proliferating hepatocytes but also of proliferating LPC on liver tissue obtained shortly after admission. We characterized improvers as patients alive with a MELD decrease of 3 points or more at three months, which is for us a meaningful surrogate marker of outcome [19]. Other baseline and dynamic models are currently under investigation [46]. When applying another more restrictive criterion (90-day mortality) as evaluated in another study [18], we also found a significant increase in proliferative LPC and hepatocyte number in patients alive compared to patients who died during this three month follow-up (Supplementary Table 5). Hepatic gene expression profile also differentiates improvers from non-improvers. A significant upregulation of genes involved in cell cycle, mitosis and proliferation pathways, as well as a high expression of SPINK1 was found in AH patients as compared to controls. In addition, the level of expression was significantly higher in improvers as compared to non-improvers. Initially named tumor-associated trypsin inhibitor or pancreatic secretory trypsin inhibitor, SPINK1 is expressed in many cancers including hepatocellular carcinoma and functional study reveals that overexpression of this liver acute-phase reactant protein [47,48] is associated with cell proliferation [49]. High SPINK1 expression was observed after

Journal of Hepatology 2015 vol. 63 j 609–621

617

Research Article

A

Non-improvers Improvers

EIF1AY RPS4Y1 DDX3Y CFHR3 UNC93A LNX1 ENPP2 ENPP2 SH3YL1 SH3YL1 EPCAM ESRP1 CPE CPE TPSAB18///8TPSB2 COL15A1 RBP7 VWF SPARCL1 SPARCL1 SPARCL1 HIGD1B GPR88 CYP2B7P1 CYP2B7P1 BBOX1 GSTA2 GSTA1 LPA ZG16 PGLYRP2 CYP2B6 CYP2B6 CYP2B6 GNMT GLS2 INSIG1 SLC38A4 RDH16 CYP2C8 CYP2C8 CYP2C8 TF PON3 DIRAS3 SLC7A2 CYP2C9 SLCO4C1 HSD17B13 HSD17B13 CYP39A1 CYP39A1 SC4MOL HPX HAO2 HAO2 CYP2C9 C19orf77 PRSS3 IGFBP1 NOTCH2 PHLDB2 IGFALS CCNE1 CDK1 CDKN3 CDC20 CCNA2 TPX2 PTTG18///8PTTG3P CCNB1 PRC1 ANLN ASPM NCAPG UBE2C TOP2A RRM2 KIAA0101 ZWINT UBE2C AURKA CCNB2 CCNE2 AURKB TNFRSF1B MIR21 CCR1 CD68 WNT11 FCN2 TNFRSF12A SLC16A12 FABP3 MORC4 ADAMTS17 TNFRSF10B CXCL14 TGFBI WISP1 ACTB ABCB4 CLDN2 SPINK1 NUMBL SERPINE2 COL1A1 COL1A1 COL1A2 COL5A1 AXIN1 NUMB XIST

% K7

K7 IHC

20 15 10 5 0 9 8 7 6 5 4 3 2 *** 1 0

**

67 R iP Ki C 67 Ki I 67 H H ep

*

Ki

Non-improvers

Ki

Improvers

67

D

K7-Ki67 IHC

Number of cells/field

B

C

*

6

% CD68

CD68 IHC

8 4 2 0

Serum HGF

Serum TGFβ

*

5000

300 200

0

0 Serum SDF-1 6000

30

pg/ml

pg/ml

5

2000

Serum IL-10

20 10

4000

0 Serum IL-6

20 10 0

***

6000

100

0

15

8000

pg/ml

10,000

Serum STNFR1

400

pg/ml

pg/ml

15,000

pg/ml

D

**

4000 2000 0

Fig. 4. Improvers are characterized by a higher number of proliferating hepatocytes and liver progenitor cells, as well as macrophage expansion. Liver sections obtained from AH patients (improvers and non-improvers) were stained with double antibodies against K7 (in pink) and Ki67 (in brown) (A and B) to identify liver progenitor cells (LPC) and proliferative cells by immunohistochemistry (IHC). Antibody against macrophages (CD68 in brown) (C) identifies Kupffer cells. Morphometric quantification of the area occupied by K7 positive cells (A) and CD68 positive cells (C) are presented. Double K7-Ki67 positive cells were classified and manually counted as cells from the ductular reaction (DR), intermediate progenitor cells (iPC) or intermediate hepatocyte-like cells (IH). Ki67 + cells with a typical hepatocyte morphology (large cuboid cells) were classified as proliferative hepatocytes (Hep) (B). Inserts show typical 400 magnification on each cell subtype. All data are mean (± S.D.) for n = 28–30/group. Serum values (D) of HGF, TGFb, STNFR1, IL-6, IL-10 and SDF-1 in AH disease at baseline in improver vs. non-improver patients. *p <0.05, **p <0.01, ***p <0.001 between groups. (This figure appears in colour on the web.)

partial hepatectomy in mice, concurrent with hepatocyte proliferation and Kupffer cell activation. Consistent with the upregulation of Fn14 reported in AH [50], we observed a significantly higher gene expression in improvers compared to non-improvers. The interaction between Fn14 and TWEAK is important to promote LPC expansion and liver regeneration, as reported in animal studies [24,51]. However, the stable TWEAK serum level in our patients prevents its use as a potential biomarker. Overall, extensive information gathered from immunohistochemistry, liver gene expression profiling analysis and serum cytokine levels enabled us to separate subgroups of patients with AH and distinct proliferation and regenerative capacities. These 618

Improvers

Non-improvers

Fig. 5. Heatmap showing hierarchical clustering analysis based on genes differentially expressed at baseline in alcoholic hepatitis between nonimprovers (n = 13) and improvers (n = 17). The intensity of the color (red: upregulation; green: downregulation) indicates differential expression. (This figure appears in colour on the web.)

characteristics translate into different outcomes in terms of liver function. Our histological results are in line with previous data in alcoholic [18,35,36], metabolic [36–38] or viral [34] liver disease, which showed that high liver total K7 expression parallel disease severity as assessed by the MELD score. Indeed, LPC are known to expand from the DR, the intrahepatic stem cell compartment, into the liver lobule when mature hepatocytes are not able to bring full regeneration following severe injury [52], particularly in the setting of ACLF [12]. However, here, we observe that baseline expansion of Ki67+/K7+ cellular compartment is associated with a favourable outcome, consistent with the contribution of proliferating LPC in the regenerative and repair processes. Indeed, some patients from this improvers group present mostly

Journal of Hepatology 2015 vol. 63 j 609–621

JOURNAL OF HEPATOLOGY

A

K7-Ki67 IHC

R = 0.479 p = 0.006

Relative liver mRNA expression

C

R = 0.470 p = 0.006

8 Cdk1

SPINK1

12 10 8 6 4 2

6 4 2

6

35

8

10 CD68

12

0

Liver SPINK in mice

30 25 20 15 10 5 0

2

7 SPINK1

12

R = 0.487 p = 0.004

8 6 4 2

2

D

7 SPINK1

12

Financial support Clinical Research Center, University Hospital and Faculty of Medicine, Geneva, the Louis-Jeantet Foundation, FLAGS (Foundation for Liver and Gut Studies) in Geneva.

*

Human pancreas

Control ***

The strength of this study is represented by a large patient population with AH with a thorough description of clinical, biological and histological characteristics measured early after hospital admission, extensive hepatic gene expression profile, and careful follow-up, allowing comparisons between subgroups with different clinical evolution. Although we acknowledge that our definition of clinical improvement can be challenged, we believe it is a meaningful surrogate marker of outcome paving the way for larger studies that will be required to study survival outcomes. In conclusion, this large study in patients with AH and liver failure identifies LPC proliferation, hepatocyte proliferation and liver macrophages expansion on baseline liver biopsy as important markers of favourable short-term outcome, together with upregulation of hepatic gene involved in several pathways including cell proliferation. These data may help to identify patients with poor regenerative capacities earlier and allow to target more specifically pro-regenerative pathways in acutely decompensated ALD.

SPINK1 ISH

Cyclin B1

B

CD68 IHC

Alcohol diet

Conflict of interest

CDE diet Control with PH

Fig. 6. SPINK mRNA expression, inflammation and proliferation. (A) Expression of K7 (pink) – Ki67 (brown) by immunohistochemical staining (IHC), expression of CD68 (brown) by IHC staining and Spink1 mRNA expression (blue) by in situ hybridization (ISH) in serial liver sections of improver AH patients. (B) Pearson correlation (and respective p value) between the expression patterns of SPINK1 and CD68, Cyclin-dependent kinase 1 (Cdk1) or cyclin B1 as assessed by Affymetrix analysis. (C) Liver SPINK3 mRNA expression in mice from various experimental conditions: ethanol-fed mice for 20 days, choline-deficient diet (CDE) for ten days and two days after partial hepatectomy. *p <0.05, *** p <0.001 vs. controls (n = 5–8/group). (D) SPINK1 mRNA expression on exocrine human pancreatic tissue served as positive control. (This figure appears in colour on the web.)

a high number of proliferative K7 cells, with a low number of proliferative hepatocytes. Furthermore, MELD amelioration at three months better correlates with baseline proliferative K7 positive cells than with baseline proliferative hepatocytes (see graphs in Supplementary Fig. 2). The fact that improvers with high proliferating K7+ cells do not present any increase in total K7+ cells in comparison with non-improvers also suggests the putative further differentiation of those proliferating K7+ cells into hepatocytes (K7). This proliferative capacity of a subset of K7 LPC was not analysed in previous studies [18,34,37,38]. The presence of infiltrating macrophages in an injured liver is not an unexpected finding in an inflammatory microenvironment, as macrophages participates both to the pro-inflammatory response and the repair mechanisms [53]. Accordingly, expansion of liver macrophages closely correlate with LPC response in experimental chronic liver injury [33]. In our patients, a high number of CD68 positive macrophages (Kupffer cells) coexisted with proliferation of LPC and resulted in a favourable outcome.

The authors who have taken part in this study declared that they do not have any conflict of interest with respect to this manuscript.

Author’s contributions NL (acquisition, analysis and interpretation of data, drafting of the manuscript), LRB (analysis and interpretation of data; critical revision of the manuscript), NLM (acquisition, analysis and interpretation of data; critical revision of the manuscript), SC (acquisition of data, analysis and interpretation of data, critical revision of the manuscript), JLF (acquisition of data, analysis and interpretation of data), NG (critical revision of the manuscript for important intellectual content), AH (study concept and design, obtained funding, interpretation of data), LS (study concept and design, acquisition and interpretation of data, drafting of the manuscript, obtained funding, study supervision).

Acknowledgments The authors are very grateful to the iGE3 Genomics Platform of the University of Geneva for the transcriptome and subsequent statistical analyses, as well as for the discussions during the study. The authors also thank Valérie Lebrun and Noémi Van Hul from the Université catholique de Louvain for their collaboration and the technical support in the mouse studies, as well as Isabelle Leclercq for critical review of the manuscript.

Journal of Hepatology 2015 vol. 63 j 609–621

619

Research Article Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2015.04. 003.

References [1] EASL guidelines. EASL clinical practical guidelines: management of alcoholic liver disease. J Hepatol 2012;57:399–420. [2] Lucey MR, Mathurin P, Morgan TR. Alcoholic hepatitis. N Engl J Med 2009;360:2758–2769. [3] Elphick DA, Dube AK, McFarlane E, Jones J, Gleeson D. Spectrum of liver histology in presumed decompensated alcoholic liver disease. Am J Gastroenterol 2007;102:780–788. [4] Altamirano J, Miquel R, Katoonizadeh A, Abraldes JG, Duarte-Rojo A, Louvet A, et al. A histologic scoring system for prognosis of patients with alcoholic hepatitis. Gastroenterology 2014;146:1231–1239. [5] Spahr L, Rubbia-Brandt L, Genevay M, Hadengue A, Giostra E. Early liver biopsy, intraparenchymal cholestasis, and prognosis in patients with alcoholic steatohepatitis. BMC Gastroenterol 2011;11:115. [6] Jepsen P, Ott P, Andersen PK, Sorensen HT, Vilstrup H. Clinical course of alcoholic liver cirrhosis: a Danish population-based cohort study. Hepatology 2010;51:1675–1682. [7] Forrest EH, Evans CD, Stewart S, Phillips M, Oo YH, McAvoy NC, et al. Analysis of factors predictive of mortality in alcoholic hepatitis and derivation and validation of the Glasgow alcoholic hepatitis score. Gut 2005;54:1174–1179. [8] Delhaye M, Louis H, Degraef C, Le MO, Deviere J, Peny MO, et al. Hepatocyte proliferative activity in human liver cirrhosis. J Hepatol 1999;30:461–471. [9] Clot P, Tabone M, Arico S, Albano E. Monitoring oxidative damage in patients with liver cirrhosis and different daily alcohol intake. Gut 1994;35: 1637–1643. [10] Horiguchi N, Ishac EJ, Gao B. Liver regeneration is suppressed in alcoholic cirrhosis: correlation with decreased STAT3 activation. Alcohol 2007;41: 271–280. [11] Moreau R, Jalan R, Gines P, Pavesi M, Angeli P, Cordoba J, et al. Acute-onchronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology 2013;144:1437. [12] Rastogi A, Bihari C, Maiwall R, Ahuja A, Sharma MK, Kumar A, et al. Hepatic stellate cells are involved in the pathogenesis of acute-on-chronic liver failure (ACLF). Virchows Arch 2012;461:393–398. [13] Espanol-Suner R, Carpentier R, Van Hul N, Legry V, Achouri Y, Cordi S, et al. Liver progenitor cells yield functional hepatocytes in response to chronic liver injury in mice. Gastroenterology 2012;143:1564–1575. [14] Furuyama K, Kawaguchi Y, Akiyama H, Horiguchi M, Kodama S, Kuhara T, et al. Continuous cell supply from a Sox9-expressing progenitor zone in adult liver, exocrine pancreas and intestine. Nat Genet 2011;43:34–41. [15] Yanger K, Knigin D, Zong Y, Maggs L, Gu G, Akiyama H, et al. Adult hepatocytes are generated by self-duplication rather than stem cell differentiation. Cell Stem Cell 2014;15:340–349. [16] Schaub JR, Malato Y, Gormond C, Willenbring H. Evidence against a stem cell origin of new hepatocytes in a common mouse model of chronic liver injury. Cell Rep 2014;8:933–939. [17] Yanger K, Zong Y, Maggs LR, Shapira SN, Maddipati R, Aiello NM, et al. Robust cellular reprogramming occurs spontaneously during liver regeneration. Genes Dev 2013;27:719–724. [18] Sancho-Bru P, Altamirano J, Rodrigo-Torres D, Coll M, Millan C, Jose LJ, et al. Liver progenitor cell markers correlate with liver damage and predict short-term mortality in patients with alcoholic hepatitis. Hepatology 2012;55:1931–1941. [19] Spahr L, Chalandon Y, Terraz S, Kindler V, Rubbia-Brandt L, Frossard JL, et al. Autologous bone marrow mononuclear cell transplantation in patients with decompensated alcoholic liver disease: a randomized controlled trial. PLoS One 2013;8:e53719. [20] Fausto N. Tweaking liver progenitor cells. Nat Med 2005;11:1053–1054. [21] Boulter L, Govaere O, Bird TG, Radulescu S, Ramachandran P, Pellicoro A, et al. Macrophage-derived Wnt opposes Notch signaling to specify hepatic progenitor cell fate in chronic liver disease. Nat Med 2012;18:572–579. [22] Garg V, Garg H, Khan A, Trehanpati N, Kumar A, Sharma BC, et al. Granulocyte colony-stimulating factor mobilizes CD34(+) cells and improves survival of patients with acute-on-chronic liver failure. Gastroenterology 2012;142:505–512.

620

[23] Spahr L, Lambert JF, Rubbia-Brandt L, Chalandon Y, Frossard JL, Giostra E, et al. Granulocyte-colony stimulating factor induces proliferation of hepatic progenitors in alcoholic steatohepatitis: a randomized trial. Hepatology 2008;48:221–229. [24] Thomas JA, Pope C, Wojtacha D, Robson AJ, Gordon-Walker TT, Hartland S, et al. Macrophage therapy for murine liver fibrosis recruits host effector cells improving fibrosis, regeneration, and function. Hepatology 2011;53: 2003–2015. [25] Kuo TK, Hung SP, Chuang CH, Chen CT, Shih YR, Fang SC, et al. Stem cell therapy for liver disease: parameters governing the success of using bone marrow mesenchymal stem cells. Gastroenterology 2008;134:2121. [26] Lebrun V, Molendi-Coste O, Lanthier N, Sempoux C, Cani PD, van RN, et al. Impact of PPAR-alpha induction on glucose homoeostasis in alcohol-fed mice. Clin Sci (Lond) 2013;125:501–511. [27] Leclercq IA, Vansteenberghe M, Lebrun VB, VanHul NK, Abarca-Quinones J, Sempoux CL, et al. Defective hepatic regeneration after partial hepatectomy in leptin-deficient mice is not rescued by exogenous leptin. Lab Invest 2006;86:1161–1171. [28] Van Hul N, Lanthier N, Suner R Espanol, Quinones J Abarca, Van Rooijen N, Leclercq I. Kupffer cells influence parenchymal invasion and phenotypic orientation, but not the proliferation, of liver progenitor cells in a murine model of liver injury. Am J Pathol 2011;179:1839–1850. [29] Kumar D, Farrell GC, Fung C, George J. Hepatitis C virus genotype 3 is cytopathic to hepatocytes: reversal of hepatic steatosis after sustained therapeutic response. Hepatology 2002;36:1266–1272. [30] Bolstad BM, Irizarry RA, Astrand M, Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 2003;19:185–193. [31] Geiss GK, Bumgarner RE, Birditt B, Dahl T, Dowidar N, Dunaway DL, et al. Direct multiplexed measurement of gene expression with color-coded probe pairs. Nat Biotechnol 2008;26:317–325. [32] Vandesompele J, De PK, Pattyn F, Poppe B, Van RN, De PA, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 2002;3, RESEARCH0034. [33] Viebahn CS, Benseler V, Holz LE, Elsegood CL, Vo M, Bertolino P, et al. Invading macrophages play a major role in the liver progenitor cell response to chronic liver injury. J Hepatol 2010;53:500–507. [34] Clouston AD, Powell EE, Walsh MJ, Richardson MM, Demetris AJ, Jonsson JR. Fibrosis correlates with a ductular reaction in hepatitis C: roles of impaired replication, progenitor cells and steatosis. Hepatology 2005;41:809–818. [35] Jung Y, Brown KD, Witek RP, Omenetti A, Yang L, Vandongen M, et al. Accumulation of hedgehog-responsive progenitors parallels alcoholic liver disease severity in mice and humans. Gastroenterology 2008;134: 1532–1543. [36] Roskams T, Yang SQ, Koteish A, Durnez A, DeVos R, Huang X, et al. Oxidative stress and oval cell accumulation in mice and humans with alcoholic and nonalcoholic fatty liver disease. Am J Pathol 2003;163:1301–1311. [37] Chiba M, Sasaki M, Kitamura S, Ikeda H, Sato Y, Nakanuma Y. Participation of bile ductular cells in the pathological progression of non-alcoholic fatty liver disease. J Clin Pathol 2011;64:564–570. [38] Richardson MM, Jonsson JR, Powell EE, Brunt EM, Neuschwander-Tetri BA, Bhathal PS, et al. Progressive fibrosis in nonalcoholic steatohepatitis: association with altered regeneration and a ductular reaction. Gastroenterology 2007;133:80–90. [39] Sakata K, Ohmuraya M, Araki K, Suzuki C, Ida S, Hashimoto D, et al. Generation and analysis of serine protease inhibitor kazal type 3-cre driver mice. Exp Anim 2014;63:45–53. [40] Fang JW, Bird GL, Nakamura T, Davis GL, Lau JY. Hepatocyte proliferation as an indicator of outcome in acute alcoholic hepatitis. Lancet 1994;343:820–823. [41] Diehl AM. Recent events in alcoholic liver disease V. effects of ethanol on liver regeneration. Am J Physiol Gastrointest Liver Physiol 2005;288: G1–G6. [42] Louvet A, Naveau S, Abdelnour M, Ramond MJ, Diaz E, Fartoux L, et al. The Lille model: a new tool for therapeutic strategy in patients with severe alcoholic hepatitis treated with steroids. Hepatology 2007;45:1348–1354. [43] Lanthier N, Terraz S, Breguet R, Chalandon Y, Rubbia-Brandt L, Spahr L. Computed tomography derived liver volume: a prognostic factor in decompensated alcoholic cirrhosis? Acta Gastroenterol Belg 2013;76:P12. [44] Katoonizadeh A, Nevens F, Verslype C, Pirenne J, Roskams T. Liver regeneration in acute severe liver impairment: a clinicopathological correlation study. Liver Int 2006;26:1225–1233. [45] Colmenero J, Bataller R, Sancho-Bru P, Bellot P, Miquel R, Moreno M, et al. Hepatic expression of candidate genes in patients with alcoholic hepatitis: correlation with disease severity. Gastroenterology 2007;132:687–697.

Journal of Hepatology 2015 vol. 63 j 609–621

JOURNAL OF HEPATOLOGY [46] Louvet A, Artru F, Lasailly G, Cannesson A, Canva-Delcambre V, Dharancy S, et al. Combinative use of baseline and dynamic models in patients with severe alcoholic hepatitis: prediction of death as a continuum in risks of mortality., 60 ed 2014. p. 268A–269A (abstract). [47] Jonsson P, Linder C, Genell S, Ohlsson K. Extrapancreatic origin of the pancreatic secretory trypsin inhibitor as an acute-phase reactant. Pancreas 1996;12:303–307. [48] Uda K, Murata A, Nishijima J, Doi S, Tomita N, Ogawa M, et al. Elevation of circulating monitor peptide/pancreatic secretory trypsin inhibitor-I (PSTI61) after turpentine-induced inflammation in rats: hepatocytes produce it as an acute phase reactant. J Surg Res 1994;57:563–568. [49] Lee YC, Pan HW, Peng SY, Lai PL, Kuo WS, Ou YH, et al. Overexpression of tumour-associated trypsin inhibitor (TATI) enhances tumour growth and is associated with portal vein invasion, early recurrence and a stage-

[50]

[51]

[52] [53]

independent prognostic factor of hepatocellular carcinoma. Eur J Cancer 2007;43:736–744. Affo S, Dominguez M, Lozano JJ, Sancho-Bru P, Rodrigo-Torres D, MoralesIbanez O, et al. Transcriptome analysis identifies TNF superfamily receptors as potential therapeutic targets in alcoholic hepatitis. Gut 2013;62:452–460. Karaca G, Swiderska-Syn M, Xie G, Syn WK, Kruger L, Machado MV, et al. TWEAK/Fn14 signaling is required for liver regeneration after partial hepatectomy in mice. PLoS One 2014;9:e83987. Fausto N. Liver regeneration and repair: hepatocytes, progenitor cells, and stem cells. Hepatology 2004;39:1477–1487. Antoniades CG, Quaglia A, Taams LS, Mitry RR, Hussain M, Abeles R, et al. Source and characterization of hepatic macrophages in acetaminopheninduced acute liver failure in humans. Hepatology 2012;56:735–746.

Journal of Hepatology 2015 vol. 63 j 609–621

621