Chronic liver disease (CLD) is characterised by a permanent damage of liver cells followed by activation of repair mechanisms and represent a major health problem worldwide. Twenty-nine million people in the European Union suffered from CLD in 2013 [1
]. Liver transplantation is the only cure for end-stage liver disease patients. Hence, it is of upmost interest to understand the mechanisms controlling liver regeneration in CLD and to identify involved pathways and biomarkers for early diagnosis and treatment. Liver disease affecting bile acid regulation is called cholestatic liver disease. It can be caused by drugs, autoimmune damage of bile ducts, genetic defects and developmental disorders. Liver regeneration in the context of a cholestatic liver disease needs to replace damaged cells, generate branching tubules and a fibrovascular stroma to preserve the new tissue. Several studies addressed the role of Notch signalling in liver development [3
] and liver regeneration [9
]. Notch signalling is an evolutionarily conserved pathway regulating numerous cellular processes including cell fate decision and differentiation [12
]. Notch is required to commit hepatic progenitor cells (HPC) to the biliary fate and to orchestrate the biliary tree remodelling. Recombination signal binding protein for immunoglobulin kappa j region (RBPJ) is the central downstream effector of Notch signalling and is essential for the activation of the Notch pathway [14
]. Liver-specific Rbpj
deletion in mice causes impaired intrahepatic bile duct development, severe cholestasis, hepatic necrosis and fibrosis [15
]. Through the capability of a cellular plasticity, the liver is able to generate intra-hepatic bile ducts (IHBD) independent of Notch signalling. Thereby the biliary lineage marker SOX9 is acquired in periportal as well as in interlobular hepatocytes to form an intermediate cell phenotype [16
Here, we show that the liver-specific deletion of Rbpj induced growth reduction, liver cholestasis and hepatic necrosis followed by a compensatory liver regeneration. Due to hepatic cholestasis, we depicted increased IQGAP1 expression and subsequent nuclear translocation of YAP, which results in hepatocyte differentiation by induction of SOX9 towards an intermediate phenotype. In addition, from our study of in vitro cholestatic liver injury, we could show that upregulation of Sox9 expression upon induction of cholestasis is independent of Notch signalling. We suppose the activation of YAP as a mechanism how the cholestatic liver can regenerate upon loss of Rbpj.
The current study provides experimental evidence that YAP activation upon Rbpj
deletion induced cholestasis is an important mechanism in liver regeneration. The biliary phenotype, which we report in our study is reminiscent of that observed in transgenic mouse models with inactivation of HES1, JAGGED1, RBPJ, or LKB1 [4
]. These studies analysed the role of Notch signalling during liver development, on biliary differentiation and morphogenesis. Rbpj
cKO mice were used in combination with different Cre recombinase driver lines to achieve liver-specific deletion of Rbpj
]. Two of the most commonly used Cre lines are the AlbCre
] and the AlfpCre
transgenic mouse line [20
]. Notch signalling was shown to control liver development by regulating the biliary differentiation and the three-dimensional architecture of intrahepatic bile ducts [8
]. Disruption of Notch signalling leads to defects in the communicating intrahepatic bile duct network [7
]. In addition, it was shown in AlbCre+ Rbpj−/−
mice that the intrahepatic bile duct regeneration does not require Notch signalling [16
]. We focused our study on 4- and 36-week-old mice to address molecular mechanisms involved in the regenerative process of the liver.
The impaired life span of Rbpj−/−
mice is due to severe cholestasis and massive necrosis, which reduces the regenerative potential in around one-third of the mice to a life-threatening event. The severity of induced necrosis is 8.2-fold higher in the AlfpCre+ Rbpj−/−
compared to the AlbCre+ Rbpj−/−
. By use of the AlbCre
line the mice did not show growth retardation and impaired survival. Necrosis was rarely detectable in 4-week-old AlbCre+ Rbpj−/−
mice. The Cre-recombinase is induced at ED16.5 under the Alb
promotor whereas the Alfp
promotor is active before ED10.5 [20
], which leads to an earlier onset of Rbpj
deletion and result in a more severe phenotype. We observed the growth defect at the suckling/weaning transition, which correlates with an increase in lethality. The severity of necrosis correlates with an impaired regeneration potential which was not reported before.
It is still under debate which mechanisms are involved in liver regeneration to restore the liver cell mass and to maintain liver homeostasis in the context of a cholestatic liver. This study shows that Rbpj
deletion results in severe cholestasis measured by increased levels of TB, ALP, ALT and AST, and thereby leads to hepatic necrosis. The importance of the Notch signalling pathway has been linked to diseases like BA and Alagille syndrome, which are two rare cholestatic diseases during early childhood [40
]. We found that the Hippo pathway enrichment signature by GiANT analysis in human disease by comparing BA and non-BA to control patients suggesting the Hippo pathway might be important in the regeneration process of neonatal cholestasis.
The liver is characterised by a cellular plasticity, which is important for liver regeneration upon liver injury [47
]. It was reported in AlbCre+ Rbpj−/−
mice that Sox9
mRNA levels decreases after Rbpj
deletion at P3 but was similar at P60 compared to wildtype mice [15
]. In the same mouse model, another group reported increased SOX9 level by IHC at P30 which disappeared by P60 [11
]. Both studies reported about parenchymal necrosis but without a correlation to regeneration. The necrosis was not significant in Rbpj−/−
]. In our study, we observed SOX9+
cells at 4 weeks and at a reduced, but still significant, level at 36 weeks of age in Rbpj−/−
mice. We observed in the liver of Rbpj−/−
cells with hepatocyte morphology, carrying the ductal and progenitor marker SOX9. SOX9+
cells were enriched in 4-week-old Rbpj−/−
mice. The occurrence of these intermediate cells positive for HNF4α and SOX9 but negative for CK19 links to the regeneration process of hepatocyte transdifferentiation rather than activation of HPCs or hepatogenic differentiation of mesenchymal stem/progenitor cells which was reported in human liver disease specimens [49
Hepatocytes and in higher amount non-hepatocytes expressed the scaffolding protein IQGAP1 which was highly increased in the liver of 4-week-old Rbpj−/−
mice. Increased bile acid levels were shown in a transgenic mouse model with a defect in bile acid homeostasis. Deletion of the nuclear receptors FXR and SHP resulted in increased IQGAP1 levels, nuclear YAP expression and liver carcinogenesis, indicating that the accumulation of bile acids leads to a nuclear YAP translocation by activating a pathway which is dependent on the induction of IQGAP1 [29
]. YAP is a key player in Hippo signalling pathway and is known to be an important regulator of organ size, cell fate and to be involved in carcinogenesis [21
]. However, the knockout of YAP leads to impaired liver regeneration in the cholestatic liver [27
]. In addition, increased IQGAP1 and nuclear YAP localisation have been reported in human biliary disorders and bile duct ligated mice, a model for experimentally induced cholestasis [27
]. The current study depicts YAP activation as an essential mechanism for liver regeneration upon Rbpj
loss induced cholestasis. Thereby, initiate a program for hepatocyte transdifferentiation, which involves upregulation of SOX9.
We performed in vitro studies by using freshly isolated hepatocytes to study the importance of Notch signalling upon cholestatic liver injury in vitro. Interestingly, we could show that Notch depletion either by the use of AlbCre+ Rbpj−/−
hepatocytes or by GSI-IX treatment on AlbCre+ Rbpj+/+
hepatocytes, which blocks Notch signalling, does not influence SOX9 expression in cholestatic liver injury in vitro. GCA treatment of hepatocytes from AlbCre+ Rbpj+/+
and AlbCre+ Rbpj−/−
mice mimics a cholestatic liver disease in both genotypes and results in an increase of Sox9
mRNA level. We do not see differences in SOX9 protein levels. This might occur due to the fact that in vitro cultivated primary hepatocytes are quiescent and do not undergo cell division [50
]. However, SOX9 expression was clearly blocked by the YAP inhibitor VP. This describes the activation of SOX9 via YAP activation, as a Notch-independent mechanism in cholestatic liver injury in vitro. In contrast to the in vitro model, the in vivo mouse model depicted a cholestatic liver disease only in Rbpj−/−
mice due to impaired IHBD maturation, Rbpj+/+
mice did not develop a cholestatic liver disease.
As reported above, we observed severe cholestasis in Rbpj−/−
mice. Severe cholestasis generates a chronic damage of the liver, leading to activation of macrophages for the clearance of necrotic areas. This might also stimulate fibrosis formation and, thereby, could be a risk factor for the initiation of hepatocarcinogenesis in the absence of Notch signalling. Hepatic stellate cells are able to orchestrate the clearance of necrotic cells by conversion of Kupffer cells to M1-like proinflammatory macrophages, which increases phagocytic activity [51
]. Here we showed that deletion of Rbpj
leads to liver fibrosis, which slightly increases, with age and the formation of liver foci in one mouse. So far, the importance of Notch signalling in hepatocarcinogenesis is addressed by overexpression of Notch pathway components [34
] and by repression of Notch signalling, which was identified in an interacting network of Hippo/Wnt/β-catenin/Notch signalling [35
]. However, the work of Kulic et al. showed loss of RBPJ in human cancer and cancer cell lines, but only one liver cancer cell line was analysed [36
]. Future work is needed to address the importance of RBPJ loss in hepatocarcinogenesis.
In conclusion, we suppose the YAP activation as a mechanism for a compensatory liver regeneration after RBPJ ablation induced cholestasis. The liver-specific deletion of Rbpj leads to loss of Notch signalling, which results in impaired IHBD formation and, thereby, an accumulation of bile acid in liver leading to cholestasis. The accumulation of bile acid generates an increased expression of the scaffolding protein IQGAP1, which facilitates nuclear YAP translocation to trigger liver regeneration. We claim this mechanism as an important driver for liver regeneration upon RBPJ loss-induced cholestasis.
4. Materials and Methods
4.1. Mouse Model
] were crossed with AlfpCre
] to generate liver-specific Rbpj
knockout mice. The following cohorts were generated and utilized in this study: AlfpCre+ Rbpjflox/flox
) and AlfpCre− Rbpjflox/flox
). Hepatocytes used for in vitro studies were isolated from AlbCre+ Rbpjflox/flox
and AlbCre− Rbpjflox/flox
]. The mice were maintained in a specific pathogen-free environment. All mice received human care and study protocols comply with the institution’s guidelines (Animal Research Centre of Ulm University). The state government of Baden-Württemberg (protocol number 35/9185.81-3/1259; date of approval 8 March 2016) approved all animal experiments.
4.2. Liver Histology
Mouse liver tissue was collected and incubated in 4% paraformaldehyde (PFA) for 16 h, processed through ethanol and xylene series, and embedded in paraffin. Sections 3.5 µm thick were used for hematoxylin and eosin (H&E) staining. The necrotic area was calculated by measuring the area of individual necrotic spots per vision field (100×) using the ImageJ software (available online: https://imagej.nih.gov/ij/
). In total 10 vision fields per liver were analysed (n
= 5–6 mice).
Paraffin-embedded mouse liver tissues were sectioned at 5 µm thickness and used for IHC or IF staining. Antigen unmasking was performed using antigen-unmasking solution (Vector Laboratories, Burlingame, CA, USA) in a steamer for 35 min. Staining was performed with the appropriate primary antibody (Table S4
) incubation overnight at 4 °C, followed by corresponding HRP-labelled secondary antibody incubation for 1 h at room temperature (RT). Nova-red (Vector Laboratories, Burlingame, CA, USA) was used for developing chromogenic staining and sections were counterstained with 20% hematoxylin. For immune fluorescence staining, sections were processed similarly to chromogenic staining. After incubation with a corresponding fluorescence-conjugated secondary antibody for 1 h at RT, the nucleus was counterstained with DAPI (Sigma-Aldrich, St Louis, MO, USA).
4.4. Protein Isolation and Western Blotting
For whole cell protein isolation, liver tissue or primary hepatocytes were homogenised and lysed in 1× RIPA buffer (50 mM TrisHCl pH 8, 150 mM, 1% NP-40, 0.5% deoxycholate, 0.1% SDS, 1 mM NaVO3, 1 mM DTT, 1 mM PMSF) containing protease inhibitor cocktail solution. Protein lysates were stored at −80 °C until analysed. For nuclear and cytoplasmic protein isolation, liver tissue was homogenised using Dignam A lysis buffer (10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM KCL, 50 mM PMSF, 1 M DTT and protease inhibitor) and centrifuged. The supernatant and the pellet were collected separately. The supernatant consists the cytoplasmic protein fraction, which was stored at −80 °C. For nuclear protein fraction, the pellet was washed in Dignam B buffer (Dignam A with 0.1% Triton X-100) and afterwards re-suspended in Dignam C buffer (20 mM HEPES pH 7.9, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl2 and 0.2 mM EDTA, 1 mM DTT, 1 mM PMSF, 1 mM Na3VO4 and protease inhibitor) and stored at −80 °C until analysed. Bradford assay was used to measure protein concentration. Standard western blotting protocol was adopted for western blot experiment.
4.5. RNA Isolation
Total RNA was isolated from liver tissue using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s guidelines and the quality of RNA was determined using the 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA samples with a RIN (RNA integrity number) value above 7.0 were used in this study.
4.6. Quantitative Real-Time PCR (qRT-PCR)
cDNA was synthesised from total RNA using the Reverse Transcription System (Promega, Madison, WI, USA) according to the manufacturer’s guidelines. cDNA was amplified by using iTaq Universal SYBR Green Supermix (BIO-RAD, Hercules, CA, USA) in a total volume of 10 µL. Used primer are listed in Table S5
4.7. Serum Parameters
Liver-specific serum enzymes such as ALT, AST, ALP and TB were measured using Reflotron test stripes (ROCHE, Penzberg, Germany).
4.8. Gene Expression Analysis
Gene expression analysis was carried out using the SurePrint G3 Mouse Gene Expression 8x60K Microarray Kit (Design ID 028005; Agilent Technologies, Santa Clara, CA, USA). Samples were labeled with the Low Input Quick Amp Labeling Kit (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer’s guidelines. Slides were scanned using a G2565CA microarray scanner (Agilent Technologies, Santa Clara, CA, USA). Expression data were extracted using the Feature Extraction software (Agilent Technologies, Santa Clara, CA, USA). All expression data were deposited in Gene Expression Omnibus (GEO accession number GSE121302). Pre-processing of expression data was performed according to Agilent’s standard workflow. Using 5 quality flags (gIsPosAndSignif, gIsFeatNonUnifOL, gIsWellAboveBG, gIsSaturated, and gIsFeatPopnOL) from the Feature Extraction software output (Agilent Technologies, Santa Clara, CA, USA), probes were labeled as detected, not detected, or compromised. Gene expression levels were background corrected, and signals for duplicated probes were summarized by geometric mean of non-compromised probes. After log2 transformation, a percentile shift normalization at the 75% level and a baseline shift to the median baseline of all probes was performed. Differentially expressed genes were calculated based on shrinkage-T statistic (false discovery rate < 0.1). Pathway enrichment analysis (KEGG pathways) was calculated by Fisher’s exact test (false discovery rate < 0.05). All computations were performed using the R statistical software framework (available online: http://www.R-project.org
4.9. Gene Set Enrichment Analysis (GSEA)
GiANT and GSEA were performed as previously reported [22
]. The Gene Expression Omnibus (GEO) dataset GSE121302 representing gene expression analysis from murine liver from Rbpj+/+
mice were used for the GiANT and the GSEA analysis of the Hippo pathway gene set (mmu04390) and GSE46995 representing gene expression analysis from human BA, non-BA and control patients, was used for the GiANT analysis of the Hippo pathway gene set (hsa04390).
4.10. Hepatocyte Isolation and Cultivation
Cells were isolated from adult mouse livers by 2-step collagenase perfusion [55
]. Hepatocytes were purified by centrifugation in 50% Percoll (50 g for 10 min at 4 °C). The cell pellet containing the viable cells were washed twice in 20 mL Dulbecco’s modified Eagle medium and centrifuged at 50 g for 5 min at 4 °C. Hepatocytes were cultivated on collagen type I- coated plates in standard Dulbecco’s modified Eagle medium containing 10% fetal bovine serum, 1× Insulin-Transferrin-Selenium X, 10−7
mol/L dexamethasone, 1× penicillin/streptomycin/L-glutamine, and 1× nonessential amino acid solution.
4.11. Cholestatic Liver Injury In Vitro
500,000 primary hepatocytes from Rbpj+/+ and Rbpj−/− mice were plated in 6-well plates. 3 h after plating the hepatocytes cells were washed with 1× phosphate-buffered saline (PBS). Immediately we started the treatment with 500 µM GCA for 6 h to induce in vitro cholestatic liver injury. In addition to GCA, cells were also incubated with 10 µM VP to block YAP/TEAD-dependent and 10 µM GSI in inhibit Notch pathway.
4.12. Statistical Analysis
The Mann–Whitney test was used to calculate statistical significance using GraphPad Prism 6 (GraphPad Software, Inc, La Jolla, CA, USA) and the data were represented as scatter dot plots (median with interquartile range).