THE ROAD TO REGENERATIVE LIVER THERAPIES: THE TRIUMPHS, TRIALS AND TRIBULATIONS

INTRODUCTION

The liver is the largest internal organ in an adult organism. It performs many important functions that sustain the organism’s vitality. Liver metabolizes nutrients from ingested food and regulates glucose levels by converting excess glucose to glycogen for storage and releasing it when the blood glucose level is low. It detoxifies xenobiotics and harmful metabolites, and synthesizes many proteins in the blood [1]. Most of these functions are provided by the parenchymal cell type, hepatocytes, that comprise approximately 70% of the adult liver mass.

Liver failure may arise from many causes, including cirrhosis, viral infections and drug overdoses. Typically, the liver has a tremendous regenerative capacity to repair itself. After partial hepatectomy to remove two-thirds of the liver surgically, it is capable of regaining its original mass over time. However, liver disorders can compromise its inherent regenerative capacity and result in complete liver failure leading to death. Although treatment of the symptoms can alleviate the severity of liver failure, organ transplantation is the only curative treatment. However, a severe shortage of donors has limited the access of liver transplants for many patients. As of 2012, there are approximately 17,000 people on the waitlist for liver transplantation in the United States alone, while only half the number of transplantations were performed annually because of the shortage of donor organs (United Network for Organ Sharing: http://optn.transplant.hrsa.gov).

Extracorporeal liver devices have been explored as a treatment to sustain patients until successful liver regeneration, or until a donor organ becomes available. These extracorporeal devices comprise of hepatocytes from a variety of cell source (porcine, human, etc.) as well as mechanical components to provide temporary assistance [2]. The mechanical components of the device employ filtration, adsorption or dialysis to remove small molecular weight toxic metabolites from the patient’s blood, while hepatic cells provide the bio-transformative and biosynthetic functions [2, 3]. Other approaches of liver failure treatments include transplantation of dissociated hepatocytes from organs and implantation of tissue engineered liver analogues to augment liver’s regenerative capacity for liver recovery [4–6].

For applications involving liver cells such as extracorporeal devices, cell transplantation and tissue engineering, primary human hepatocytes have been the preferred cell source because of its low risk of immunogenicity. The use of isolated liver cells can potentially expand the pool of donor organs, as even organs unsuitable for transplantation may be suitable for use in hepatocytes transplant. However, difficulties in expanding and maintaining primary hepatocytes in culture still remain a major hurdle in this field. Even with expanded pools of donor organs as the source of hepatocytes, the need still exceeds the availability of hepatocytes. Furthermore, functional capabilities decrease rapidly during in vitro culture [4]. In addition to maintaining our desired cell type, we must also address the need for large quantities of primary cells are needed for the treatment of even a single patient.

Hepatocytes isolated from other species, primarily porcine, may provide an alternative source, however, these cells also suffer from rapid decrease of functional activities when cultured in vitro similar to primary human hepatocytes. Moreover, the differences in their drug metabolism and other hepatic functions with human hepatocytes, along with potential immunogenic concerns, render these xenogeneic hepatocytes less than desirable compared to human sources [3].

For future medical applications of liver cells, including cell therapy and extracorporeal liver assist devices, in vitro cultivation is most likely to be employed to expand the supply of human cells. These expanded cell population can then be guided to differentiate to the desired cell type for specific applications. In the past few years, stem cell research has made significant advances; stem cells and progenitors cells can now be isolated from various sources, and expanded and differentiated towards the liver lineage. This has brightened the prospect of generating large numbers of functional hepatocytes for applications in hepatic cell transplantation, extracorporeal liver-assist devices and liver tissue engineering. In this article, we will highlight those advances and the path forward for transforming these protocols into standard clinical therapies.

Embryonic Liver Development-the guide for in vitro culture processes

In this section, we will describe the development of mouse liver, as an example of mammalian development, being cognizant that the development in mouse and man differs in certain aspects.

In early embryo development, the blastocyst consists of an inner cell mass and an outer layer of trophoblast cells. As the primitive blastocysts become polarized and exposed to a number of signaling pathway cues, they will give rise to the inner cell mass [7, 8]. During this developmental stage, embryonic stem cells can be isolated from the inner cell mass which can give rise to all three germ layers. The inner cell mass will further differentiate to two specialized cell type, hypoblast and epiblast cells [9]. Hypoblast cells will give rise to extraembryonic tissues, while epiblast cells will make up all the tissues in an adult by first differentiating to ectoderm and mesoendoderm [10]. During this time, the formation of the primitive streak will set in place the bilateral symmetry and anterior-posterior axis indicating the start of gastrulation [11]. This event marks the beginning of the delineation between the three germ layers, ectoderm, endoderm, and mesoderm, leading up to organogenesis.

The differentiation of the intermediate stage, mesoendoderm, is driven primarily by Nodal signaling, a member of the transforming growth factor (TGF-β) family, as demonstrated in an explant model [12–14]. The Nodal protein acts as a morphogen; high levels promote endoderm formation and low levels promote mesoderm formation. In the developing mouse embryo, Nodal is produced at the anterior region of the primitive streak where it can exert its effect through a number of downstream transcription factors, including Foxa2, Sox17, Gata4-6, Mixl1 and Eomesodermin [15]. The expression level of the genes regulated by these transcription factors delineates the difference between endoderm and mesoderm. The importance of Nodal signaling for endoderm commitment was demonstrated in multiple transplantation studies that showed ectopic regions expressing Nodal signaling can induce cells to express endoderm markers and differentiate further into endodermal derivatives [16, 17].

The epithelial layer of endoderm, in close contact with a thick layer of mesenchymal cells called the septum transversum, will give rise to the digestive and respiratory organs, including the liver. The processes by which cells undergo a massive transformation into a multi-layered group of cells from the blastula are regulated by several signaling pathways. The most widely studied pathway, FGF signaling, causes cells to undergo an epithelial to mesenchymal (EMT) transition by decreasing the amount of cell-cell adhesion [18, 19]. The decrease allows cells to expand and spread out to form new layers. The initial stage of gastrulation involves the invagination of the epithelium which results in the cell movement to subdivide the gut tube into foregut, midgut and hind gut regions. In the ventral region of foregut, the cardiogenic mesenchymal cells adjacent to the endoderm secrete several fibroblast growth factors (FGFs), in conjunction with bone morphogenic proteins (BMP)(also members of TGF-superfamily) produced by septum transversum mesenchyme (STM) cells, to induce hepatic specification [20–24]. Fast proliferating hepatoblasts then emerge and commit this segment of endoderm to develop into the liver bud.

Hepatic endoderm cells or hepatoblasts are bi-potential and can undergo differentiation to hepatocytes or biliary epithelial cells that line the lumen of the intrahepatic bile ducts. Hepatoblasts are capable of proliferating extensively and invading the surrounding septum transversum. Endothelial cells then interact with these hepatoblasts by providing specific growth factors needed for hepatoblasts maintenance and proliferation [25]. Between E10–15 in mice, the formation of the liver bud undergoes rapid growth and vascularization. By day E14–15, the liver bud is a highly vascularized tissue. During this stage of liver development, the STM continues to provide BMP, and additionally expresses hepatocyte growth factor (HGF), while the hepatoblasts express c-Met, the HGF receptor [26]. HGF acts as a suppressor for the apoptosis of hepatoblasts by promoting hepatoblasts proliferation through Wnt3a [27]. It has been speculated that both HGF and BMP provide growth signals, perhaps through parallel pathways. In addition to HGF, Oncostatin is released by the haematopoietic cells which promotes hepatocyte differentiation and maturation by the JAK/Stat3 signaling pathway through activation of the gp130 receptor [28]. The importance of this finding will be illustrated in sections below when we discuss the selective growth factors for in vitro differentiation.

The functional cells slowly undergo maturation and the biliary network is formed as the liver attains the appropriate tissue architecture and functional capability [1, 29, 30]. The detailed mechanisms behind liver bud development and subsequent maturation of the hepatocytes have been investigated further using microarray analysis on samples taken through various timepoints of mouse embryonic development [31].

Fetal liver cells

Progenitor cells derived from the fetal liver may be a promising source of hepatocytes for liver cell based therapy. In embryonic liver development, hepatoblasts, a common liver lineage committed cell type, can give rise to both hepatocytes and cholangiocytes. These hepatoblasts exhibit a very much larger proliferative potential compared to primary hepatocytes because of its less differentiated state. Furthermore, they are less prone to rapid de-differentiation in culture than primary hepatocytes and are more suitable for transplantation [32]. Thus there has been a long sustained interest in their isolation and in vitro cultivation. However, because these cells are isolated from fetal livers, their study has mostly been performed in animal models.

Hepatic progenitor cells have been isolated from fetal rodent livers around E11 to 14.5 by flow cytometry or magnetic activated sorting of dissociated liver cells based on the expression of one or a combination of surface markers. Early cell sorting studies used the absence of CD45 and TER119 along with low c-kit expression to sort for these types of early progenitor cells [33, 34]. Later, Delta-like protein 1 (DLK1), which begins its expression in the mouse liver bud around E10.5, was used as a surface marker to enrich for hepatoblasts from rat [35] and mouse fetal livers [36, 37]. E-cadherin, a cell adhesion protein [38] and epithelial cell adhesion molecule (EPCAM), a cell surface glycoprotein [39] have also been used to isolate hepatoblasts. These fetal liver derived progenitor cells also express cytokeratin 19 (CK19) and albumin. In culture, these cells also possess the capacity of proliferation with the supplementation of growth factors such as HGF and EGF, and are capable of differentiating to both hepatic lineage and cholangenic lineage in culture [38].

These progenitor cells have also been isolated from fetal human liver. The Fausto laboratory isolated clonal progenitor cells from clusters of small cells arising from long term primary cultures established from fetal livers between the gestational stage of 74 and 108 days [40]. These cells exhibited many markers reported in rodent hepatoblasts, including CK19, EPCAM, C-kit. However, maintained on feeder layers of irradiated NIH 3T3 cells, they appear to be pre-hepatoblasts as they express neither albumin nor α-fetal protein. They also lacked the transcripts of a number of liver transcription factors (HNF1α, HNF3α and HNF4α) and are capable of differentiating to not only hepatic lineage but also mesenchymal lineage. Thus, they claimed that the characteristics of these cells made them possibly closer to the more primitive mesendoderm stage of development.

Using EPCAM and ICAM as markers, hepatic progenitors or hepatic stem cells were isolated from both human fetal and postnatal (neonatal, pediatric and adult) liver [41, 42]. In the absence of a supporting feeder layer, EPCAM⁺NCAM⁺AFP⁻ICAM1⁻ cells became predominant. These cells were termed human hepatic stem cells (hHpSC). When cultured on STO feeders (a mouse embryonic fibroblast line) cells, hHpSC will differentiate into EPCAM⁺NCAM⁻AFP⁺ICAM1⁺ hepatoblasts [41]. The proliferative capacity of these hepatoblast cells were demonstrated when the authors showed the maintained phenotype for over 150 doublings in culture.

Adult liver stem cells

Hepatocytes in an adult liver are capable of being highly regenerative when the liver experiences acute organ failure. Upon acute liver injury or partial hepatectomy, they replicate to generate the lost cell mass. Through serial transplantation of cells into the liver of metabolically deficient transgenic mice, hepatocytes isolated from adult mouse liver have been shown to proliferate up to 70 doublings and continue to contribute to the liver mass [43]. However, hepatocyte replication is not the only mechanism by which the liver can regenerate. In situations of chronic insult or severe liver injury, the regenerative capability of adult hepatocytes is overwhelmed and the resident hepatic stem cells will become the key contributor to the repair of the damaged hepatic tissue in a process known as ductal reaction [32, 42]. These proliferative cells are often referred to as oval cells due to their shape in rodent livers.

Thus, both hepatocytes and stem/progenitor cells in adult liver are capable of proliferation and reconstitution of lost liver mass. Such capability can potentially be harnessed for in vitro culture and clinical applications. However, as briefly mentioned before, hepatocytes in culture lose their high proliferative capacity. In contrast, many have focused our ability to isolate and culture liver stem cells or progenitor cells in hopes of utilizing these specialized cell types towards clinical applications.

Study of liver stem/progenitor cells often employs partial hepatectomy or treatment with drugs like acetylaminofluoren in rodents to enrich the cell population in the liver [44]. By BrdU labeling of proliferative cells in regenerating livers of rodents after induced acetaminophen injury, the region that stem cells reside and their niches were identified [45]. Four differently labeled stem cell populations were identified, namely periductal mononuclear cells and peribiliary hepatocytes along with cells in the canal of Hering and intralobular bile ducts [45]. These findings suggest that there may be more than one cell type that gives rise to the regenerative process. Instead, multiple cell populations may possess this ability and can respond to their specific niche cues differently allowing for their participation in the liver repair process.

The availability of genomic data and the enhanced capability to identify genes expressed differentially under various differentiation states has helped identify markers suitable for cell isolation. The progenitor/stem cell markers identified in rodents has been used to enable for selective isolation of their counter parts in human liver. For example, EPCAM has been a key marker for enriching both fetal and adult liver stem/progenitor cells in rodents and human [41, 46]. However, through the detection of high level expression of aldehyde dehydrogenase using a fluorescent substrate, adult liver stem cells from both mice and human liver were isolated without using any chemical stimuli to enrich their population [47]. This methodology could possibly provide a readily robust method of isolating a relatively homogenous population of hepatic progenitor cells.

A recent review by Turner et al. provides a more in-depth look at the current understanding of the localisation of liver stem cells in the liver lobule and their characteristics [42]. They comprehensively describe the progression of liver stem cells during the development of the liver. It is postulated that the hepatic stem cells reside in the periportal region of the liver, specifically the canal of Hering. As these cells proliferate, “older” cells migrate along the cell plate toward the central vein and grow more committed and mature; first becoming bipotential, then becoming proliferative hepatocytes, and mature hepatocytes. As they move toward the central vein, they also become larger in size. Further details regarding the origin and repair mechanisms are described elsewhere [32, 42, 44].

Figure 1 summarizes the current stem cell/progenitor cells populations which have been isolated from stem cell hepatic differentiations, fetal liver, and adult liver. Mulitpotent cells from human fetal liver reported by the Fausto lab [40] are more primitive than the bipotential hepatic stem cells characterized by the Reid lab owing to the former’s capability to differentiate to mesenchymal cells.

Figure 1.

Stem cells and progenitor cells of liver lineage isolated from fetal and adult liver and from in vitro directed differentiation of pluripotent stem cells. The various surface antigens used for cell isolation, when applicable, are also shown. Studies involving human systems are in black while those involving rodent systems are in grey.

The hepatic stem cells have been isolated from both fetal liver and adult liver; whereas the mesonendodermal cells have only been reported from human fetal liver. Both precede the hepatoblast; although both share many markers of hepatoblasts including EPCAM, E-cadherin, CK8, CK18, and CK19, both are also AFP and albumin negative. A distinction between hepatic stem cells and hepatoblast cells is their NCAM/ICAM-1 expression; ++/− and −/++ for hepatic stem cell and hepatoblast respectively.

Hepatic stem cells isolated from human adult liver are about 7–10 μm in diameter [42, 47]; somewhat smaller than hepatoblasts in both adult and fetal liver (10–12 μm). They exhibit a high nucleus to cytoplasm ratio. Hepatoblasts are abundant in fetal liver (~80%), but decreases after birth to less than 0.01% in adult liver. In contrast, hepatic stem cells remain relatively stable comprising of about 0.5 to 1.5% of the liver cell mass throughout [48].

Hepatic growth factors HGF, EGF are often included in the medium for hepatoblasts. In addition to growth factors, surface properties also affect the outcome of cell fate. Schmelzer et al. reported that both hepatic stem cells and hepatoblasts grew out from EPCAM sorted cell population when plated on plastic surface, but only hepatic stem cell colonies emerge when plated on STO feeder layer [41]. Turner et al report that angioblast feeder layers or collagen III aid in proliferation of hepatic stem cells while stromal feeder cells or collagen IV aid in hepatoblasts proliferation [42].

Differentiation of hepatic progenitor cells isolated from liver

Hepatoblasts and liver stem cells isolated from fetal and adult livers have been shown to differentiate to hepatocyte-like cells in culture. A combination of growth factors, HGF, EGF, Oncostatin and Dexamethasone administered over one week was used to induce hepatic differentiation for mouse hepatoblasts [47]. Hepatic progenitor cells isolated from humans can be expanded and differentiated to hepatocyte-like cells with many liver functions including urea synthesis, albumin secretion and CYP activity. Decellularized liver biomatrix was used as an ECM for human differentiations, thereby reducing the need of commonly used growth factors HGF and EGF [47, 49].

DIFFERENTIATION TO HEPATIC LINEAGE FROM STEM CELLS

Due to the liver’s high regenerative capacity, there has been a long history of research on liver stem/progenitor cells and on their in vitro expansion. Those attempts aim to expand cells already committed to a developmental path. In contrast another set of efforts seek to direct uncommitted embryonic stem cells, to the hepatic lineage. Embryonic stem cells have unlimited self-renewal and differentiation capacity thereby resulting in a potentially unlimited supply of hepatocytes. The recent emergence of iPSCs only amplified efforts towards generating functional hepatocytes from pluripotent stem cells.

IN VITRO DIFFERENTIATION TO LIVER LINEAGE

Differentiation of ES and iPS cells to hepatic lineage using soluble growth factors

The development of liver in vivo entails the specification to definitive endoderm (about E8.5 in mouse), followed by commitment to hepatoblast and formation of liver bud, and finally fully differentiated hepatocytes and the emergence of bile ducts. The progression of the development is guided by a number of inductive signals dynamically described briefly above. The early differentiation to mesoendoderm and subsequent distinction of mesenchyme and endoderm is driven primarily by Nodal, BMPs and Activin signaling [50–52]. Further signaling from the FGF and BMP family, specifically BMP4, FGF2, FGF4 induce differentiation to hepatoblasts. After liver bud formation and expansion, the inductive signals of HGF and Oncostatin stimulate the hepatoblasts to differentiate towards hepatocytes.

Most hepatocyte differentiation protocols start with Activin and Wnt3a for a 3–5 day period to induce definitive endoderm commitment [50, 53–55]. It has been shown in mice that Wnt3 signaling can maintain Nodal expression and Activin signaling can replace the Nodal/Crypto signal needed for endodermal formation [56, 57]. FGF and BMP play a significant inductive role in promoting endodermal progenitor maintenance and expansion and will subsequently be used in the second stage of directed differentiation [20, 23, 58, 59]. Further treatment with a mixture of FGFs facilitates the commitment to hepatic fate or the equivalent of the hepatoblast state in liver development based on the in vitro findings of Sekhon et al [60]. The last stage often entails the use of oncostatin or follistatin and HGF mixed along dexamethasone. It was previously reported that oncostatin-treated stem cells would lead to an increased activity level of STAT3 [61]. The use of follistatin or oncostatin is used to favor hepatic differentiation over cholangiocytes, whereas HGF is used universally to mimic the hepatic environment [62]. HGF, on the other hand, was shown to contribute to liver development in a STAT3-independent manner [61]. Depending on the protocol, dexamethasone is often used as a mature hepatic inducer [63].

Most efforts in guiding the differentiation of PSCs to hepatic lineage utilize multiple soluble growth factors to mimic the temporal dynamics of cues in hepatogenesis in vivo. However, it should be noted that the development of the liver and any other tissue in vivo is a continuous process, whereas our efforts to replicate these processes tends to be in discrete stages. Various protocols for guiding the differentiation of stem cells to hepatic lineage are segmented into separate differentiation stages. Thus, each protocol may implement different numbers of stages, with varying duration and medium composition. However, most protocols share common motifs across the duration of their differentiation which is depicted in Figure 2 that illustrates several representative protocols for differentiation to hepatic lineage in vitro.

Figure 2.

Representative protocols for the differentiation of pluripotent stem cells to hepatoci lneage. The stages of differentiation, the corresponding guiding constituents (growth factors and signaling molecules etc.) are shown. The number of stages and the duration of each stage are indicated. Abbreviations: FBS-Fetal Bovine Serum, DEX-Dexamethasone, OSM-Oncostatin M, DMSO–Dimethyl Sulfoxide, KO SR-Knockout Serum, LY294002-inhibitor of PI3K, SB431542-inhibitor of activin/nodal receptor.

Transdifferentiation to live lineage from other tissue cells or stem cells

In addition to the PSCs, a number of extrahepatic stem cells isolated from adult tissues, including adult mesenchymal stem cells (MSC) derived from human bone marrow [64, 65] and adipose tissues [66, 67], were shown to have the capacity to differentiate to hepatic lineage. A protocol calls for treating MSCs with demethylating agent 5′-azacytidine and then culturing them in hepatocyte maintenance medium containing HGF and EGF [65]. After 3 weeks the differentiated cell exhibited hepatic markers including albumin, CYP1A1, CYP3A4. Engraftment was shown upon transplantation into immune-compromised mice which underwent partial hepatectomy. A similar protocol for adipose tissue-derived MSC underwent hepatic differentiation without the demethylation step [67], however, comparison of the effect of demethylation has not been available. The differentiation status of hepatocyte-like cells derived from MSC has not been characterized as well as the hepatocyte-like cells derived from PSCs. Nevertheless the results indicate that the future source of hepatocyte can still potentially be expanded to other adult stem cells in addition to the conventional PSCs.

Directing to liver cell fate through gene transfection

A number of laboratories have demonstrated the differentiation of iPSCs to hepatocyte-like cells using the protocols developed for ES cells [68–70]. It appears that the protocols developed for embryonic stem cells are mostly applicable to iPSCs. This opens the possibility of personalized hepatocyte-like cells for drug screening and even for therapy. However, deriving iPSC is a lengthy process. Recently the feasibility of directly inducing fibroblastic cells to hepatic lineage through gene transfection mediated reprogramming was demonstrated in mice models. A combination of Hnf4α with Foxa1, Foxa2 or Foxa3 was used to transfect and reprogram adult mouse fibroblasts to hepatocytes-like cells which are called iHEP cells [71]. The efficiency was about 1 in 1000 of initially transfected embryonic fibroblasts. Interestingly the resulting iHep cells possess a very high proliferative potential, whereas such proliferative potential has not been reported for hepatocyte-like cells derived from from pluripotent cells. In another study the over-expression of Gata4, Hnf1α and Foxa3 along with p19^Arf inactivation enabled mouse fibroblasts to become iHep cells [72]. In this case inactivation of p19 ^Arf suppresses senescence of fibroblasts to enhance the success of reprogramming. The reprogrammed iHep cells expressed key hepatic genes and other liver functions and are successful engrafted in animal models. The obvious question then is why do the resulting iHep cells have a higher proliferative potential compared to the hepatocyte-like cells derived from pluripotent cells. The differences and similarities between the different types of hepatocyte-like cells derived from stem cells and primary hepatic stem cells have yet to be studied extensively. However, the approach will likely see many in vitro applications if it can be shown to work in reprogramming human cells.

Expansion and differentiation of in vitro derived progenitor cells

All reported directed differentiation of ES or iPS cells typically span a 20–25 day period and give rise to a mixture of differentiated cells that include a high percentage of hepatocyte-like cells at different degrees of maturity as assessed by the various liver specific markers. In the transition of pluripotent state to definitive endoderm stage, typically a significant degree of cell death is seen. Some cell expansion is typically seen in the endoderm stage and in the bipotential hepatic progenitor state. At the end of directed differentiation the differentiated cells typically are nearly confluent on the culture surface [73].

In embryo development, cell expansion accompanies endoderm development, hepatic lineage commitment as well as in the hepatoblast stage. Directed hepatic differentiation of pluripotent stem cells passes through the corresponding differentiation stages of in vitro equivalents of the endoderm and hepatoblast stages. In these stages they are also likely to be capable of proliferation although their proliferative potential has not been quantified. Zhao et al. isolated hepatic progenitor cells from hepatic differentiation of ES cells, using FACS sorting of N-cadherin positive cells [74]. These colony forming cells were cultured on mouse embryonic stromal feeders using serum free DMEM based medium which promoted their expansion. These cells, like their counter parts isolated from fetal and adult livers, were bipotential and could be passaged several times in culture.

Recently endoderm progenitor cells were isolated based on their expression of CXCR4 and CD117 from cultures of directed human ES or iPS cells undergoing directed differentiation to the hepatic lineage [58]. These cells are apparently in an earlier differentiation stage than the hepatoblast stage and are capable of differentiating to both pancreatic and liver lineage. They possess a very large proliferative potential and can be maintained in medium containing BMP4, FGF, VEGF and EGF. These endoderm progenitor cells are at an earlier developmental state than the hepatic stem cells and hepatoblasts; but are not as early as the cell line reported earlier by Fausto’s group that is capable of differentiating to mesenchyme in addition to endoderm.

PROPERTIES OF THE IN VITRO DIFFERENTIATED HEPATOCYTE-LIKE CELLS

The hepatocyte like cells derived from directed differentiation of pluripotent stem cells all have distinct epithelial morphology, express liver specific genes including albumin, alpha-antitrypsin, and a number of cytochrome P450 enzymes. A direct comparison of different protocols is not easy because different combinations of assays were used in each report and the functionality is often evaluated at different levels, transcripts or protein or functional activities. Typically the maturity of the differentiated cells is evaluated by transcript and/or protein levels of marker genes, and assessed by representative characteristic markers of different aspects of liver functions, including drug metabolism, bile secretion, gluconeogenesis, urea synthesis, and protein synthetic functions (serum albumin, alpha antitrypsin). As the genomic tools become widely employed and data on transcriptome, miRNA and epigenetic state of cells at different differentiation states become available, we should gain a better understanding of cell differentiation under different protocols. The hepatocyte-like cells obtained from directed differentiation of pluripotent cells are not mature hepatocytes. While they express higher level of transcripts of liver specific gene, the levels generally fall short of that of primary hepatocytes, especially for cytochrome P450 enzymes. As hepatocytes mature, many enzymes in metabolic pathways switch from the dominant infant isoform to an adult isoform. Gluconeogenesis is one key metabolic pathway that is observed primarily in the liver. Using microarray analysis, we can quantitatively observe that the transcript-level of a fetal-specific isozyme, ADH1A, was significantly higher than its adult isozymes, ADH1C (data not published). Detailed transcriptome analysis then suggests that our current methodology of deriving hepatocytes from stem cells is more representative of the fetal state than the adult tissue.

The capability of in-vitro cultured hepatocyte-like cells to progressing to mature hepatocytes have been demonstrated in animal models as illustrated in several studies [69, 75]. However, the cues necessary to achieve that in vitro have yet to be unveiled.

To increase the percentage of cells differentiated to hepatic lineage and enhance their maturity, transient forced gene expression of hepatic transcription factors was used. Lentiviral tranfection of Foxa2, HNF4α and C-EBPα in adult liver derived progenitor cells increased hepatic maturity and functional capabilities of albumin secretion and glycogen storage [76]. Similarly, by infecting differentiating ES cells with Sox17 at Stage 1, Hex at Stage 2 and HNF4 at stage 3 of hepatic differentiation and further incubation for 11 days, hepatic cells expressing higher transcript levels of all major genes examined was observed [77]. In a follow up study, adenoviral vector with Foxa2 was used to infect differentiating pluripotent stem cells at the definitive endoderm conversion, hepatic specification and hepatic expansion stages, followed by a combination of adenoviral vector-Foxa2 and HNF1 infection to further increase maturation [78]. The cytochrome P450 activities were markedly increased, albeit still lower than those in primary hepatocytes and the level desired for toxicity testing. Nevertheless, the work represents encouraging progress for in vitro generation of mature hepatocytes by directed differentiation.

FROM STEM CELL SCIENCE TO TECHNOLOGY

Deriving hepatocytes for liver disease treatment

The advances in the past few years have demonstrated the potential of deriving functional hepatocytes through cells derived from fetal and adult liver or through directed differentiation from pluripotent stem cells. These possible routes of obtaining hepatocytes are summarized in Figure 1. Those advances enable us not only to direct differentiation to liver cell fate, but also allow us to isolate (and clone if necessary) progenitor cells for further expansion in culture. The expanded cells can then be further directed to differentiate to hepatocytes. Along the path of differentiating from stem cells to hepatocytes, stable progenitor states exist that allow cells to exit the progression toward hepatocytes and be maintained and expanded at those states. Interestingly, despite their different origins stem cells isolated from fetal, adult livers and from differentiating pluripotent cells all seem to exhibit similar stable intermediate states (endoderm progenitor, hepatoblast, early hepatocytes).

Prospect of application of culture derived hepatocytes

Therapeutic applications cell transplantation and bioartificial liver

Because of the prevalence of liver diseases, the severity of liver failure and its lack of curative treatments, it is imperative that we further advance stem cell based regenerative medicine for liver disease treatment. In addition to liver failure treatment, many congenital liver diseases will also benefit from stem cell based therapy. Recently iPSCs derived from patients with alpha1-anti trypsin deficiency were genetically corrected for its mutation using zinc finger technology and differentiated to hepatocyte-like cells with restored α1AT activity [79]. Such work along with the advances in the isolation, cultivation, expansion of liver-lineage-capable stem cells, has raised the hope for clinical applications of these technologies.

Possible applications for therapy include transplantation of liver lineage cells or tissue engineered liver tissue analogues, and extracorporeal bioartificial liver devices. Cell transplantation of exogenous hepatocytes or hepatic precursor cells may be directed to reconstitute liver tissue. They may be engineered prior to transplantation to provide a trophic environment for resident hepatocytes to reconstitute the liver. Many reports have demonstrated the feasibility of cell transplantation, engraftment and even correction of metabolic disorders using hepatocyte-like cells derived from stem cells. However, given the complexity of the functions of liver, the large quantities of cells required and the possible long duration from transplant to recovery, the realization of stem cell based liver cell therapy still faces major challenges. For ex vivo applications, such as extracorporeal bioartificial liver as bridge to transplantation or regeneration, the technical barrier may be easier to overcome. A major challenge is the capability of producing cells to deliver high liver specific activities in a large quantity and at a low cost.

Disease model and drug metabolism testing

Stem cell derived hepatocytes have great potential as in vitro models for studying liver disease mechanisms and drug metabolism. The primary tools to date for such studies are animals and primary human hepatocytes; the former are not ideal model for human disease while the latter is constrained by cell availability and representation of genetic diversity. As experimentation on animal models faces increasing ethical challenges, hepatocytes generated from cultured stem cells are becoming even more attractive. Furthermore, with iPSC and iHep, the investigation can be performed using derived hepatocytes of different genetic backgrounds.

A notable example of a cell-based liver model is the replication of hepatitis C using hepatocyte-like cells derived from stem cells [80–83]. The model should facilitate the screening of inhibitors of HepC virus replication pathways. Given the influence of host genotypes on hepatitis C virus entry, assembly and replication, cell based models representing diverse genetic background will be a welcoming development. Another example is the use of hepatocyte-like cells obtained from iPSCs derived from patients with familial hypercholesterolemia. These cells with a mutation in the low density lipoprotein receptor serve as a model of the deficiency in LDL-cholesterol uptake [84].

Such in vitro applications as a liver analogue for disease studies and drug screenings require the reproduction of biological functions or even the structure. Like other in vitro applications the maturity of the derived hepatocytes will still need to be enhanced substantially.

Quantity of cell and means of acquisition

The adult liver consists of about 10¹¹ hepatocytes. Any clinical cellular therapy or bioartificial liver applications will likely employ at least 5–10% of that cell number in order to to provide sufficient functional activities, or to shorten the time frame of regeneration. With a typical bioreactor, cultivating cells as aggregates [85] or on microcarriers [86], a cell concentration of at least 10⁹ cell/L is easily achievable. Assuming a manufacturing process of preparing 100 doses per batch, the size of the reactor is in the range of 1000 L. Reactors with thousands of liters in volume are commonly seen in the production of viral vaccines and protein biologics [87]. Therefore, process scale up does not appear to pose insurmountable technical challenge for the applications of stem cell derived hepatocytes. The implementation of stem cell based hepatocyte production will involve both expansion and directed differentiation. The challenges are likely to arise in the control of cell quality, in serial cell transfer from smaller to larger reactors, and in the final isolation of cells after cell differentiation. Scale up aspects of stem cell process has been reviewed recently. Readers are referred to those articles for more discussion [88–90].

CONCLUDING REMARKS

The availability of an expandable source of human hepatocytes will have profound implications in the treatment of liver failure and congenital liver diseases. Sources of hepatocytes include hepatic progenitors from fetal or adult liver, differentiation of pluripotent stem cells or mesenchymal stem cells and direct reprogramming to fibroblasts. Hepatocytes like cells obtained with current differentiation protocols are still far away from their primary counterparts in their functional maturity. While these cells may not be suitable for clinical transplantation currently, this may not limit their use in bioartificial liver devices or as in vitro models for drug toxicity and diseases. This has been demonstrated in proof of concept studies in establishing liver disease models. Finally, prior expertise in cellular bioprocessing technology can be harnessed for the development of a bioprocess catering to the large scale production of stem cell derived hepatocytes. Thus, one can envision the translation of stem cell research to the clinic in the not too distant future.