Clinical Science (1997) 92, 103-1 12 (Printed in Great Britain)
I03
Editorial Review
Pathogenesis of liver fibrosis
R. ALCOLADO, M. J. P. ARTHUR and J.
P. IREDALE
University Medicine, University of Southampton, U.K.
1. Liver fibrosis is a common sequel to diverse liver
injuries. It is characterized by an accumulation of
interstitial collagens and other matrix components.
The hepatic stellate cell is pivotal to the pathogenic
process. Fibrotic liver injury results in activation of
the hepatic stellate cell which undergoes a phenotypic change to a proliferative myofibroblast-like
cell which synthesizes excess interstitial collagens
and other matrix components.
2. The process of initiation of activation and its
perpetuation result from complex, often interrelated
series of signalling mechanisms which converge on
this effector cell. Such mechanisms include alterations in matrix resulting in changed cell-matrix
interactions and stimulation by cytokines released
from damaged hepatocytes, infiltrating inflammatory cells, Kupffer cells and matrix. Foremost
among the profibrotic cytokines is transforming
growth factor PI.
3. Once the hepatic stellate cell is activated the
preceding matrix changes and recurrent injurious
stimuli will perpetuate the activated state.
4. Despite the accumulation of excess collagens, the
liver retains a capacity for matrix degradation. This
capacity may be overwhelmed and any secreted
matrix remodelling enzymes may be inhibited by the
concurrently secreted tissue inhibitors of metalloproteinase-1 and az-macroglobulin.
5. Our understanding of the molecular pathogenesis of liver fibrosis is increasing. It is anticipated
that this knowledge will provide novel therapeutic
avenues to treat this disease process.
changes in the nature of the matrix components and
their distribution within the liver. Recent research
has highlighted several key areas of the pathological
process which has enhanced our understanding not
only of liver fibrosis, but of the fibrotic process in
other solid organs.
The key cellular and molecular events involved in
the pathogenesis of liver fibrosis include activation
of hepatic stellate cells (stellate cells, lipocytes,
fat-storing cells or Ito cells) to a myofibroblast-like
phenotype, their production of excess matrix
proteins and their ability to regulate matrix degradation. The mechanisms controlling stellate cell activation and the factors regulating the synthesis and
degradation of matrix proteins have thus become
the principal focus of research activity in liver
fibrosis. This work encompasses analysis of interactions between stellate cells and both parenchymal
cells (hepatocytes) and non-parenchymal cells
(Kupffer cells and endothelial cells) as well as
examining the cell-matrix interactions that determine stellate cell and hepatocyte phenotype and
behaviour, and the role of soluble cell messengers
(cytokines, growth factors and other soluble messengers) released by inflammatory cells and resident
liver cells.
The purpose of this article is to review the current
state of knowledge of this complex pathological
process and to describe the principal events that
occur as the previously ordered liver microanatomy
progresses to the distorted architecture that characterizes fibrosis.
INTRODUCTION
MICROANATOMY, STRUCTURE AND CELLS OF THE
HEPATIC SlNUSOlD
Liver fibrosis and cirrhosis represent the final
common pathological pathway for a variety of liver
diseases, irrespective of the underlying aetiology. In
the fibrotic liver there is not only an increase in the
total content of collagens and other matrix proteins
compared with normal, but there are also qualitative
The functional unit of the liver is the acinus which
consists of a central portal tract containing an
efferent biliary radical and terminal branches of the
portal and arterial blood supplies. The blood
supplies feed into the hepatic sinusoid, which are
bounded by plates of hepatocytes, and finally drain
Key words: cell-matrix interaction, hepatic stellate cells, liver fibrosis, matrix, matrix regulation, tissue inhibitor of metalloproteinase-I, transforminggrowth factor 8.
Abbreviations: IGF, insulin-likegrowth factor; M6P, mannose-6-phosphate: MMP, matrix metalloproteinse; PDGF, platelet-derivedgrowth factor; TGF-/I, transforminggrowth
factor-8; TIMP-I, tissue inhibitor of metalloproteinase-I;uPA, urokinase plaminogen activator.
Correspondence: Professor M. J. P. Arthur, University Medicine, Level D, South Block. Southampton General Hospital, Trrmona Road, Southampton. Hampshire SO1 6YD, U.K.
R. Alcolado et al.
104
into the efferent central vein. Hepatocytes make up
over 80% of the volume of the human liver and are
the site of almost all liver synthetic and degradative
functions. Normal hepatocytes are cuboidal in shape
and are arranged in palisades along the sinusoid, in
close contact with sinusoidal blood.
The vascular sinusoid is lined by a highly specialized fenestrated endothelium and contains within its
lumen Kupffer cells and pit cells. Kupffer cells are
the resident tissue macrophages found predominantly in the periportal areas of the sinusoid and are
responsible for removing bacterial antigens and
endotoxin from the portal blood. Endothelial cells in
the liver are highly specialized, possessing many
fenestrae which allow free passage of macromolecules from sinusoidal blood into the space of Disse,
thus permitting contact with hepatocytes. Pit cells
bear a strong resemblance to the natural killer cells
of the spleen which play a major role in clearing
opsonized particles and tumour cells from the blood.
Between the sinusoidal endothelium and the
brush-bordered sinusoidal surface of the hepatocyte
lies the space of Disse. This contains a specialized
matrix with the same principal matrix components
as a basement membrane (type IV collagen, laminin,
proteoglycans and the stellate cells) (Fig. 1).
Stellate cells, the major focus of this review, are
located in the space of Disse and are believed to be
the liver equivalent of pericytes, with many
cytoplasmic processes encircling the sinusoid. They
contain many lipid-rich droplets which store over
75% of the body’s store of retinoids, largely as
retinyl esters. During liver injury stellate cells
undergo activation to a myofibroblast-like phenotype, proliferate and express the matrix proteins
which characterize fibrosis. The pathogenic role of
stellate cell activation and its relationship to liver
matrix and soluble cell messengers is considered in
detail below.
m-0
QUIESCENT HSC
PROFIBROTIC INJURY
0
0
LIVER MATRIX IN NORMAL LIVER
In normal liver, matrix proteins are distributed
throughout the organ, but two principal types of
matrix may be identified. Fibrillar or interstitial
matrix proteins (collagens I, I11 and V) are located
in the liver capsule, large and medium-sized blood
vessels and the portal tracts. This contrasts with the
basement membrane-like matrix located in the
functionally important sinusoidal space of Disse.
The latter matrix is not electron dense but is
comprised of type IV collagen, laminin and proteoglycans as determined by immunohistochemical
studies [l-51.
Of the many collagen subtypes described to date
only five have been detected in liver; they are types
I, 111, IV, V and VI. The fibril-forming collagens
make up the majority of collagen in normal liver [6]
and are found predominantly in the perivascular
spaces, around portal tracts, and subcapsular areas.
Type I and I11 collagens form hybrid fibrils [7]
condensing with type V fibres which forms the fibril
core.
Type IV collagen is located in the space of Disse
[8], together with type VI collagen [9], as well as in
vascular and bile duct basement membranes. Type
IV collagen together with laminin and proteoglycans
forms a flexible lattice structure [lo, 111 while type
VI collagen monomers interact with each other to
form chains of molecules within the space of Disse
t121.
Periductal and perivascular areas are rich in fibronectin and vitreonectin. The basement membranelike matrix in the space of Disse contains fibronectin
[2, 8, 131 and laminin in the same distribution as
type IV collagen. Undulin is found in small amounts
in association with the fibril-forming collagens [14].
Heparan sulphate, the commonest proteoglycan in
normal liver [13], is found predominantly at the
portal end of the sinusoid, with heparin proteoglycan at the venous end [15]. Dermatan and
chondroitin sulphate are found in very small
amounts in normal liver [15, 161.
mpATocm
Normal baxmcnlmcmbmc-like
matrix rich in urllagen N
1-
and synlhcrk collagens I and 111
Fig. I. Major changes that occur in the sinusoid with profibrotic
injury. The hepatic stellate cells (HSCs) become activated, and proliferate
and synthesize collagens I and 111 which replace the normal basement
membrane-like matrix. In consequence, hepatocyte and endothelial cell
phenotype is altered and the sinusoid becomes capillarized.
MATRIX CHANGES IN DISEASED LIVER
The early response to liver injury is characterized
by the activation of stellate cells. This event is
common to a wide range of liver insults. A major
consequence of stellate cell activation to a myofibroblast-like phenotype is that the synthetic activity of
these cells alters and they secrete excess collagen I
and I11 in addition to other matrix molecules. These
matrix proteins are laid down in the space of Disse,
initially causing an alteration in the appearance of
the sinusoid. With further progression of fibrosis
there are alterations in the phenotype of the
endothelial cells which lose their fenestrae, so-called
capillarization of the sinusoid, and perturbation of
hepatocyte function (Fig. 1).
Further progression of the fibrotic process is
Pathogenesis of liver fibrosis
associated with a gross distortion of the sinusoidal
architecture as the vascular structures at either end
become linked by dense connective tissue. Connective tissue septae can also be visualized extending
into the parenchyma. The final result of this process
is the end-stage cirrhotic liver in which the whole
anatomy is macroscopically distorted by thick bands
of collagen, within which nodules of hepatocytes,
often with regenerative foci, are surrounded.
Cirrhotic liver contains approximately six times
the amount of matrix as normal liver [6]; there is an
increase in the amounts of types I, I11 [8] and IV [l]
collagen in the Space of Disse, but for type I
collagen this increase is disproportionate so that the
ratio of type I [17, 181 to types I11 and IV increases.
There is an increase in laminin [6] within the space
of Disse and alterations in both the total quantity
and the type of fibronectin produced [6]. The
spectrum of glycoproteins expressed changes with a
relative decrease in heparin and heparan sulphate
proteoglycans and an increase in dermatan and
chondroitin sulphate proteoglycans. The latter
contain the core proteins decorin and biglycan which
have a potentially important role in cytokine binding
within the extracellular matrix and are discussed in
more detail below. Hyaluronan, which is only
present in small quantities in normal liver, is
increased more than 8-fold in diseased liver and is
secreted by hepatic stellate cells [19].
CELLULAR SOURCES OF MATRIX
Early studies of matrix secretion by isolated cells
indicated that hepatocytes were a significant source
of matrix components. More recently, access to
rigorously characterized cell populations has cast
doubt on this finding, suggesting that it may have
resulted from overgrowth of contaminating stellate
cells, and has given a clearer picture of the role of
individual cell types in matrix synthesis. These
findings have been confirmed by studies using in situ
hybridization or analysis of cell fractions extracted
from normal and injured liver to confirm the origin
of matrix components.
Hepatocytes contain the mRNAs for collagens I
[20] and IV although not all workers [21] have been
able to reproduce these results. Cultured hepatocytes produce plasma fibronectin [22] and the
proteoglycan heparan sulphate [4].
Current studies indicate that stellate cells are the
major source of matrix proteins in the diseased liver
[21-231. They produce a broad range of matrix
components - mRNA for collagen type I, I11 and IV
[24-261 has been isolated along with mRNAs for
laminin [5, 251, fibronectin [22, 241, heparan
sulphate [4] and dermatan and chondroitin sulphate
[4, 271 proteoglycan and the collagen synthetic
enzyme, prolyl hydroxylase.
mRNA transcripts for type IV collagen [21, 281
and fibronectin [29] have been identified in endo-
I05
thelial cells, while Kupffer cells and Pit cells do not
appear to elaborate any matrix components on the
basis of present data. Doubt has been cast on earlier
data suggesting that hepatocytes are a major source
of matrix after it was demonstrated that hepatocyte
cultures are often contaminated with stellate cells
~301.
BIOLOGICAL EFFECT OF MATRIX CHANGES IN
FIBROSIS AND THEIR RELATIONSHIP TO STELLATE
CELL ACTIVATION
In addition to the described architectural changes,
the shift in matrix constitution to one rich in
collagen I and I11 is associated with specific changes
in the phenotype of parenchymal and non-parenchymal cells.
The evidence for a biological activity of matrix lies
in a series of elegant experiments which have
demonstrated that hepatocytes retain a differentiated phenotype in which they continue to express
albumin and cytochrome P-450, when cultured on a
model basement membrane. This biological effect of
the whole-model matrix is greater than each of its
constituent parts, when these are used as individual
culture substrata. In contrast, culture of hepatocytes
on type I collagen or tissue culture plastic results in
a flattening and spreading of the cells, in association
with which, tissue-specific gene expression is lost
[31, 321. Interestingly, more simple matrices which
result in cultured hepatocytes retaining their threedimensional shape, including a gelatin sandwich,
have been demonstrated to maintain gene-specific
function, suggesting that the maintenance of normal
shape is important for the persistence of a differentiated phenotype [33].
The implications of these studies are that hepatocyte function may be disrupted comparatively early
in the progression of liver fibrosis and before gross
changes occur, by virtue of altered cell-matrix interactions. In addition, the findings are consistent with
the hypothesis that hepatocellular dysfunction may
remain a feature of liver injury even after
withdrawal of the injurious agent in the absence of
adequate matrix remodelling.
Evidence also exists for important cell-matrix
interactions involving endothelial cells [16] and,
pertinent to this article, stellate cells. Of importance
for both the initiation and progression of fibrosis,
stellate cells respond to a change in substratum by
undergoing a phenotypic change [34]. Stellate cells
in normal liver lie in the space of Disse. They are
rich in retinoids and fat droplets and can be readily
recognized. During liver injury these cells become
activated to a myofibroblast-like phenotype [35].
They shed their retinoids after hydrolysing the esters
and start to express a-smooth muscle actin [36-391.
This phenotypic change is central to the pathogenesis of fibrosis because it is in this activated phenotype that stellate cells express and secrete excess
I06
R. Alcolado et al.
matrix proteins during fibrosis and downregulate
their subsequent degradation.
The activation of stellate cells is a general
response to injury and can be demonstrated to be
temporally related to an insult (after cc14 intoxication, stellate cells can be demonstrated to pass
through stages of activation, so-called transitional
cells). However, activation is also related to the site
of an injury; thus after cch intoxication, stellate cell
activation will occur in association with hepatocyte
degeneration pericentrally, while bile duct ligation is
associated with periportal stellate cell activation. In
the presence of recurrent or progressive injury,
stellate cell activation becomes more generalized
[21, 37, 38, 40-421.
By culture on a model basement-like matrix,
stellate cells can be maintained in a normal quiescent, retinoid-rich phenotype. When type I collagen
or uncoated tissue culture plastic is substituted the
stellate cells rapidly undergo a phenotypic transformation to activated cells in a manner that is very
similar to the process in vivo [25, 34, 431. These
findings have led to the hypothesis that alterations
in the matrix substratum, specifically the degradation of the normal basement membrane-like matrix
and its replacement with a matrix rich in interstitial
collagens, may be important in both the activation
and perpetuation of activation of stellate cells.
Furthermore, the demonstration that activation can
be mimicked in tissue culture, by plating primary
cultures of stellate cells onto plastic, has greatly
facilitated the study of stellate cell expression and
synthesis of matrix proteins and matrix remodelling
enzymes.
Another important matrix change which occurs
early in the course of experimental liver injury and
is important for stellate cell activation is the production, by endothelial cells, of fibronectin spliced to
contain the EIII-A variable region [29]. This isoform
is virtually absent from normal liver. The peak in
EIII-A fibronectin production occurs within 12 h of
experimental bile duct ligation and remains elevated
for 7 days. Stellate cells cultured with endothelial
cells isolated from bile duct-ligated livers become
activated more rapidly than equivalent controls. This
action of the endothelial cell can be blocked by
antibody directed against the EIII-A region of fibronectin [29], suggesting that the endothelial cell
response to injury may directly activate stellate cells
via changes in matrix environment.
The mechanisms of cell-matrix interactions
involving stellate cells, hepatocytes and endothelial
cells are likely to be mediated via integrin and
non-integrin receptors. Integrins are transmembrane
heterodimeric proteins which mediate cell adhesion
and cell signalling and may be important in the
secretion of ordered matrix by synthetic cells [44,
451. Normal basement membrane constituents
contain well-characterized integrin recognition
sequences. The repeated Arg-Gly-Asp motif which
occurs in the triple helical region of the type I [46]
and type IV [47] collagen molecule is recognized by
the a$l integrin receptor which occurs on hepatocytes [47]. Fibronectin [48] bears a number of
integrin recognition sequences for the M5Pl [49]
(Arg-Gly-Asp) and C(4Pl [50] (Leu-Asp-Val, or
Arg-Glu-Asp-Val)
receptors. It also contains a
number of recognition sequences for other matrix
components as well as for the transmembrane
proteoglycan syndecan. Laminin [49] has eight cell
adhesion or recognition sequences, many of which
recognize other matrix components, especially
fibrillar collagens, and are thought to be responsible
for the arrangement of collagen into fibrils. Laminin
also contains epidermal growth factor-like peptide
repeats which, when hydrolysed, are capable of
stimulating cellular epidermal growth factor receptors [51]. Proteoglycans interact with other matrix
components via non-integrin receptors but their
major interaction with cells may be mediated via
their ability to bind cytokines. Heparan sulphate
proteoglycan, found in normal liver, is capable of
binding /3-fibroblast growth factor [52], which
prevents its rapid degradation by membrane-bound
plasmin and enhances binding to specific P-fibroblast growth factor receptors. Decorin and biglycan,
the protein core of chondroitin and dermatan
sulphates, are capable of binding TGF-PI reversibly
[531*
OTHER FACTORS INFLUENCING STELLATE CELL
ACTlVATlO N
In addition to alterations in the matrix
substratum, there is evidence that stellate cell activation is influenced by a wide variety of cytokines and
cell messengers. Kupffer cells are probably central
to the response of the liver to many injuries. Indeed,
experimental liver injury can be ameliorated by prior
depletion of Kupffer cells. Foremost among the
bioactive substances released by Kupffer cells is the
cytokine TGF-PI (Fig. 2). Extensive studies of this
cytokine elsewhere in the body indicate that its
expression is fundamental to the scarring response.
Transforming growth factor P, so-called because
of its ability to transform a non-proliferating fibroblast cell line in soft agar culture [54], inhibits the
proliferation of rat stellate cells [55] but promotes
the platelet-derived growth factor (PDGF)-mediated
proliferation of human stellate cells [56]. TGF-PI is
one of a family of related peptides, TGF-Pl-TGFPs, of which only TGF-PI, P 2 and P 3 have been
detected in liver [57].
In disease models including CC14 intoxication, bile
duct ligation, iron and alcohol intoxication, streptococcal cell wall administration and murine schistosomiasis [58-621, an increase in the level of mRNA
for TGF-PI has been documented. Fibrotic liver
from patients with a variety of chronic liver diseases
has been shown to contain increased levels of
TGF-Dl mRNA [63]. In chronic liver injury TGF-PI
I07
Pathogenesis of liver fibrosis
PDGF
-
ACTIVATION
-
TGFP,
ACTIVATED
MYOFIBROBLASTLIKE HSC
PROLIFERATION
COLLAGEN-I
SYNTHESIS
Fig. 2. Major cytokines and mechanisms believed to be important in
the process of hepatic stellate cell (HSC) activation to the proliferative myofibroblast-like phenotype observed in fibrogenic injury
is produced by a number of non-parenchymal cells
[64]. TGF-PI may be derived from activated stellate
cells [65], Kupffer cells [55] and endothelial cells
[60]. In addition, TGF-PI may be produced by
migrating mononuclear cells [58].
TGF-Pl has profound effects on matrix production in a number of well-characterized cell lines [66,
671 including fibroblasts, osteoblasts and, to a lesser
extent, myofibroblastic stellate cells [24, 681. TGF-PI
not only causes an increase in the production of type
I collagen but also in types I11 and IV [24]. In
addition, fibronectin [69, 701 and proteoglycan [55,
711 synthesis is enhanced. In primary stellate cell
culture, addition of TGF-PI 24 h after plating results
in an accelerated transformation to the activated
phenotype [69].
TGF-PI also alters the balance of matrix turnover
against matrix degradation and in favour of matrix
accumulation, by inhibiting production of interstitial
collagenase [72, 731, stromelysin and plasminogen
activators, and promoting the production of the
potent collagenase inhibitor, tissue inhibitor of
metalloproteinase-1 (TIMP-1) [72] and plasminogen
activator inhibitor-1 [74]. Fibroblasts will also
respond to TGF-PI by upregulating gelatinase A
[73]. However, this has not been demonstrated in
stellate cells [75]. A further significant effect of
TGF-PI on stellate cells is the stimulation of a
positive autocrine loop [24, 651. TGF-PI increases
the facility for stellate cell-matrix interactions by
increased transcription of mRNA for the a and P
subunits of integrin receptors and increased
assembly of these subunits to form the active receptors 1761.
Relative levels of mRNA cannot be directly correlated with active TGF-PI because the protein is
released in a latent form. The mechanism of activation of TGF-PI in liver is probably via the insulin-
like growth factor II/mannose-6-phosphate (IGFII/M6P) receptor. Moreover, it has been demonstrated that stellate cells extracted from CC14treated rats but not normal liver bear the IGF-III
M6P receptor on the cell surface [77].
The balance of activated versus inactivated
cytokine must be of central importance as TGF-Pl is
detectable in normal liver [77] and almost all normal
liver cells exhibit receptors for active TGF-PI, while
mRNA levels for receptor types 1, 2 and 3 are all
increased in activated stellate cells [78]. Latent
TGF-PI may also exist in a pool within the liver
matrix. The type 3 TGF-Pl receptor has long been
known to be a non-signalling receptor: it has
recently been identified as membrane proteoglycan
P-glycan [79, 801. In addition, the matrix component
decorin binds and inactivates TGF-PI. The binding
of TGF+ to these matrix constituents may both
prevent activation and protect the latent form from
breakdown. The characteristics of this binding in
vivo are not understood, particularly the conditions
required for the release of latent TGF-Pl from these
receptors. Nonetheless, decorin has been demonstrated to have an anti-fibrotic effect in renal fibrosis
in vivo [Sl]. Moreover, the position of TGF-Pl as a
central cytokine to the fibrotic/scarring process has
now been demonstrated in a series of mechanistic
studies in which fibrosis was reduced by the addition
of decorin and neutralizing antibodies to cell culture
and in vivo models [81-841. Perhaps predictably, a
transgenic line of TGF-PI overexpressor animals
developed evidence of hepatic and renal fibrosis
~51.
Another very important cytokine is PDGF [86]
(Fig. 2). This consists of A and B peptide chains.
Three forms exist, the homodimeric AA and BB
forms or the heterodimer. PDGF is a potent stellate
cell mitogen [87] which is produced by endothelial
cells [SS] and by activated stellate cells [89] in
culture. Quiescent stellate cells express the a-PDGF
receptor subunit, but do not express the b-subunit
until activated. This is observed in both inflammatory liver injury and after bile duct ligation [90]. The
b-receptor subunit only interacts with the B chain of
PDGF, and the PDGF dimers containing at least
one B-chain are significantly more potent to
activated stellate cells than the PDGF-AA dimers.
Expression of the PDGF b-receptor subunit is
enhanced in response to TGF-PI. Furthermore,
there is evidence that Kupffer cells may direct
stellate cell PDGF receptor expression by means of
an as yet uncharacterized soluble factor [91].
PDGF, like TGF-PI, is able to induce transcription of its own mRNA and PDGF BB has been
shown to induce transcription of both A and B
chains in culture-activated stellate cells [89].
While Kupffer cell-derived factors may. be
important to the PDGF- and TGF-P-mediated
effects on stellate cells, recent work has suggested
that substances released directly from damaged
hepatocytes are mitogenic for cultured stellate cells
'
I08
R Alcolado et al.
[92]. Moreover, IGF and IGF-binding protein
released from injured hepatocytes may act in a
paracrine manner to activate stellate cells locally.
Studies of stellate cell IGF-I and I1 receptors
indicate that there is a diminution of quantity
(IGF-Ir) and binding affinity (IGF-IIr) with transformation to activated phenotype. This suggests that
any IGF-mediated effect is only likely to be
important early on in the process of stellate cell
activation [93].
Tumour necrosis factor-a is a known mitogen [68,
691 for stellate cells and also has an effect on transcription of mRNA for chondroitin sulphate [24],
interstitial collagenase [94, 9-51 and TIMP-1 [95].
Cytokines and cell messengers released during
injury and inflammation may also moderate stellate
cell activation, proliferation and synthetic function.
The addition of interferon y to a culture medium of
myofibroblastic stellate cells results in reversion to a
more quiescent phenotype in which the cell regains
its lipid droplets. It also affects collagen production
[96] with falls in procollagen I and type IV collagen
and significant reduction in the mRNA level for
total fibronectin [96].
Addition of retinoids to stellate cells in culture
can maintain a quiescent phenotype, and when
added to activated stellate cells causes a reversion of
phenotypic changes together with a decrease in the
rate of cell proliferation [97, 981. Retinoids also
modulate collagen production by culture-activated
lipocytes - total collagen production is reduced in a
dose-dependent fashion [98]. Further important
effects of retinoids on culture-activated stellate cells
include the inhibition of both TGF-PI mRNA
production [98] and PDGF-mediated proliferation
[99]. However, a paradox exists: athough retinoids
display this potent cytoprotective role in models of
fibrosis, the ingestion of large amounts of retinoids
is known to induce liver fibrosis [loo].
It is apparent that there are many factors which
play a role in the activation of stellate cells to the
proliferating fibrogenic phenotype seen in chronic
liver disease. There is a complex interaction between
each individual cell type, the local matrix milieu and
the production and activation of many cytokines, all
of which converge on the stellate cell. This is
summarized in Fig. 2. Key to the development and
progression of fibrosis is the activation and persistence of activation of the stellate cell. The presence
and persistence of particular matrix components is
probably fundamental to the perpetuation of stellate
cell activation and the progression of fibrosis. The
role of matrix remodelling is therefore considered
below.
MATRIX REGULATION
The control of matrix turnover is central to any
understanding of the pathogenesis of liver fibrosis,
since in profibrotic liver injury either the regulatory
mechanisms are overwhelmed by collagen production or the normal response to increasing collagen
secretion is altered such that the excessive matrix is
not degraded. Accumulating evidence indicates that
progressive liver fibrosis is characterized by changes
in the pattern of matrix degradation in addition to
synthesis.
The turnover and remodelling of matrix by
mesenchymal cells is mediated by a family of zincdependent neutral proteases, the matrix metalloproteinases (MMPs) [loll. In addition, during
inflammatory liver injury, matrix degradation may be
mediated by matrix-degrading cathepsins.
In the analysis of matrix degradation during liver
fibrosis two major groups of MMPs are believed to
be important. The first, comprising the gelatinases
and stromelysins, have activity against type IV
collagen and other components of the normal
basement membrane-like matrix [loll. These
enzymes may therefore significantly alter stellate cell
activation through degradation of the normal matrix.
There is evidence that members of these groups are
expressed by activated Kupffer cells (gelatinase B)
[102], stellate cells (gelatinase A and stromelysin)
[103-1051 and endothelial cells (stromelysin) [106].
Moreover, expression of both stromelysin and gelatinase A has been described in acute liver injury in
vivo in addition to studies of non-parenchymal cells
in culture [106, 1071.
The second group comprises the collagenases
[loll. These are derived from two sources: interstitial collagenase, derived from fibroblastic cells
including activated myofibroblastic stellate cells and
neutrophil collagenase, released from neutrophil
phagosomes in areas of inflammation [loll. Because
none of the other MMPs can initiate breakdown of
interstitial collagens it is axiomatic that interstitial or
neutrophil collagenase must be expressed for matrix
remodelling to occur in established fibrosis [loll.
Studies of cultured stellate cells indicate that these
cells express interstitial collagenase in response to
specific cytokines and early in models of activation
[94, 95, 1081. One group suggested recently that
gelatinase A has degradative activity against type I
collagen [109]. Moreover, studies of mRNA expression in fibrotic human liver and in animal models of
liver fibrosis indicate that transcripts for interstitial
collagenase are present and indeed may be upregulated in certain pathologies [95, 108, 1101. These
studies have also indicated that fibrosis is accompanied by a significant upregulation in mRNA
transcripts of gelatinase A [110].
There is a relatively large literature suggesting
that MMP activity in liver decreases as fibrosis
progresses [lll-1131. This may in part result from
the expression of the powerful MMP inhibitors the
TIMPs and az-macroglobulin by stellate cells and
hepatocytes [95, 108, 110, 114-1181. Detailed studies
of culture-activated stellate cells have demonstrated
that these cells express TIMP-1 and 2, and that
these inhibitors exert a profound inhibition of
Pathogenesis of liver fibrosis
co-expressed gelatinase activity [114]. Expression of
mRNA for TIMPs in diseased human liver and
animal models of fibrosis indicates that transcripts
for both TIMP-1 and 2 are significantly upregulated
[lo81 and, in the case of TIMP-1, are probably
stellate cell derived [108].
TIMPs have been immunolocalized in diseased
liver and co-localize with cells expressing interstitial
collagenase [110]. In addition there are several
studies which indicate that serum TIMP levels are
elevated in fibrotic liver disease [119], and that
TIMP-1 may be expressed by hepatocytes as a
feature of the acute phase response [115]. crz-Macroglobulin has also been demonstrated to be expressed
and secreted by stellate cells and hepatocytes [120].
Stellate cells may also influence both the degradation of matrix and the activation of pro-MMPs by
manipulating the plasmin/plasminogen cascade.
Stellate cells express urokinase plasminogen
activator (uPA) during activation and through
expression of this protein can generate active
plasmin from plasminogen [121]. The plasmin in
turn cleaves the propiece from pro-stromelysin,
thereby also initiating the activation of interstitial
collagenase. By simultaneously expressing uPa
receptors, stellate cells are able to activate plasmin
in a focused pericellular manner [121]. As activation
of stellate cells becomes more advanced, expression
of uPA decreases and is replaced by the plasminogen activator inhibitor with a resulting decrease in
uPA activity [121].
Taken together, these data provide evidence that
the potential for matrix degradation exists, even in
comparatively advanced fibrosis. However, the
process of matrix degradation is either overwhelmed
by synthesis, and/or that matrix degradation is
inhibited as a result of TIMP and plasminogen
activator inhibitor expression.
CONCLUSION
There has been significant progress in our understanding of the cellular and molecular events which
characterize the response of the liver to injury and
the consequent development of liver fibrosis.
Central to the process is the stellate cell which
undergoes a phenotypic change to a proliferative
myofibroblast-like cell which then synthesizes excess
interstitial collagens and other matrix components.
The process of initiation of activation and its perpetuation result from a complex, often interrelated,
series of signalling mechanisms which converge on
this important effector cell. These mechanisms
include alterations in matrix resulting in changed
cell-matrix interactions and stimulation by cytokines
released from damaged hepatocytes, infiltrating
inflammatory cells, Kupffer cells and matrix. Once
the stellate cell is activated the preceding matrix
changes and any recurrent injurious stimuli will
perpetuate the activated state. The activated stellate
I09
cell is highly synthetic and secretes interstitial collagens which accumulate initially in the space of
Disse, but ultimately result in gross liver fibrosis.
Despite these effects, the liver retains a capacity/
potential for matrix degradation/remodelling.
However, this capacity may be overwhelmed and any
secreted matrix remodelling enzymes may be
inhibited by the concurrently secreted TIMPs and
crz-macroglobulin.
Future work in each of the areas highlighted in
this review will ultimately provide a detailed dissection of the molecular mechanism of stellate cell
activation and the process of matrix secretion and
assimilation. Specific targets are likely to include
mechanisms to reverse stellate cell activation or
inhibit matrix synthetic activity and intervention to
promote matrix degradation. This may be accomplished directly, i.e. by enhancing expression of
collagenase and other MMPs, or indirectly by
reducing the level of MMP inhibitory factors such as
TIMPs. Such an approach, based on control of
matrix degradation, is attractive because the
majority of patients with chronic liver disease will
present clinically when fibrosis is already established
and may be relatively advanced. Tools to intervene
directly in the process of experimental fibrosis (e.g.
transgenic and gene knockout mice) are now
becoming available, and it is anticipated that these
data will highlight suitable avenues for therapeutic
intervention.
REFERENCES
I. Hahn EG, Wick G, Pencev D, Timpl R Distribution of basement membrane
proteins in normal and fibrotic human liver: collagen type N laminin and
fibronectin. Gut 1980; 21: 63-71.
2. Martinez-Hernandezk The hepatic extracellular matrix. I. Electron
immunohistochemical studies in normal rat liver. Lab Invest 1984; 51: 57-69.
3. Abrahamson DR, CaulfieldJP. Distribution of laminin within rat and mouse
renal, splenic, intestinal, and hepatic basement membranes identified after the
intravenous injection of heterologousantilaminin IgG. Lab Invest 1985; 5 2 169.
4. Arenson DM, Friedman SL, Bissell DM. Formation of extracellular matrix in
normal rat liver: lipocytes as a major source of proteoglycan. Gastroenterology
1988; 95: 441-7.
5. Maher Jj,Friedman SL, Roll Fj, 8issell DM. lmmunolocalizationof laminin in
normal rat liver and biosynthesis of laminin by hepatic lipocytes in primary
culture. Gastroenterology 1988; 9 4 1053-62.
6. Schuppan D. Structure of extracellular matrix in normal and fibrotic liver.
collagens and glycoproteins. Semin Liver Dis 1990; 10: 1-10.
7. Geem Schuppan D, h e r o m s S,De-Zanger R, Wisse E. Collagen type I and
111 occur together in hybrid fibrils in the space of Disse of normal rat liver.
Hepatology 1990; I2 233-4 I.
8. Burt AD, Griffiths MR, Schuppan D, Voss B, MacSween RN. Ultrastructural
localizationof extracellular matrix proteins in liver biopsies using
ultracryomicrotomyand immuno-gold labelling. Histopathology 1990; I 6 53-8.
9. Yurchenco PD. Assembly of basement membranes. Ann N Y Acad Sci 1990;
5 8 0 195-214.
10. Siebold B, Deutzman R, Kuhn K. The arrangement of intra and intermolecular
disulfide bonds in the carboxy terminal non collagenous aggregation and cross
linking domain of basement membrane type IV collagen. Eur J Biochem 1988;
I76 6 I7-24.
I I. Yurchenco PD, Furthmayr H. Self-assembly of basement membrane collagen.
Biochemistry 1984; 23 1839-50.
12. Griffiths MR Ferrier R, Schuppan D, Burt AD. Collagen type VI: a component
of the perisinusoidal extracellular matrix in human liver. In: Wisse E, Knook DL.
I10
R Alcolado et al.
McCuskey RS, eds. Cells of the hepatic sinusoid. Leiden, The Netherlands:
Kupffer Cell Foundation, 1991: vol. 3, 155-60.
13. Geerts A, Geuze HJ, Slot J-W, et al. lmmunogold localizationof procollagen 111,
fibronectin and heparan sulfate proteoglycanon ultrathin frozen sections of the
normal rat liver. Histochemistry 1986; 8 4 355-62.
14. Schuppan D, Cantaluppi MC, Becker J, et al. Undulin, an extracellular matrix
glycoprotein associated with collagen fibrils. J Biol Chem 1990; 265: 8823-32.
IS. Reid LM, Fiorino AS, Sigal SH, Brill S, Holst PA Extracellular matrix gradients in
the space of Disse: relevance to liver biology [Editorial]. Hepatology 1992; 1 5
1198-203.
16. McGuire RF. Bissell DM, Boyles J, Roll FJ. Role of extracellular matrix in
regulating fenestrations of endothelial cells isolated from normal rat liver.
Hepatology 1992; 15: 989-97.
17. Rojkind M, Giambrone M-A, Biempica L. Collagen types in normal and cirrhotic
liver. Gastroenterology 1979; 7 6 7 10- 19.
18. Seyer JM, Huhenon ET, Kang AH. Collagen polymorphism in normal and
cirrhotic human liver. J Clin Invest 1977; 5 9 241-8.
19. Gressner AM, Haarmann R Hyaluronic acid synthesis and secretion by rat liver
fat storing cells (perisinusoidal lipocytes) in culture. Biochem Biophys Res
Commun 1988; 151: 222-9.
20. Clement B, Laurent M, Guguen-Guillouzo C, Lebeau G, Guillouzo A Types I
and IV procollagen gene expression in cultured rat hepatocytes. Coll Relat Res
1988; 8: 349-59.
21. Maher JJ, McGuire RF. Extracellular matrix gene expression increases
preferentially in rat lipocytes and sinusoidal endothelial cells during hepatic
fibrosis in vivo. J Clin Invest 1990; 8 6 I641-8.
22. Ramadori G, Rieder H, Knittel T, Dienes HP, Meyer zum Buschenfelde K-H. Fat
storing cells (FSC) of rat liver synthesize and secrete fibronectin. J Hepatol
1987; 4 190-7.
23. Friedman SL, Roll FJ, Boyles J, Bissell DM. Hepatic lipocytes: the principal
collagen-producing cells of normal rat liver. Proc Natl Acad Sci USA 1985; 8 2
868 1-5.
24. Weiner FR, Giambrone MA, Czaja MJ, et al. Ito-cell gene expression and
collagen regulation. Hepatology 1990; II: I I 1-17.
25. Friedman SL, Rockey DC, McGuire RF, Maher JJ, Boyles JK, Yamasaki G.
Isolated hepatic lipocytes and Kupffer cells from normal human liver:
morphologicaland functional characteristics in primary culture. Hepatology
1992; IS: 234-43.
26. Kawase T, Shiratori Y, Sugimoto T. Collagen productionby rat liver fat storing
cells in I degree culture. Exp Cell Biol 1986; 5 4 183-92.
27. Schafer S, Zerbe 0, Gressner AM. The synthesis of proteoglycans in fat-storing
cells of rat liver. Hepatology (Baltimore) 1987; 7: 680-7.
28. ltving MG, Roll FJ, Huang S, Bissell DM. Characterisationand culture of
sinusoidal endothelium from normal rat liver, lipoprotein uptake and collagen
phenotype. Gastroenterology 1984; 87: 1233-47.
29. Jarnagin WR, Rockey DC, Koteliansky VE, Wang SS, Bissell DM. Expressionof
variant fibronectins in wound healing: cellular sources and biological activity of
the ElllA segment in rat hepatic fibrogenesis. J Cell Biol 1994; 127: 2037-48.
30. Maher JJ,Bissell DM, Friedman SL, Roll FJ. Collagen measured in primary
cultures of normal rat hepatocytes derives from lipocytes within the monolayer.
J Clin Invest 1988; 8 2 450-9.
3 I. Bissell DM, Arenson DM, Maher JJ, Roll FJ. Support of cultured hepatocytes by
a laminin-rich gel. J Clin Invest 1987; 79: 801- 12.
32. Bissell DM, Caron JM, Babiss LE, Friedman JM. Transcriptional regulation of the
albumin gene in cultured rat hepatocytes. Role of basement-membrane matrix.
Mol Biol Med 1990; 7: 187-97.
33. lredale JP, Arthur MJP. Hepatocyte-matrix interactions. Gut 1994; 35: 729-32.
34. Friedman SL, Roll FJ, Boyles J, Arenson DM, Biaell DM. Maintenanceof
differentiated phenotype of cultured rat hepatic lipocytes by basement
membrane matrix. J Biol Chem 1989; 264 10756-62.
35. Friedman SL. The cellular basis of hepatic fibrosis: mechanisms and treatment
strategies. N Engl J Med 1993: 328 1828-35.
36. Horn T, lunge J, Christoffenen P. Early alcoholic liver injury. Activation of
lipocytes in acinar zone 3 and correlation to degree of collagen formation in
Disse space. J Hepatol 1986; 3:333-40.
37. Mak KM, Lieber CS. Lipocytes and transitional cells in alcoholic liver disease: a
morphometric study. Hepatology (Baltimore) 1988; 8 1027-33.
38. McGee 10, Patrick Rs. The role of perisinusoidal cells in hepatic fibrogenesis.
An electron microscopic study of acute carbon tetrachloride liver injury. Lab
Invest 1972; 2 6 429-40.
39. Rockey DC, Boyles JK, Gabbiani G, Friedman SL. Rat hepatic lipocytes express
smooth muscle actin upon activation in vivo and in culture. J Submicrosc Cytol
Pathol 1992; 2 4 193-203.
40. Mak KM, Leo AM, Lieber CS. Alcoholic liver injury in baboons: transformation
of lipocytes to transitional cells. Gastroenterology 1984; 87: 188-200.
41. Leo MA, Mak KM, Savolainen ER, Lieber CS. Isolation and culture of
myofibroblastsfrom rat liver (42193). Proc Soc Exp Biol Med 1985; 180
382-9 I.
42. Shiratori Y, lchida T. Geerts A, Wisse E. Modulation of collagen synthesis by
fat-storing cells, isolated from CC14- or vitamin A-treated rats. Dig Dis Sci 1987;
3 2 1281-9.
43. Geerts A, Vrijsen R, RauterbergJ, Burt A, Schellinck P, Wise E. In vitro
differentiation of fat-storing cells parallels marked increase of collagen synthesis
and secretion. J Hepatol 1989; 9 59-68.
44. Ruoslahti E. Integrins. J Clin Invest 1991; 87: 1-5.
45. Wu C, Keivens VM, 07ooIe TE, McDonald JA, Ginsberg MH. lntegrin
activation and cytoskeletal interaction are essential for the assembly of a
fibronectin matrix. Cell 1995; 83: 715-24.
46. Dedhar S, Ruoslahti E, Pierschbacher MD. A cell surface receptor complex for
collagen type I recognizes the Arg-Gly-kp sequence. J Cell Biol 1987; 104
585-93.
47. StamatoglouSC, Bawumia S, JohanssonS,Forsberg E, Hughes RC. Affinity of
integrin alpha-I beta-I from liver sinusoidal membranes for type IV collagen.
FEBS Lett 1991; 288: 241-3.
48. Ruoslahti E, Pierschbacher MD. New perspectives in cell adhesion: RGD and
integrins. Science (Washington DC) 1987; 238: 491-7.
49. Argraves WS, Suzuki S, Arai H, Thompson K, Pienchbacher MD, Ruoslahti E.
Amino acid sequence of the human fibronectin receptor. J Cell Biol 1987; 105
I 183-90.
50. Wayner EA, Garcia-Pardo A, Humphreys MJ, McDonald ]A, Carter WG.
Identification of the T-lymphocyte adhesion receptor for an alternative cell
attachment domain (CS-I) in plasma fibronectin. J Cell Biol 1989; 109:
1321-30.
5 I. Panayotou G, End P, Aumailley M, et al. Domains of laminin with growth factor
activity. Cell 1989; 48: 989-96.
52. Klagsbrunn M, Baird A. A dual receptor system is required for basic fibroblast
growth factor activity. Cell I99 I; 67: 229-3 I.
53. Andres JL, Stanley K, Cheifea S, Massague J. Membrane-anchored and soluble
forms of betaglycan, a polymorphic proteoglycan that binds transforming growth
factor-beta. J Cell Biol 1989; 109 3 137-45.
54. Roberts AB, Anzano MA, Lamb LC, Smith JM, Sporn MB. New class of
transforminggrowth factor potentiated by epidermal growth factor. Proc Natl
Acad Sci USA 1981; 78: 5339-41.
55. Meyer DH, Bachem MG, Gressner AM. Modulation of hepatic lipocyte
proteoglycan synthesis and proliferation by Kupffer cell-derived transforming
growth factors type beta I and type alpha. Biochem Biophys Res Commun
1990 171: 1122-9.
56. Win KM, Charlotte F. Mallat A et al. Mitogenic effect of transforming growth
factor-beta1 on human It0 cells in culture: evidence for mediation by
endogenous platelet-derived growth factor. Hepatology 1993; 18: 137-4s.
57. Jakowlew SB, Mead JE, Damelpour D, Wu J, Roberts AB, Fausto N.
Transforminggrowth factor beta (TGF-beta) isoforms in rat liver regeneration:
messenger RNA expression and activation of latent TGF-beta. Cell Regulation
1991; 2 535-48.
58. Nakatsukasa H, Nagy P, E m RP, Hsia CC, Marxlen E, Thorgeinson SS.
Cellular distribution of transforming growth factor-beta I and procollagen types
I, 111, and IV transcripts in carbon tetrachloride-induced rat liver fibrosis. J Clin
Invest 1990; 85: 1833-43.
59. Manthey CL, Allan JB, Ellingsworth LR, Wahl SM. In situ expression of
transforminggrowth factor-beta in streptococcal cell wall-induced
granulomatous inflammation and hepatic fibrosis. Growth Factors 1990; 4
17-26.
60. Czaja MJ, Weiner FR, Flanders KC, et al. In vitro and in vivo association of
transforminggrowth factor beta-I with hepatic fibrosis. J Cell Biol 1989; 108:
2477-82.
6 I. Tsukamoto H, Horne W, Kamimura S, et al. Experimental liver cirrhosis
induced by alcohol and iron. J Clin Invest 1995; 9 6 620-30.
62. Bissell DM. Wang SS, JarnaginWR, Roll FJ. Cell-specificexpression of
transforminggrowth factor-beta in rat liver. Evidence for autocrine regulation of
hepatocyte proliferation. J Clin Invest 1995; 9 6 447-55.
63. Castilla A, PrietoJ, Fausto N.Transforming growth factors beta-I and alpha in
chronic liver disease: Effects of interferon therapy. N Engl J Med I99 I; 324
933-940.
64. Milani S, Schuppan D, Herbst H, Surrenti C. Expression of transforminggrowth
factor beta, in normal and fibrotic human liver. In: Gressner A, Ramadori G,
eds. Molecular and cell biology of liver fibrogenesis. Dordrecht, The
Netherlands: Kluwer Academic Publishers, 1992 254-62.
Pathogenesis of liver fibrosis
65. Bachem MG, Meyer D, Melchior R, Sell K-M, Gressner AM. Activation of rat
liver perisinusoidal lipocytes by transforming growth factors derived from
myofibroblast-like cells. J Clin Invest 1992; 8 9 19-27.
66. Masague J.The transforming growth factor beta family. Annu Rev Cell Biol
1990; 6 597-641.
67. Roberts AB, Spom MB. The transforming growth factors-beta. In: Sporn MB,
Roberts AB. ed. Handbook of experimental pharmacology. Peptide growth
factors and their receptors. Heidelberg, Germany: Springer-Verlag, 1990: vol.
95/L, 419-72.
68. Matsuoka M, Pham NT, Tsukamoto H. Differentialeffects of interleukin I alpha,
tumour necrosis factor alpha, and transforming growth factor beta-I on cell
proliferationand collagen formation by cultured fat storing cells. Liver 1989; 9
71-8.
69. Bachem MH, Sell KM, Melchior R, Kropf J. Eller T. Gressner AM. TNF-alpha
and TGF beta-I stimulate fibronectin synthesis and the transdifferentiation of fat
storing cells in rat liver to myofibroblasts. Virchows Archiv B Cell Pathol 1993;
6 3 123-30.
70. Casini A, Pinzani M, Milani S, et al. Regulation of extracellular matrix synthesis
by transforminggrowth factor beta I in human fat-storing cells.
Gastroenterology 1993; 105: 245-53.
71. Bachem MG, Riess U, Melchior R, Sell KM, Gressner AM. Transforming growth
factors (TGF alpha and TGF beta I)stimulate chondroitin sulfate and hyluronate
synthesis in cultured rat liver fat storing cells. FEBS Lett 1989; 257: 134-7.
72. Edwards DR, Murphy G, Reynolds JJ,et al. Transforming growth factor beta
modulates the expression of collagenase and metalloproteinaseinhibitor. EMBO
J 1987: 6 1899-904.
73. Overall CM, Wrana JL, Sudek J. Independent regulation of collagenase, 72 kD
progelatinase, and metalloendoproteinase inhibitor expression in human
fibroblasts by transforming growth factor-beta. J Biol Chem 1989; 264 1860-9.
74. Kojima S, W i n DB. Mechanism of retinoid-induced activation of latent
transforminggrowth factor beta in bovine endothelial cells. J Cell Physiol 1993;
155: 323-32.
75. Blazejewski S, Preaux AM, Mallat A, et al. Human myofibroblast like cells
obtained by outgrowth are representativeof the fibrogenic cells in the liver.
Hepatology 1995; 22: 788-97.
76. Roberts CJ. Birkenmeir TM, McQuillan JJ* et al. Transforming growth factorbeta stimulates the expression of fibronectin and of both subunits of the human
fibronectin receptor by cultured human fibroblasts. J Biol Chem 1988; 263:
4586-92.
77. De Bleser PJ, Jannes P, van Bud-Offers SC, et al. Insulin-likegrowth factorIl/mannose 6-phosphate receptor is expressed on CCI4-exposed rat fat-storing
cells and facilitates activation of latent transforminggrowth factor-beta in
cocultures with sinusoidal endothelial cells. Hepatology 1995; 21: 1429-37.
78. Friedman SL, Yamasaki G. Wong L. Modulation of transforming growth factor
beta receptors of rat lipocytes during the hepatic wound healing response. J Biol
Chem 1994: 269 I055 1-8.
79. Cheifetz S, Andres JL, Massague J. The transforming growth factor-beta
receptor type 111 is a membrane proteoglycan. Domain structure of the
receptor. J Biol Chem 1988; 263: 16984-91.
80. Lopez-Casillas F. Cheifetz 5, DoodyJ.Andres JL, Lane WS, Massague J.
Structure and expression of the membrane proteoglycanbetaglycan, a
component of the TGF-beta receptor system. Cell 1991; 67: 785-95.
81. Border WA, Noble NA, Yamamoto T, et al. Natural inhibitor of transforming
growth factor-beta protects against scarring in experimental kidney disease.
Nature (London) 1992; 3 6 0 361-4.
82. Shah M, Foreman DM, Ferguson MWJ. Control of scarring in adult wounds by
neutralising antibody to transforminggrowth factor beta. Lancet 1992; 339
213- 14.
83. Border WA, Okuda 5, Languino LR, Sporn MB. Ruoslahti E. Suppression of
experimentalglomerulonephritis by antiserum against transforming growth
factor beta I. Nature (London) 1990; 346 371-4.
84. lsaka Y, Brees DK. lkegaya K, et al. Gene therapy by skeletal muscle expression
of decorin prevents fibrotic disease in rat kidney. Nature Med 1996; 2: 418-23.
85. Sanderson N, Factor V, Nagy P, et al. Hepatic expression of mature
transforminggrowth factor beta I in transgenic mice results in multiple tissue
lesions. Proc Natl Acad Sci USA 1995; 9 2 2572-6.
86. Ross R, Raines EW, Bower-Pope DF. The biology of platelet derived growth
factor. Cell 1986; 4 6 155-69.
87. Pinzani M, Gesukado L, Sabbah GM, Abboud HE. Effects of platelet derived
growth factor and other polypeptide mitogens on DNA synthesis and growth of
cultured rat liver fat storing cells. J Clin Invest 1989; 84: 1786-93.
88. Rieder H, Ramadori G, Meyer zum Buschenfelde K-H. Sinusoidal cells in vitro
secrete PDGF like factors stimulating the proliferationof cultured liver
myofibroblasticcells. In: Gressner AM, Ramadori G, eds. Molecular and Cell
Ill
biology of liver fibrogenesis. Dordrecht, The Netherlands: Kluwer Academic
Publishers, I992 172-6.
89. Marra F. Choudhury GG, Pinzani M, Abboud HE. Regulation of platelet-derived
growth factor secretion and gene expression in human liver fat-storing cells.
Gastroenterology 1994; 107: I I 10- 17.
90. Wong L, Yamasaki G, Johnson RJ, Friedman SL. Induction of beta-plateletderived growth factor receptor in rat hepatic lipocytes during cellular activation
in vivo and in culture. J Clin Invest 1994; 9 4 1563-9.
91. Friedman SL. Arthur MJP. Activation of cultured rat hepatic lipocytes by Kupffer
cell conditioned medium. Direct enhancement of matrix synthesis and
stimulation of cell proliferationvia induction of platelet derived growth factor
receptors. J Clin Invest 1989; 84: 1780-5.
92. Gressner AM, Hoffmann C, Lahme B, Scheckel L. Hepatocyte injury stimulates
the mitogenic activation of cultured fat storing cells. In: Wise E, Wake K,
Knook DL, eds. Cells of the hepatic sinusoid. Leiden, The Netherlands: The
Kupffer Cell Foundation, 1995 vol. 5, 380-2.
93. Brenzel A, Weiner OH, Gressner AM. Stage-dependentexpression of insulinlike growth factor (IGF)-I- and IGF-ll-bindingsites in rat liver fat storing cells
(FSC) during in vitro transformation to myofibroblasts(MFB). In: Wise E, Wake
K, Knook DL, eds. Cells of the hepatic sinusoid. Leiden, The Netherlands: The
Kupffer Cell Foundation, 1995 vol. 5, 386-90.
94. Emonard H, Guillouzo A, Lapiere CM, GrimaudJA Human liver fibroblast
capacity for synthesizing interstitialcollagenase in vitro. Cell Mol Biol 1990; 3 6
46 I-7.
95. lredale JP, Goddard S, Murphy G, Benyon RC, Arthur MJP. Tissue inhibitor of
metalloproteinase-I and interstitialcollagenase expression in autoimmune
chronic active hepatitis and activated human hepatic lipocytes. Clin Sci 1995:
89: 75-8 I.
96. Rockey DC, Maher JJ, JarnaginWR, Gabbiani G, Friedman SL. Inhibition of rat
hepatic lipocyte activation in culture by interferon-gamma. Hepatology 1992;
1 6 776-84.
97. Geerts A, Vrijsen R, Schelliack P, Wisse E. Retinol affects the phenotype and
protein synthesis of fat storing cell derived myofibroblastsin vitro. In: Wisse E,
Knook DL, Decker K, eds. Cells of the hepatic sinusoid. Leiden, The
Netherlands: The Kupffer Cell Foundation, I989 vol. 2, 20-4.
98. Davis BH, Kramer RT, Davidson NO. Retinoic acid modulates rat Ito cell
proliferation. collagen, and transforming growth factor beta production. J Clin
Invest 1990; 8 6 2062-70.
99. Davis BH, Rapp UR, Davidson NO. Retinoic acid and transforminggrowth
factor beta differentially inhibit platelet-derived-growth-factor-inducedito-cell
activation. Biochem J 1991; 278 43-7.
100. Hendriks HFJ, Bosma A, Brouwer A. Fat-storingcells: Hyper- and
hypovitaminosis A and the relationships with liver fibrosis. Semin Liver Dis 1993;
13: 72-80.
101. Murphy G, Docherty AJP, Hembry RM, Reynolds JJ. Metalloproteinases and
tissue damage. Br J Rheumatol 1991; 30 (Suppl. I):25-3 I.
102. Winwood PJ, Schuppan D, lredale JP, Kawser CA, Docherty AJP. Arthur MJP.
Kupffer cell-derived 95 kDa type IV collagenase/gelatinase B characterisation
and expression in cultured cells. Hepatology 1995: 22: 304- 15.
103. Vyas SK, Leyland H, Gentry J, Arthur MJP. Transin (stromelysin) is expressed in
early rat lipocyte primary culture. Gastroenterology 1995; 109 889-98.
104. Arthur MJP, Friedman SL, Roll FJ, Bissell DM. Lipocytes from normal rat liver
release a neutral metalloproteinasethat degrades basement membrane (type IVj
collagen. J Clin Invest 1989; 8 4 1076-85.
105. Arthur MJP, Stanley A, lredale JP, Rafferty JA,Hembry RM, Friedman SL.
Secretion of 72 kDa type IV collagenase/gelatinase by cultured human lipocytes:
analysis of gene expression, protein synthesis and proteinase activity. Biochem
J 1992; 287: 701-7.
106. Herbst H, Heinrichs 0, Schuppan D, Milani S, Stein H. Temporal and spatial
patterns of transin/stromelysin RNA expression following toxic injury in rat liver.
Virchows Archiv B Cell Pathol 1991; 60: 295-300.
107. lredaleJP, Winwood PJ, Kawser CA, Arthur MJP. Lipocyte proliferationin the
C. porvum model of macrophage-induced liver injury is associated with
expression of 72 kDa type IV collagenase/gelatinase. In: Knook DL, Wisse E,
eds. Cells of the hepatic sinusoid. Leiden, The Netherlands: Kupffer Cell
Foundation, 1993: vol. 4, 105-8.
108. lredale JP, Ferris WF, Goddard S, et al. TIMP-I expression is enhanced relative
to interstitialcollagenase messenger RNA in experimental liver injury and
fibrosis. Hepatology 1996; 2 4 176-84.
109. Aimes RT. Quigley JP. Matrix metalloproteinase-2is an interstitialcollagenase Inhibitor-free enzyme catalyzes the cleavage of collagen fibrils and soluble native
I12
R Alcolado et al.
type I collagen generating the specific 3/4-and I/4-length fragments. J Biol Chem
1995; 2 7 0 5872-6.
I 10. Benyon RC, lredale JP, Goddard S,Winwood PJ, Arthur MJP. Increased
expression of tissue inhibitor of metalloproteinases-l and -2 relative to
interstitialcollagenase in fibrotic human liver. Gastroenterology I996 II0
821-3 I.
I I I. Okazaki I, Maruyama K. Collagenase activity in experimental hepatic fibrosis.
Nature (London) 1974; 252: 49-SO.
I 12. Perez-Tamayo R, Montfort I, Gonzalez E. Collagenolytic activity in experimental
cirrhosis of the liver. Exp Mol Pathol 1987; 47: 300-8.
I 13. Maruyama K, Feinman L, Fainsilber Z, Nakano M, Okazaki I, Lieber CS.
Mammalian collagenase increases in early alcoholic liver disease and decreases
with cirrhosis. Life Sci 1982; 3 0 1379-84.
I 14. lredale JP, Murphy G, Hembry RM, Friedman SL, Arthur MJP. Human hepatic
lipocytes synthesize tissue inhibitor of metalloproteinases- I (TIMP- I):
implicationsfor regulation of matrix degradation in liver. J Clin Invest 1992; 9 0
282-7.
I 15. Kordula T, Guttgemann I, Rosejohn S, et al. Synthesis of tissue inhibitor of
metalloproteinase-I (TIMP-I) in human hepatoma cells (HepG2) -
up-regulationby interleukin-6 and transforminggrowth factor-betal. FEBS Lett
1992; 313: 143-7.
I 16. Roeb E, Graeve L, Hoffmann R, Decker K, Edwards DR Heinrich PC.
Regulation of tissue inhibitor of metalloproteinases- I gene expression by
cytokines and dexamethasone in rat hepatocyte primary cultures. Hepatology
1993; 18: 1437-42.
I 17. Roeb E, Rosejohn S, Erren A, et al. Tissue inhibitor of metalloproteinases-2
(TIMP-2) in rat liver cells is increased by lipopolysaccharide and prostaglandin
E(2). FEBS Lett 1995; 357: 33-6.
I 18. Truden JL, Boros DL. Detection of alpha-2-macroglobulin. alpha-I-protease
inhibitor, and neutral protease-antiprotease complexes within liver granulomas
of schistosoma mansoni-infected mice. Am J Pathol 1988 130 281-8.
119. Murawaki Y, Yamarnoto H, Kawasaki H, Shima H. Serum tissue inhibitor of
metalloproteinases in patients with chronic liver disease and with hepatocellular
carcinoma. Clin Chim Acta 1993; 218: 47-58.
120. Andus T, Gross V. Tran-Thi T-A, Schreiber G, Nagashima M, Heinrich PC. The
biosynthesisof acute-phase proteins in primary cultures of rat hepatocytes. Eur
J Biochem 1983; 133: 561-71.
I 2 I. Leyland H, Gentry J. Arthur MJP, Benyon RC. The plasminogen activating
system in hepatic stellate cells. Hepatology I996 2 4 I 172-8.