Article Text


Adult liver stem cells: bone marrow, blood, or liver derived?
  1. H A CROSBY,
  1. School of Biosciences, University of Birmingham
  2. and Liver and Hepatobiliary Unit, University Hospital
  3. Birmingham, UK
  1. Professor A J Strain, Liver Research Labs, Queen Elizabeth Hospital, Edgbaston, Birmingham, B15 2TH, UK. a.j.strain{at}

Statistics from


Another striking development has recently been made in the hepatic stem cell debate,1 with the publication of the two reports highlighted above. Since the appearance of papers demonstrating the potential of haematopoietic stem cells to “transdifferentiate” into liver epithelium in two different rodent bone marrow transplant models,2 3 several questions have been posed. For example, are all liver stem cells derived from a common haematopoietic precursor or are oval cells, now widely accepted as having bipotentiality1 (that is, the ability to differentiate into either hepatocytes or biliary epithelial cells (BEC)), present in the liver throughout development and into adulthood? Does this transdifferentiation process represent a pathophysiologically significant phenomenon and if so does it occur in humans? Does the effect depend on the degree of liver damage inflicted?

In both studies,4 5 two groups of patients were examined; female bone marrow transplant (bmtx) recipients receiving male donor cells and males transplanted with female donor livers. The fate of the donor marrow derived cells in the former and the host haematopoietic stem cells in the latter were followed by identifying male derived cells in histological sections of livers using molecular probes specific for the Y chromosome. In parallel, phenotypic markers were used to confirm that the haematopoietic cells were indeed differentiating into either hepatocytes or BEC. In the first study liver biopsies were examined from two bmtx and four liver transplant cases.4 In most cases a mild inflammatory reaction was reported in the livers. In the second,5 samples were from 11 males whose transplanted livers were later removed because of recurrent disease, and from nine bmtx patients, although little detail of the liver pathology (that is, degree, if any, of damage or inflammation) was included.

Significantly, the overall findings were consistent with the previous rodent studies.2 3 In all cases examined, hepatocytes which expressed the Y chromosomal marker were identified. Only in the study of Theise and colleagues4 however was it suggested that BEC were also derived from the haematopoietic cells. Alisonet al reported that the frequency of Y positive hepatocytes ranged from 0.5 to 2%,5 a value that correlates well with the two experimental studies.2 3 In the study of Theise et al however, the numbers after a technical “adjustment” were much more variable and in some cases very high indeed; in the range 10–40% for hepatocytes and 4–38% for BEC.4

The conclusion which can be drawn from these studies is undoubtedly consistent with the hypothesis that (some) haematopoietic cells can give rise to liver epithelium and therefore represent extremely important findings in the age old study of liver repair processes. However, these developments raise a number of points which will require further clarification. Firstly, how widespread is this cellular response? Taken at face value, the extrapolated values reported by Theise et al seem high. If these reflect the true response which can occur in vivo, it clearly indicates that this is a significant phenomenon and one which could surely be exploited in vitro as a novel source of human hepatocytes (and biliary cells) for other studies, including cell transplantation, toxicology testing, or the development of a viable bioartificial liver system. However, the presence of a large number of haematopoietic cells trafficking through the liver6 may lead to overestimation of the numbers of transdifferentiated cells. Secondly, what is the relationship between this cellular response and the degree of liver damage? At present, this is difficult to assess as in the study of Theise et al, in only one patient was any significant degree of liver damage indicated and in the brief report from Alisonet al details were lacking. Conceptually it seems unlikely that this transdifferentiation phenomenon will occur without liver damage as in those circumstances there is no requirement for cell replacement. On the other hand, in one of the two experimental rodent models, no direct hepatic injury was invoked.3

Another question which requires careful consideration is whether there is a discrete subpopulation of specialised blood derived cells with this plasticity or can all haematopoietic stem cells, if exposed to the correct environment, differentiate into liver cells? As yet little is known about the specific phenotype of the cells responsible. However, in recent work, CD34 or c-kit positive cells (markers expressed on haematopoietic stem cells), isolated from both normal and diseased human livers, have been shown to differentiate into BEC in vitro.7 The next step will be to induce differentiation into the hepatocyte phenotype. Finally, do these observations help resolve the controversy surrounding the question of the origin of hepatic stem (oval) cells? The answer to this remains open, although the current consensus view is that there may well be more than one progenitor cell population capable of this cellular phenotypic transition and therefore more than one cellular repair pathway leading to liver cell repopulation.

Probably the most challenging issue will be to elucidate the mechanisms regulating the cellular differentiation processes involved. The answer is likely to be complex but must surely lie in the microenvironment of the cells, the signals which originate from the extracellular matrix through adhesion related events,8 and the cocktail of soluble (or membrane anchored) ligands now known to control cell growth, differentiation, and morphogenesis.9 Progress, which depends on the design of suitable cell transplantation and cell culture based models, is moving rapidly. As well as the liver, haematopoietic stem cells can give rise to muscle and brain and vice versa.10 The parallel between the adult liver and brain, two organs traditionally recognised to have a very low cell turnover rate but which both harbour cells with stem cell properties, is remarkable. It is certainly true that the plasticity of the cells in question goes way beyond that which was originally believed to be the case, at least in the adult.

These latest reports by Theise and colleagues4 and Alison and colleagues5 extend the current haematopoietic/liver stem cell controversy to the clinical setting and so intensify the search for a novel source of human hepatocytes, which are much in demand. They also suggest an alternative mechanism of liver regeneration and repair following injury for the patient with liver failure. Although we are still some way off a situation where the hepatologist refers the patient with acute or chronic liver failure to the haematology clinic for a “shot” of the magiccellular potion, the capacity to accelerate the liver repair process by this means may be a realistic possibility in the foreseeable future.

Note added in proof

Another important recent report11 indicating that in mice with a fatal hereditary tyrosinaemia, transplantation of bone marrow derived haematopoietic stem cells can rescue the animals and lead to liver cell repopulation. This addresses some of the issues raised here.





It has been shown in animal models that hepatocytes and cholangiocytes can derive from bone marrow cells. We have investigated whether such a process occurs in humans. Archival autopsy and biopsy liver specimens were obtained from 2 female recipients of therapeutic bone marrow transplantations with male donors and from 4 male recipients of orthotopic liver transplantations from female donors. Immunohistochemical staining with monoclonal antibody CAM5.2, specific for cytokeratins 8, 18, and 19, gave typical strong staining of hepatocytes, cholangiocytes, and ductular reactions in all tissues, to the exclusion of all nonepithelial cells. Slides were systematically photographed and then restained by fluorescence in situ hybridization (FISH) for X and Y chromosomes. Using morphologic criteria, field-by-field comparison of the fluorescent images with the prior photomicrographs, and persistence of the diaminiobenzidene (DAB) stain through the FISH protease digestion, Y-positive hepatocytes and cholangiocytes could be identified in male control liver tissue and in all study specimens. Cell counts were adjusted based on the number of Y-positive cells in the male control liver to correct for partial sampling of nuclei in the 3-micron thin tissue sections. Adjusted Y-positive hepatocyte and cholangiocyte engraftment ranged from 4% to 43% and from 4% to 38%, respectively, in study specimens, with the peak values being found in a case of fibrosing cholestatic recurrent hepatitis C in one of the liver transplant recipients. We therefore show that in humans, hepatocytes and cholangiocytes can be derived from extrahepatic circulating stem cells, probably of bone marrow origin, and such “transdifferentiation” can replenish large numbers of hepatic parenchymal cells.

(2) Alison M, Poulsom R, Jeffery R,et al. Hepatocytes from non-hepatic adult stem cells. Nature2000;406:257.

View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.