Hepatic stem cells: from inside and outside the liver?

INTRODUCTION

Perhaps born out of necessity from the plethora of potentially cell‐damaging xenobiotics that assail the liver, plus a myriad of other cellular insults, e.g. hepatotropic viruses, the mammalian liver can invoke not just one, but at least three apparently distinct cell lineages to contribute to regenerative growth after damage (Fig. 1).

HEPATOCYTES

In response to parenchymal cell loss, the hepatocytes are the cells that normally restore the liver mass, rapidly re‐entering the cell cycle from the G₀ phase (Fig. 2). However, even after a two‐thirds partial hepatectomy, the remaining cells have to cycle on average only 1.4 times to restore the pre‐operative cell number, as all remaining hepatocytes traverse the cell cycle at least once in young adult rats. This seemingly modest response lead to the incorrect assumption that hepatocytes had only limited division potential, and thus were not true stem cells. A crucial property that defines a stem cell is its ability to give rise to a large family of descendants, i.e. be clonogenic (Alison et al. 2002), and, importantly, at least some hepatocytes can do this. Hepatocyte transplantation models have shown that the transplanted cells are capable of significant clonal expansion within the diseased livers of experimental animals and probably humans (see below).

In the diseased human liver there may not be the substantial selective growth advantage for transplanted cells that is operative in many of the rodent models, and it therefore becomes of interest to determine if it is possible to enrich for true stem cells that would continue to expand in the recipient liver in the absence of a major growth stimulus. The foetal rat liver is potentially a rich source of bipotential stem cells for liver transplantation (Shafritz & Dabeva 2002; 2003a, 2003b) and has been the focus of considerable attention. For example, Kubota & Reid (2000) have described a population of bipotential progenitors from ED13 foetal rat liver that lacked expression of major histocompatability complex (MHC) class I and had modest ICAM‐1 expression, features that may allow hepatoblasts to escape from the immune system when transplanted into an MHC‐incompatible host. These cells were clonogenic in culture with some descendants expressing phenotypic markers of hepatocytes [α‐fetoprotein (AFP) and albumin] and others of cholangiocytes [cytokeratin (CK)‐19].

In a similar vein, Shafritz and colleagues (Sandhu et al. 2001) have clearly demonstrated that foetal liver epithelial progenitors (FLEPs) from ED14 rats are more clonogenic than normal adult hepatocytes; when wild‐type FLEPs are injected into recently hepatectomized syngeneic dipeptidyl peptidase (DPPIV–)‐deficient F344 rats they proliferate for at least 6 months and constitute 7% of the recipient liver at this time compared with colonization of only 0.06% of the liver by wild‐type adult hepatocytes. DPPIV is an exopeptidase expressed in the bile canalicular surface of hepatocytes in contrast to diffuse cytoplasmic expression in bile duct epithelia. Thus, DPPIV‐positive hepatocytes are readily detected in the recipient mutant liver either by enzyme histochemistry or immunohistochemistry (IHC). Much greater colonization of the mutant liver was observed when the recipient rats were given prior administration of the DNA‐binding pyrrolizidine alkaloid retrorsine (usually two injections of 30 mg/kg each, 2 weeks apart); at 6 months after transplantation 60–80% of the recipient liver was occupied by DPPIV+ hepatocytes (Dabeva et al. 2000). Likewise, when these FLEPs were transduced with lentiviral vectors expressing green fluorescent protein (GFP) under the control of the albumin promoter, the colonizing cells co‐expressed DPPIV and GFP (Oertel et al. 2003). Progress is being made to identify the genes that are exclusively expressed by these FLEPs with a view to their specific identification and isolation (Petkov et al. 2000).

Clonogenic cells can also be isolated from the foetal mouse liver (ED13.5); hepatocytes expressing the integrins α6 (CD49f) and β1 (CD29), but not c‐kit, CD45 or Ter119 (erythroid precursor antigen) had the greatest colony forming ability (Suzuki et al. 2000). Designated hepatic colony forming units in culture (H‐CFU‐C) this sorting achieved a 35‐fold enrichment of H‐CFU‐C over total foetal liver cells. In a recent development, further selection based on c‐Met‐positivity, enriched for H‐CFU‐C and these cells, could produce both hepatocytes (albumin‐positive) and biliary cells (cytokeratin‐19‐positive) in culture (Suzuki et al. 2002). EGFP‐marked cells from these clonally derived H‐CFU‐C also produced hepatocytes and biliary cells when injected into mice, and more surprisingly were found to apparently differentiate into pancreatic ducts and acini and duodenal mucosal cells when injected directly into these organs.

Many other studies have examined the transplantation potential of adult hepatocytes in the DPPIV– mutant rat combining retrorsine treatment with a mitogenic stimulus such as partial hepatectomy or triidothyronine (T3) leading to rapid replacement of DPPIV– cells by DPPIV+ donor cells (Laconi et al. 1998; Oren et al. 1999); even in the absence of a mitogenic stimulus near total replacement by donor cells occurs within 12 months (Laconi et al. 2001).

Intriguingly, when retrorsine‐treated adult rats are given a two‐thirds partial hepatectomy, regeneration is accomplished by the activation, expansion and differentiation of so‐called small hepatocyte‐like progenitors (SHPCs) (Gordon et al. 2000). These cells showed phenotypic traits of foetal hepatoblasts, oval cells and fully differentiated hepatocytes, but they were morphologically and phenotypically distinct from all three. Cytochrome (CYP) P450 enzymes have a pivotal role in hepatocyte biology (Mugford & Kedderis 1998), but typically these cell clusters lacked CYP enzymes that are usually readily induced by retrorsine, and this probably accounted for their resistance to the anti‐proliferative effects of retrorsine. When such cells (H4‐positive) were isolated, established in short‐term culture and then transplanted into syngeneic rats, they gave rise to differentiated hepatocytes as evidenced by expression of albumin and transferrin, but lack of AFP (Gordon et al. 2002).

The clonogenic potential of transplanted adult hepatocytes has been most impressively demonstrated in the Fah null mouse, a model of hereditary type 1 tyrosinaemia, where there exists a profoundly strong positive selection pressure on the transplanted wild‐type cells, as Fah‐deficient mice will die as neonates unless rescued by 2‐(2‐nitro‐4‐trifluoro‐methylbenzoyl)‐1,3‐cyclohexanedione (NTBC), a compound that prevents the accumulation of toxic metabolites in the tyrosine catabolic pathway. When 10⁴ normal hepatocytes from congenic male wild‐type mice are intrasplenically injected into mutant female mice, these cells will quickly colonize the mutant liver (Overturf et al. 1997). Moreover, serial transplantations from the colonized liver to other mutant livers indicated that at least 69 doublings would have been necessary from the original hepatocytes for six rounds of liver re‐population. This estimate is likely to be a minimal figure as it assumes that all injected hepatocytes migrate to the liver from the spleen and take part equally in the cycles of regeneration. In fact, probably at best only 15% of intrasplenically transplanted hepatocytes migrate to the liver, and if all these participated equally in repopulation, a minimum of 86 doublings would be required. This figure may be even higher if not all the cells that migrated to the liver actually took part in re‐population, and the authors suggested that maybe there is a subpopulation of hepatocyte stem cells which they designated as regenerative transplantable hepatocytes (RTHs). We could speculate that perhaps these RTHs are analogous to the SHPCs described by Gordon et al. (2000). The Fah null mouse can also be rescued by pancreatic cells; though most Fah‐deficient mice withdrawn from NTBC treatment and transplanted with pancreatic cells will die, a small proportion do survive with 50–90% replacement of the diseased liver by pancreatic cell‐derived hepatocytes (Wang et al. 2001). Given that animals fed a copper‐deficient diet undergo pancreatic exocrine cell atrophy and re‐feeding induces the surviving ducts to give rise to hepatocytes (Rao & Reddy 1995), it was surprising that pancreatic cell suspensions enriched for pancreatic ducts were poorer than unfractionated pancreatic cells at reconstituting the diseased Fah‐deficient liver with functional hepatocytes.

Other cell lines with hepatocyte potential can also be isolated from adult rat liver (Nagai et al. 2002); these non‐hepatocyte cells were AFP‐, albumin‐ and CK19‐negative, but after co‐culture with hepatic stellate cells they expressed albumin, transferrin and α ‐1 anti‐trypsin. It is well known that the WB‐F344 rat liver diploid epithelial cell line readily differentiates into hepatocytes when transplanted into syngeneic rats, but biliary differentiation [CK19+, BDS7+, gamma glutamyl transpepetidase (GGT)+] can be induced by in vitro culture on Matrigel (Couchie et al. 2002). Tateno & Yoshizato (1996) have defined the conditions for the long‐term expansion of cells isolated from adult Fisher 344 rats.

Hepatic stem cells, from whatever source, may be therapeutically useful for treating a variety of diseases that affect the liver. This would include a number of genetic diseases that produce liver disease such as Wilson's disease (copper accumulation), Crigler Najjar syndrome (lack bilirubin conjugation activity) and tyrosinemia, and cases where there is extrahepatic expression of the disease, e.g. Factor IX deficiency. In terms of therapeutic potential, we have already noted the rescue of the Fah null mouse and the DPPIV‐negative rat by hepatocytes, and there are now other examples. Transplantation of adult rat hepatocytes has also been effective in normalizing bilirubin levels and improving bilirubin conjugation activity in Gunn rats (a model of Crigler–Najjar syndrome) (Tada et al. 1998; Guha et al. 2002). These cells were reversibly immortalized and transduced with the bilirubin‐uridine 5′‐diphosphoglucuronate glucuronsyltransferase gene (UGT1A1) and engraftment was improved by prior irradiation and partial hepatectomy of the recipient rats. Following on from this, an infusion of isolated hepatocytes through the portal vein equivalent to 5% of the parenchymal mass to a patient with Crigler Najjar syndrome, achieved a medium‐term reduction in serum bilirubin and increased bilirubin conjugate levels in the bile (Fox et al. 1998). Hepatocyte transplantation has also been successful in the treatment of human glycogen storage disease type 1a (Muraca et al. 2002), but not in the treatment of severe ornithine transcarbamylase deficiency, where rejection of the transplanted cells was thought to be the reason for only temporary (11 days) relief (Horslen et al. 2003).

CHOLANGIOCYTES

When either massive damage is inflicted upon the liver or regeneration after damage is compromised, a potential stem cell compartment located within the smallest branches of the intrahepatic biliary tree is activated. This so‐called ‘oval cell’ or ‘ductular reaction’ amplifies the biliary population before these cells differentiate into either hepatocytes or cholangiocytes (1998, 1996a, 1997a, 1997b; Alison et al. 1998). Careful studies in rats indicate that oval cells are pre‐dominantly derived from the canal of Hering, and thus this is the location of a stem cell niche (Paku et al. 2001). In rats and mice, the canals of Hering barely extend beyond the limiting plate, but the resultant oval cell proliferation can result in arborizing ducts that express AFP (Fig. 3a), that stretch to the midzonal areas (Fig. 3b) before these cells differentiate into hepatocytes (Fig. 3c). A wide range of markers has been used to identify ovals cells including GGT and glutathione‐S‐transferase (GST‐P) activity along with a host of monoclonal antibodies raised against cytoskeletal proteins and unknown surface antigens (see Table 1) (Agelli et al. 1997; Hixson et al. 1997; Hixson et al. 2000). Moreover, antigens traditionally associated with haematopoietic cells can also be expressed by oval cells, including c‐kit, flt‐3, Thy‐1 and CD34 (Omori et al. 1997; Lemmer et al. 1998; Petersen et al. 1998; Baumann et al. 1999). This may be no more than coincidental, but it has given support to the notion that at least some hepatic oval cells are directly derived from a precursor of bone marrow origin, particularly when the biliary tree is damaged (Petersen 2001), though other studies indicate that many/most oval cells are derived from the direct intrahepatic proliferation of cells already located within the biliary tree (Alison et al. 1996b). Regarding the contribution of bone marrow (see below), Hatch et al. (2002) have suggested that only when the rat liver is severely damaged do hepatocytes up‐regulate the chemokine, stromal‐derived factor‐1 alpha (SDF‐1α), and both oval cells are activated and bone marrow cells recruited through SDF−1α/CXCR4 inter‐actions.

The identification of molecular markers that best facilitate the isolation and characterization of stem cell populations has long been a challenge. In studies of the haematopoietic system, both experimental and clinical, the expression of the sialomucin CD34 has traditionally been exploited to enrich for cells with long‐term marrow re‐population capacity. However, in 1996, Goodell et al. reported on a new method for the isolation of haematopoietic stem cells (HSCs) based on the ability of the HSCs to efflux a fluorescent dye. Like the activity of the P‐glycoprotein (encoded by the mdr1 gene), this activity was verapamil‐sensitive. Cells subjected to Hoechst 33342 dye staining that actively efflux the Hoechst dye appeared as a distinct population of cells on the side of a flow cytometry profile, hence the name the ‘side population’ (SP) has been given to these cells (Alison 2003). Numerous studies now point to the fact that the SP phenotype of HSCs in mice and humans is largely determined by the expression of a protein known as the ABCG2 transporter [ATP‐binding cassette (ABC) subfamily G member 2, also known as BCRP1] (Scharenberg et al. 2002). Perhaps not surprisingly, we now have reports of the up‐regulation of several ABC transport proteins in damaged human liver, particularly in regenerating ductules (Ros et al. 2003) and of ABCG2/BCRP1 in rat oval cells (Shimano et al. 2003), adding support to the belief that ABC transporter proteins are intimately involved in the biology of stem/progenitor cells in many tissues.

In the liver, bile secretion is driven largely by ABC‐type proteins that are located in the canalicular membrane and effect ATP‐dependent transport of bile acids, phospholipids and non‐bile acid organic ions. Canalicular ABC‐type proteins belong to two subfamilies (see http://www.nutrigene.4t.com/humanabc.htm); members of subfamily B including MDR1 (ABCB1), MDR3 (ABCB4) and SPGP (sister of P‐glycoprotein, ABCB11) and subfamily C including MRP1‐3 (ABCC1‐3). Subfamily C also contains the cystic fibrosis transmembrane conductance regulator (CFTR, ABCC7) that is expressed by bile duct epithelia (cholangiocytes). However, is the expression of these ABC transporters in liver ductules a genuine marker of stem cells/progenitors or merely a reflection of the protective role these proteins undoubtedly perform within the biliary tree against toxic bile constituents (Scheffer et al. 2002)? The correct answer is probably the latter as it is highly unlikely that all the reactive ductular cells are stem cells or even progenitors, and many are imminently going to differentiate into hepatocytes and cholangiocytes. Nevertheless, the expression of ABC transporters adds incrementally to the battery of already established markers for this stem cell response (Table 1). A further factor that might afford oval cells protection in a hostile environment is the expression of ‘deleted in malignant brain tumour 1’ (DMBT1), a molecule significantly up‐regulated in rat oval cells (Bisgaard et al. 2002); DMBT1 is a putative receptor for the mucus‐associated trefoil peptides that may have a general role of cytoprotection throughout the gastrointestinal tract (Taupin & Podolsky 2003).

Most models of oval cell activation have employed potential carcinogens to inhibit hepatocyte replication in the face of a regenerative stimulus; these have been reviewed elsewhere (1996a, 1997b, 1998). In the rat, protocols have included administering 2‐acetylaminofluorene (2‐AAF) prior to a two‐thirds partial hepatectomy (AAF/PH) or a necrogenic dose of carbon tetrachloride, feeding a choline‐deficient diet supplemented with ethionine (CDE), or simply treating animals with the likes of 3′‐methyl‐diaminobenzidine (3′‐Me‐DAB), galactosamine or furan. Cells derived by such procedures are clearly not relevant to human studies, but Sell and co‐workers (1999, 2002) have demonstrated that it is possible to derive bipotential liver progenitor cells (LPCs) from rat livers without resorting to mutagenic chemicals. Allyl alcohol causes periportal necrosis, and the resultant oval cells can be isolated and propagated, producing a number of clonally derived cell lines, capable of at least 100 doublings in the presence of a feeder layer of lethally irradiated fibroblasts. These cell lines were able to differentiate into hepatocytes after removal of the feeder cells; supplementation with either a combination of oncostatin M and dexamethasone or bFGF‐promoted differentiation. Similarly, such factors have proved instrumental in causing the transdifferentation of pancreatic acinar cells to hepatocytes (oncostatin M and dexamethasone) (Shen et al. 2000) and multipotent adult progenitor cells (MAPCs) to hepatocytes (bFGF) (Schwartz et al. 2002).

We have little idea of the precise cues that promote the in vivo hepatocytic differentiation of oval cells. GGT is a useful oval/biliary cell marker, however, GGT gene expression is driven by a number of different promoters and in the rat, differentiation down either the hepatocytic or biliary lineages is associated with specific promoter extinction (Holic et al. 2000). Under the CDE protocol, oval cell differentiation towards hepatocytes (expression of albumin, CYP4A1 and AFP) can be promoted by an activator of the transcription factor ‘peroxisome proliferator activated receptor (PPARα) (Kaplanski et al. 2000). In terms of therapeutic potential, one of the most impressive demonstrations of the hepatocytic potential of oval cells has come from experiments in which oval cells were isolated from Long Evans Cinnamon (LEC) rats (a model of Wilson's disease); transduced ex vivo with a reporter gene, β‐galactosidase, and then transplanted into LEC/Nagase analbuminemic double mutant rats, these cells differentiated into albumin‐expressing hepatocytes (Yasui et al. 1997).

Oval cells also occur in human liver. Elegant 3‐dimensional reconstructions of serial sections of human liver immunostained for cytokeratin‐19 have shown that the smallest biliary ducts, the canals of Hering, unlike those in rodents normally extend into the proximate third of the lobule (Theise et al. 1999), and it is envisaged that these canals react to massive liver damage (akin to a trip‐wire), proliferating and then differentiating into hepatocytes (Fig. 4). Oval cell numbers in human liver rise with increasing severity of liver disease (Lowes et al. 1999) and this ductular reaction is widely accepted to be a stem cell response rather than a ductular metaplasia of ‘damaged’ hepatocytes. Notwithstanding, Falkowski et al. (2003) have recently still felt compelled to formally dispel such a notion in human liver, showing that ‘cholestatic’ hepatocytes were very largely not associated with ductular reactions and moreover, that in cirrhosis most (94%) intraseptal buds of hepatocytes were associated with ductular reactions. Non‐parenchymal epithelial cells from adult porcine liver, essentially ductal epithelia, are also capable of differentiating into both biliary epithelia and hepatocytes in vitro (Kano et al. 2000).

BONE MARROW

Within an adult tissue, the locally resident stem cells were formerly considered to be capable of only giving rise to the cell lineage(s) normally present. However, adult haematopoietic stem cells (HSCs) in particular appear to be much more flexible: removed from their usual niche they are capable of differentiating into all manner of tissues including skeletal and cardiac muscle, endothelia, and a variety of epithelia including neuronal cells, pneumocytes and hepatocytes. Some oval cells/hepatocytes were first revealed to be derived from circulating bone marrow cells in the rat (Table 2a): Petersen et al. (1999) followed the fate of syngeneic male bone marrow cells transplanted into lethally irradiated female recipient animals whose livers were subsequently injured by a regime of 2‐acetylaminofluorene (which blocks hepatocyte regeneration) and carbon tetrachloride (which causes hepatocyte necrosis) designed to cause oval cell activation. Y chromosome‐positive oval cells were found at 9 days after liver injury and some Y chromosome‐positive hepatocytes were seen at 13 days when oval cells were differentiating into hepatocytes. Additional evidence for hepatic engraftment of bone marrow cells was forthcoming from a rat whole liver transplant model. Lewis rats expressing the MHC class II antigen L21‐6 were recipients of livers from Brown Norway rats that were negative for L21‐6. Subsequently, ductular structures in the transplants contained both L21‐6‐negative and L21‐6‐positive cells indicating that some biliary epithelium was of in‐situ derivation and some was of recipient origin, presumably from circulating bone marrow cells.

Using a similar gender mismatch bone marrow transplantation approach in mice to track the fate of bone marrow cells, Theise et al. (2000a) reported that, over a 6‐month period, 1–2% of hepatocytes in the murine liver may be derived from bone marrow in the absence of any obvious liver damage, suggesting bone marrow contributes to normal ‘wear and tear’ renewal (Table 2b). In two contemporaneous papers, Alison et al. (2000) and Theise et al. (2000b) demonstrated that hepatocytes can also be derived from bone marrow cell populations in humans (see Table 2c). Two approaches were adopted – firstly, the livers of female patients who had previously received a bone marrow transplant from a male donor were examined for cells of donor origin using a DNA probe specific for the Y chromosome, localized using in situ hybridization. Secondly, Y‐positive cells were sought in female livers engrafted into male patients but which were later removed for recurrent disease. In both sets of patients, Y‐chromosome‐positive hepatocytes were readily identified, but the degree of hepatic engraftment of HSCs into human liver was highly variable; most likely related to the severity of parenchymal damage, up to 40% of hepatocytes and cholangiocytes appeared to be derived from bone marrow in a liver transplant recipient with recurrent hepatitis (Theise et al. 2000b). Subsequent human investigations with G‐CSF mobilized peripheral blood CD34+ stem cells have shown that these cells are also apparently able to transdifferentiate into hepatocytes, with 4–7% of hepatocytes in female livers being Y chromosome‐positive after the cell transplant from male donors (Korbling et al. 2002). However, it is worth noting that some other studies, examining the contribution of recipient cells to liver allografts in humans, have failed to register any real engraftment in the allografted liver (Fogt et al. 2002; Ng et al. 2003; Wu et al. 2003; see Table 2c).

In undoubtedly the most convincing ‘proof of principle’ demonstration of the potential therapeutic utility of bone marrow, mice with a metabolic liver disease have been cured (Lagasse et al. 2000). Female mice deficient in the enzyme fumarylacetoacetate hydrolase (fah–/–, a model of fatal hereditary tyrosinaemia type 1), a key component of the tyrosine catabolic pathway, can be rescued biochemically by a million unfractionated bone marrow cells that are wild‐type for Fah. Moreover, only purified HSCs (c‐kit^highThy^lowLin⁻Sca‐1⁺) were capable of this functional repopulation, with as few as 50 of these cells being capable of hepatic engraftment when haematopoiesis was supported by 2 × 10⁵ fah–/– congenic adult female bone marrow cells. The salient point to arise from this powerful demonstration of the therapeutic potential of bone marrow cells was that, though the initial engraftment was low, approximately one bone marrow cell for every million indigenous hepatocytes, the strong selection pressure exerted thereafter on the engrafted bone marrow cells resulted in their clonal expansion to eventually occupy almost half the liver (Fig. 5). This positive selection was achieved by cycles of withdrawal of NTBC, the compound that blocks the breakdown of tyrosine to fumarylacetoacetate (FAA) in the Fah‐deficient mice, so protecting against liver failure. In the absence of NTBC, FAA accumulates and destroys the hepatocytes; thus the ensuing regenerative stimulus promotes the growth of the engrafted cells. Furthermore, in the absence of NTBC, no engraftment was seen (Wang et al. 2002).

However, it now turns out that the new healthy liver cells in the fah–/– mouse contain chromosomes from both the recipient and donor cells, with presumably the donor haematopoietic cell nuclei being reprogrammed when they fused with the unhealthy fah–/– hepatocyte nuclei to become functional hepatocytes (Vassilopoulos et al. 2003; Wang et al. 2003b). The key to this discovery was to perform the gender mismatch bone marrow transplantation experiments the other way round from the usual, i.e. instead of male to female and looking for the Y chromosome in apparently transdifferentiated, say epithelial cells, bone marrow transplants were performed between female donors and male recipients – a cell with a donor‐specific marker and a putative transdifferentiated marker is acceptable, but if it also has a Y chromosome then you know it is cell fusion (Fig. 6). In one experiment, a million donor bone marrow cells (fah+/+) from Fanconi anaemia group C (fancc–/–) homozygous mutant mice were serially transplanted into lethally irradiated fah–/– recipients (Wang et al. 2003b). The usual repopulation (∼50%) of the mutant liver by Fah‐positive hepatocytes was noted, but Southern blot analysis of the purified repopulating cells revealed that they were heterozygous for alleles (fah+/+; fancc–/–) that were unique to the donor marrow – fusion with host liver cells must have occurred. To confirm this conclusion, in a second experiment, fah+/+ bone marrow from ROSA26 female mice was transplanted into male fah knockout mice. Cytogenetic analysis of the LacZ‐positive, sorted bone marrow‐derived hepatocytes revealed that most, if not all had a Y chromosome, thus confirming fusion (Fig. 6). Before bone marrow transplant, most host hepatocytes had a karyotype of either 40,XY or 80,XXYY, but after transplantation with fah+/+ bone marrow, the commonest karyotype of Fah‐positive heptocytes was either 80,XXXY suggesting fusion between a diploid female donor cell and a diploid male host cell, or 120,XXXXYY, suggesting fusion between a female donor diploid blood cell and a tetraploid male host hepatocyte. However, a substantial proportion of bone marrow‐derived hepatocytes were aneuploid, suggesting fusion had created some sort of genetic instability with the hybrid cells randomly shedding chromosomes. In the companion paper (Vassilopoulos et al. 2003) to that of Grompe and colleagues, lineage‐depleted wild‐type bone marrow was transplanted into lethally irradiated female fah–/– mice, and, following withdrawal of NTBC, the usual Fah‐positive nodules emerged 4–5 months later. When restriction enzyme‐digested genomic DNA from these nodules was probed for fah sequences, then the mean level of donor DNA was found to be only 26%, again leading to the conclusion that the donor haematopoietic cells had fused with the host fah–/–, generating polyploid hepatocytes. Fusion of bone marrow cells has also been found to occur in the normal mouse, not only with hepatocytes, but also with Purkinje cells and cardiomyocytes (Alvarez‐Dolado et al. 2003). These were very elegant studies in vivo and in vitro in which a reporter gene was activated only when cells fused. However, unlike the Fah null mouse, no selection pressure (liver damage) was operative, and even after 10 months only 9–59 fused cells/5.5 × 10⁵ hepatocytes were found: importantly, they also found evidence that with time either donor genes had been inactivated or eliminated, again suggestive of genetic instability in heterokaryons.

Mouse bone marrow‐derived hepatocytes can also be expanded selectively if they are engineered to over‐express Bcl‐2, and then the indigenous cells are targeted for destruction by an anti‐Fas antibody (Mallet et al. 2002); it will be interesting to know whether cell fusion operates in this model. One could also add that if fusion was responsible for all these observations made in the liver, then clearly these hybrids have a selective growth advantage turning unhealthy hepatocytes into metabolically competent hepatocytes and would not negate the therapeutic potential of bone marrow cells in the liver. Expressing a similar sentiment, Blau (2002) has suggested that if cell fusion was responsible for the apparent reprogramming of certain adult cells, then there is something ‘exciting’ about rescuing damaged cells through fusion, with, for example, bone marrow‐derived cells providing a healthy and entire genetic complement, even one that has been manipulated for gene therapy. As in the human arena, not all murine studies are in complete accord, e.g. Terai et al. (2003) report an impressive 25% contribution of bone marrow to the parenchyma after CCL4, but Kanazawa & Verma (2003) failed to find any evidence for bone marrow engraftment in three models of chronic liver injury, including CCL4 (see Table 2b). However, many studies have testified as to the ability of human cord blood cells to transdifferentiate into hepatocytes in the liver of the immunodeficient mouse (Table 2d), albeit at a low level.

While it seems logical to believe that parenchymal damage is a stimulus to hepatic engraftment by HSCs, the molecules that mediate this homing reaction to the liver are not well understood. Petrenko et al. (1999) speculated that in mice the molecule AA4 (murine homologue of the C1q receptor protein) may be involved in the homing of haematopoietic progenitors to the foetal liver – maybe this receptor protein is expressed on HSCs that engraft to the damaged liver? Another alternative is that cells in the liver express the stem cell chemoattractant ‘stromal derived factor‐1’ (SDF‐1) for which HSCs have the appropriate receptor known as CXCR4 (Whetton & Graham 1999). Hatch et al. (2002) have persuasive evidence that SDF‐1 is involved in oval cell activation, furthermore speculating that this chemokine may secondarily recruit bone marrow to the injured liver. More definitive proof was provided by Kollet et al. (2003) who observed increased SDF‐1 expression (particularly in biliary epithelia) after parenchymal damage, and concomitant with such damage was increased HGF production, a cytokine that was very effective in promoting protrusion formation and CXCR4 up‐regulation in human CD34+ haematopoietic progenitors.

CONCLUSIONS

This review has highlighted recent progress in identifying cells with hepatic potential in rodents and humans. The chronic shortage of livers for orthotopic liver transplantation has provided a strong impetus for the search for alternative sources of cellular therapy, in particular hepatocyte or even bone marrow transplants. The clinical need for healthy functioning hepatocytes is very clear, not only for the correction of inherited metabolic liver disease, but also for acute liver failure, hepatocellular carcinoma, cirrhosis, bioartificial liver support, hepatotropic viral studies and drug toxicity testing. Hepatocyte transplants have been moderately successful in humans, but clearly there is a need to enrich for cells with potent clonogenic potential; studies in rodents have gone some way to the identification of such cells in these species. Clearly, the key to the success of hepatocyte transplants will be to selectively enhance the growth of the transplanted cells, but unfortunately treatments such as retrorsine (used with conspicuous success in rats) are not clinically acceptable.

Turning to the vexacious subject of bone marrow stem cells as a source of hepatocytes, then the jury is still very much out. Impressive functional re‐population has been achieved in the Fah‐deficient mouse liver, though this has been achieved through cell fusion between bone marrow cells and deficient hepatocytes. If deficient cells are re‐programmed in this way then that per se is not necessarily a bad thing, but if the formed heterokaryons are genetically unstable, then this has significant pathological implication. It also begs the question whether this process is going on all the time in healthy individuals without experimental manipulation such as irradiation and bone marrow transplantation? Like hepatocyte transplants, the key to success will be the selective amplification of bone marrow‐derived hepatocytes; studies in the Fah‐deficient mouse indicate that the level of initial engraftment is independent of damage, though subsequent clonal expansion is very much dependent on the induction of liver failure. Alternatively, other models of chronic liver injury have failed to detect bone marrow‐derived hepatocytes; such conundrums will exercise investigators in this exciting field over the next few years – watch this space.