Stem cells compose a “reservoir” of potential cells at various stages of development that can be used for the restoration and regeneration of damaged tissues and organs. Under proper conditions, stem cells may differentiate into specialized tissues and organs. They are self-sustaining and can replicate themselves for a long time. These unique features make them a promising tool for studies on therapy for diseases such as chronic liver disease, heart stroke, spinal injuries, stroke, Parkinson's disease, Alzheimer's disease, retinal degeneration, and diabetes mellitus.
Stem cells can be classified into two major categories according to their developmental status: embryonic and adult (postnatal). Each represents a diverse differentiation potential status and a different potential application. The liver is one target for which the development of stem cell-based therapy is of great significance. Even though an injured liver is highly regenerative, many debilitating diseases lead to hepatocyte dysfunction and organ failure. Treatments such as resection are usually arrested because of too little remaining liver function. Liver transplantation is the only effective treatment for severe liver injuries. However, because of organ rejection and the limited number of donors, alternative therapeutic approaches are needed. Stem cells could offer a potentially unlimited and minimally invasive source of cells for hepatocyte replacement and liver regeneration.
Many types of stem cells have been differentiated in vivo and/or in vitro into hepatocyte-like cells (Fig. 1) using different induction strategies. However, there is no defined strategy to produce hepatocytes from stem cells, and more study is needed.
In this review, we present the current status of information regarding hepatocyte differentiation protocols using stem cells from different sources. We also highlight the superior strategies regarding the functions of generated hepatocytes and the best types of stem cells for liver regeneration.
AAT α-1-antitrypsin AFP α-fetoprotein ALB albumin AT adipose tissue BM bone marrow BMP bone morphogenic protein CCl4 carbon tetrachloride CD cluster differentiation CK cytokeratin CYP cytochrom p450 DEX dexamethasone Dlk-1 delta-like protein 1 DMN dimethylnitrosoamine DMSO dimethylsulfoxide EB embryoid body ECM extracellular matrix EGF epidermal growth factor EPCAM epithelial cell adhesion molecule ES embryonic stem FACS/MACS fluorescence/magnetically activated cell-sorting FBS fetal bovine serum FGF fibroblast growth factor FISH fluorescence in situ hybridization GATA GATA binding protein GFP green fluorescent protein GVHD graft-versus-host disease G6P glucose-6-phosphatase HAT histone acetyltransferase HDAC histone deacetylase HGF hepatocyte growth factor HIFC hepatic induction factor cocktail HNF/FOXA hepatocyte nuclear factor/forkhead box HSC hematopoietic stem cells ICG indocyanin-green IGF insulin-like growth factor ITS insulin/transferrin/selenium LDL low density lipoprotein LIF leukemia inhibitory factor MSC mesenchymal stem cell NCAM neural cell adhesion molecule NGF nerve growth factor OC oval cell OsM oncostatin M RA retinoic acid STM septum transversum mesenchyme TAT tyrosine aminotransferase TDO2 tryptophan-2,3-dioxygenase TGF transforming growth factor TNF tumor necrosis factor TTR transthyretin UCB umbilical cord blood.
EMBRYONIC DEVELOPMENT OF THE LIVER
Based on an actual understanding of embryonic development of the liver (Lemaigre and Zaret,2004; Zhao and Duncan,2005), many studies in vivo and in vitro have indicated the therapeutic potential of stem cells for liver regeneration. Knowledge of embryonic liver development basically includes studies on rodent embryos, but, recently, many studies have been done on chicks, zebrafish, and frogs. During mouse development, the induction of hepatic genes occurs in a segment of the definitive endoderm at about 8.5 days of gestation (Lemaigre and Zaret,2004; Zhao and Duncan,2005) (Fig. 2). The induction requires signaling cues from cardiogenic mesodermal cells in the form of fibroblast growth factors (FGFs) and from septum transversum mesenchyme (STM) in the form of bone morphogenic proteins 2 and 4 (BMP2, BMP4). Afterwards, the endodermal cells start to proliferate and bud into the STM, where there is essential interaction with endothelial cells. However, the role of endothelial cells is yet unknown. Hematopoietic stem cells (HSCs) also have a significant role; these cells migrate into the liver bud and proliferate there, apparently emitting the signal for further liver development. When the hepatic endoderm is specified and the liver bud is growing, the cells within the so-called hepatoblasts are bipotential and capable of differentiating into hepatocytes and bile duct cells (cholangiocytes).
There are numerous signals required for the growth of the fetal liver and the prevention of apoptosis (Lemaigre and Zaret,2004; Zhao and Duncan,2005). Each of those cues generated by STM, cardiac mesoderm, endothelial cells, HSCs, and a microenvironment rich in extracellular matrix (ECM) is crucial for liver development.
DEVELOPMENT OF DIFFERENTIATION CONDITION MEDIA FOR HEPATOCYTE INDUCTION IN VITRO
Numerous cytokines and growth factors have been shown to have a potent effect on hepatic growth and differentiation in vitro. They include the hepatocyte growth factor (HGF), epidermal growth factor (EGF), transforming growth factor (TGF)-β, acidic FGF/FGF1, insulin, insulin-like growth factor (IGF), and oncostatin M (OsM). It is noteworthy that for each of the cytokines/growth factors, the timing, dosage, and combinations must be carefully chosen according to the stem cell type and desired effect. Among the non-proteinaceous chemical compounds that are known to promote hepatic differentiation and/or maintenance in vitro are dexamethasone (DEX), retinoic acid (RA), sodium butyrate, nicotinamide, norepinephrine, and dimetylsulfoxide (DMSO). Sequential treatment with growth factors that mimic in in vivo development of the liver has been demonstrated. Lastly, the important factor during liver development, growth, and regeneration is ECM, which also undergoes extensive remodeling under pathological conditions. Hence, the introduction of appropriate ECM (collagen, fibronectin) during in vitro differentiation may enhance the differentiation process.
Specific transcription factors of early endoderm, bipotential cells, and, finally, hepatocytes are expressed at specific stages of development. Early differentiation markers include the hepatocyte nuclear factor (HNF)-3beta/forkhead box A2 (FOXA2), GATA binding protein 4/6 (GATA4/6), α-fetoprotein (AFP), albumin (ALB), and cytokeratin 8/18 (CK8/18), and there are other factors, such as α-1-antitrypsin (AAT), tyrosine aminotransferase (TAT), transthyretin (TTR), tryptophan-2,3-dioxygenase (TDO2), glucose-6-phosphatase (G6P), and a sort of cytochrom P450 (CYP)-metabolizing enzyme. When specific markers are expressed, the biochemical and metabolic functions are of importance. These include albumin and urea synthesis, glucose production, glycogen storage ability, and CYP activities.
HEPATIC STEM CELLS: EXTRINSIC AND INTRINSIC SOURCES
The liver involves three distinct sources of cells that participate in regeneration. The first source is mature hepatocytes, which respond rapidly to liver injury. They represent unipotent stem cell–like properties, although their division potential is limited. It is important to mention here that, for stem cell–based therapy, the primary culture of hepatocytes does not replicate sufficiently in vitro to produce the number of cells necessary for transplantation and does not maintain its properties in vitro.
The cell source, which is activated when extensive and chronic damage occurs, is located within the biliary tree and gives rise to bipotential so-called oval cells (OCs), which can differentiate into hepatocytes and biliary epithelial cells (cholangiocytes) (Beltrami et al.,2007; Evarts et al.,1987; Farber,1956; Fausto,2004; Fougere-Deschatrette et al.,2006; Herrera et al.,2006; Kojima et al.,2005; Shafritz et al.,2006; Wang et al.,2003; Yovchev et al.,2007; Zhou et al.,2007). The term “oval cell” was introduced by Farber (1956) who found non-parenchymal cells with the characteristic morphological appearance after treating rats with carcinogenic agents. Farber believed that OCs are not the hepatocyte progenitors. However, the group of Thorgeirsson showed evidence of the bipotentiality of OCs (Evarts et al.,1987), and other authors confirmed their observation (Fougere-Deschatrette et al.,2006). The existence and profile of OCs in adult liver is still debated. The isolation of such progenitors is difficult due to the lack of surface markers. OCs express the markers of biliary epithelial cells (CK 7, CK 19, and OV-6), immature fetal hepatoblasts (AFP, ALB, gamma-glutamyltranspeptidase [GGT]), hematopoietic stem cells (CD34, CD90, Sca-1, CD117, and the flt-3 receptor), and markers not expressed on other types of cells, for example, chromogranin A, the neural cell adhesion molecule (NCAM), and delta-like protein 1 (Dlk-1). Information relative to surface markers on progenitors within the adult liver is constantly increasing (Herrera et al.,2006; Yovchev et al.,2007). Herrera et al. isolated hepatic progenitors from adult human liver expressing mesenchymal stem cell markers (CD29+, CD73+, CD44+, CD90+, CD34−, CD45−, CD11−, and CD133−) together with hepatocyte-specific markers (ALB+, AFP+, CK8low+, and CK18low+) (Herrera et al.,2006). Beltrami et al. isolated multipotent stem cells from adult human liver which had the characteristics of pluripotent stem cells (Beltrami et al.,2007). Rogler's research group identified transit bipotential cells (NCAM+, CK 19−, and HepPar 1−); when these cells became committed to cholangiocytes, they expressed: NCAM+, CK 19+, and HepPar 1−, and while committed to hepatocytes, they expressed: NCAM−, CK 19−, and HepPar 1− (Zhou et al.,2007). Miyajima's research group has documented several important molecules responsible for the differentiation and maturation of fetal hepatocytes (Kojima et al.,2005; Watanabe et al.,2007). These findings make the isolation of liver progenitors easier. However, the number of OCs in the normal liver is extremely low; therefore, many researchers have characterized the hepatoblasts from fetal liver (Dan et al.,2006; Nierhoff et al.,2007). Fausto's research group determined the immunophenotype of fetal liver stem cells as CD34+, CD90+, CD117+, the epithelial cell-adhesion molecule (EPCAM)+, c-met+, SSEA-4+, CK18+, CK19+, ALB−, AFP−, CD44h+, and vimentin+ (Dan et al.,2006).
A third source is exogenous stem cells, which may be derived from bone marrow (BM) or other organs or tissues. How these cells from different sources integrate to achieve a homeostatic balance remains unexplained. Thus, the question of whether those stem cells transdifferentiate into hepatocytes under patho-physiological conditions or only in experimental strategies remains to be answered (Fausto,2004; Fausto et al.,2006; Shafritz et al.,2006). In addition, the question of how homeostasis is disrupted by pathologic conditions also needs to be answered. Seeking a proper inner balance is also important. Answering these questions is essential for understanding how the body functions. At present, the answers to the above questions remain elusive; however, below are discussions of the various stem cells and their hepatogenic potentials.
STEM CELLS FOR HEPATOCYTE DIFFERENTIATION
Embryonic Stem Cells
Embryonic stem (ES) cells were first isolated from mouse embryos (Evans and Kaufman,1981). Pluripotent ES cells isolated from the inner cell mass of blastocysts are capable of giving rise to cells found in all three germ layers of the embryo. They are considered to have the greatest range of differentiation potential. The isolation of human ES cells several years ago expanded the potential of ES cells as a source of cells for not only developmental studies but also stem cell–based therapy (Thomson et al.,1998).
The pluripotency of mouse ES cells has been proven in vivo and in vitro. In vivo, an injection of ES cells generates teratomas harboring derivatives of all three embryonic germ layers. In vitro, after removal from the feeder layer or from the leukemia inhibitory factor (LIF), mouse ES cells aggregate in a suspension to form spheroid clumps of cells called embryoid bodies (EBs). The cells within the EBs differentiate spontaneously and express molecular markers specific for the three embryonic germ layers.
Here, we present examples of hepatic induction (in vitro and in vivo) using mouse and human ES cells utilized by different groups (Fig. 3) (Lavon and Benvenisty,2005; Teramoto et al.,2005). In vitro strategies include differentiation through EB formation, co-culture, and mono-culture systems, while in vivo strategies utilize animals with liver injury or liver regeneration. All progress regarding ES cell differentiation is summarized in Table 1.
Table 1. Differentiation Potential of ES Cells Towards Hepatocyte-Like Cellsa
Hepayocyte specific markers/functions
GST, glutathione -S-transferase; CPS-1, carbamyl phosphate synthetase 1; PEPCK, phosphoenolpyruvate carboxykinase; LST1, liver specific organic anion transporter 1; C3, complement 3; PXMP1-L, peroxisomal membrane protein 1-like protein; dHGF, HGF with deletion of 5 amino acids; ABCG2, ATP-binding cassette subfamily G member 2; Dlk-1, delta-like protein 1; Cx32, Gap junction protein; c-Met, Met-protooncogene/HGF receptor; ApoE, apolipoprotein E.
Mouse ES cells in vivo
EBs into hepatectomized mice, FGF1+FGF2, HGF, OsM+DEX
ES cells have a propensity to develop teratomas when implanted into animals. Teratomas form tumors and finally cause the death of the host animal, which severely limits their clinical use. Chinzei et al. (2002) demonstrated that cells isolated from EBs nine or more days after LIF removal expressed a panel of hepatic markers and were capable of producing albumin and urea. After transplantation into partial hepatectomy of female mice pretreated with 2-acetylaminofluorene, ES cell–derived cells survived and expressed ALB, whereas teratomas were found in mice transplanted with ES cells or EBs up to day six. These authors demonstrated that, while ES cells always developed teratomas in recipient mice, the incidence was decreased with implantation of EBs; but this improvement depended on the culture period of the EBs. The in vivo differentiation of ES cells carrying green fluorescent protein (GFP) in the AFP locus was achieved by Yin et al. (2002). They selected a subpopulation of GFP-positive and AFP-expressing cells from differentiating in vitro ES cells. After transplantation into partially hepatectomized lacZ-positive ROSA26 mice, GFP-positive cells engrafted and differentiated into lacZ-negative and ALB-positive cells. In this case, no teratomas were observed. Furthermore, using an animal with an injured liver (regenerative condition), Yamamoto et al. (2003) reported efficient differentiation of ES cells into hepatocytes with therapeutic properties. They harvested a GFP+ fraction of in vivo differentiated ES cells and characterized them. Those cells revealed hepatocyte-specific markers and therapeutic potential (Table 1). Similar studies with in vivo transplantation of GFP+ ES cells were conducted by Moriya et al. (2007). These authors reported the presence of ES cells within the liver on days 10 and 20, after liver injury; however, on day 30 they could not detect any. In addition, until day 30, no tumors were detected, and fibrosis decreased.
In vitro differentiation.
EB formation. EB cells mature by the processes of spontaneous differentiation and cavitation and acquire markers for a variety of differentiated cell types. Dissociating EBs and plating the differentiated cells as a monolayer yields many cell lineages. Several growth factors and transcription factors have been shown to be capable of directing the differentiation of mouse ES cells. Different matrix proteins may dramatically influence the generation and survival of these cells (Flaim et al.,2005). Usually, collagen is used as the matrix for culturing the cells towards a hepatic lineage because the liver bud proliferates and migrates into STM, which is composed of loose connective tissue containing collagen. Hamazaki et al. (2001) demonstrated that mouse EBs can be differentiated into hepatocyte-like cells, when cultured on collagen-coated plates, with early (FGFs), middle (HGF), and late (OsM, DEX, and insulin+transferring+selenium [ITS]) differentiation stage factors. Jones et al. (2002) confirmed these observations by culturing ES cells carrying a gene trap vector insertion into an ankyrin-repeat-containing gene. This modification induces beta-galactosidase expression when hepatocyte differentiation begins. Kuai et al. (2003) reported that the nerve growth factor (NGF)-β also promotes hepatic differentiation, which is increased in the presence of HGF and RA.
Miyashita et al. (2002) also demonstrated in vitro hepatic differentiation through the formation of EBs without using hepatocyte-specific cytokines. Yamada et al. (2002), using an ES cell line carrying the enhanced GFP gene, identified indocyanin-green (ICG) uptake by cells differentiated from mouse EBs and reported the presence of liver-specific markers using RT-PCR and immunocytochemistry. Ishizaka et al. (2002) demonstrated that when transfected with HNF-3β, mouse ES cells were able to differentiate into hepatocytes with liver-specific metabolic functions after stimulation with FGF2, DEX, L-ascorbic-2-phosphate, and nicotinamide. The same genetically modified ES cells were differentiated through EB formation into hepatic-like cells by Kanda et al. (2003) using an attached culture system. Importantly, later on, they discovered that HNF-3β transfected ES cell–derived hepatic-like cells have infinite proliferating potential, resulting in tumor formation and, finally, the death of the animals after transplantation.
EBs offer the advantage of providing a three-dimensional (3D) structure, which enhances cell–cell interactions that may be important for hepatocyte development. However, the complexity of EBs is a problem because of the cytokines and inducing factors generated within these structures that induce differentiation of other cell lineages as well. Until now, none of the hepatic differentiation systems based on EB formation has revealed sufficiently efficient induction of functional hepatocytes for an experimental therapeutic study.
A direct co-culture.
A co-culture strategy was used by Ishii et al. (2005). They produced in vitro mature hepatocytes from ES cells (carrying GFP in the AFP locus) entirely via isolation of AFP-producing cells and subsequent maturation of these cells by co-culture with Thy1-positive (CD90) mouse fetal liver cells. These hepatic-like cells produced and stored glycogen and neutralized ammonia.
In a co-culture system, differentiation takes place in contact with assisting cells (fetal liver cells or stromal cells) because of the presence of differentiation factors; however, undefined factors produced by these supportive cells may influence the differentiation of ES cells into other than hepatocyte cell types. An additional disadvantage is the difficulty entailed in separating the ES cell–derived hepatocytes from assisting cells.
An adherent mono-culture.
Teratani et al. (2005) showed that ES cells can differentiate into functional hepatocytes without requiring EB formation, in vivo transplantation, or a co-culture system. By comparing the genes in CCl4-treated and untreated normal mouse liver, their group identified a hepatic induction factor cocktail (HIFC). ES cell–derived hepatocytes, after HIFC induction, expressed multiple liver-specific markers as well as hepatocyte nuclear factors. In addition, AFP expression appeared in the early and TDO2 at the late stage of differentiation, which means that ES cell–derived hepatocytes mimic normal liver development. The functionality shown by their ability to produce glucose and clear ammonia and to synthesize urea displays the characteristics of mature hepatocytes. In addition, in this case, no teratomas were observed, and karotype analysis showed a normal chromosome number. Most importantly, the transplantation of ES cell–derived hepatocytes in mice with cirrhosis generated by dimethylnitrosoamine (DMN) showed a significant therapeutic effect. On the basis of the HIFC differentiation system, Yamamoto et al. (2005) compared the gene expression profile of ES cell–derived hepatocytes with adult mice liver and found significant similarities in a gene expression profile. Using small interfering RNA (siRNA) technology, they found that HNF-3β/FOXA2 is essential for in vitro hepatic differentiation, which also indicates that this system progresses via endoderm differentiation, imitating hepatic development in vivo: step (0) pluripotent ES cells, step (1) endoderm specification (HNF-3β), step (2) immature hepatocytes (AFP, ALB), and step (3) mature hepatocytes (ALB, TDO2).
Selection of ES cell-derived hepatic progenitors.
BMP-4 has been found to be crucial in the hepatic specification of the mouse ES–derived definitive endoderm (Gouon-Evans et al.,2006). When treated with Activin A in serum-free condition, mouse ES cells generate an endoderm progenitor population that co-expresses T Brachyury, HNF-3β, c-kit/CD117, or c-kit/CD117 and CXCR4. Those progenitors, when treated with BMP-4, FGF2, and Activin A, developed into a highly enriched population (45–75%) of functional hepatocyte-like cells, positive for AFP and ALB, producing albumin and storing glycogen. The authors demonstrated that BMP-4 is required to induce a hepatic fate in the GFP-Brachyury+/CD4-FOXA2high/c-kithigh embryoid body population. Those progenitors, when treated with FGF2 and Activin A, developed into a highly enriched population of functional hepatocyte-like cells, producing albumin. After transplantation into injured livers, they engrafted and proliferated within the parenchyma. Another study concerning hepatic progenitors derived from ES cells was presented by Heo et al. (2006). The authors used mouse ES cells transfected with the GFP reporter gene regulated by the ALB promoter. After culturing these cells in a serum-free chemically defined medium without growth factors or in feeder layers, they observed the formation of EBs and hepatic differentiation. The FACS-sorted GFP+ fraction developed into functional hepatocytes without evidence of cell fusion and participated in the repair of liver injury. Moreover, the GFP+ fraction also differentiated into biliary epithelial cells. This event confirms the success of their selection of progenitor cells. Impressive studies have been published by Sato-Gutierrez et al. (2006). Their differentiation protocol (specifically described by Sato-Gutierrez et al.,2007) utilized EB formation and then treatment with Activin A and FGF2. After that, it required indirect co-culturing with human liver non-parenchymal cell lines (cholangiocytes, stellate cells, and liver endothelial cells) and treatment with HGF, DMSO, and DEX (Sato-Gutierrez et al.,2006). The final step involved sorting of the most hepatocyte-reminiscent cell population (GFP+ cells under the control of the ALB promoter). Those cells displayed metabolic capacity (albumin production and ammonia detoxification) and rescued animals with 90%-hepatectomized livers.
Human ES Cells.
Experiments involving human ES cells are limited because of obvious ethical concerns. Nonetheless, the number of published reports using human ES cells is rising. Schuldiner et al. (2000) showed the potential of human ES cells to differentiate into three embryonic germ layers after stimulation with different growth factors. EBs were dissociated and plated onto fibronectin-coated dishes and treated with growth factors, none of which induced differentiation into any specific cell type (Schuldiner et al.,2000). Rambhatla et al. (2003) used sodium butyrate to induce hepatocyte differentiation in human ES cells through EB formation. The characteristics of hepatocyte morphology and glycogen accumulation have been observed; however, sodium butyrate induced significant cell death. Levenberg et al. (2003) used biodegradable scaffolds of PLGA-poly(lactic-co-glycolic acid) and PLLA-poly(L-lactic acid) to induce tissue-like structures after seeding ES cells or EBs, and they found hepatocyte differentiation after stimulation with Activin A and IGF. Fourteen days after the implantation of 2-week-old constructs into SCID mice, immunostaining analysis of CK and AFP indicated that the implanted constructs continued to express these human proteins. Cai et al. (2007) used a strategy of sequential treatment (Activin A+ ALB/Activin A + ITS/FGF4 + BMP-2/HGF/OsM + DEX) under a serum-free condition. The generated hepatocytes revealed hepatocyte-specific functions. In addition, they were readily infected by the human immunodeficiency hepatitis C virus pseudotype. Another report also showed effective sequential treatment of human ES cells using a conditioned medium supplemented with FGF2, an unconditioned medium supplemented with DMSO, then HCM supplemented with HGF and EGF, and finally HCM supplemented with HGF and OsM (Hay et al.,2007).
Despite the great potential of ES cells, their use has been limited because of difficulty regarding their availability, problems related to histocompatibility, and ethical issues.
In addition, there is a problem with post-transplantation uncontrolled differentiation followed by tumorigenesis. Such handicaps might be sidestepped in the future by somatic cell nuclear transfer of a patient's own skin cells into donated oocytes (Fu,2007; Gurdon et al.,2003). Notably, genetic manipulations on mouse fibroblasts in order to achieve pluripotency have succeeded (Takahashi and Yamanaka,2006). The authors transfected mouse embryonic fibroblasts with Oct3/4, Sox2, c-Myc, and Klf4. Afterwards, they generated ES cells and cloned animals from fertilized mouse eggs (Okita et al.,2007). The use of ES cell–like cells (induced pluripotent stem iPS cells) instead of human ES cells might provide wide applications in the field of regenerative medicine. Additionally, donor oocyte mitochondria may present another technical hurdle since they will put non-self antigens in the nuclear transfer cells (Bowles et al.,2007; Harvey et al.,2007).
Further investigations concerning genomic stability, differentiation fidelity, and cellular “reprogramming” need to be performed in order to verify the safety of using ES cell.
Adult Stem Cells
Nearly all postnatal organs and tissues contain populations of stem cells, which have the capacity for renewal after trauma, disease, or ageing. In adults, there is a spectrum of stem cells with a different scale of quantity and potentiality (uni-, di-, tri-, and multi-potent). They are ready to receive signals from circulating blood to control homeostasis. Some of the adult stem cells have already shown their ability for hepatocyte differentiation (Fig. 1). Initially, research mainly focused on bone marrow (BM) as a source of stem cells for liver regeneration when its contribution to liver regeneration in vivo was described. Petersen et al. (1999) showed that the transplantation of unfractionated male BM into the livers of lethally irradiated female rats, whose livers had been injured by 2-acetylaminofluorene and CCl4, rescued the animals from radiation-induced BM ablation and simultaneously produced small numbers of BM-derived hepatic stem cells. Later on, many studies were performed to confirm the contribution of these cells to liver regeneration on unfractionated and fractionated BM-derived stem cells (Alison et al.,2000; Lagasse et al.,2000; Petersen et al.,1999; Theise et al.,2000a,b). However, some researchers postulate that this event is not transdifferentiation but, rather, a nuclear transfer via cell fusion between BM-derived stem cells and recipient hepatocytes (Vassilopoulos et al.,2003; Wang et al.,2003). At the present time, some researchers have difficulty differentiating BM stem cells into liver cells (Rountree et al.,2007; Vig et al.,2006) and others claim that HSCs transdifferentiate towards hepatocytes without cell fusion (Jang et al.,2004). A relevant summary describing the contribution of BM-HSCs in liver regeneration can be found in a review by Thorgeirsson and Grisham (2006).
An adherent fraction of cells, first found in the stroma of BM, so-called mesenchymal stem cells (MSCs), colony-forming units-fibroblasts (CFU-F), and stromal cells, emerged as a remarkable tool for regenerative medicine (Pittenger et al.,1999). Such cells have also been found in other than BM sources, such as amniotic fluid (De Coppi et al.,2007), placenta (In't Anker et al.,2004), adipose tissue (AT) (Zuk et al.,2001,2002), umbilical cord blood (UCB) (Bieback et al.,2004), and many fetal tissues and organs (Campagnoli et al.,2001; In't Anker et al.,2004). Even though isolated from different tissues, these cells share similar surface markers with small variability (SH3+, CD29+, CD44+, CD71+, CD90+, CD105+, CD106+, CD120a+, CD124+, CD14−, CD31−, CD34−, and CD45−) (Kern et al.,2006; Wagner et al.,2005). Such small variations are related to the tissues from which they have been obtained, the donors, and the age of the donors. The ability of MSCs to differentiate toward hepatocytes has been confirmed in many independent studies (Table 2). The possibility for their future application in the therapy of liver diseases is very likely (Fig. 4). MSCs can be easily obtained from a patient's own tissues, isolated ex vivo, expanded, differentiated toward hepatocytes, and transplanted back into the patient. In the future, after the development of tissue-engineering technologies, they might be implanted back into the patient as 3D cultures (Huang et al.,2006) or liver devices (Sato-Gutierrez et al.,2006). Such a possibility sidesteps the limits regarding ethical and immunocompatible problems. Currently, attention is being given to adipose tissue (AT) as a source of MSCs for regenerative medicine (Banas et al.,2007; Seo et al.,2005; Talens-Visconti et al.,2006; Zuk et al.,2001,2002). From AT, a sufficient number of stem cells for stem cell–based therapy may be obtained without invasiveness or damage to a patient's health. Below, we present the current strategies of hepatocyte differentiation from MSCs from different sources, emphasizing those strategies that utilize MSCs from sources of high availability, as well as pointing out the functions of generated hepatocytes.
Table 2. Differentiation Potential of Mesenchymal Stem Cells Towards Hepatocyte-Like Cellsa
Schwartz et al. (2002) showed that rat, mouse, and human BM–derived multipotent adult progenitor cells (MAPCs), cultured with FGF4 and HGF on Matrigel, can differentiate into cells expressing several liver-specific markers. These cells had functional characteristics of hepatocytes, e.g., albumin and urea secretion and CYP activity. Sato et al. (2005) showed that human BM-MSCs xenografted into liver of a rat treated with allylalcohol differentiated into human hepatocytes, which express liver-specific markers, without cell fusion. These studies excluded spontaneous fusion between human MSCs and rat hepatocytes, by identification of both human and rat chromosomes using fluorescence in situ hybridization (FISH). Lee KD et al. (2004) demonstrated the evidence of hepatogenic induction of BM-MSCs using sequential treatment with factors. They showed the characteristics of hepatocyte-like cells, including albumin and urea secretion, glycogen storage, low-density lipoprotein (LDL) uptake, and CYP activity. Using the same strategy, they obtained similar results by differentiating umbilical cord blood (UCB) MSCs (Lee KD et al.,2004). Other studies using sequential treatment have been performed by Snykers et al. (2006). They have used HGF, FGF4, DEX, and ITS and achieved functional hepatocytes. Other interesting studies have been conducted by Ong et al. (2006). Prior to treatment with HGF and OsM, these researchers co-cultured human BM-MSCs with rat liver slices derived from gadolinium chloride (GdCl3)-treated rats. They detected a significant alteration in hepatocyte functions, such as albumin and urea production, after co-culture with liver slices derived from intoxicated animals. This fact may suggest the use of some pro-inflammatory, such as tumor necrosis factor (TNF) α, in order to upgrade the differentiation ratio. The next two reports show the therapeutic effect of BM-MSCs on liver injury (Oyagi et al.,2006; Zhao et al.,2005). Oyagi et al. (2006) showed that, after treatment with HGF, transplanting BM-MSCs into a rat with CCl4 intoxication resulted in improved liver function (albumin and glutamic-oxaloacetic transaminase [GOT] levels) and decreased fibrosis. Similarly, Zhao et al. (2005) have demonstrated a protective effect of MSCs isolated from rat BM on fibrosis caused by CCl4 and DMN.
Human UCB has already been shown to be applicable to the clinical treatment of various hematopoietic diseases. As well as HSCs, UCB contains stem cells with the properties of MSCs (Kern et al.,2006; Wagner et al.,2005). It has been reported that, after stimulation with HGF, DEX, and OsM, UCB-MSCs differentiate into hepatocyte-like cells (Hong et al.,2005; Lee OK et al.,2004). Functional hepatocytes, together with BM-MSCs, were generated by another group (Lee KD et al.,2004). Kang et al. (2005) used FGF4 and HGF in a 1% (fetal bovine serum) FBS supplemented medium. They generated functional hepatocytes as well. Wang et al. (2005) sorted the β2m−, c-Myc+ fraction of UCB-MSCs, which, after co-culture with non-parenchymal liver cells in the presence of HGF, differentiated into functional hepatocytes secreting ALB and urea. Since UCB is normally discarded at birth, the highly proliferative UCB-MSCs can be obtained easily from newborns and used for therapy.
Adipose Tissue Mesenchymal Stem Cells (AT-MSCs).
AT represents an attractive source of MSCs for future stem cell–based therapy (Fig. 4) (Zuk et al.,2001,2002). Large numbers of human AT-MSCs can be obtained with minimal invasiveness. The ability of AT-MSCs to differentiate into hepatocytes was first observed by Seo et al. (2005). They used HGF, OsM, and DMSO and demonstrated that generated hepatocytes had the ability to uptake LDLs and synthesize urea. Next, the protocol of differentiation of BM-MSCs and UCB-MSCs toward hepatogenic lineage by Lee KD et al. (2004) was also confirmed on AT-MSCs (Talens-Visconti et al.,2006). Our group recently reported hepatogenic differentiation of AT-MSCs from cancer patients' own CD105+ AT-MSCs (Banas et al.,2007). Although AT-MSCs are a more heterogeneous fraction than BM-MSCs, after selection/sorting, we obtained a homogeneous, highly proliferative, and potent population of stem cells. The hepatic differentiation protocol was based on previously established ES cell differentiation by Teratani et al. (2005). The generated cells revealed hepatocyte-specific morphology (Fig. 5B), markers detected by RT-PCR, immunostaining (Fig. 6B), and functions such as ammonia detoxification and albumin production (Banas et al.,2007). Xenobiotic metabolic enzymes, such as CYP1A1, CYP3A4, CYP2C9, and NADPH P-450 reductase, were detected by Western blot analysis. Adipose tissue is an attractive source of MSCs for liver therapy. Important to note is the fact that undifferentiated AT-MSCs express some hepatocyte-specific markers, which raises the possibility that AT-MSCs might contain a large number of hepatic progenitors or that they are predisposed toward a hepatic lineage because they are in close vicinity to the liver. These speculations need evaluation; however, they seem very intriguing and promising.
Mesenchymal Stem Cells From Other Sources.
Of interest are recent studies demonstrating the evidence of hepatogenic differentiation of MSCs from amniotic fluid (De Coppi et al.,2007). Those multipotential stem cells were identified by immunoselection of c-kit/CD117, and they expressed both mesenchymal (CD29, CD105) and embryonic (SSEA-4, Oct3/4) stem cell markers, while hematopoietic (CD45, CD34) stem cell markers were not expressed. The studies showed evidence of hepatogenic differentiation, which expressed the hepatic marker ABC transporter MDR1 and HNF-4α and revealed the ability to synthesize urea. Until now, amniotic fluid similar to UCB has been a medical waste; instead, it might be used as a stem cell source to cure liver dysfunction.
In addition, MSCs from placenta, by culturing in a hepatocyte culture medium, have been differentiated into hepatocyte-like cells (Chien et al.,2006).
Mesenchymal Stem Cells Are an Ideal Source for Stem Cell Therapy.
Because it is easy to obtain from patients, the autologous MSCs are a particularly promising source of cells for many clinical applications in the evolving field of regenerative medicine. Furthermore, MSCs represent an advantageous cell type for allogenic transplantation since MSCs are immune-privileged, with low MHC I and no MHC II expression, therefore reducing the risks of transplant rejection. MSCs have been found to be immunosuppressive, through a mechanism thought to involve paracrine inhibition of T- and B-cell proliferation and as such have been used in trials investigating their effects on autoimmune diseases and graft-versus-host disease (GVHD) (Ucelli et al.,2007). Co-infusion of donor-derived MSCs together with HSCs has been shown to reduce the incidence and severity of GVHD in sibling allografts (Le Blanc et al.,2007). It was reported that a patient suffering from severe therapy-resistant GVHD was treated with human AT- MSC and revealed a complete recovery (Fang et al.,2007). The hypo-immunogenic properties of MSCs are considered to be sufficient to allow transplantation even between individuals who are not HLA-compatible. Therefore, MSCs represent a population of cells with the potential to contribute to future treatments for a wide range of acute or degenerative diseases including liver.
STEM CELL DIFFERENTIATION TOWARDS CHOLANGIOCYTES
The liver consists of two types of endodermal epithelial cells, hepatocytes and cholangiocytes. Both cells originate from bipotential hepatoblasts (OCs) (Beltrami et al.,2007; Evarts et al.,1987; Farber,1956; Fausto,2004; Fougere-Deschatrette et al.,2006; Herrera et al.,2006; Kojima et al.,2005; Shafritz et al.,2006; Wang et al.,2003; Yovchev et al.,2007; Zhou et al.,2007). Although cells derived from bone marrow cells (rat), MSCs (mouse), or HSCs (mouse) can differentiate into hepatocytes, whether these non-hepatic cells can directly differentiate into cholangiocytes remains controversial (Krause et al.,2001; Moritoki et al.,2006; Theise et al.,2000a,b). Little is known regarding the molecular pathways required for biliary epithelial cell development. However, it has been reported that the Notch signaling pathway is involved in the development of intrahepatic bile ducts (Kodama et al.,2004). In fact, a primary culture of hepatoblastic cells, which overexpressed a constitutively active form of Notch, resulted in preferential differentiation into cholangiocytes (Tanimizu et al.,2004). Recently, the same group reported that hepatoblastic cells can be converted into cholangiocyte-type epithelial polarity in a matrigel 3D culture (Tanimizu et al.,2007). Despite confidence that hepatoblastic cells and extra-hepatic stem cells may differentiate into cholangiocytes, the mechanistic details of this process remain poorly understood. Moreover, the physiological importance and therapeutic utility of this phenomenon need to be investigated.
THE ROLE OF “EPIGENETIC MODIFIERS” AND “ENVIRONMENTAL NICHES” IN HEPATOGENIC DIFFERENTIATION
The pool of stem cells in the human body has a highly diverse developmental status (Fig. 7). The change from pluripotent to multipotent and to more developmentally restricted states is accompanied by global changes in gene expression. Genes active in earlier progenitors are gradually silenced at developmentally later stages, and a subset of cell type-specific genes is turned off (Dennis and Charbord,2002) (Fig. 7).
This progression is the result of the selective expression of transcription factors together with chromatin remodeling and modification (involvement of histone acetyltransferases [HATs] and histone deacetylases [HDAC]) and DNA methylation of CpG dinucleotides, localization of chromatin to specialized nuclear domains, and the recently discovered miRNA involvement. These epigenetic mechanisms allow cells to maintain their identity even when exposed to signals that would induce alternative cell fate in less determined cell lineages.
None of the studies we have discussed address epigenetic mechanisms in liver stem cell biology. Some progress on the mechanisms involved is beginning to be published. For example, the role of epigenetic modifiers (hypomethylating agent 5-aza-2-deoxycytydine [5azadC] and deacetylating agent-histone deacetylase inhibitor Trichostatin A [TSA]) has been shown in neurogenic differentiation (Alexanian,2007). In hepatogenic differentiation, factors such as TSA (Snykers et al.,2007), DMSO (Sato-Gutierrez et al.,2006,2007; Seo et al.,2005) and sodium butyrate (Rambhatla et al.,2003), which are the inhibitors of HDAC, are often used to manipulate cell fate. Importantly, a report has been published regarding the alteration of hepatic differentiation of MSCs using TSA (Snykers et al.,2007). Nevertheless, generalizations about the developmental mechanisms of epigenetic change associated with potency restriction have not emerged; at present this important area of study is enigmatic.
A parallel important area of research is the interaction of stem cells with their physical environment as they travel to and/or arrive at their final destination. Specifically, the connective tissue/matrix environs of niche compartments must be intensively examined. The cells that make up and create the niche are as important as the stem cells that occupy and respond to the niche environment. Here again, generalizations regarding developmental and molecular mechanisms do not exist.
The pool of mesenchymal stem cells (MSCs) within the human body may be a highly effective tool to realize the potential of regenerative medicine using autologous cells. To do this, an understanding of the molecular basis of the stability of the differentiated state and of niche compartments, along with elucidation of stem cell mobilization and homing, must be in place. The interesting unknown is whether what works to reconstitute the hepatic system with stem cells will yield insight into the reconstitution of any other organ, or that, much like the hoped-for cure for cancer, regenerative medicine with stem cells will be done one organ system at a time.
CONCLUSIONS AND FUTURE PERSPECTIVES
The field of hepatic stem cell study has undergone tremendous growth during the past decade. Despite many reports on the differentiation potential of stem cells, there is little understanding of the molecular basis of stem cell plasticity. Array-based gene expression analyses of “stemness” have been reported (Boquest et al.,2005; Ramalho-Santos et al.,2002; Urs et al.,2004). Under normal conditions, the expression of “stemness genes” is tightly regulated by a dynamic array of mediators, including the spatial and temporal expression of inhibitors and the epigenetic modulation of the genome. Once stem cells are exposed to microenvironmental cues of tissue or organ injury and regeneration, the balance of regulatory mediators is restored, with the plasticity of stem cells being induced towards differentiation into a specific cell lineage. In the natural milieu, the hepatic differentiation of stem cells involves multiple pathways. This may be mimicked in vitro by using a combination of various factors and techniques. It is anticipated that, over the next few years, we will see profound investigations of hepatic stem cells and, particularly, of their mechanisms, transduction pathways, and epigenetic modulations. Hepatic differentiation may certainly be enhanced by further studies and a combination of various techniques, including tissue-engineering technologies. However, while stem cell plasticity is certainly a hot issue these days, its potency for clinical application is even hotter. In considering stem cell–based therapy, MSCs have emerged with great potential. An assortment of studies has documented their contribution in hepatogenic generation in vivo and in vitro. Preliminary results of only a few clinical studies on the administration of bone marrow stem cells to liver cirrhotic patients seem to be very promising, but additional well-designed and controlled studies are needed.
This work was supported in part by a Grant-in-Aid for the Third-Term Comprehensive 10-Year Strategy for Cancer Control; Health Science Research Grants for Research on the Human Genome and Regenerative Medicine from the Ministry of Health, Labor, and Welfare of Japan; and a Grant from Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists. Y.Y. is a Research Fellow of Japan Society for the Promotion of Science (JSPS).