Methods

The PubMed database was searched using the following keywords: fibrosis, collagen, cytokine, growth factor, laminin, liver, matrix metalloproteinase (MMP), proteoglycan, posttranslational modification, ECM, and neoepitope. Each author further selected key publications according to his/her specific expertise.

Clinical Significance of the ECM in Liver Fibrosis

The common denominator of most chronic liver diseases is altered remodeling of the ECM initiated by inflammation, with both quantitative and qualitative alterations of its composition, thereby gradually disrupting normal function and finally leading to increased morbidity and mortality attributable to organ failure. Figure 2 highlights the main cellular and structural components of and differences between healthy and fibrotic tissue. The ECM is now considered a biologically active system that tends to perpetuate inflammation and fibrosis, rather than a passive consequence of chronic liver injury. This, together with accumulating evidence that fibrosis and even cirrhosis are reversible conditions, has changed the field and paved the path for novel optimism and interest in antifibrotic therapies (287). The novel insights into the pathophysiology and function of the ECM in chronic fibroproliferative diseases in general, and in liver fibrosis in particular, are already beginning to be translated into the clinic (26, 101, 287, 334). Early diagnosis for high-risk populations combined with targeted, individualized, and more rational interventions, including specific antifibrotic drugs, to prevent progressive disease or induce its regression and thus the morbidity, mortality, and excessive cost associated with symptomatic treatment of advanced stage fibrosis is urgently needed. Thus a central question to be addressed in this review is whether and how far our fairly advanced knowledge on the pathophysiology and molecular biology of the ECM can be translated into clinical practice.

Open in a separate window

Fig. 2.

Illustration depicting the changes that occur in the extracellular matrix (ECM) remodeling during the development of fibrosis. The ECM may be divided into the loose basement membrane and the compact interstitial matrix. The basement membrane consists mainly of type IV collagen, laminins, entactin, and proteoglycans and functions to anchor cells and connect to the interstitial ECM. The interstitial matrix has a different composition and consists mainly of fibrillar collagens, numerous noncollagenous glycoproteins, proteoglycans, and elastin. In healthy tissue, ECM remodeling is tightly controlled to ensure homeostasis; the synthesis of ECM proteins by fibroblasts has a relatively slow metabolic turnover, and the proteolytic activity is limited. During fibrogenesis, however, this process is disturbed. As a result of chronic wound healing, immune cells infiltrate the interstitial matrix, which helps drive the profibrotic response. Fibroblasts and fibroblast precursors like hepatic stellate cells differentiate into myofibroblasts, which deposit excessive levels of interstitial collagens. Initially, they may release enhanced levels of ECM-degrading matrix metalloproteinases (MMPs), whereas, at later stages, most MMPs are downregulated. This in turn results in a change in the composition of the ECM, where high levels of ECM degradation products and increased collagen cross-linking can be observed.

Fibrotic liver diseases are usually silent with limited clinical symptoms until development of cirrhosis with portal hypertension and complications like ascites, bleeding varices, or hepatocellular carcinoma (HCC). Furthermore, the heterogeneity within different subtypes of liver diseases is a separate challenge. Fewer than 20% of chronic alcohol abusers develop fibrosis, and of these only few will progress to cirrhosis (293, 319). Similarly, ~50% of patients with hepatitis C virus infection will never develop significant fibrosis (169), and only 10–20% of patients with nonalcoholic fatty liver disease will develop nonalcoholic steatohepatitis (NASH); of these, 10–20% will progress toward cirrhosis within 5–15 yr (290). Although the individual predictors for fast fibrosis progression are diverse, including genetic predisposition, increased age at onset of disease, and especially (hepatic) comorbidities (second hits), there is still a lack of sensitive noninvasive biomarkers to predict the individual risk of fibrosis progression (288). However, the population at risk for chronic liver diseases is enormous and poses a global health challenge. Thus the prevalence of obesity alone is 10–30% in most Western and developing countries, as is the consumption of harmful quantities of alcohol (1, 61, 232, 393). On the other hand, with the advent of potent antivirals, the global epidemic of chronic viral hepatitis will likely be harnessed in the near future causally (180, 233). However, even in chronic viral hepatitis, antifibrotics will be useful in those patients with advanced disease in whom fast regression is desirable to further decrease the risk of hepatic decompensation and HCC.

Despite the emergence of more effective antifibrotic agents, there is a lack of agents that specifically derive from the ECM (287), the fibrotic structure itself. The dynamics of the ECM likely harbor important diagnostic information but, more importantly, also clues and pathway for a targeted intervention. Thus biomarkers of ECM remodeling are the most likely candidates to predict fibrosis progression or regression or to permit noninvasive monitoring of antifibrotic interventions (162). During tissue remodeling, proteases release small protein fragments into the circulation (52, 162, 259, 288) (Fig. 2) that may serve as biomarkers of the fibrotic process. Furthermore, the dynamic nature of ECM measures vs. current static measures obtained by imaging or histology should synergistically improve diagnostic and prognostic information on fibrotic liver diseases (52, 162, 259, 288).

Fibrosis and the ECM

The matrix composition itself is particularly important for the regulation of fibrosis. Because the ECM composition regulates the behavior and phenotype of cells housed in the ECM, any ECM remodeling will in turn influence adjacent cells and modify their behavior/phenotype. Consequently, abnormal ECM remodeling may be a prerequisite for fibrogenesis and/or fibrolysis.

Common to all fibrotic disorders is the characteristic alterations in ECM remodeling, which results in the excessive accumulation of ECM components in a given organ that can ultimately lead to organ malfunction (117, 151, 352). Fibrosis is also a dynamic condition with accelerated ECM turnover in which both tissue formation and tissue degradation are highly upregulated (15, 16, 297, 350), resulting in pathological collagen deposition and altered MMP expression profiles (353). Moreover, modifications to amino acids or proteolytic cleavage at specific locations by specific PTMs result in both immunologically and functionally different proteins (161).

Fibrogenesis and fibrolysis are driven by many cell types and molecular events. Fibrogenesis is intimately linked to wound healing, serving to prevent tissues from disassembly during inflammation, apoptosis, necrosis, and release of lytic enzymes. In the liver, the major downstream effectors of fibrosis are activated stellate cells and (portal) fibroblasts with a phenotype of wound myofibroblasts (175), which produce excessive amounts of ECM molecules, such as the prominent collagens, mainly the abundant fibrillar type I and III collagens (131), which are a hallmark of all fibrotic diseases increasing up to tenfold in advanced fibrosis of organs, such as lungs and liver (205, 370). Type IV collagen is also subject to extensive remodeling during fibrosis, and its quantity has been found to increase more than 10-fold during liver fibrogenesis (116, 289). Variants of type IV collagen together with isoforms of laminin, nidogen, and perlecan are the main components of all basement membranes and form a sheet-like scaffold at the basal site of epithelia and endothelia and around interstitial cells (135, 289) that maintain viability and differentiation of these cells (157, 231). Other functions of this network are the provision of interaction sites for other ECM components, inflammatory cells, chemokines, cytokines, and important functions in cellular signaling (394). The collagen formation observed during fibrosis is also accompanied by an increase in type IV collagen degradation and unfavorable remodeling of the basement membranes, resulting from an increased expression of proteases in the affected tissue (129, 194, 351). This favors myofibroblast activation and a net increase in the deposition of, for example, fibrillar collagens by myofibroblasts (379). This is in part accompanied by a deregulated MMP activity toward basement membranes and increased expression of tissue inhibitors of matrix metalloproteinases (TIMPs) by several cell types in the fibrotic liver, especially TIMP-1, which blocks local interstitial collagen degradation by, for example, MMP-1, -8, and -13 and further promotes myofibroblast activation directly (31, 129, 309). Importantly, collagen properties and quality are altered in fibrosis by protein modifications, including increased cross-linking, that result in enhanced tissue stiffness, which contributes to fibroblast activation and compromised hepatocyte function. Although there presently is no consensus on the most important proteins of the ECM that should be specifically addressed in fibrosis, type I and III collagens are the most abundant ones, followed by type IV, V, and VI collagens, nidogen, laminin, fibronectin, biglycan, mimican, versican, decorin, lumican, and elastin, to name a few (117–119, 195, 287, 289), all clearly being altered in quantity and quality in fibrosis. Although myofibroblasts may be the central ECM-producing cells, macrophages are emerging as important upstream regulators of hepatic fibrogenesis and direct effectors of fibrosis resolution, controlled by specific subtypes, such as M1 and variant M2 macrophage subsets (20, 37, 84, 91, 223, 265, 266, 284, 332, 349, 379–381). Specific functional macrophage subtypes that are highly dependent on the ECM and cellular environment and soluble mediators in the host tissue can also phagocytose cellular debris, which removes potential proinflammatory signals and can cause increased MMP expression and enhanced matrix degradation (260). In a simplified view, M1 macrophages are rather proinflammatory and antifibrotic, in contrast to (subtypes of) M2 macrophages, which can assume profibrotic properties.

The Importance of the ECM for Tissue Function: The ECM Is Controlling Cell Phenotype

The first evidence of a central functional role of the ECM in controlling cellular behavior was obtained when a malignant cellular genotype and phenotype could be repressed by a normal mouse embryonic ECM (81, 214). Generally the ECM is a 3-D scaffold that supports or encapsulates sessile or migrating cells and defines their microenvironment (11). It consists of a meshwork of proteins and nonprotein components (glycosaminoglycans) to which soluble factors, such as growth factors and cytokines, can bind, which regulates accessibility of common nutrients. The importance and the role of ECM in cell phenotype, tissue-specific differentiation is exemplified by the fact that cells grown as 2-D monolayers on top of either a plastic substrate or a glass coverslip, with or without ECM ligand, fail to assemble into the same tissue-like structures as those growing in or on top of the normal 3-D ECM. Cells grown on plastic or glass are also unable to express differentiated proteins upon stimulation (247) or to respond to growth factors or protease inhibitors in the same way as cells growing in a 3-D setting (163). These phenotypic disparities can be explained, at least in part, by ECM-derived signals that living tissues in 3-D emit and that are transmitted into the cell via ECM receptors such as integrins and via receptors for ECM-bound proteins such as growth factors. This temporospatial 3-D signaling is absent or altered in 2-D substrata. This has been convincingly shown for both epithelial and interstitial cells (122). The architecture of the interstitial matrix in vivo also differs substantially from that offered to or found typically in cells cultured on plastic (163). As an example, osteoblasts grown on plastic in 2-D do not rely on MMPs for survival, whereas osteoblasts embedded in an interstitial 3-D matrix containing type I collagen are critically dependent on MMP activation of latent transforming growth factor (TGF)-β for their survival (163). Moreover, the orientation and function of cells and collagen fibers are lost when cells are grown in 2-D compared with 3-D, which critically regulate cell and tissue behavior (238, 251, 252). Taken together, in vitro models need to replicate the naturally occurring 3-D environment, encompassing a sufficient physiological and pathophysiological variety of ECM components. Thus the effect of 2-D vs. 3-D ECM environments on central biological features of fibroblasts and myofibroblasts, e.g., contraction, migration, proliferation, ECM synthesis, and degradation, is remarkable, usually with a much higher fibrogenic activation under 2-D conditions (44, 71, 76, 149).

The Relation of Structural Proteins to Pathologies

Important information on the functional role of structural components of the ECM has been obtained from mutations in ECM genes that lead to pathologies. Table 1 contains a summary of key structural proteins and their known mutations leading to matrix and tissue failure. With the caveat that some of these components fulfil key functions only in development but may be dispensable later in life, these disease phenotypes provide pivotal information on ECM molecules important for tissue function and thus give insight into their function and dysfunction in the pathology of nongenomic disorders.

Mutations within these structural proteins clearly suggest that these proteins have capacities that are important for the maintenance of a healthy ECM phenotype.

The ECM as Regulator of Cytokine and Growth Factor Activities

In addition to the direct effect of ECM structural molecules on cell phenotype and tissue function, the ECM also serves as a storage site for multiple otherwise soluble cytokines and growth factors (306). In particular, the small leucine-rich proteoglycans (SLRPs), which act as matricellular proteins, are active components of the ECM with a specific role in direct or indirect modulation of the cell-matrix crosstalk. These molecules are modulators of growth factor and cytokine functions, such as TGF-β1 (considered the most potent profibrotic cytokine), tumor necrosis factor (TNF)-α, Wnt-1 induced secreted protein 1, and bone morphogenic proteins (BMPs) (211), but they possess other signaling properties, both as whole proteins and as protein fragments. Specifically, decorin and biglycan are antifibrotic molecules, which by binding active TGF-β1 can interfere with its signaling and neutralize its activity (106). The biological activity of TGF-β1 is attenuated by binding to the core proteins of the SLRPs decorin and biglycan (280–282). The SLRPs are also potent antiapoptotic molecules of tubular epithelial and endothelial cells, acting through binding to the insulin-like growth factor (IGF) type I receptor (211, 280). Notably, biglycan degradation was recently shown to be highly correlated to fibrosis in carbon tetrachloride (CCl₄) and bile duct ligation models of liver fibrosis (106). The soluble factors are usually bound to specific ECM components via low-affinity noncovalent interactions, which create stable concentration gradients around the ECM-embedded cells that produce these cytokines and growth factors, guiding, for example, inflammatory cells but also (myo)fibroblasts toward the target site of inflammation and often excess ECM production. Moreover, ECM binding serves as acellular storage sites, from which these factors will be released during tissue injury, when inflammatory cells degrade the ECM, promoting tissue regeneration and remodeling. Many of the heparan sulfate (HS) (proteoglycan) binding growth factors promote angiogenesis [e.g., fibroblast growth factor (FGF)-1 and vascular endothelial growth factor (VEGF)] or epithelial cell proliferation and survival [e.g., epidermal growth factor (EGF), hepatocyte growth factor (HGF), and keratinocyte growth factor]; others that bind to the abundant collagens regulate fibroblast or immune cell activation [e.g., platelet-derived growth factor (PDGF)-B, oncostatin-M, interleukin-2, and HGF]. Most of these interactions have been studied in detail (289, 291, 315–317), which is schematically drawn in Fig. 3 and listed in Table 2 and highlights that the ECM is a key specific storage facility for potent signaling molecules. Other important examples of molecules sequestered in the ECM are 1) the proinflammatory cytokine TNF-like weak inducer of apoptosis (TWEAK) that induces proliferation of hepatic progenitor cells (HPCs) directly through its receptor fibroblast growth factor-inducible 14, known to be overexpressed in chronic liver diseases (148, 243, 333). Lastly, the hedgehog and Wnt/β-catenin pathways together with Notch signaling have been documented to be important in HPC activation and differentiation (37, 154, 307, 320), as HPCs maintain viability via autocrine and/or paracrine Hedgehog signaling (307).

Open in a separate window

Fig. 3.

Binding of certain growth factors and cytokines to the extracellular matrix. The figure highlights prominent interactions. Shown is a selection of relevant ECM binding factors (also discussed in the text) and their association with either heparan sulfate or collagen and their target cells. The liberation of ECM-stored biologically active growth factors and cytokines can either trigger (inflammatory) cells to further degrade ECM or promote excess ECM deposition by (myo)fibroblasts (as is the case with decorin/biglycan-bound transforming growth factor, TGF-β1). EGF, epidermal growth factor; FGF, fibroblast growth factor; HGF, hepatocyte growth factor; IL-2, interleukin-2; KGF, keratinocyte growth factor; OsM, oncostatin-M; PDGF, platelet-derived growth factor; VEGF, vascular endothelial growth factor.

Tissue Turnover Generates Posttranslational Modifications with Signaling Function Affecting Cell Phenotype

To maintain healthy tissue, the ECM must regenerate itself by normal remodeling, in which aged or damaged proteins are broken down in a specific sequence of proteolytic events and replaced by new proteins. However, during pathological conditions, such as fibrosis and inflammation, the delicate repair-response balance is disturbed (142, 289). The constituents of the “aged” ECM are degraded, which in these disease states results in an array of protein fragments and altered proteins, some of which have documented signaling potential (216, 230, 268, 305, 358, 388), which will be discussed in depth on an individual protein level in the following sections. The original proteins of the ECM are replaced by different constituents, and, consequently, the composition and quality of the matrix are altered. This transformation from a healthy tissue to a pathological one may cause the matrix to become stiff, which has been shown to enhance tumor cell migration, myofibroblast activation, and collagen deposition (19, 21, 29, 66, 143, 171, 219, 253, 271), thereby linking the actual matrix quality to disease progression and changing cell phenotypes. Figure 4 illustrates the steps of abnormal ECM remodeling in fibrosis. Healthy ECM consists of a network of fibers organized in a highly ordered fashion, with binding of key growth factors and signaling molecules at specific interaction sites in the network. During high matrix turnover, the ECM is degraded, leaving fragments of the ECM in the matrix and releasing other fragments into the circulation. Multiple enzymes are released into the matrix by both resident and invading cells, modulating the ECM and generating an altered microenvironment. Consequently, the altered cell-ECM interactions result in altered cellular phenotypes. Different steps during ECM remodeling and fibrogenesis are likely characterized by unique patterns of protein formation, deposition, and degradation, which generates unique protein fingerprints, part of which should be released into the circulation to be exploited as stage-specific serum markers of liver fibrogenesis and/or fibrolysis (259, 288).

Open in a separate window

Fig. 4.

Schematic representation of the high rate of extracellular matrix remodeling in fibrosis. Healthy ECM consists of a network of fibers organized in a highly ordered fashion. During high matrix turnover, the ECM is degraded, leaving fragments of the ECM in the matrix and releasing other fragments into the circulation, while an accumulation of both new and already existing proteins occur, macroscopically described as fibrosis. ECM remodeling is a delicate equilibrium and a prerequisite for maintenance of a healthy tissue, in which senescent proteins are continuously degraded and replaced by new ones. This delicate balance is disturbed in fibrotic diseases, resulting in an increased ECM turnover (both formation and degradation). Thus a subset of pathological proteases is overexpressed in the affected tissue, resulting in the release of protease-specific fragments of signature proteins of the fibrotic ECM, referred to as protein fingerprints or neoepitopes. These fragments may be used as early diagnostic or prognostic serological markers of tissue degradation and in part formation. PTMs, posttranslational modifications. [From Karsdal et al. (164).]

One group of modifications affecting protein quality is PTMs. PTMs are non-DNA-encoded modifications and are a consequence of tissue physiology and pathophysiology (63, 64). Examples of physiological PTMs are isomerization of aspartate (seen in tissue aging), citrullination of arginine, nitrolysation of cysteine (occurring during inflammation), protease degradation at cleavage hotspots (observed in fibrosis and inflammation), glycosylation (in glycemia, type II diabetes), and the fibrosis-specific modifications made by polysialic acid that modulate the profibrotic ductular reaction in liver injury (62, 64, 339). Each of these modifications may change the function and signaling of the modified protein. Several lines of independent evidence suggest that PTMs to specific proteins contribute to abnormal cellular proliferation, adhesion, and morphology (185) and may cause many of the differences in fibrotic compared with normal tissue (36, 125, 185, 209, 279, 321). These specific PTMs made to specific proteins, in particular to structural proteins including collagens, have been shown to be an integrated part of disease progression, exemplified by citrullination of type II collagen in rheumatoid arthritis, cross-linking of collagens in fibrosis and acetylation, citrullination, isomerization, and phosphorylation of myelin basic protein in systemic lupus or multiple sclerosis (63, 164). Similar PTMs made to other collagens or key structural proteins could be important determinants of fibrosis progression, and evidence has recently been obtained for this role, as exemplified by modifications to type XVIII collagen, as discussed later.

There is a growing list of ECM molecules with documented effect on tissue function. This effect may be referred to as a newly discovered paracrine function of ECM proteins. A nonexhaustive list of ECM proteins (as this list continues to grow) and their exerted effects is shown in Table 3. These examples highlight those ECM proteins that serve as paracrine signaling molecules, often revealed during pathological processes that, in addition to cytokines, growth factors, and hormones, become essential players in tissue homeostasis, apart from their roles to anchor cells and transmit positional information and differentiation signals. Notably, some proteins do not change the cellular phenotypes in their native conformation, whereas, subsequent to a specific PTM, a highly potent and novel function of the same protein is revealed.

Posttranslationally Modified Proteins Affect Cell Phenotype: Examples of a Paracrine Function of the ECM in Relation to Fibrosis

Type VI collagen as regulator of fibrogenesis.

Type VI collagen forms microfilaments that traverse interstitial connective tissues. These microfibrils are among the first ECM structures to be degraded upon tissue injury, resulting in larger fragments that, via engagement of integrins and perhaps NG2 proteoglycan, prevent apoptosis and induce fibrogenic activation of surrounding (myo)fibroblasts via activation of proproliferative mitogen-activated kinases and PDGF receptor-α (9, 273, 274). This mechanism likely serves to trigger an immediate wound-healing response, with type VI collagen as a sensor for ECM destruction. Modulation of type VI collagen degradation or blocking its receptors may prevent an overshooting wound-healing response, as occurs in fibrotic tissue remodeling, which is regularly coupled to deregulated ECM proteolysis, as illustrated in Fig. 5.

Open in a separate window

Fig. 5.

Type VI collagen as autoparacrine and paracrine activator of fibrogenesis. Type VI collagen microfibrils serve as sensor of early tissue injury and ECM destruction, usually initiated by release of ECM-degrading proteases like the MMPs by inflammatory cells. The generated proteolytic fragments activate type VI collagen receptors on (myo)fibroblasts that promote their fibrogenic activation, in part via integrins, focal adhesion kinase, and mitogen-activated kinase (Erk1 and Erk2) activation. There is also coactivation of the mitogenic PDGF-α receptor. The activated myofibroblasts then enhance their deposition of ECM components, including intact type VI collagen, whose ongoing degradation maintains a fibrogenic wound-healing response.

The ECM as Regulator of Cell Function: The ECM Interacts with Cells Through Integrins and DDRs and Induces Specific Signaling

Integrins and discoidin domain receptors (DDRs) are part of a network that enables the cell to sense and interact with the microenvironment and adapt to changes (102), through specific interactions and signaling, as outlined in Table 4. Integrins are involved in cell-cell interactions and the attachment of cells to the ECM. They are vital in development and tissue homeostasis, where they transmit “outside-in signals” and “inside-out signals,” which are important in the regulation of cellular proliferation, apoptosis, adhesion, migration, and growth (294). Their extracellular domain is responsible for sensing the microenvironment by ECM binding, which also modulates hepatocyte differentiation (283). Several proteins of the ECM, also the abundant collagens, signal through integrins and DDRs. In hepatic stellate cells (HSC), collagen type I signals through integrin-α₁β₁ (and to a minor degree through α₂β₁), whereas collagen type III engages mainly DDR1 (Table 4). The DDRs are receptor tyrosine kinases that specifically recognize collagens as their ligands (276). Similar to collagen-binding β1-integrins, the DDRs bind to specific secondary structural motifs within the collagen triple helix (383). DDRs are involved in tissue homeostasis and transduce signals regulating cell polarity, tissue morphogenesis, and cell differentiation (345). Here, DDR1 and DDR2, in particular, have demonstrated important functions in tissue homeostasis and cancers (243, 385), in which lack or inhibition of DDR1 attenuated fibrogenesis (35, 97, 120, 272, 276), whereas DDR2 deficiency promoted experimental liver fibrosis (101).

Liver Pathologies Associated with the ECM

Common to all underlying causes of fibrosis is the disruption of the normal ECM pattern attributable to activation of (myo)fibroblasts (activated HSCs) in the portal tracts initiated by tissue injury and inflammation. This enables proliferation and migration of these fibrogenic effector cells into the parenchyma along the sinusoids, which is the hallmark of incipient septum formation (289). The activated HSCs deposit excess ECM, serving to “close the wound” and to provide a scaffold for orderly liver regeneration in case of a short-term insult. However, with ongoing (inflammatory) insults, the excess ECM, which becomes dominated by fibrillar collagen (mainly type I and type III), generates scar tissue that merely maintains organ integrity but has lost the guiding function for orderly tissue regeneration (267). The deposition of ECM is dependent on etiology and consequently on activation of different subtypes of cells, including mainly myofibroblasts of variant activation state, suggesting that “fibrosis is not just fibrosis.” During fibrosis progression, the quantity and quality of the hepatic ECM changes with an up to 10-fold increase in collagenous and noncollagenous components followed by a shift in matrix composition from the low-density basement membrane-like matrix to an interstitial matrix containing mainly fibril-forming collagens (289). A specific characteristic of fibrosis development in the liver is the presence of specific growth points of the scar tissue, i.e., portal zone vs. central zone, and space of Disse, determining the development of portal, central, or pericellular fibrosis, with different clinical consequences as to development of liver failure, i.e., with relative preservation of liver function in portal fibrosis despite massive collagen accumulation (285). Thus the developmental pattern of fibrosis is dependent on the underlying etiology causing the fibrosis (51, 285).

The structural representation of fibrosis depends on the cell types and injury involved. Today, several fibroblast-like cell types have been identified in the liver, all of which also contribute to the development of fibrosis, including 1) septal myofibroblasts present in the inner part of fibrous septa, 2) activated HSCs located in capillarized sinusoids adjacent to expanded portal tracts, 3) interface myofibroblasts located at the edge of fibrous septa derived from activated HSCs (or portal fibroblasts) recruited at the site of injury where the ECM turnover (synthesis and degradation) and accompanying cell damage and inflammatory infiltration are highest (329), and 4) smooth muscle cells localized in the larger vessel walls (2, 99, 115, 198). Consequently, the activity of these different cell types results in distinct fibrosis patterns with formation of 1) portal-portal, 2) portal-central, or 3) central-central septa (Table 5).

ECM and HCC

Worldwide, HCC is the fifth most common cancer and the third most common cause of cancer-related death (304). In the US, the occurrence of HCC is increasing, mainly because of the rising prevalence of end-stage hepatitis C and NASH (160, 236). HCC is strongly correlated to fibrosis severity (203), with 90% of HCC cases arising in cirrhotic livers (299). Thus more than 80% of patients with hepatitis B and C that have HCC are cirrhotic (140). Similarly, HCC prevalence and progression is related to cirrhosis (93) in alcoholic steatohepatitis and NASH (8), with a yearly HCC incidence of 1.7% (93) and 2.6%, respectively (8). Mechanisms linking fibrosis and HCC are in need of further exploration, whereas there is increasing evidence that the fibrotic/cirrhotic and associated (immunosuppressive) inflammatory liver microenvironment appears to be a major player in HCC pathology, with several molecular pathways being shared between fibrogenesis and carcinogenesis (340). Thus, apart from direct carcinogens such as hepatitis B virus, microenvironmental stimuli include intrahepatic cell subpopulations, such as progenitor, immune and stellate cells, proliferative or differentiation-modulating cytokines, and growth factors like TGF-β1, EGF, IGF-1, and VEGF and altered ECM molecules and ECM receptors like the integrins. These interact and increase protease activities directed at the ECM or cell-associated structures, facilitating ECM remodeling, cancer cell proliferation, and migration. An example is MMP-2, which is able to degrade numerous components of the ECM that are closely correlated with tumor invasion and metastasis. Together with hypoxia-inducible factor 1α, which regulates proteolytic and angiogenic activities, MMP-2 is overexpressed and its activity enhanced in HCC compared with healthy tissue (359). Additionally, MMP-7 and MMP-26 levels are significantly higher in HCC tissue compared with adjacent healthy hepatic cells from patients with HCC (361) and strongly correlated to phosphorylated FGF receptor 2, supporting a direct link to angiogenesis (361).

Another class of enzymes that is linked to HCC development is LOXL2, which catalyzes collagen cross-linking, enhancing matrix stiffness and resistance to proteolysis, and promotes HCC metastasis (374). Induction of LOXL2 occurs via multiple regulators, including hypoxia, TGF-β1, and microRNAs (374). Matrix stiffness alone appears to determine fibrosis and HCC progression, as expression of procollagen, LOXL2, VEGF and the endothelial marker CD31 is highly correlated (82). Therefore, the ECM remains a yet insufficiently exploited target for HCC therapies.

Do We Need to Target the ECM?

We have highlighted the importance of the ECM in cell differentiation and function, fibrosis, and HCC development. The normal and especially fibrotic ECM affects cell phenotype through cell-ECM receptor interactions and liberated and stored growth factors/cytokines/chemokines and vice versa. Moreover, we have highlighted specific PTMs made to the ECM that themselves affect cell phenotype and fate differently than the parent molecules. Consequent to these observations, it should be possible to ameliorate progression of and reverse established fibrosis without crucially disrupting beneficial cell-ECM interactions by differentially targeting pathogenic ECM structures, PTMs, and protein fragments.

Clearly, interfering with the usual suspects TGF-β, PDGF, or VEGF may ameliorate fibrogenesis in certain settings, only in preclinical studies, to date. Direct targeting of the ECM and cells whose fibrogenic activation or fibrolytic activity depends on specific ECM structures that are prevalent in fibrosis would render a fibrosis-specific and potentially fibrolysis-inducing therapy. This appears feasible because the ECM is continuously remodeled, even in cirrhosis (289, 291, 301). Such therapy would ideally be coupled to elimination or suppression of the major cause of fibrosis, such as viral hepatitis B or C. Although this seems a long way to go, the advanced understanding of the pathophysiology of the ECM in fibrosis is already beginning to translate into the clinic, as exemplified in an ongoing clinical study with a LOXL2-blocking antibody.

Conclusions

There is a growing body of evidence that modifications made to the structural proteins of the matrix may be both a consequence of the disease as well as drivers of disease progression. Consequently, PTMs within specific ECM proteins (such as degradation products) may be more integrated in pathogenesis than previously thought. Evidence for the use of biomimetic peptides from the ECM, such as types IV and XVIIII collagen, as anti-ECM therapeutics, which block fibrogenesis and ECM remodeling, is emerging (189, 358, 388). These examples begin to suggest that matrix molecules themselves may be antifibrotic agents in addition to kinase inhibitors and receptor blockers. A further example on how PTMs regulate ECM function is cross-linking of collagens or elastin by LOXL2, which prevents degradation and increases stiffness of the tissue. These combined observations clearly suggest that the ECM is more than just a structural framework for tissues and may perform paracrine functions affecting cell phenotype and fate. It is of key interest to understand the role of each of the major ECM components and their PTMs as well as their signaling potential to grasp the full potential of the ECM and its importance in pathogenesis.