Stem cells for liver tissue repair: Current knowledge and perspectives

Abstract

Stem cells from extra- or intrahepatic sources have been recently characterized and their usefulness for the generation of hepatocyte-like lineages has been demonstrated. Therefore, they are being increasingly considered for future applications in liver cell therapy. In that field, liver cell transplantation is currently regarded as a possible alternative to whole organ transplantation, while stem cells possess theoretical advantages on hepatocytes as they display higher in vitro culture performances and could be used in autologous transplant procedures. However, the current research on the hepatic fate of stem cells is still facing difficulties to demonstrate the acquisition of a full mature hepatocyte phenotype, both in vitro and in vivo. Furthermore, the lack of obvious demonstration of in vivo hepatocyte-like cell functionality remains associated to low repopulation rates obtained after current transplantation procedures. The present review focuses on the current knowledge of the stem cell potential for liver therapy. We discuss the characteristics of the principal cell candidates and the methods to demonstrate their hepatic potential in vitro and in vivo. We finally address the question of the future clinical applications of stem cells for liver tissue repair and the technical aspects that remain to be investigated.

INTRODUCTION

Orthotopic liver transplantation (OLT) is the gold standard treatment for end-stage liver failure and for numerous liver based inborn errors of metabolism. However, organ shortage remains a major limiting factor and alternative solutions are being examined in the liver therapy field. Liver cell transplantation (LCT) is emerging with heartening success[1,2], but is still limited by cell viability, modest engraftment and limited tissue availability. Increasing interest is carried to stem cells regarding the recent demonstration of their plasticity[3]. Theoretical advantages of stem cells for tissue regenerative medicine are multiple: ease of harvest, proliferation capacity, efficiency of in vitro transfection and potential use of autologous cells. Different types of stem cells are eligible for liver cell therapy according to their hepatic potential, for instance mesenchymal stem cells (MSCs), hematopoietic stem cells (HSCs) and adult liver stem/progenitor cells. Despite encouraging results, key pitfalls remain while using stem cells-derived hepatocyte-like cells: lack of tissue-specific functionality and, up to now, no evidence of strong liver repopulation level in animal models. Moreover, the demonstration of the acquired hepatocyte-like phenotype is impaired by technical restraints. To reach clinical application, stem cell therapy requires further development to become competitive regarding LCT. Here, we present the current knowledge on the use of stem cells for hepatic tissue engineering and describe the future orientations in the field.

IDENTIFICATION OF THE STEM CELL CANDIDATES

Mesenchymal stem cells

Firstly described in 1970 from bone marrow (BM) isolates[4], these cells can currently be obtained from various tissue sources as BM, cord blood (CB), adipose tissue, umbilical cord blood (UCB). Their in vivo function and behavior remain largely unknown[5]. MSCs have the characteristics to be highly proliferative in vitro, easily transfectable, resistant to cryopreservation[6] and to display a large in vitro and in vivo differentiation potential[7]. The first description of their hepatic potential was provided by Lee et al[8] and confirmed by other studies[9–26] (Table 1). While some studies evidenced the in vivo potential of MSCs[10,21], this was recently questioned by other authors who did not detect engrafting cells after syngeneic transplantation[20]. However, experimental protocols were quite different rendering comparison hazardous. A key advantage of MSCs is their immunological properties making them lesser immunogenic and possibly able to induce tolerance as highlighted by promising in vitro studies and clinical trials[27–32].

Table 1.

Representative studies investigating in vitro and/or in vivo hepatocyte differentiation of MSCs

Cell source	In vitrodifferentiation	In vivodifferentiation	References
Human BM&UCB immunoselection	Cocktail: EGF-bFGF (2 d), HGF-bFGF-nicot (7 d), OSM-dexa-ITS Characterization: RT-PCR (8), ICC (2), urea production, LDL uptake, PROD	ND	Lee et al[8]
Human UCB	ND	Host: preimmune fetal sheep Injury: - Cell characterization: Alb+, HepPar1+ Fusion: microdissection + PCR: - Repopulation rate: > 20%	Kögler et al[16]
Human adipose tissue	Cocktail: DMSO-HGF-OSM Characterization: RT-PCR (4), ICC (1), urea production, LDL uptake	Host: NOD/SCID mice Injury: CCl4 TRP: 1.106 cells, tail vein Cell characterization: Alb+/CM-Dil+ Repopulation rate: up to 0.13% Alb+ at 10 d	Seo et al[22]
Human BM MSCs, CD34+ cells, non MSCs/CD34- cells	ND	Host: WT rats (+CsA) Injury: AA TRP: 1.106 cells, liver Cell characterization: Alb+, CK18+, CK19+, αFP+, AGPR+, serum Alb+ Fusion: FISH hum Y/rat chr 12: -	Sato et al[21]
Human BM	Cocktail: 5'aza (1 d) - EGF-EGF (14 d) Characterization: RT-PCR (11), ICC (4), WB (4), urea production, PAS, reporter gene construct for PEPCK activity	Host: Pfp/Rag2-/- mice Injury: propanolol, partial hepatect TRP: 1.106 cells, liver Cell characterization: Alb+, HepPar-1+, PEPCK+, CX32+, PAS+ Fusion: FISH hum alu/mouse satellite: -	Aurich et al[10]
Human BM	Cocktail: HGF-FGF4-ITS-nicot (7 d), OSM-dexa-ITS (15 d) Characterization: RT-PCR (17), ICC (8), flow cytometry, urea&Alb production, PAS, G6Pase activity, neoglucogenesis	Host: SCID mice Injury: partial hepatect or retrorsine/AA TRP: 1.106 cells, spleen or liver Cell characterization: Alb+, αFP+, CK18+, fibronectin+, vimentin+, hum alu+	Lysy et al[106]

Alb: Albumin; AA: Allyl alcohol; αFP: Alpha-fetoprotein; CCl₄: Carbon tetrachloride; CB: Cord blood; CK: Cytokeratin; CsA: Ciclosporin A; d: Days; dexa:Dexamethasone; FGF: Fibroblastic growth factor; G6Pase: Glucose-6-phosphatase; hepatect: Hepatectomy; HGF: Hepatocyte growth factor; hum: Human; ITS: Insulin-transferrin-selenium; nicot: Nicotinamide; OSM: Oncostatin M; PAS: Periodic acid-schiff; PEPCK: Phosphoenolpyruvate carboxykinase; PROD: Pentoxyresorufin-O-dealkylase; WT: Wild-type; chr: Chromosome; FISH: Fluorescent in situ hybridization; ICC: Immunocytochemistry; ND: Not documented; RT-PCR: Reverse transcription-polymerase chain reaction; TRP: Transplantation; WB: Western blot. Amounts of analyzed markers are provided between brackets.

Hematopoietic stem cells

These cells constitute the paradigm of stem cells: cells capable of self-renewing, proliferation and production of a progeny of committed cell lineages, permitting the regeneration of the tissue, even after transplantation. These can be isolated from BM, UCB and peripheral blood[33] making them highly available while they are still difficult to expand in vitro[34,35]. In this cell population, cell identity and phenotype are mainly determined by a selection procedure whose criteria remain debated[36]. HSCs have recently aroused much enthusiasm in the hepatology field after the demonstration of their hepatic differentiation potential as described in Table 2. A majority of studies used mononuclear cell preparations[37–63], which hamper the identification of the cells involved in the differentiation process. For that reason, we included in Table 2 only studies using defined subpopulations[64–80]. However, HSCs’ liver potential is increasingly debated[40,65,81–84] and the possibility of cell fusion as an explanation of ‘plasticity’ make these cells less considerable for a clinical use[85].

Table 2.

Representative studies investigating in vitro and/or in vivo hepatocyte differentiation of HSCs

Cell source	In vitrodifferentiation	In vivodifferentiation	References
Mouse BM	ND	Host: irradiated FAH-/- mice Injury: NTBC withdrawal TRP: 10 to 1000 cells; retroorbital plexus Cell characterization: β-gal+, Alb+, DPPIV+, e-cadherin+, FISH Y+, liver function, FAH activity Repopulation rate: 30%-50% βgal+ cells	Lagasse et al[73]
Mouse BM	ND	Host: irradiated CD45.1+ mice Injury: hepatect or anti-Fas Ab treatment TRP: 106 cells; vein Cell characterization: CK8+, CK18+, CK19+, CD45.2+, FISH Y+, PCR Bcl-2+ Repopulation rate: up to 0.05%-0.8% with anti-Fas Ab treatment	Mallet et al[74]
Human UCB&BM	ND	Host: irradiated NOD/SCID mice ± HGF Injury: CCl4 TRP: 2000 to 105 cells; vein Cell characterization: Alb+, CD45-, human alu+, CK19+, Alb mRNA+, serum Alb+ Repopulation rate: < 1%	Wang et al[77]
Human UCB&PB	ND	Host: irradiated NOD/SCID mice Injury: CCl4 TRP: 2 to 6 × 105 CD34+ cells; tail vein Cell characterization: CD45+, Alb+, α1-AT+, hum DNA+, hum Alb mRNA+&prot+ Repopulation rate: ≤ 0.01% hum Alb mRNA in mice vs human liver	Kollet et al[70]
Mouse BM	Cocktail: co-culture with damaged liver tissue Characterization: RT-PCR (8), ICC (10), FISH Alb mRNA	Host: mice ± irradiation Injury: CCl4 TRP: 105 cells; tail vein Cell characterization: Alb+, FISH Y+, E-cadh+, PKH26+, serum liver function Repopulation rate: up to 7.6% at 2 d Fusion: FISH X-Y: -	Jang et al[69]
Mouse BM	ND	Host: mice ± irradiation ± rhG-CSF Injury: partial hepatect or CCl4 TRP: 5 × 103 selected cells or 5 × 105 MNCs; spleen Cell characterization: no co-eGFP/DPPIV staining, PCR eGFP- Repopulation rate: 0/6.9 × 107 host hepatocytes	Cantz et al[65]

α1-AT: Alpha-1-antitrypsin; Ab: Antibody, Alb: Albumin; β-gal: β-galactosidase; BM: Bone marrow; CCl₄: Carbon tetrachloride; CK: Cytokeratin; d: Days; dexa: Dexamethasone; DPPIV: Dipeptidyl peptidase IV; eGFP: Enhanced green fluorescent protein; FAH: Fumarylacetoacetate hydrolase; hepatect: Hepatectomy; HGF: Hepatocyte growth factor; hum: Human; MNCs: Mononuclear cells; ND: Not documented; NTBC: 2-2nitro-4-trifluoromethylbezoyle-1,3-cyclohexanedione; PB: Peripheral blood; UCB: Umbilical cord blood; E-cadh: E-cadherin; rhG-CSF: Recombinant human Granulocyte-Colony Stimulating Factor; TRP: Transplantation. Amounts of analyzed markers are provided between brackets.

Adult liver stem cells

As organ shortage is limiting the availability of this cell population, the conditions for their use in cell therapy are governed by two main characteristics, a high proliferation rate and/or a robust cell banking capacity. Regarding their promising hepatocyte-like functionality, some cell compartments could be promptly considered for toxicological assays. Even if identity and in vivo function of this cell population are currently under controversy, four main types of hepatic progenitors are described: oval cells, small hepatocytes, liver epithelial cells and mesenchymal-like cells. Oval cells are generated from the biliary tree in response to hepatic injury. They display a bipotent differentiation potential (hepatic and biliary cells) and can be expanded in vitro. There is no consensus on whether these cells are originating in BM or not[48,86]. However, similar cell types were recently isolated from healthy liver[87]. The first description of small hepatocytes isolation was made by Mitaka et al[88] from a non-parenchymal fraction after centrifugation of isolated liver cells. These cells are smaller than hepatocytes, possess an in vitro proliferation capacity[89] and can differentiate into mature hepatocytes in vitro[90]. Liver epithelial cells is a population firstly described by Tsao et al[91] and more recently in healthy adult human liver[92]. Although different from oval cells, these cells are bipotential and able to differentiate into hepatocyte-like cells in vivo[92]. A mesenchymal-like cell population has also been isolated from adult human liver[93,94]. This population depicts high level of proliferation and, as MSCs, possesses a broad differentiation potential. How these cells share MSCs’ immunological properties has to be determined. Other cell types were also isolated from manipulated livers[95], but the possibility of culture artefacts[96,97] should always be considered before describing new ‘liver progenitor cells’.

Other cell types

Embryonic stem cells constitute, at present, the best in vitro model for hepatocyte differentiation. However, ethical restrictions[98] and the possibility of malignancies development[99] mainly limit their use in the clinical setting[100,101]. Other more committed cell lineages as monocytes[102,103] or fibroblasts[104] were evaluated for hepatic differentiation. In this context, recent data about nuclear reprogramming[105] provided possible explanation of the differentiation potential of cells thought to be restricted to a cell lineage.

IN VITRO STUDY AS ARGUMENT FOR HEPATOCYTE COMMITMENT

Up to now, studies illustrating in vitro hepatic features of stem cells-derived hepatocyte-like cells were essentially confined at the phenotypic rather than the functional level. Acquisition of specific markers is a tool for evidencing a cell commitment while hepatocyte-like functionality is required to consider a cell for therapy.

Hepatocyte-like phenotype

While a morphological change is common after in vitro cell conditioning, the morphology is rarely similar to mature hepatocytes and authors should rather talk about a ‘hepatocyte-like morphology’ in differentiating cells, more especially as ultrastructural data are mostly not provided. How the morphological changes are evocative of the acquisition of an epithelial phenotype is questionable as similar morphological changes can be obtained when incubating hBM-MSCs with a classical differentiation medium or with a control medium[106]. Using specific protein and mRNA marker expression for the description of a hepatocyte-like phenotype could also be questioned as many studies were based on restricted panel of markers and as some authors described the presence of hepatocyte-specific markers in undifferentiated cells[8,14,16,22,82,107,108]. The broadest amount of markers for phenotypic characterization should be used. In this manner, we observed an incomplete and erratic acquisition of hepatocyte-specific features in hBM-MSCs[106], cf Table 1. Moreover, these cells partially maintained the expression of native mesenchymal markers, suggesting a chimerical phenotype. This phenotype could be the result of the artificial in vitro differentiation process.

Hepatocyte-like functionality

The functional characterization of stem cells-derived hepatocyte-like cells is, up to now, principally achieved by commercially available assays (ELISA, colorimetric tests, PROD or EROD assays) whose handling are often poorly reproducible and quite difficult to interpret, or by other tests (as lipoprotein uptake) with low specificity. As clearly stated by Hengstler et al[109] in a recent report, there is a crucial need for standardized and normalized techniques for characterizing hepatocyte-like functionality in differentiating cells. Regarding clinical application, one should provide the evidence of a specific functional activity after differentiation before considering stem cells for cell therapy.

IN VIVO STUDY AS PROOF OF HEPATOCYTE DIFFERENTIATION POTENTIAL

Direct detection of differentiated donor cells in situ

Most studies about stem cells in vivo hepatic repopulation and differentiation are performed in a xenogeneic setting (e.g. human cells transplanted into immunocompromised rodents). The impact of cell-to-cell interactions in an interspecies microenvironment on cell adhesion and differentiation[110] is poorly documented. Moreover, this model does not reflect clearly the clinical reality and data must be carefully interpreted. In the same way, a drawback of allogeneic cell transplantation in rodents is that stem cell behavior was described to be quite different between species making extrapolation to human perilous[111]. Besides a huge diversity in the studies design, almost all in vivo models are based on the analysis of a repopulation potential after cell infusion in a diseased liver. Putative mechanisms of hepatic repopulation and tools designed to enhance it are described in Figure 1. It arises from the literature that hepatocyte differentiation from extrahepatic stem cells does not occur (or at very low level) in a non-injury model, and that the injury has to be strong enough and give advantage to the donor cells to seed and proliferate into the liver parenchyma. This was elegantly suggested by Stadtfeld et al[83] who demonstrated in a transgenic mouse model the minor role of the hematopoietic compartment in the adult liver ontogenetic development. Drug-induced liver injury has to be considered only as a tool for studying in vivo hepatocyte differentiation potential and no consensual protocol exists to date. Many questions are still opened: type of liver injury, time dependency of liver stimulus (acute vs chronic liver failure), use of drugs blocking host cell regeneration, relevance of injury after cell transplantation (compatibility with donor cell viability), etc. Globally, liver repopulation after stem cell transplantation remains limited. It is noteworthy that these data were obtained after single cell injection while for clinical purpose one would consider serial injections. Engraftment level has to be determined in this situation. By comparison, LCT can potentially reach higher repopulation levels[112] and, for that reason, is currently more valuable for clinical purpose. Conversely, some authors reported amazing repopulation levels with HSCs[73] or with UCB stem cells[16], but in both cases animal models were highly permissive and not representative of common cell transplant procedures. Methods for evaluation of cell engraftment are multiple and deal each with proper pitfalls[113]. For example, fluorescent in situ hybridization for detecting Y chromosome in a sex-mismatched transplantation can be difficult to interpret and, at low positivity levels, could be hampered by female microchimerism occurring after several litters[114]. The evaluation of in vivo cell differentiation is prominently performed by objectifying protein markers expression whereas immunoassays are often technically limited by antibodies affinity or specificity. For instance, the demonstration of in vivo expression of HepPar-1 (OCHE15 clone) should be questioned because of the antibody’s cross-reactivity between species. For that reason, the use of combined techniques is recommended for the identification of in vivo differentiating cells, as staining assays performed on serial sections and association of various techniques (ex: in situ hybridization, immunostaining, cell tracking with fluorescent molecules, flow cytometry, RT-PCR, cellular magnetic resonance) (Figure 2).

Figure 1.

Mechanisms of liver repopulation after stem cell transplantation. A: Stem cell migration into the liver parenchyma after spleen infusion is guided mainly by the splenic and the portal vein flow. Homing of the cells is tightly regulated by chemoattractive agents (as SDF-1, HGF and SCF) secreted by liver cells under liver damage conditions[136], and acting respectively through the interaction with CXCR4, c-met and c-kit receptors of stem cells. MMP-9, expressed by host hepatocytes after injury, is implicated in the homing process through its action on the extracellular matrix. This process can be triggered by various types of liver injury; B: While reaching the portal vein, 70% to 80% of the infused cells are lost and phagocytosis by macrophages/phagocytes can be observed in the portal areas[137]. After 16-20 h post-infusion, the remaining cells enter the liver sinusoid after having crossed over the endothelial barrier rendered permeable by a local secretion of VEGF[137], and by the combined action of various MMPs[138]. The permeabilization of the endothelium by specific drugs owning toxicity against endothelial cells (as monocrotaline[139] or doxorubicin[140]) or acting on the permeation (VEGF[141,142]) was described to enhance the repopulation rate of the infused cells. The penetration of the parenchyma is facilitated by the action of MMPs and the disruption of gap junctions[143]. The cell implantation creates an alteration of venous blood flow which can hamper the progression of the cells inside the parenchyma but can be alleviated by the use of vasodilators[144]. Globally, inducing liver injury facilitates cell integration by remodeling the liver architecture and by creating a local microenvironment favorable for the proliferation of the transplanted cells. This is mediated by the local secretion of cytokines/growth factors (HGF, FGF, TGFα). Cell implantation implicates the reformation of gap junctions that can be observed 3 to 7 d after transplantation. The implanted cells display variable cell phenotype parallel to the level of their hepatocyte differentiation; C: After integration, some cells will be cleared from the hepatic parenchyma. Dead cells lacking implantation stimulus or cell-to-cell interactions will be phagocytes by Kupffer cells. This can be inhibited by the use of gadolinium chloride[145] or by monocrotaline[139] which was demonstrated in animal models to improve the kinetics of cell implantation. Other cells will be rejected by the host immunological system (effector cells) and tools to alleviate this rejection are multiple (immunosuppression protocols, encapsulation of the infused cells, co-transplantation with cells owning immunosuppressive properties). FGF: Fibroblastic growth factor; HGF: Hepatocyte growth factor; MMP: Matrix metalloproteinase; SCF: Stem cell factor; SDF: Stromal derived factor; TGF: Transforming growth factor; VEGF: Vascular endothelial growth factor.

Figure 2.

Analysis of liver engraftment of transplanted stem cells. Human BM-MSCs were transplanted into the liver of partially hepatectomized SCID mice and livers were screened after one month for analysis. A: Staining on serial sections for fibronectin, albumin and CK18 by immunohistochemistry revealed a co-expression of mesodermal and hepatocyte markers and confirmed the chimerical phenotype of hepatocyte differentiating MSCs (400 ×); B: The use of combined techniques for revealing the expression of human alu probe by in situ hybridization and of fibronectin by immunohistochemistry in a cell cluster (800 ×).

Demonstration of donor cell functionality through their metabolic activity

The gold-standard for evaluating in vivo differentiation is the detection of cell functionality. To this end, some authors reported albumin secretion in transplanted mice serum. Other reports on ‘in vivo functionality’ are argued on basis of the reduction of mortality rate, improvement of liver regeneration or of liver fibrosis[26,59,81,115–118]. It has to be determined how these findings are the fact of indirect cell-to-cell interactions or of direct cell functionality. However, Lagasse et al[73] historically demonstrated the improvement of tyrosinemia disease in mouse after cell infusion whereas the selective pressure in this model is high and elicits fusion events.

Implication of cell fusion in the differentiation process

The fusion process implies that a cell inserts its genetic content into another cell to form a resulting unit that acquire the ‘host’ phenotype. The resultant product creates a heterokaryon in which the nuclei do not always fuse. The concept of fusion has emerged after experiments on co-culture of BM cells with ESCs[119,120]. Fusion between hematopoietic cells and hepatocytes has been demonstrated[43,121–125] and invalidated[69,126,127]. This is explained by the conceptual diversity of these studies and perhaps the complexity of the process itself. For example, fusion events can be acted by HSCs[128] or require homing of these cells and implication of progeny, as highlighted by studies showing fusion between myelomonocytic lineages and hepatocytes[122,125]. Interestingly, studies providing strong data about fusion used the FAH^-/- model which produces mitotic and chromosomal abnormalities[129] that could strengthen the fusion process. It appears that selective pressure is necessary to induce relevant fusion processes which happen rarely in a non-injury model. How the cell fusion and plasticity phenomenon are parts of the same process or vary in importance according to the population used (stem cells versus committed cells[130]), has to be explored by tracing the donor cells in their route and studying signalling pathways. Interestingly, fusion events have, to date, never been described with MSCs[10,21].

FROM BENCH TO CLINICAL SIDE

Literature concerning the use of stem cells for clinical therapy is unceasingly growing in many fields (including hematology, cardiology, neurology and orthopedics)[131]. However, to date, there is no clear documentation about long-term safety of such procedure[132,133]. Multi-center consensual state-of-the-art about several clinical aspects is mandatory. First is the election of the appropriate clinical indication. In the hepatology field, liver-based inborn errors of metabolism are the most attractive candidate because of the possibility for elective treatment and, in some cases, the absence of global liver function impairment making technical procedures more feasible. The lack of liver injury may limit stem cells engraftment and differentiation. However, stem cells therapy remains relevant since small increases of the deficient enzyme activity can improve patients’ quality of life. In that case, the acquisition of a specific enzyme activity has to be demonstrated in vitro before cell transplantation. Liver cirrhosis/fibrosis has been considered: besides encouraging results, there is, to date, no evidence about a direct or a mediated cell involvement, nor about how the fibrosis could hamper cell implantation[116]. Before considering acute liver failure for stem cell therapy, it would be necessary to achieve a performing banking of functional hepatocyte-like cells. Globally, the possibility of repeated cell infusions could render cell therapy advantageous over OLT in patients whose disorder is not life-threatening (ex: Criggler-Najar patients). In this condition, whether stem cells could have an advantage on liver cells according to their proliferative capacity needs further assessment. Other points requiring consensus is the technique and route of cell delivery (surgically placed catheter vs transcutaneous injection), the amount of delivered cells and timing of infusions, and the protocol of immunosuppression.

CONCLUSION

The first challenge for considering stem cells for therapy is the proof of their safety at long term by in vitro and in vivo assays. The other major goal is the determination of their in vivo functionality in small animal models combining a metabolic disorder allowing objective screening of cell function and a permissive immunological context (SCID or knock-out mice, immunosuppression therapy in allotransplantation). Large animal models would provide definitive arguments about safety and functionality and allow an ideal setting for designing immunosuppression protocol and cell injection program. These models will be useful to attest the place of stem cells in liver cell therapy by comparison with LCT. Further research is also mandatory to examine the involvement of stem cells in the reconstitution of the non-parenchymal cell compartment after transplantation. Indeed, some cell populations have the potential to differentiate into endothelial cells[93,134,135] or could differentiate into mesodermal lineages similar to their phenotype (as hepatic stellate cells or fibroblasts) and as such interfere with the liver tissue architecture (Figure 3). Regarding the promising results obtained using MSCs as third-party for HSCs transplantation[29,31] or in immunological pathologies[30], the immunomodulatory role of these cells as adjuvant in LCT or in OLT should be determined.

Figure 3.

Involvement of stem cells in the regeneration of the liver parenchymal and non-parenchymal cell lineages. Schema shows the liver architecture and details the topography of the cell populations. In the table are resumed the cell fractions that could potentially be the targets of differentiation after transplantation of stem cells. Some of these differentiation potentials were documented with stem cells while other were contradicted and were thus exposed as controversial. Other differentiation pathways are purely hypothetical based on the embryonic origin of the stem cells. BC: Bile canalicule; CLV: Centrolobular vein; GJ: Gap junctions.