Liver injury and regeneration mechanisms

Generally, liver injury can be classified into two categories, acute liver injury

and chronic liver injury, which have different modes of regeneration (Itoh and Miyajima, 2014). Acute liver injury causes rapid deterioration of liver functions, often resulting from liver transplantation surgery, drug overdose or virus infection.

Partial hepatectomy (PHx) has long been used as a model of acute liver injury.

Upon PHx, about 70% of liver tissue is removed, while the remaining tissue is still functional. The regeneration is achieved by hepatocytes hypertrophy and following proliferation, restoring liver to its original size. In this case, LPCs, the progenitor of hepatocytes, are not involved in regeneration (Miyaoka et al., 2012).

On the contrary, chronic liver injuries, which occur more in human, are generally caused by prolonged hepatitis virus infection, alcohol assumption, obesity or drug administration. It is a long-period process of liver destruction and may lead to fibrosis, cirrhosis and hepatocellular carcinoma. The PHx model described above cannot imitate the conditions of chronic liver disease due to restrained proliferation ability of hepatocytes and the occurrence of inflammation during chronic liver injury. Instead, immature LPCs emerge and expand, a process known as ductular reactions (DRs) (Roskams et al., 2004). Afterwards, LPCs would differentiate into both hepatocytes and cholangiocytes to restore liver functions. Several injury models have been applied to activate LPC population,

al., 1999), choline-deficient ethionine-supplemented (CDE) diet (Akhurst et al., 2001), bile duct ligation (Tag et al., 2015), and multiple carbon tetrachloride (CCl4) injections (Weber et al., 2003). Using the models above, the role of LPCs during liver regeneration and how LPCs are activated can be further illustrated.

1.2. Liver progenitor cell (LPC)

1.2.1. LPC and its role in liver regeneration

Liver progenitor cells (LPCs) are adult liver stem cells that are able to differentiate toward hepatocytes and cholangiocytes. LPCs are also called facultative stem cells since they are activated under liver damage but are inapparently observed during normal homeostasis (Yanger and Stanger, 2011).

LPCs often emerge in the periportal region and they can compensate the fact that mature hepatocytes and cholangiocytes are insufficient to restore the population during severe damage. Various markers have been used to identify LPCs, such as EpCAM, CD133, Sox9 (Furuyama et al., 2011; Okabe et al., 2009; Rountree et al., 2007). By utilizing these markers, LPCs can be isolated or traced to further study their roles in liver regeneration.

1.2.2. Signaling pathways promoting LPC activation and expansion

The regenerative process of LPCs is regulated by adjacent cells in their

microenvironment. In addition to hepatocytes and cholangiocytes, other non-parenchymal cells including hepatic stellate cells (HSCs), Kupffer cells, endothelial cells also reside in liver tissue. These niche cells provided signals to mediate LPCs activation and expansion and orchestrate the LPC response in injured liver. Thy1⁺ mesenchymal cells were found to produce FGF7 and acted on FGFR2b on LPCs to induce LPC response upon DDC-induced injury (Takase et al., 2013). Macrophage-derived TNF-related WEAK inducer of apoptosis (TWEAK) was another important factor that can facilitate regeneration through downstream NF-κB to stimulate LPC proliferation (Tirnitz-Parker et al., 2010).

Other demonstrated that Wnt signaling was enhanced by DDC diet and contributed to LPC activation in mice (Hu et al., 2007). Cross-talks between different cell types indicated that stem cell niche are indeed important to mediate LPC response, but more thorough investigation should be further conducted.

1.2.3. LPC origin

LPCs usually emerge from the canal of Hering, a structure linking bile canaliculi and bile duct, which comprise hepatocytes and cholangiocytes respectively (Paku et al., 2001). This location seems reasonable to be the niche of progenitor cells since they could immediately repopulate the two cell lineages.

tissue were regarded as LPCs origin in different researches, including mature hepatocytes, cholangiocytes, hepatic stellate cells (HSCs) and extrahepatic bone marrow cells.

Cholangiocytes have phenotypic resemblance as LPCs and they share similar molecular markers, suggesting that cholangiocytes may be the origin of LPCs (Itoh and Miyajima, 2014). Indeed, studies using lineage tracing indicated that LPCs generated during liver injury were derived from cholangiocytes (Furuyama et al., 2011). Additionally, HSCs were activated and transformed into myofibroblasts upon liver injury, which produced hepatocytes and cholangiocytes by mesenchymal-epithelial transition (MET) (Michelotti et al., 2013). However, recent studies demonstrated that mature hepatocytes also had the potential to reprogram into LPCs. Activation of Notch-Hes1 signaling was shown to be responsible for the conversion of hepatocytes into primitive ductules (Sekiya and Suzuki, 2014). Others suggested that hepatocytes, in response to injury, could give rise to progenitor cells and re-differentiate into hepatocytes to restore liver functions (Tarlow et al., 2014). Moreover, some groups illustrated that hepatocytes could even transdifferentiate into cholangiocytes during injury, implying that hepatocytes have the plasticity to reprogram (Jeliazkova et al., 2013;

Michalopoulos et al., 2005; Yanger et al., 2013). Several signaling pathways have

been shown to involve in hepatocyte reprogramming, such as Wnt/β-catenin, Notch, Hippo/YAP, TGF-β. For instance, increasing YAP expression could facilitate hepatocyte de-differentiation (Yimlamai et al., 2014).

1.3. Hypoxia and Hypoxia-inducible factor 1 (HIF-1)

1.3.1. Hypoxia and HIF-1

Hypoxia-inducible factor 1 (HIF-1) is a heterodimeric transcription factor consisting of an O2-sensitive α subunit and a constitutively expressed β subunit (Wang et al., 1995). HIF-1 is a key regulator in response to hypoxia and can further govern the transcription of hundreds of genes (Semenza, 2007). Under normoxic condition, HIF-1α subunit is hydroxylated by prolyl hydroxylase domain-containing protein 2 (PHD2), bound by von Hippel-Lindau protein (VHL), leading to ubiquitination and degradation of HIF-1 (Epstein et al., 2001; Jaakkola et al., 2001). On the contrary, hypoxic condition can stabilize HIF-1α, which then dimerizes with HIF-1β. HIF-1 can translocate into nucleus and bind onto hypoxia responsive element (HRE) of its downstream target genes. By regulating the transcription of target genes, HIF-1 can make changes to the cell characteristics, enabling cells to survive under hypoxic stress. Besides, HIF-1 also modulates

differentiation and metastasis (Semenza, 2007). Moreover, HIF-1 was shown to maintain undifferentiated phenotype in various types of stem cells and sustain quiescence to escape the damages (Mathieu et al., 2011; Mohyeldin et al., 2010) (Figure 1.1).

1.3.2. HIF-1α in liver diseases

During chronic liver disease, hepatic stellate cells (HSCs) are activated and transform into myofibroblasts which can secrete numerous extracellular matrix (ECM). Accumulation of ECM could lead to liver architectural alteration and eventually causes fibrosis (Lee and Friedman, 2011). Upon fibrosis, fibrotic septa would resist blood flow and decrease oxygen delivery, thereby create hypoxic environment (Cannito et al., 2014) (Figure 1.2). HIF-1α could alter several signaling pathways during chronic liver injury. A research using BDL as chronic liver injury model showed that HIF could modulate NF-κB and TGF-β to stimulate inflammation and collagen synthesis (Moczydlowska et al., 2016).

HIFs are also responsible for HCC aggressiveness, leading to a more metastatic tumor phenotype. By increasing stemness of liver tumorous tissue, HIF-1α could enhance tumor progression, metastasis and drug resistance (Bogaerts et al., 2015). Also, Wnt/β-catenin could promote HIF-1α-induced EMT in HCC and inhibit cancer cell death (Zhang et al., 2013). Additionally, some

studies indicated that HIF-1α could contribute to the appearance of tumor-initiating cells in tumor by increasing stem cell marker. For example, in PHD2 haplodeficient mice with HCC, which HIF-1α could continuously express, the phenotype of tumor was changed from HCC into mixed HCC-CC due to increased stemness marker (Heindryckx et al., 2012).

Figure 1.1. Under normoxia, the presence of oxygen leads to HIF-1α proteasomal

degradation. Under hypoxia, HIF-1α dimerizes with HIF-1β and binds onto hypoxia responsive element (HRE) to promote the transcription of its target genes.

Figure 1.2. Upon liver injury, hepatic stellate cells are activated and transform into

myofibroblasts, which can secrete extracellular matrix (ECM). Accumulation of ECM resists blood flow and decreases oxygen delivery, leading to hypoxia.

1.3.3. Fibrosis-induced hypoxia and LPC activation

We previously used CCl4 injections to simulate chronic liver injury in mice, fibrosis appeared and it was accompanied by hypoxia and HIF-1α up-regulation.

Moreover, HIF-1α was primarily expressed in LPCs since it was detected to co-localize with LPC markers (Figure 1.3). Also, Wnt signaling was found to be critical for HIF-1α-induced LPC emergence (Figure 1.4). Thus, these previous results demonstrated that CCl4-induced chronic liver injury could result in LPCs activation and proliferation through hypoxia and Wnt signaling.

However, the origin of those LPCs were still unknown. From immunofluorescence staining results, LPCs were surprisingly emerged around central vein instead of the common residence, Canal of Hering, which is periportal (Figure 1.5). Since mature hepatocytes are the major cell types around central

vein, we suggested that hepatocytes might serve as the origin of LPCs through de-differentiation. However, more investigations are needed to clarify the origin of LPCs during chronic liver injury.

Figure 1.3. 6-week CCl4 injections induced HIF-1α expression and EpCAM⁺ LPCs emergence. Co-localization of HIF-1α and EpCAM showed that HIF-1α endowed certain cells with stemness and made them transform into LPCs.

Figure 1.4. 6-week CCl4 injections promoted nuclear translocation of β-catenin in EpCAM⁺/CK19⁺ LPC-like cells, indicating the importance of Wnt signaling during HIF-1α-induced LPCs emergence.

Figure 1.5. 6-week CCl4 injections-induced injury and following LPCs emergence appeared near central vein, which mainly consists of mature hepatocytes. PV, portal vein;

CV, central vein.

1.4. Wnt/β-catenin signaling pathway

1.4.1. Wnt/β-catenin signaling pathway

Wnt/β-catenin pathway plays vital roles in the development of various organs by directing cell proliferation, cell fate decision and cell polarity (Logan and Nusse, 2004). During tissue homeostasis, Wnt signaling was also shown to be critical to stem cell maintenance (Lowry et al., 2005; Reya et al., 2003). Therefore, dysregulation of Wnt signaling may cause cancers or other diseases. For instance, mutations of Wnt pathway were often observed in colon cancer (Lammi et al., 2004).

β-catenin is the major effector of Wnt signaling. In the absence of Wnt ligand, β-catenin forms a complex with GSK-3β, Axin and APC in the cytoplasm, leading to its ubiquitination and proteasomal degradation. When Wnt ligand binds onto its receptor, Frizzled (Fzd), β-catenin is released from the complex and then translocates into the nucleus. By binding with DNA-binding factors, T-cell factor/lymphoid enhancer factor (TCF/LEF), β-catenin initiates transcription of downstream target genes involved in various biological functions (Clevers, 2006) (Figure 1.6).

Figure 1.6. Scheme of canonical Wnt/β-catenin signaling.

1.4.2. Wnt/β-catenin signaling in liver diseases

Growing evidences have shown that Wnt signaling is also crucial to liver diseases as well as liver regeneration (Huch et al., 2013). Recent studies indicated β-catenin could stimulate cell-cycle entry of hepatocytes during liver regeneration after PHx treatment (Sekine et al., 2007). Likewise, accumulated β-catenin were shown to be involved in HCC progression by promoting cell proliferation (Thompson and Monga, 2007). Other researches further examined the role of Wnt signaling in the regeneration of chronic liver injury since it could regulate cell proliferation after PHx. A study using DDC diet to induce LPCs emergence and proliferation depicted that Wnt/β-catenin signaling was a part of niche signals that regulated LPCs behaviors (Itoh et al., 2009). Also, resection specimens from patients of liver injuries revealed an important role of Wnt signaling in human LPCs proliferation (Spee et al., 2010). As mentioned above, Wnt signaling could maintain stem cell properties, and it was proved to drive LPCs specification into hepatocytes (Boulter et al., 2012). These evidences implied the significance of Wnt signaling in liver regeneration, especially LPCs activation.

1.4.3. HIF-1α and Wnt/β-catenin signaling

In previous studies, HIF-1α has been proved to promote the expression of

β-regeneration. For example, hypoxia led to β-catenin accumulation in HCC and further stimulated EMT of cancer cells (Liu et al., 2010). Moreover, hypoxia could promote self-renewal in ESCs through Wnt signaling (Mazumdar et al., 2010).

Though, whether similar mechanisms work in other stem cell populations are still unknown.

Some recent researches further examined the direct interaction between 1α and β-catenin. A study showed that β-catenin would directly interact with HIF-1α in HCC, increasing the transcription activity of HIF-HIF-1α and facilitating EMT (Zhang et al., 2013). Another study found out that β-catenin could serve as coactivator of HIF-1α and improve hepatocyte survival using ischemia and reperfusion to induce liver injury (Lehwald et al., 2011). However, the relationship between HIF-1α and β-catenin during LPC activation has not been explored.

1.5. Epithelial cell adhesion molecule (EpCAM)

1.5.1. Epithelial cell adhesion molecule (EpCAM)

EpCAM is a transmembrane glycoprotein, containing an extracellular domain (EpEx), a transmembrane domain and a small intracellular domain (EpICD) (Schnell et al., 2013) (Figure 1.7). It was first characterized to function as an adhesive factor, which located on cell membrane and mediate cell-cell

adhesion (Litvinov et al., 1997). Later, EpCAM was found overexpressed in various carcinoma, but the mechanisms and its function in cancer cells are still not well known (Maetzel et al., 2009). Also, EpCAM was frequently detected in stem/progenitor cells and was lost in differentiated cells (Schmelzer et al., 2006).

Recent studies revealed that EpCAM can engage in signal transduction, which is associated with cell proliferation, migration, de-differentiation, maintenance of stem cell population (Gonzalez et al., 2009; Munz et al., 2004). The signal transduction begins by proteolysis of EpCAM, releasing EpICD into cytoplasm.

EpICD would form a complex with FHL2, β-catenin and Lef-1, translocate into nucleus, and then stimulate the transcription of downstream target genes (Munz et al., 2009) (Figure 1.8).

1.5.2. EpCAM in liver disease

As mentioned above, EpCAM may serve as an oncogenic signaling protein and was highly expressed in carcinomas. Some studies indicated that EpCAM could be a biomarker of HCC, and that the expression level of EpCAM was correlated with poor prognosis (Yamashita et al., 2008). EpCAM is also a marker of LPCs, thereby suggesting that EpCAM may be an essential factor to maintain cancer stem cells and promote their growth. As mentioned above, EpCAM is able

Figure 1.7. EpCAM is a transmembrane glycoprotein, containing an extracellular domain

(EpEx), a transmembrane domain (TM) and an intracellular domain (EpICD).

Figure 1.8. EpICD releases from plasma membrane and form a complex with FHL2,

β-catenin and Lef. The complex then stimulates the transcription of downstream target genes in the nucleus.

showed that EpCAM⁺ cells increased during liver injury (Okabe et al., 2009;

Yoon et al., 2011). Based on these results, it is reasonable to suspect that EpCAM might mediate LPCs proliferation during liver injury and promote regeneration.

1.5.3. HIF-1α and EpCAM

The relationship between HIF-1α and EpCAM has not been confirmed yet.

However, a recent study discovered that EpCAM increased under hypoxia condition and contributed to renal regeneration (Trzpis et al., 2008). Additionally, by examining specimens from HCC patients, HIF-1α and EpCAM were significantly highly expressed during recurrence after radiotherapy (Yamada et al., 2014). The results suggested that HIF-1α increment may subsequently induce the expression of EpCAM, a cancer stem cell marker, leading to tumor aggressiveness.

In silico analysis of EpCAM promoter discovered several regulatory enhancers

including hypoxia responsive element (HRE). Thus, HIF-1α may probably bind onto the HRE of EpCAM and regulate its expression.

1.5.4. Crosstalk between Wnt/β-catenin and EpCAM

EpCAM was characterized as a hepatic stem cell marker and Wnt/β-catenin signaling is critical to stem cell maintenance. A study found that EpCAM is a

EpCAM induction would be important to maintain cancer stem cell growth (Yamashita et al., 2007). Later, a nuclear complex formed by EpICD, FHL2, β-catenin and Lef-1 was detected. The complex could induce gene transcription by binding to Lef-1 consensus sites and promote cell proliferation (Maetzel et al., 2009). Other researches revealed the crosstalk between Wnt signaling and EpCAM in a variety of carcinomas, including HCC, colon cancer, thyroid cancer and ovarian cancer (Ralhan, 2010; van der Gun et al., 2011; Yamashita et al., 2009;

Zhou et al., 2015). However, whether Wnt signaling and EpCAM could synergistically mediate liver regeneration and LPCs activation is still unknown.

Chapter 2. Motivation and Aim

Our previous results showed that HIF-1α was up-regulated during liver fibrosis and that HIF-1α would lead to LPCs activation through Wnt signaling. However, the origin of LPCs and the mechanism of LPCs activation during liver regeneration remains controversial. Since LPCs were detected around central vein in our previous result, I suggested that LPCs might be generated from mature hepatocytes. Therefore, I cultured primary hepatocytes under hypoxia to investigate whether they could de-differentiate into LPCs. If so, the relationships between HIF-1α, Wnt signaling and the LPC marker, EpCAM, during LPCs activation would be further examined.

Chapter 3. Materials and Methods

3.1. Primary hepatocytes isolation and culture

Primary hepatocytes were harvested from six-week old male C57BL/6 mice (BioLASCO) with following procedure. Briefly, mice were anesthetized and their portal vein were first perfused with PBS followed by Liver Perfusion Medium (Life Technology) and Liver Digestion Medium (Life Technology). Perfused liver was minced, suspended with Dulbecco’s modified Eagle’s medium (DMEM, Life Technologies) supplemented with 10% fetal bovine serum (FBS, GeneDireX) and 1% penicillin/streptomycin/glutamine (PSG, Life Technologies) and passed through a 40-μm cell strainer (BD Falcon). Cells were washed several times with DMEM and re-suspended with primary cell culture medium [DMEM supplemented with 10% FBS, 1% PSG, 1% glutamine (Life Technologies), 1% ITS (Life Technologies), 1% NEAA (Life Technologies) and 0.1% dexamethasone (Life Technologies)]. Cells were then centrifuged through a discontinuous Percoll (GE Healthcare) gradient and cultured on 0.1% gelatin-coated dish at 37^oC in a humidified incubator with 5% CO2. Hypoxic culture was conducted with same culture medium in the incubator with 5% CO2 and 1% O2.

3.2. Cellular RNA extraction

Cellular RNA was extracted with TriPure isolation reagent (Roche). 200 µL

chloroform was added into 1 mL cell extraction. The extraction was mixed by vortexing for 15 seconds and centrifuged at 12000 g at 4^oC for 15 minutes. RNA was extracted and then was precipitated using 0.7x aqueous phase volume isopropanol, followed by centrifugation at 12000 g at 4^oC for 10 minutes.

Precipitated RNA was washed twice with 70% ethanol and air-dried. After dissolving in 30 µL DEPC/ddH2O at 55^oC for 15 minutes, the quality and concentration of RNA was determined by Nanodrop 2000 (Thermo).

3.3. Quantitative reverse transcription polymerase chain reaction (qRTPCR) 2 µg total RNA was reverse transcribed into cDNA using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystem) according to the protocol manual. For quantification, 1 ng cDNA was used for qPCR with 200 nM forward and reverse primers in total 10 µL reaction volume by SYBR Green system Rad). qPCR was performed using Bio-Rad CFX96 Real-Time PCR Machine (Bio-Rad) with following protocol: denaturation at 95^oC for 3 minutes, 40 cycles of elongation at 65^oC for 30 seconds and denaturation at 95^oC for 10 seconds. The expression level of target genes was assayed in triplicate and normalized to internal control. The primer used for qPCR were listed in Table 1.

3.4. Cellular protein extraction

buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 1 mM EDTA) supplemented with 1X cOmplete Cock Tail Protease Inhibitor (Roche) on ice. The total cell lysates were centrifuged at 12000 g at 4^oC for 10 minutes to remove the pellet. Protein concentration was quantified by Protein Assay Dye Reagent (Bio-Rad). Protein samples were denatured with 5x protein sample buffer (50% glycerol, 10% SDS, 0.25 M Tris-HCl, 0.05% Bromophenol blue, 5% β-mercaptoethanol) at 100^oC for 10 minutes.

3.5. Western blotting

Protein samples were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). SDS-PAGE gel (10% separation gel and 4 % stacking gel) and TGS buffer system (50 mM Tris-HCl, pH 8.0, 380 mM Glycine, 0.1% SDS) were used to separate total protein. Electrophoresis was performed by following protocol: 100 V for 20 minutes and 150 V for 1.5 hour. Proteins were then transferred onto PVDF membrane (Roche) by Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad) at 1.5 mA/cm² for 90 minutes. The membrane was washed with TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Tween-20) and blocked with 4% skim milk in TBST for 1 hour at room temperature. After washed with TBST, primary antibody [diluted in gelatin-NET (0.25% gelatin, 0.15 M NaCl, 5 mM EDTA, 0.05% Tween-20, 50 mM Tris-HCl ,pH 8.0) or 4% skim milk] was added

and incubated in 4^oC overnight. The membrane was washed thrice and second antibody (diluted in gelatin-NET) was added for 1 hour at room temperature. After washed with TBST twice and once in ddH2O, membrane was rinsed with Luminate^TM Classico Western HRP Substrate (Millipore) or Immobilon^TM Western Chemiluminescent HRP Substrate (Millipore) for 1 minute and analyzed by Biospectrum MultiSpectral Imaging System (UVP BioSpectrum 800). The antibodies and their dilution condition were listed in Table 2.

3.6. Cell immunofluorescence staining

Cells were fixed with 4% paraformaldehyde (PFA) at room temperature for 10 minutes and then permeabilized with 0.5% Triton X-100 at room temperature for 10 minutes. After blocked with 3% FBS (diluted in PBS) at room temperature for 1 hour, cells were incubated with primary antibody at 4^oC overnight. Second antibody

Cells were fixed with 4% paraformaldehyde (PFA) at room temperature for 10 minutes and then permeabilized with 0.5% Triton X-100 at room temperature for 10 minutes. After blocked with 3% FBS (diluted in PBS) at room temperature for 1 hour, cells were incubated with primary antibody at 4^oC overnight. Second antibody