SEARCH

SEARCH BY CITATION

Keywords:

  • Bone marrow;
  • Cell differentiation;
  • Cell fusion;
  • Hematopoietic stem cells Hepatocytes;
  • Liver regeneration;
  • Stem cells

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

Adult stem cell plasticity raised expectations regarding novel cellular therapies of regenerative medicine after findings of unexpected plasticity were reported. In this review, reports of hematopoietic stem cells (HSCs) contributing to hepatocytic lineages are critically discussed with reference to rodent and human models. In particular, the role of liver injury and the potential contribution HSCs make to hepatic regeneration in both injury and physiological maintenance is reviewed. The relative contributions of genomic plasticity and cell fusion are studied across different model systems, highlighting possible factors that may explain differences between often conflicting reports. Insights from experimental studies will be described that shed light on the mechanisms underlying the migration, engraftment, and transdifferentiation of HSCs in liver injury. Although it appears that under differing circumstances, macrophage fusion, HSC fusion, and HSC transdifferentiation can all contribute to hepatic epithelial lineages, a much greater understanding of the factors that regulate the long-term efficacy of such cells is needed before this phenomenon can be used clinically.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

Presently, orthotopic liver transplantation is the major therapeutic option for patients with acute and chronic end-stage liver disease. However, a shortage of suitable donor organs and requirement for immunosuppression restrict its usage, highlighting the need to find suitable alternatives.

A novel and exciting approach could be offered through the current developments in the field of stem cell biology. In the past few years, multiple studies have demonstrated that adult stem cell plasticity is far greater and complex than previously thought, raising expectations that it could lay the foundations for new cellular therapies in regenerative medicine. In this review, the evidence for adult stem cell plasticity will be discussed with respect to hepatology, covering experimental models in animal and human tissues along with a discussion of the factors, putative mechanisms involved, current controversies, and potential clinical implications.

Hematopoietic Stem Cells: Multilineage Plasticity

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

Stem cells can be defined as cells capable of unlimited self-renewal, multilineage differentiation, and in vivo functional reconstitution of a given tissue with flexibility in the timing of this differentiation [1]. Until relatively recently, pluripotent stem cells were thought to derive only from embryonic sources. Embryonic stem (ES) cells derive from totipotent cells of the early postimplantation embryo and are capable of unlimited, undifferentiated proliferation in vitro while maintaining the potential to form cell types of all three germ layers [2]. Subsequently, primordial germ cell cultures were found to give rise to cells with characteristics of ES cells and were designated embryonic germ cells [3]. Although the use of ES cells has raised new possibilities for therapies in regenerative medicine, there are many associated ethical and legal issues. Consequently, other sources of stem cells have been sought. In the adult, many tissues are known to contain lineage-restricted stem cells that possess the ability for lifelong renewal and functional maintenance [47]. The most widely studied example of adult stem cells is hematopoietic stem cells (HSCs), which sustain formation of the blood and immune systems throughout life. The bone marrow compartment is largely made up of committed progenitor cells, non-circulating stromal cells that have the ability to develop into mesenchymal lineages (termed mesenchymal stem cells), and HSCs [8,9]. This latter group has been the predominant focus of research examining the stem cell compartment of bone marrow, its identification relying largely on the expression of cell-surface markers to define a subpopulation enriched for HSCs. Although this method of identification can be easily performed in the laboratory, there are several caveats, most notably that there is no assessment of stem cell function inherent to it. However, complete characterization of HSCs before their use invokes the biological equivalent of Heisenberg's uncertainty principle; by the time the cell has been isolated and demonstrated to differentiate down multiple lineages, it is no longer a stem cell [1].

Nevertheless, surface marker expression is used in standard experimental practice to identify a population that is enriched for HSCs. Although expression of the CD34 antigen is generally used in human studies as a surrogate marker for these progenitors [10], there is a population that lacks this surface marker that acts as a more primitive form of HSCs [11]. Although the CD34+ population is used in clinical practice for patients undergoing stem cell transplantation, it should be noted that in murine settings, CD34+ expression is a dynamic phenomenon and may not truly reflect stem cell content [11]. In humans, within the CD34+ population, the monoclonal antibody AC133 identifies a CD34bright subpopulation that has greater hematopoietic-reconstituting properties in xenotransplantation models [12] (compared with the CD34dim population).

By staining HSCs with Hoechst 33342 dye, a selection of side population (SP) of cells with the highest efflux capacity has been demonstrated to identify a primitive population [13]. Furthermore, it is the expression of the ATP-binding cassette (ABC) transporter, ABCG2, that mediates the SP phenotype [14,15]. Notably, ABC transporters have been shown to be upregulated in rodent hepatic oval cells [16] and human hepatic oval cells [17], implying an overlap with this lineage-restricted stem cell compartment. This overlap of putative surface markers across different stem cell populations is a common feature and can result in difficulties discerning respective cellular populations when they admix in end organs such as liver and muscle.

It had been assumed that adult stem cells, unlike ES cells, were lineage restricted, but recent observations demonstrating that bone marrow–derived myogenic progenitors participate in regeneration of damaged skeletal muscle [18] and ischemic myocardium [1921] challenged this. In addition, presumptive muscle stem cells were also shown to contribute to hematopoiesis [22,23], although this could relate to their common embryological origin, because both blood and muscle cells derive from the mesoderm. However, reports of participation of hematopoietic cells in neurogenesis [24,25], the conversion of adult neural stem cells into hematopoietic cells [26], and the demonstration that adult mouse neural stem cells can give rise to cells of all germ layers [27] seemed to confirm that differentiation could occur out with the original germ layer. This challenged the conventional trilaminar view of organ development and implied that adult stem cells may exhibit as much pluripotentiality as ES cells. However, these findings have not been universally confirmed, with many groups either unable to reproduce these findings or suggesting that fusion rather than transdifferentiation underpins these phenomena [2830]. This review will critically appraise the reported contribution of HSCs to hepatic reconstitution and the conditions that are required.

Does Liver Injury Upregulate the Contribution of HSCs?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

The adult liver has a remarkable regenerative capacity, with mature hepatocytes proliferating rapidly in response to most mild-to-moderate liver injuries and an intrahepatic stem cell compartment contributing in more severe injury. Experimental models of partial hepatectomy [31] or carbon tetrachloride (CCl4) toxicity [32] indicate that regeneration can occur wholly through proliferation of mature hepatocytes. However, in models of more severe injury, such as CCl4 combined with the hepatocarcinogen 2-acetylaminofluorene (2-AAF), which inhibits mature hepatocyte proliferation, additional cellular components contribute. After such injury, large numbers of small oval cells [33] appear, which are able to differentiate into both hepatocytes and ductular cells [34]. Similarly, addition of the DNA alkylating agent retrorsine to CCl4 gives rise to progenitors that express phenotypic characteristics of, but are morphologically distinct from, oval cells [35]. This has led investigators to establish the concept that liver regeneration occurs on three separate levels: hepatocytes, intrahepatic stem cells, and extrahepatic stem cells [36].

In the first account of hepatic transdifferentiation, lethally irradiated rats underwent cross-sex or cross-strain bone marrow transplantation followed by administration of 2-AAF to suppress hepatocyte proliferation and CCl4 to induce hepatic injury. Examination of the host liver demonstrated hepatic cells that were donor derived, as detected by expression of dipeptidyl peptidase IV enzyme (DPPIV+) in DPPIV rats or the Y chromosome in female animals [37]. This led to the conclusion that marrow-derived cells could act as progenitors for hepatic cells, albeit in a model of injury in which the replicative capacity of mature host hepatocytes was impaired.

However, liver repopulation by marrow-derived cells has also been seen in the absence of any intentional liver injury. Bone marrow from male donors was infused into irradiated female mice, and the liver tissue was analyzed by fluorescent in situ hybridization (FISH) for the Y chromosome and albumin mRNA, demonstrating significant levels of donor-derived hepatocytes [38]. Further work, again in mice, indicated that a single male HSC transplanted into an irradiated female recipient demonstrated diverse differentiative potential. In addition to hepatic engraftment, epithelial cells from throughout the gastrointestinal tract, bronchus, and skin were donor derived [39]. Notably, in multiorgan engraftment, highest levels and most diffuse clustering of donor cells were seen in alveolar epithelium. The authors postulated that the observed differences in engraftment may relate to the injury induced by irradiation, because lung tissue is known to be more radiosensitive [40]. It remains unclear whether irradiation injury is relevant in the hepatic models, because although radiation-induced liver damage is known to occur [41], it is usually seen with larger doses than those used in such preparative regimens. Additionally, there was no histological evidence of tissue damage in either study, raising the possibility that a contribution of bone marrow cells may occur even in physiological maintenance or minimal liver injury.

Other groups have been unable to reproduce these findings. Chimeric animals were generated by transplantation of a single green fluorescent protein (GFP)–marked HSC into sublethally irradiated mice in the absence of specific tissue injury. Single HSCs resulted in significant hematopoietic engraftment, but hepatocytes were produced at a frequency of only approximately 1 in 70,000 cells [28]. Furthermore, GFP+:GFP parabiotic mice with a common circulatory system were created, surgically enabling evaluation of circulating stem cells and HSC engraftment in a model that does not even require irradiation. Despite successful hematopoietic cross-engraftment, there was no engraftment of non-hematopoietic tissue, leading the authors to conclude that HSCs played no role in the production of nonhematopoietic cells under physiological conditions. A different group used various liver injury models to assess hepatic regeneration after gender-mismatched bone marrow transplantation. No significant contribution was demonstrated [42].

Nevertheless, several other studies seem to corroborate adult stem cell plasticity, with moderate to severe injury increasing the level of hepatic transdifferentiation, a finding augmented additionally in model systems in which the donor HSCs have a survival advantage [43]. The significance of such a survival advantage is confirmed in a transgenic model based on the protective effect of the antiapoptotic gene, Bcl-2, against Fas-mediated cell death. Bone marrow from mice expressing this transgene, under the control of a liver-specific promoter, was infused into normal mice. Some mice underwent repeated injections with a Fas-agonist antibody to induce liver injury, whereas others did not. Only those that had received antibody injections showed mature hepatocytes expressing Bcl-244, implying that transdifferentiation is inefficient under physiological conditions and that tissue injury such as accumulation of toxic catabolites [43] or apoptotic challenge [44] is required to generate a more robust response. It is interesting to note that a recent study of the contribution of HSCs in a model of liver fibrosis demonstrated higher than previously reported levels of marrow-derived hepatocytes. Transgenic mice expressing GFP were used as a source of bone marrow, and up to 26% of the recipient liver was repopulated by 4 weeks [45]. For the first time, donor bone marrow cells without a survival advantage resulted in robust and efficient regeneration, although there has been some concern raised regarding ambiguous identification and unusual architecture of the reporter cells [46].

Although the presence and severity of liver injury may be important in regulating the extent of stem cell plasticity and engraftment, the reported variation in the literature is still marked (Table 1). Detailed analysis of the models used demonstrates that different subpopulations of stem cells may have different levels of functional plasticity. Although initial studies mainly used unfractionated bone marrow, subsequent work highlighted the capacity of CD34+ cells for hepatic engraftment [38]. More recently, the SP fraction of marrow cells was demonstrated to contain cells with similar ability [47], and yet it is rich in CD34 stem cells [11,13]. Indeed, the authors of these many studies have recognized that supposedly minimal differences in experimental methods may be responsible for the observed discrepancies. Such factors have been discussed in more detail in recent commentaries [48,49].

Table Table 1.. Stem cell–derived hepatocytic differentiation in animal models
  1. a

    Abbreviations: 2-AAF, 2-acetylaminofluorene; FAH, fumarylacetoacetate hydrolase; FISH, fluorescent in situ hybridization; IHC, immunohistochemistry; ISH, in siti hybridization; RT-PCR, reverse transcription–polymerase chain reaction.

StudySpecies/StrainIntentional Hepatic InjuryIrradiation (Gy)Detection MethodHepatic Repopulation (%)
Petersen et al. [37]F-344 rats2-AAF/CCl4Dose not statedPCR, ISH, IHC0.14–0.16 at day 13
Theise et al. [38]B6D2 miceNo12FISH0.2 at 7 days; 2.2 at 6 mos
Wagers et al. [28]C57BL miceNoDose not statedImmunofluorescence0.000014 at 4–9 mos
Lagasse et al. [43](FAH−/−) 129SvJ miceFAH mutants12IHC, FISH30–50 at 6 mos
Mallet et al. [44](L-PK-Bcl2) C57BL mice(1) no;9.5RT-PCR, IHC, FISH(1) 0.000001at 14 wks;
  (2) apoptotic challenges  (2) 0.05–0.8 at 14 wks
Dahlke et al. [50]LEW ratsRetrorsine/CCl47IHCNot seen
Terai et al. [45](act-EGFP) C57BL miceChronic CCl4No irradiationImmunofluorescence26 at 4 wks

Do Stem Cell–Derived Hepatocytes Have Functional Significance?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

Although marrow-derived cells can, under defined conditions, contribute to hepatic regeneration, therapeutic potential depends on these cells conferring sufficient function. In a model using fumarylacetoacetate hydrolase (FAH)–deficient mice (a model of fatal hereditary tyrosinemia liver disease), adult bone marrow cells from wild-type animals were infused into lethally irradiated mutants. The mutants develop progressive liver failure unless treated with 2-(2-nitro-4-trifluoro-methylbenzoyl)-1,3-cyclokexanedione (NTBC). NTBC feeding was stopped after transplantation to facilitate selection of liver-repopulating cells. Upon euthanasia, bone marrow–derived cells were seen to reconstitute large portions of liver mass (up to 30%–50%), and large donor-derived nodules consisting of morphologically normal hepatocytes expressing the FAH enzyme were visualized. Assessment of biochemical function revealed that marrow-derived cells had restored these parameters almost back to normal, including expression of the missing liver hydrolase, leading to long-term survival of the animals [43]. Furthermore, by isolating HSCs from bone marrow by fluorescence-activated cell sorter, these were seen to be the only marrow cell types able to give rise to hepatocytes. Infusion of small numbers (10 to 1,000) of HSCs, along with congenic adult bone marrow cells as a radioprotective dose, was sufficient to generate a cluster of functioning hepatocytes.

In contrast, a recent study used a rodent model of CCl4 and retrorsine to determine whether marrow-derived cells could contribute to functional hepatic regeneration. Stable bone marrow chimeras were generated, and migration of donor cells into the liver was traced. Notably, infusion of bone marrow cells after injury did not result in any functional benefit in terms of animal survival [50]. Analysis of hematopoietic cells infiltrating the liver revealed that most of them were granulocytic, representing an inflammatory response to hepatic necrosis. A small number of small bile duct–associated cells were also identified, but there was no evidence of hepatocytic differentiation.

In rodents, therefore, it seems that the contribution of HSCs to liver repair and survival is, in general, limited to defined models of liver injury in which there is a survival advantage of infused stem cells.

Are Humans Simply Large Mice?

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

The characterization of stem cell biology in animal work is clearly important, but clinical application requires convincing evidence that human stem cells also share the properties demonstrated by adult rodent stem cells. Cellular phenotypes of human hepatic stem cells and tissue reactions similar to those seen in animal models have been described in a variety of human acute and chronic liver diseases [5153].

The first reports implying transdifferentiation in human cells used archival biopsies, in which liver specimens from recipients of sex-mismatched bone marrow or liver transplants were analyzed for marrow-derived hepatic cell types. Using immunohistochemistry with FISH staining for the Y chromosome, analysis of liver and bone marrow samples revealed that hepatic cell types had arisen from a marrow-derived population [54,55]. Such differentiation was also reported in patients who had undergone transfusion of peripheral-blood stem cells for the treatment of hematological malignancy, suggesting that such progenitor cells circulate in the blood [56]. In this study, epithelial cells of the skin and gastrointestinal tract were also seen to derive from cells of donor origin.

There is some concern that Y chromosome–positive cells identified in the female livers in these studies could occur as a result of the transplacental passage of male fetal blood cells during pregnancy [57], because fetal–maternal microchimerism has previously been documented [58,59]. However, this cannot be entirely responsible, because these male cells have been seen in at least one human liver from a nulliparous female [54,57] along with a female with no history of male childbearing [55]. In addition, in the murine models, female recipients were all nulliparous [38].

Additional analysis of this archival work allows interesting comparisons with the animal models. For example, specimens from human liver allografts revealed varying degrees of injury, with mild biliary obstruction in most and fibrosing cholestatic hepatitis in one. Injury severity correlated directly with hepatic engraftment frequency [54], yet recipients of bone marrow transplantation also demonstrated significant hepatic engraftment despite the absence of overt hepatic injury, in keeping with the murine model previously discussed [38]. Additionally, histology revealed that in most allografts, hepatocyte engraftment was isolated and scattered, whereas the liver with severe recurrent disease showed more extensive ductular clustering. Again, this is analogous to the animal data suggesting that, as in animals, there may be several modes of human hepatic regeneration with a potential role for adult stem cells in both physiological maintenance and acute injury.

Indeed, this highlights the need for an in vivo system in which to study human stem cells directly (Table 2). Human cord blood cells were infused into sublethally irradiated nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice, and at euthanasia, murine livers were analyzed for the presence of human hepatic cells. Despite the absence of intentional liver damage, human-derived hepatocytes were detected in murine liver [60]. Using a similar xenogeneic transplantation model, engrafted human cells were seen to produce human albumin mRNA, as detected by reverse transcription–polymerase chain reaction [61].

Table Table 2.. Hepatocyte differentiation of hematopoietic cells in human cord blood in immunodeficient mice
  1. a

    Abbreviations: 2-AAF, 2-acetylaminofluorene; β2M, β2 microglobulin; FISH, fluorescent in situ hybridization; 5-FU, 5-fluorouracil; IHC, immunohistochemistry; NOD/SCID, nonobese diabetic/severe combined immunodeficient; PH, partial hepatectomy; rhHGF, recombinant human hepatocyte growth factor; RT-PCR, reverse transcription–polymerase chain reaction.

StudyMice UsedHuman ProgenitorsLiver InjuryMarrow PreparationGrowth FactorDetection MethodHepatic Repopulation (%)
Newsome et al. [60]NOD/SCIDUnsortedNo2.5-Gy irradiationNoIHC, FISH0.011
Ishikawa et al. [61]NOD/SCID/β2M-nullCD34+ or CD45+No5-FU and antibody to c-kitNoIHC, FISH, RT-PCRNot quantified; lower than Danet et al.
Kakinuma et al. [62]C.B17/SCIDUnsorted cord blood2-AAF/PHNoNoIHC, RT-PCR, FISH0.1–1.0
Wang et al. [63]NOD/SCID or NOD/SCID/β2 M-nullCD34+ or CD34+/CD38+CD7CCl43-Gy irradiationrhHGFIHC, RT-PCR, FISH<1
Danet et al. [64]NOD/SCIDC1QrpNo3.75-Gy irradiationNoIHC, PCR0.05–0.1
Beerheide et al. [68]SCIDFibroblastoid cells from cord bloodNoNoNoIHC, RT-PCRNot quantified
Kollet et al. [66]NOD/SCIDCD34+CCl43.75-Gy irradiationNoIHC, RT-PCR0.003–0.012

The role of liver injury in such models has also been reported. Hepatic injury, induced by one-third partial hepatectomy and 2-AAF after infusion of human cord blood cells, led to the identification of functional hepatocytes [62]. Notably, in another xenogeneic model comparing injured and noninjured mice, only those who had been administered CCl4 were seen to express human albumin [63]. Additionally, this study demonstrated that it was CD34+ HSCs that gave rise to this phenomenon. However, as in murine models, this remains controversial. In another study, human stem cells expressing the receptor for the complement molecule C1q (C1qRp) were infused [64]. This is a human homologue of a mouse stem cell antigen, AA4, recognized by a monoclonal antibody used to define murine multipotent hematopoietic progenitors [65]. Notably, C1qRp is expressed on both CD34+ and CD34 cells, suggesting that this marker may define an even more primitive subpopulation. Perhaps differences in the differentiative potential of the two populations are responsible for the conflicting data regarding the role of injury.

Furthermore, recent studies in the NOD/SCID model demonstrate the role of the stromal cell–derived factor-1 (SDF-1)/CXC chemokine receptor 4 (CXCR4) axis in regulating the migration of HSCs to damaged liver [66]. Intriguingly, hepatocyte growth factor, which is upregulated in liver injury [67] and has been shown to augment hepatocytic differentiation of engrafted HSCs [63], was shown to play a key role in recruiting stem cells via its interaction with SDF-1 [66]. Some of the engrafted human cells differentiated into albumin-producing hepatocyte-like cells. Another group has used the SCID model to examine different mechanistic factors. Consistent downregulation of β2 microglobluin (β2M), an integral part of the major histocompatibility complex, was observed early after stem cell transplantation in liver tissue in which human albumin expression was upregulated [68]. It was suggested that this switching off of β2M may be an important mechanism in escaping the host immune system. Although there is clearly much to be established regarding these mechanistic factors, human blood-derived cells can, under strictly defined circumstances perhaps similar to previous animal studies, transdifferentiate into hepatic cell types, and liver injury influences the observed differentiation. As yet, there are no data to demonstrate that human stem cells have therapeutic potential, but this xenogeneic transplantation model will be useful in studying human adult stem cell biology further.

The Role of Cell Fusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

Somatic stem cell plasticity has recently been challenged by work suggesting that any transdifferentiation seen may in fact be the product of cellular fusion between infused stem cells and pre-existing differentiated cells. In vitro, mouse bone marrow cells labeled with distinct genetic markers (GFP and puromycin-resistance protein) were cultured together with ES cells in a medium containing puromycin. Surviving colonies were GFP-positive but were similar to ES cells in their morphology and growth kinetics. In addition, they expressed ES-cell proteins and were able to differentiate into various morphologies, including cardiac myocytes [69]. In a similar approach, neural stem cells were used, and once again the colonies recovered expressed GFP but had also become ES-like [70]. However, in the neural cell study, the ES cells had also been labeled with a transgene, which was subsequently expressed on recovered cells [70]. In addition, genetic analysis in both studies revealed that the derived colonies had supra-diploid DNA content, and the bone marrow–derived progeny were shown to have a hybrid genotype [69]. This suggests that the alteration in phenotype had arisen through generation of hybrids, that is, cellular fusion.

Subsequent in vivo studies have also supported the view that fusion contributes to perceived transdifferentiation. In one study, FAH-expressing liver nodules were generated by transplantation of marrow cells from wild-type males into irradiated FAH-deficient females. Genomic DNA was isolated from dissected liver nodules and probed for FAH sequences and Y chromosome sequences to measure the amounts of mutant host and wild-type donor alleles. Each hepatic nodule contained low levels of donor DNA (approximately 26%), whereas bone marrow contained the expected high proportion (>90%), which implied that donor bone marrow could not have transdifferentiated into hepatocytes [71]. Another group using the same murine model with serial transplantation of bone marrow–derived hepatocytes studied hepatocyte alleles. Analysis of DNA from the tertiary recipients revealed that although massive liver repopulation had occurred, only a fraction of the original donor genotype was preserved in the repopulating cells [72], which is less than would have been expected if they had derived solely through transdifferentiation of HSCs.

After transplantation of marrow from FAH female wild-types into male mutants, analysis of metaphase chromosomes revealed that although the nuclei of controls, and up to half of derived hepatocytes, had the expected normal male diploid (40, XY) or tetraploid (80, XXYY) karyotype, most had karyotypes predicted to result from fusion. Additionally, only a small proportion of cells contained only X chromosomes, showing that the original donor female karyotype had been lost, again suggesting fusion had taken place [72]. Similarly, calculations based on the normal ploidy values for murine hepatocytes were used to predict percentage of donor DNA that should be present within regenerating nodules in models of both fusion and nonfusion. Every nodule analyzed revealed DNA values outside the range predicted for transdifferentiation alone [71]. It is well documented that liver injury and impairment of DNA repair mechanisms, including exposure to ionizing radiation, can induce advanced hepatic polyploidy in association with terminal differentiation and cell senescence [73,74]. It has been suggested that hepatic polyploidy originates from fusion of multinuclear cells, which may well account for some of these results. However, beyond the liver, reports of fusion of bone marrow cells with Purkinje neurons and cardiac myocytes in vivo also support the idea that cellular fusion is the mechanism by which these cells become multinucleated or polyploid [30].

It remains unclear, however, which cells fuse with the host liver. Initially, it was presumed that it was the donor-derived HSCs that were fusing with host tissues, but it has been suggested that it may in fact be hematopoietic progeny of HSCs that are responsible. Among them, macrophages (Kuppfer cells) may be responsible [71], given their abundance, location, and propensity to fuse with cells under other conditions [75]. Indeed, it has been postulated that plasticity may be explained by the establishment of a myeloid lineage after homing of transplanted HSCs to the recipient bone marrow [76]. In response to liver injury, circulating myeloid cells could then be recruited to the site of damage as part of the inflammatory response and in this fusogenic environment become incorporated into the hepatic parenchyma. Indeed, this has recently been demonstrated experimentally. By transplanting lymphocyte-deficient cells and mapping the fate of the myeloid lineage, HSC-derived hepatocytes derived primarily from myeloid cells [77]. The myeloid lineage rather than the stem cells themselves would be responsible for fusion.

So is fusion important? In contrast to these reports, human studies have suggested that fusion plays little or no role in the observed transdifferentiation. In xenogeneic transplantation models, the pattern of nuclear staining, along with FISH analysis for murine and human DNA, allowed clear distinction of nuclear origin and revealed no evidence of cellular fusion [60,61]. This suggests that even if fusion occurs, it is at very low rates and does not contribute significantly to adult human stem cell plasticity in these models. In a recent study, a novel in vitro model system was used to attempt to discriminate between plasticity and fusion. HSCs were cocultured with either normal or damaged liver tissue separated by a trans-well membrane. Immunofluorescence revealed that HSCs cocultured with damaged liver lost their hematopoietic phenotype as they began to express albumin. Additionally, temporal detection of tissue-specific markers that are normally expressed during the differentiation of the liver provided additional evidence of the conversion of HSCs into hepatocytes. Fusion could not have been responsible for the plasticity observed in this system for several reasons. First, there was no contact between the HSCs and liver in the coculture system. Second, all liver tissue was taken from female donors, and HSCs were of male donor origin. Cytogenetic analysis revealed that although tetraploid cells were identified amongst the HSCs, the resulting karyotype (XYXY) was derived from male cells only and was not that expected from fusion with the female hepatocytes (XXXY). The authors conclude that HSC conversion occurs as an early event, that it occurs through genuine plasticity rather than fusion, and that microenvironmental cues are responsible for this germ-layer switch [78].

Indeed, it seems likely that differences in experimental model system may be responsible for the variations in reported mechanisms of genomic plasticity. Given the absence of cell fusion in other models of stem cell plasticity [79,80] and the compelling evidence supporting in vitro plasticity of HSCs [81], it is possible that the observed fusion in the FAH-null model reflects the extreme architectural disruption and hepatocyte membrane instability that occurs [82]. Because neither study reported that all of the marrow-derived hepatocytes were formed by fusion [71,72], and given that hepatocytes themselves are known to fuse in pathological conditions, it is important to recognize that although fusion may occur, it does not explain most new hepatocytes. The role of fusion-derived hepatocytes is unclear, but in the FAH−/− model, such cells do result in survival of the animal with no evidence of subsequent carcinogenesis [43]. Even if fusion is significant in vivo, this could be a means by which cells could be supplied with corrective genetic material as a basis for gene therapy [83].

Reports of multipotent adult progenitor cells, cells derived from adult human bone marrow [81], which can differentiate in vitro into various mature cell types, including functional hepatocyte-like cells [84], suggest that true adult stem cell plasticity occurs. Fusion cannot be responsible in this model, because there are no differentiated cells present. Taken together, it seems likely that there are a variety of mechanisms that are responsible for genomic plasticity (Fig. 1). Cellular fusion may occur in certain systems, and transdifferentiation without fusion may occur in others, but at present, it remains unclear as to the factors that regulate this.

thumbnail image

Figure Figure 1.. Models of differentiation of HSCs into hepatocytes. (A): Transdifferentiation. Upon exposure to the hepatic environment, donor HSCs undergo genetic reprogramming, switch lineage, and generate hepatocytes. The frequency at which this occurs depends on several factors, including the type and extent of liver injury. (B): Fusion. Donor HSCs, or indeed other hematopoietic cells such as macrophages, fuse with mature hepatocytes. These eventually redifferentiate into terminally differentiated hepatocytes. It should be noted that although most fusion cells contain multiples of the normal karyotypes, approximately 30% of cells are aneuploid. (C): Two-step process. Donor HSCs engraft and differentiate into hepatocytes, which then undergo fusion with mature native hepatocytes. Abbreviation: HSC, hematopoietic stem cell.

Download figure to PowerPoint

Clinical Implications

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References

Despite uncertainty surrounding the mechanisms underlying adult stem cell plasticity, there is much speculation regarding potential clinical implications. Enthusiasm has been heightened by pioneering clinical trials in other disciplines. For example, in cardiology, the delivery of marrow-derived cells into the coronary circulation of human subjects was reported to improve blood flow and cardiac function in ischemic myocardium [85,86]. In hepatology, the data presented here provide hope that somatic stem cells could eventually be used in tissue replacement protocols for the treatment of inherited and acquired end-stage liver diseases.

Use of adult stem cells overcomes many of the moral and ethical barriers of ES cell manipulation, and if somatic cells genuinely can switch lineage barriers, then HSCs are an ideal source. There is already considerable experience in their handling, and they are relatively accessible. Additionally, their administration may induce immunological tolerance, because they may potentially induce hematopoietic microchimerism, thus obviating the need for immunosuppression. Rather than relying on cadaveric organs from deceased donors who are often immunologically disparate, HSCs offer ready availability of liver-repopulating stem cells or progenitors obtained from living donors.

Despite such enthusiasm, at present there remains significant uncertainty as to what such cells would accomplish in the clinical setting, and there are many issues to be addressed before translation into clinical practice. First, a better understanding of the mechanisms that modulate the role of HSCs in physiological maintenance and liver injury is required. Specifically, more work is required to establish whether HSCs may play a therapeutic role in chronic liver disease, which is a more clinically relevant target. Second, the mechanism of genomic plasticity needs to be further defined. If transdifferentiation truly is responsible, then there could be wide-ranging utility for a range of acquired liver diseases, whereas if fusion is responsible, then this could be exploited to deliver corrective genes for hepatic metabolic disorders, as long as genetic stability in the reprogrammed cells could be assured. However, given the great disparity in levels of transdifferentiation reported and controversy regarding the mechanism responsible, these issues seem far from resolved and bring into question therapeutic strategies based on this idea. Additionally, much work is needed to assess practical issues; if HSCs are capable of engrafting directly as hepatocytes, then direct intraorgan injection may facilitate a more robust response and obviate the need for irradiation to deplete the recipient's bone marrow. Alternatively, if marrow engraftment is a prerequisite, then peripheral administration with marrow preparation and manipulation of the hepatic microenvironment to maximally attract HSCs will likely be required.

In conclusion, the new findings in adult stem cell biology are transforming our understanding of tissue repair with promising hopes of regenerative hepatology. However, perhaps we should remain cautious at present. Adult stem cell plasticity does occur, but it is a rare event even under selective pressure, and it remains to be seen whether this will be clinically significant in the human context.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Hematopoietic Stem Cells: Multilineage Plasticity
  5. Does Liver Injury Upregulate the Contribution of HSCs?
  6. Do Stem Cell–Derived Hepatocytes Have Functional Significance?
  7. Are Humans Simply Large Mice?
  8. The Role of Cell Fusion
  9. Clinical Implications
  10. References