Stem cell therapy for chronic liver disease-choosing the right tools for the job

[Guest guest]

Stem cell therapy for chronic liver disease-choosing the right tools

for the job

Commentaries

Gut Feb 2008;57:153-155

Stuart J Forbes

Correspondence to:

Professor Stuart J Forbes, MRC/University of Edinburgh Centre for

Inflammation Research, The Queen's Medical Research Institute, 47

Little France Crescent, Edinburgh EH16 4TJ, UK; stuart.forbes@...

The liver has a fantastic regenerative capacity but, following

chronic liver damage, this begins to fail, and then fibrosis, and

eventually cirrhosis, develops.1 Currently the only curative

treatment for advanced liver cirrhosis is liver transplant. Although

liver transplant has become a procedure with a relatively good 5-year

survival, organ donation has not kept up with demand, which has

resulted in an increasing number of patients on the liver transplant

waiting list waiting longer for a donor organ, which leads to

increased morbidity and mortality.2 Furthermore it is estimated that

over the next few years there will be a 5-fold increase in the need

for liver transplantation in the UK.3 Although there is emerging

evidence that extending the donor organ criteria may impact on this

mortality rate,4 there is clearly still an urgent need to develop

alternative strategies for the treatment of advanced liver disease,

and numerically cirrhosis is the most important target. It is with

this background that there has been understandable enthusiasm for the

development of stem cell therapies for liver regeneration.

Bone marrow (BM) stem cells have been intensively investigated as a

potential source of liver stem cells and as a means to regenerate the

cirrhotic liver, and it is worth briefly outlining why they have

attracted this attention. There are a population of intrahepatic

progenitor cells, termed oval cells in rodents, which can take over

liver regeneration when the usual mode of regeneration, via division

of mature hepatocytes, begins to fail. It was suggested some time ago

that these cells expressed the haematopoeitic stem cell marker THY-15

(though this is now disputed), and it was therefore postulated that

these oval cells were originating from the BM. Indeed initial

experiments supported this hypothesis.6 Furthermore, following

analysis of both mouse BM transplant models and tissue from liver and

BM transplant patients, it was suggested that the BM could contribute

to the mature hepatocyte population.7 8 Real excitement resulted from

the demonstration that BM transplant could rescue a mouse model of

tyrosinaemia, a hereditary defect of a hepatocytic enzyme.9 The BM

transplant resulted in repopulation of the liver with apparently

normal hepatocytes. Given that this model was otherwise lethal to the

mice, it was a powerful demonstration of the therapeutic potential of

stem cells, and stimulated further studies using different animal

models of liver disease. Unfortunately, these studies have mostly

been negative and little evidence has been found for a significant

repopulation of liver parenchyma by BM-derived cells.10-13

Furthermore, subsequent analysis of the tyrosinaemia model showed

that the rescue was a result of cell fusion of BM-derived monocytes

with the diseased hepatocytes, resulting in a form of cellular gene

therapy. Because of the unique selection pressure seen in this model,

the fused cells divided to repopulate the liver.14 A selection

pressure of this strength is unlikely to be available for most

clinical applications.

Mesenchymal stem cells (MSCs) can be derived from various tissues

including bone, fat and dental tissue. Their definition includes the

ability to differentiate into osteoblasts, adipocytes and

chondroblasts.15 There are now several studies that show that MSCs

can be coaxed in vitro into many cell types, including cells

with " hepatocyte-like " properties, based on a combination of

morphology, gene expression and metabolic/synthetic activity.16-20

The cells are often described as being " hepatocyte-like " as they

often fall short of being identical to hepatocytes. Adipose tissue-

derived MSCs (AT-MSCs) appear capable of differentiation into a

hepatocyte phenotype in vitro.21-23 For example, Banas et al found

that a particular fraction of human AT-MSCs (CD105+) could become

hepatocyte-like, with gene expression, morphology and metabolic

activity similar to hepatocytes.23 With the current excess of human

adipose tissue, these cells could be a readily available source with

many willing donors! Data on the in vivo capabilities of MSCs are

less clear, with some studies assessing functional performance for as

little as 24 h post-transplant,23 or using limited criteria to define

hepatocyte-like cells. Of the studies where longer term engraftment

was studied, the data are still conflicting. For example, Sgodda et

al found that pre-differentiated rat AT-MSCs were capable of

integrating into recipient livers for at least 10 weeks and appeared

to form appropriate connections in the recipient liver.22 In an

apparently robust study, Sato et al found that human MSCs could

engraft immunosuppressed rat livers and form relatively large numbers

of donor-derived cells with a hepatocyte phenotype based on gene

expression and protein criteria.24 Furthermore, a recent study in

fetal sheep used direct intrahepatic injection of clonal populations

of MSCs to produce a significant number of human hepatocyte-like

cells (12.5%) that were both long lasting (56-70 days) and able to

secrete human albumin.25

Conversely, in an equally carefully conducted study, Popp et al found

that MSCs did not differentiate into hepatocytes within the liver of

rats undergoing prolonged liver injury.26 Unfortunately, such

apparently contradictory results have been commonly seen in this

field and probably reflect a combination of factors, including the

use of varying cell derivation and differentiation protocols, and the

fact that the models of liver injury are often differing and rarely

repeated. Notwithstanding these reservations, it appears that MSCs

can be induced to demonstrate at least some hepatocyte functions in

vitro, but more in vivo research is definitely needed as it appears

that both the exact derivation and pre-conditioning of the starting

cells and the host liver environment will determine the resulting

phenotype of the transplanted cells.

It is in this context that the current work by Valfrè di Bonzo et

al27 makes interesting reading (see page 10.1136/gut.2006.111617).

Using the differentiation protocol published by Lee et al,16 they

confirmed that cells could be found to express human albumin mRNA

weakly. However, when tested in a NOD-SCID model, that importantly

had chronic liver injury, very few human hepatocyte-like cells were

identified. Worse still (ie, if you want to " make hepatocytes " ), they

found the transplanted MSCs had a propensity to form myofibroblast-

like cells (scar-forming cells) in the areas of hepatic injury, and

by the end of the study period the transplanted cells were >10 times

more likely to form myofibroblasts than hepatocytes. This less

optimistic but undoubtedly realistic study is, however, perhaps not

too surprising. We and others have shown that BM-derived cells can

form myofibroblasts within various damaged organs including the

liver.28 Furthermore, the data from murine models have implicated

MSCs as the major source of these BM-derived myofibroblasts.29

This report is particularly timely as clinical studies of stem cell

therapy for cirrhosis and liver disease are now beginning to be

undertaken in earnest. To date there are at least four studies

reported of BM therapy for liver disease. The first was in patients

undergoing resection for liver cancer; here autologous CD133+ BM stem

cells were used to stimulate the liver's regenerative capacity.30 The

protocol involved portal vein embolisation to induce contralateral

lobe hypertrophy and thereby increase the size of the future remnant

liver volume prior to an extensive partial hepatectomy. Autologous

CD133+ BM cells were infused into the portal vein supplying the

future remnant and appeared to stimulate the resulting liver

regeneration. In another study by Gordon et al, a population of CD34+

cells was identified that in vitro appear to have multilineage

potential. In a study of five cirrhotic patients, autologous CD34+

cells were isolated using leucopharesis and re-injected via the

hepatic artery or portal vein. Again this largely appeared to induce

an improvement in the liver function of these patients.31 Similarly,

studies by Terai et al and Lyra et al have used autologous BM-derived

monocytes.32 33 The cells were concentrated ex vivo and re-applied to

the patient's hepatic artery or portal vein. All the protocols appear

to be safe (including the use of granulocyte colony-stimulating

factor (GCSF)), and indeed the data may even be encouraging; however,

none of these studies was a definitive controlled study-which are

likely to follow. The cell populations used in these clinical studies

to date are either enriched for haematopoietic stem cells (HSCs) or-

like monocytes-are derived from HSCs. This may seem odd given that

the majority of the data cited suggest that it is the population of

MSCs that are likely to be inducible into a hepatocyte-like

phenotype, and there are few hard data to suggest HSCs have this

potential to any useful degree. There is, however, logic to this

approach as the monocyte-macrophage population appears to have the

potential to remodel the scars of the cirrhotic liver. In a recent

study into the mechanism of liver fibrosis formation and resolution,

hepatic scar-associated macrophages were selectively deleted in

either the phase of fibrosis progression during injury or the phase

of fibrosis resolution post-injury.34 If they were deleted during the

injury phase, there was less fibrosis formation-demonstrating they

can have a pro-fibrogenic role; however if they were deleted during

the phase of resolution, then there was less scar resolution-thereby

demonstrating their role in scar resolution and tissue remodelling.

This report adds weight to the growing feeling that MSCs can have

their downsides as well as upsides, especially if applied in vivo to

damaged organs, and is a timely caution for anyone contemplating

clinical trials of MSC therapy for liver cirrhosis. This is not to

say that MSCs have no future in this area. It may be that if MSCs can

be differentiated efficiently into a hepatocyte phenotype, then these

cells can be used for ex vivo liver support, drug testing, etc. For

transplantation protocols, one would need to be confident that the

cells were pre-programmed either to differentiate into or to remain

as hepatocytes, and it appears we are some way short of this mark.

Finally, this paper highlights a key conceptual issue in the

understanding of regenerative therapy for the cirrhotic liver. Is it

likely to be fruitful to add hepatocyte-like cells to an already

cirrhotic liver, where the environment is extremely harsh, and even

professional hepatocytes find the going tough, or should one seek to

modify the cellular and extracellular milieu, to allow the endogenous

hepatocytes and progenitor cells to regenerate the liver? The latter

may well be a more realistic medium-term goal for cell therapy. In

other words, spending some time improving the soil may bring forth

more regeneration than simply throwing more seeds onto rocky ground.

In the end, the humble rake may be a better tool for this job than

that fancy double-edged sword.