Sinusoidal remodeling and angiogenesis: A new function for the liver-specific pericyte?

Abstract

Pericyte is a term for vascular mural cells that make specific focal contacts with the endothelium within the microvasculature.1 For a long time, the existence and role of pericytes were neglected, but during recent years these cells have gained increasing attention, not only as contractile cells but also as obligatory regulators of vascular development, stabilization, maturation, and remodeling. Intimate interactions between pericytes and endothelial cells are reflected by observations that impairment of either of these vessel wall cell types inevitably affects the other. Pericytes also reside in liver, where they maintain specialized functions. Hepatic stellate cells (HSCs) are thought to be the pericyte equivalent in the liver.2 Much of our understanding of HSC function is derived from the broader function of pericytes in vascular beds outside the liver. In liver, HSC are well established as collagen-producing cells but are being increasingly recognized for their role in angiogenesis and vascular remodeling, processes that are particularly important in cancer and portal hypertension. This review highlights new advances in pericyte biology that are relevant to angiogenesis and vascular remodeling and places them in the context of HSC and liver pathobiology. EC, endothelial cells; HSC, hepatic stellate cells; PDGF, platelet-derived growth factor; TGF-β, transforming growth factor beta. Although pericytes were described more than 100 years ago as perivascular cells that wrap around blood capillaries, the term pericyte (peri, around; cyte, cell) was first introduced by Zimmerman in 1923.1 Because pericytes have some morphological features of smooth muscle cells, and a smooth muscle layer is absent from capillaries and postcapillary venules, Zimmerman attributed the property of contractility to pericytes. Morphologically, pericytes possess long processes embracing the abluminal endothelium wall in precapillary arterioles, capillaries, and postcapillary venules. Long cytoplasmic processes that extend along and encircle the endothelial tube ideally position these cells for paracrine signaling with endothelial cells (EC) (Fig. 1). Close apposition of pericytes and endothelial cells in microvasculature. (A) Longitudinal and cross-sectional cartoons depict pericytes wrapping around endothelial cells as occurs in precapillary arterioles, capillaries, and postcapillary venules. (B) Complementary scanning electron micrograph of regenerating liver with active angiogenesis 72 hours after partial hepatectomy depicts HSC (arrows) wrapping around endothelial cells (Reproduced from www.sciencephoto.com). (C) Transmission electron micrograph depicts HSC (SC) in close contact (arrows) with the hepatic sinusoidal EC within the hepatic sinusoid, providing anatomic evidence of the concept of HSC as a liver-specific pericyte. Arrowhead indicates fenestrae in sinusoidal endothelial cell. H: hepatocyte (Reproduction by permission from HEPATOLOGY 2001;33: 363–379). Pericyte density differs with respect to the function of vessels and organs in which they are found. For example, pericytes in brain contribute to blood–brain barrier regulation, whereas in kidney they contribute to glomerular function, and in pancreas they may contribute to fibrosis.3-5 The pericyte coverage of the abluminal endothelial surface is only partial and varies extensively (10%-40%) between capillary beds of different tissues,6 likely reflecting specific functional features of microvessels in different organs.7 Pericytes have an intermediate phenotype between vascular smooth muscle cells and fibroblasts with a capacity to differentiate into a myofibroblast phenotype. These varied phenotypes have translated into diverse functions, many of which are facilitated by paracrine interplay with neighboring EC and are implicated in a variety of disease processes. Anatomic characteristics of HSC have led to the distinction of HSC as a liver-specific pericyte.2, 8 HSCs are located in the space of Disse in close contact with sinusoidal EC, supporting their pericytic function (Fig. 1).9, 10 HSCs are dispersed along the sinusoid at relatively fixed distances, and with their spatial extensions, they may be sufficient to cover the entire hepatic sinusoidal microcirculatory network.11 Although the most conspicuous ultrastructural feature of HSC in normal adult liver is the presence of cytoplasmic lipid droplets,11 HSC do share some cell phenotypic markers with other pericytes. This includes expression of smooth muscle actin, desmin, NG2, and glial fibrillary acidic protein.12, 13 Advances in culture methods have generated a wealth of cellular and molecular information about HSC function. This includes (1) the function of retinol transport and storage in the liver14-16; (2) transforming growth factor beta (TGF-β)–dependent extracellular matrix regulation17-22; (3) mitogenic and motogenic responses to platelet-derived growth factor (PDGF),23, 24; and (4) the process of transdifferentiation or activation.25 However, the discovery of the vasomotor activity of HSC in response to nitric oxide and endothelin25, 26 began to reinforce the emerging concept of these cells as liver-specific pericytes.8, 27 More recent studies have extended this concept by defining a new function of HSC in the process of tumor angiogenesis and sinusoidal remodeling.28, 29 These new functions are recapitulated by studies of pericytes in other organs that identify the protean role and relevant mechanisms of this cell type for vascular development and function. The complementary processes of angiogenesis (vascular growth from preexisting vessels) and vasculogenesis (de novo blood vessel development) regulate vascular development and homeostasis (Fig. 2). Although EC are the cell type most recognized in vasculogenesis and angiogenesis, pericytes also contribute to these processes by stabilizing and maintaining blood vessels. Similar to EC, pericytes also may be derived from a hemangiopoietic cell and may be capable of guiding sprouting processes required for angiogenesis30, 31 (although, in this respect, HSC that may be derived from mesenchymal stem cells may differ from non-liver pericytes) (discussed later). Mechanisms of formation of new blood vessels. New blood vessels can be formed by different mechanisms: Progenitor cells form rudimentary vascular tubes during vasculogenesis, which eventually develop into mature vessels (top). This de novo formation of new blood vessels from progenitor cells contributes to liver organogenesis and probably contributes to changes in vessel structure postnatally as well. Angiogenesis (bottom), the formation of new vessels from pre-existing vasculature by sprouting or intussusception, can occur in physiological situations (e.g., liver regeneration) or in pathological settings (e.g., chronic liver disease, tumor angiogenesis in hepatocellular carcinoma, and liver metastases). Some of our understanding of the role of HSC in hepatic angiogenesis and remodeling comes from work using the rodent partial hepatectomy model, a surgical model of liver regeneration.32-37 Angiogenesis and sinusoidal remodeling is requisite to supply blood flow to newly replicating avascular islands of hepatocytes during the regeneration process.38 HSC along with EC migrate into these parenchymal cell clusters, resulting in the formation of new sinusoidal branches. Although the precise role of HSC in this process is not fully defined, matrix changes and production of growth factors and cytokines that occur in response to HSC activation do regulate hepatocyte proliferation.38, 39 Furthermore, HSC activation also regulates sinusoidal structural changes in a precise manner, because excess HSC activation as well as inadequate HSC activation are both associated with suboptimal sinusoidal revascularization during liver regeneration.40-43 Pericyte-driven angiogenesis and vascular remodeling involve both the pericyte recruitment and proliferation to the vascular wall and the secretion of angiogenic factors by pericytes that attracts endothelial cells. These two related and probably codependent processes are described below. A number of signaling pathways mediate pericyte recruitment to vessels in the process of vascular remodeling and angiogenesis, including PDGF, TGF-β, angiopoietins, and nitric oxide, are described later and are illustrated schematically in Fig. 3. PDGF may be the most critical growth factor pathway in the recruitment of pericytes to newly formed vessels.44 Knockout of PDGF or PDGF receptor leads to perinatal lethality.45 However, detailed analysis of the microvasculature in the embryos of these mice demonstrates abnormalities in the diameter of the pericyte-deficient vessels with endothelial hyperplasia and vascular leak,46 supporting the concept that pericytes work closely with EC and are important in the process of liver organogenesis.46 During angiogenesis, PDGF ligand is expressed by the sprouting endothelium, and PDGF receptor is expressed by the pericyte,44, 47 again supporting a paracrine mode of interaction between the two cell types.48, 49 Furthermore, the level of PDGF production is also a key determinant of pericyte recruitment. Indeed, available evidence strongly suggests that endothelial PDGF signals control pericyte recruitment to angiogenic vessels through specific chemotactic gradients that are generated and regulated by EC.49-52 Angiogenic signaling pathways in pericytes. Pericytes (green) secrete vascular factors that contribute to endothelial cell (brown)–based angiogenesis such as VEGF and angiopoietins. Pericytes also respond to vascular factors secreted by endothelial cells and parenchymal cells. These factors, which include sphingosine-1 phosphate, PDGF, nitric oxide, and TGF-β, have diverse effects on pericytes that contribute to pericyte-based angiogenesis and vascular remodeling. Another key step in the process of vessel formation and remodeling is the regulation of vessel stability. The angiopoietins Ang1 and Ang2 regulate vessel stability by activating (Ang1) or antagonizing (Ang2) signaling via the Tie 2 receptor. The receptor tyrosine kinase Tie2 is expressed in EC53 and is bound by Ang-1 or (Ang-2), secreted by surrounding pericytes. Ang-1 is involved in vessel maturation, as supported by recent data that Ang-1 promotes maturation of pericyte-deficient blood vessels in the retina.54 Ang-1 also stabilizes vessels by promoting pericyte recruitment.55 The mechanism for this effect is unclear but may involve the release of pericyte recruitment factors (e.g., PDGF) from EC. Ang-1 also may have direct effects on pericytes or pericyte progenitor cells. Mice genetically deficient in Tie2 or Ang-153, 56 have severe vascular defects and are unable to recruit pericytes.7 Ang-2, which is a natural antagonist of Tie2, is implicated in vessel destabilization and is expressed mainly by EC at sites of active angiogenesis.57 Expression of Ang-2 has been negatively correlated with pericyte coverage in both spontaneous and experimental tumors.58, 59 Thus, Ang-1 and Ang-2 may mediate opposing effects on pericyte recruitment and vessel stabilization,60 although this topic continues under intensive investigation. Sphingosine-1-phosphate is a secreted sphingolipid engaged in cell communication through its cognate G-protein coupled receptors. Sphingosine-1-phosphate triggers cytoskeletal, adhesive, and junctional changes affecting cell migration, proliferation, and survival.61 Although sphingosine-1-phosphate receptor expression was originally described in EC, these receptors may be expressed widely in mesenchymal cells as well.62 Disruption of the sphingosine-1-phosphate receptor gene in mice causes embryonic lethality because of vascular defects characterized by aberrant recruitment of vascular smooth muscle cells and pericytes to the developing aorta.62 Thus, sphingosine-1-phosphate signaling further highlights the important role of EC–pericyte cross-talk in vascular structures.63-68 The roles of some of these pathways in the specific context of HSC and their liver-specific functions are discussed further in later sections. Pericytes are rich sources of polypeptides, eicosanoids, and various other small molecules with paracrine, juxtacrine, autocrine, or chemoattractant functions. In the case of HSC, these include (1) polypeptides that enhance proliferation in an autocrine or paracrine way [HGF, vascular endothelial growth factor (VEGF), endothelin, IGF-II, and possibly TGF-α, epidermal growth factor, and acidic fibroblast growth factor]; (2) TGF-β superfamily members; (3) neurotrophins; and (4) hematopoietic growth factors such as erythropoietin.69 This concept was exemplified in a recent study in which the angiogenic cytokine leptin was found to up-regulate proangiogenic cytokine release from HSC. Because angiogenesis may be a prerequisite step for fibrosis in liver, similar to that observed in lung, such pathways may have significant implications for the wound healing response and ensuing fibrosis in liver, as discussed later.70 Angiogenesis in tumors leads to a chaotic, poorly organized vasculature with tortuous, irregularly shaped, and leaky vessels that are often unable to support efficient blood flow. Because of the imbalanced expression pattern of angiogenic factors, tumor vessels appear to be in a constant state of remodeling, which involves simultaneous formation and regression of vascular tubes.71, 72 Just as tumor EC differ from the normal, quiescent EC, tumor pericytes also differ from normal pericytes. In general, pericytes in tumors appear to be more loosely attached to the vasculature, and their cytoplasmic processes can extend into the tumor tissue. They seem to be less abundant in some tumor tissues in comparison with the respective normal tissue and can change their expression profile.73, 74 The exact causes of abnormal pericyte behavior are still unknown but may include imbalanced EC/pericyte signaling circuits or a limited pool of recruitable pericytes.75 Hematopoietic cells from bone marrow that expressed the pericyte marker NG2 were recently identified in close contact with blood vessels in xenograft Bl6-F1 melanoma tumor models76 and in the transgenic model of pancreatic islet carcinomas.77 This suggests that recruitment of bone marrow–derived cells to sites of a growing vasculature is not limited to EC, but can also include pericytes. The functional significance of PDGF receptor signaling in pericytes from tumors was evident in a recent study with PDGF retention mice (PDGF-Bret/ret), which lack the C-terminal retention motif in PDGF that mediates PDGF binding to proteoglycans at the cell surface and in the extracellular matrix. These mice evidence disruption of the PDGF chemotactic gradient in vivo. These mice are viable but have fewer pericytes, because they lack proper recruitment and integration of pericytes within the vessel wall.52 Implanted tumors in PDGF-Bret/retmice were hemorrhagic and contained few pericytes around the tumor blood vessels, which were hyperdilated. Ectopic expression of PDGF in those tumor cells was able to increase pericyte density but failed to cause pericytes to attach more firmly to blood vessels, indicating that localized PDGF from the endothelium is essential for proper pericyte adhesion to the vessel wall.51 These data suggest that tumors may use mechanisms for angiogenesis that are similar to those used in developmental angiogenesis. These results also indicate that tumor pericytes, though less abundant and more loosely attached than normal pericytes, still regulate vessel integrity, maintenance, and function.78 Current anti-angiogenic strategies for cancer have focused on the EC. Can the combination of anti-endothelial and anti-pericyte agents act synergistically in antiangiogenic therapy? Glioblastomas or fibrosarcomas that overexpress PDGF exhibited an increased pericyte density around blood vessels.79 Blocking PDGFR signaling in a transgenic mouse model of pancreatic islet carcinogenesis (Rip1Tag2) with the receptor tyrosine kinase inhibitor SU6668 caused regression of blood vessels, which was attributable to the detachment of pericytes from tumor vessels and thereby restricted tumor growth.80 Similarly, SU6668 caused detachment of pericytes in xenotransplant tumors, thereby restricting tumor growth.81, 82 The fact that tumor vessels without pericytes appear more vulnerable suggests that they may be more responsive to anti-endothelial cell drugs. Indeed, combinations of receptor tyrosine kinase inhibitors that target ECs and pericytes by blocking VEGF and PDGF signaling, respectively, more efficiently diminished tumor blood vessels and tumor size than either inhibitor alone.80 Similar effects were observed when PDGF pathway inhibitors were combined with an antiangiogenic chemotherapy regimen that targeted EC.83 Thus, targeting pericytes in combination with anti-EC approaches may collectively destabilize the existing tumor vasculature more potently than targeting of either cell type individually. The liver is one of the first sites of primary or secondary oncogenesis, and numerous reports also have described the development of activated HSCs concomitantly with HCC, supporting the concept that HSC serve a pericytic function in liver tumor angiogenesis.29 Tumor-activated HSCs are responsible for the remodeling and deposition of tumor-associated extracellular matrix84-86 and have been involved in the migration and growth of hepatoma and metastatic cells.29, 87-90 HSCs also synthesize VEGF on in vitro activation and in response to hypoxia,91 suggesting their additional contribution to the angiogenic needs of developing hepatic tumors through secretion of vascular factors. Thus, the hypoxic induction of VEGF in tumor-activated HSCs may create a proangiogenic microenvironment, facilitating EC recruitment and survival during hepatic metastasis transition from an avascular to a vascular stage.28 Therefore, activation and migration of HSC into tumor nodules appear to be essential requirements for hepatic tumor angiogenesis. In cirrhosis, the syndrome of portal hypertension occurs in large part through changes in hepatic resistance. Prior work has helped to elucidate the role of sinusoidal vasoconstriction in the genesis of portal hypertension, especially mechanisms by which HSC operate as a “contractile machinery” and relax in response to nitric oxide.92-94 Although good evidence supports a contraction-based vasoconstrictive function of HSC in the setting of cirrhosis and portal hypertension,8, 25, 95 the role of HSC as a regulator of vascular tone in normal liver is less established. For example, although some studies have demonstrated that modulation of vasoactive agents such as endothelin and carbon monoxide cause changes in sinusoidal diameter at locations where HSC reside,96, 97 other studies have suggested that these vasoregulatory changes actually occur outside of the hepatic sinusoids.98, 99 HSC density and coverage of the sinusoidal lumen is enhanced in cirrhosis (Fig. 4). Although other structural changes occur within the sinusoids as well in cirrhosis, the enhanced coverage of sinusoidal vessels by HSC, because of the contractile nature of HSC, will likely allow this process of sinusoidal remodeling to contribute to a high-resistance, constricted sinusoidal vessel. This process of sinusoidal remodeling is distinct from the more characterized role of HSC in the process of collagen deposition and fibrosis.17, 100 How this contractile machinery is actively assembled around the sinusoidal channel in a manner necessary to generate vasoconstriction is not well understood. However, HSC motility and migration are no doubt required to promote enhanced coverage of HSC around an EC-lined sinusoid. At the cellular level, recent studies indicate that the changes in HSC membrane structures that promote their motility and migration are important steps in the process of sinusoidal remodeling.101 Although TGF-β is largely recognized for its contribution to HSC-based collagen deposition, significant crosstalk occurs between TGF-β and PDGF in the process of HSC motility. Indeed, these signals may converge at the level of the c-abl tyrosine kinase.102 Recent studies have inhibited both of these pathways through use of compounds such as imatinib, thereby limiting vascular remodeling in lung.102 Similar pathways may be active in hepatic sinusoidal remodeling as well. Furthermore, inducing HSC apoptosis also may represent a strategy to prevent aberrant sinusoidal remodeling in portal hypertension. HSC contribute to vascular remodeling in chronic liver disease. (A) Chronic liver disease leads to profound changes of the sinusoidal vascular network with development of portal hypertension. The scanning electron microscopy of a vascular cast from rats shows normal architecture of sinusoidal vessels (top left, bar = 100 μm). Experimental fibrosis in CCl4-treated (top right, bar = 100 μm) leads to remodeled and abnormal vessels of varying diameter (arrow) separated into micronodules (arrowheads). (Scanning electron microscopy with permission from: Onori P, Morini S, Franchitto A, Sferra R, Alvaro D, Gaudio E. Hepatic microvascular features in experimental cirrhosis: a structural and morphometrical study in CCl4-treated rats. J Hepatol 2000;33:555-563). (B) Several processes have been recognized to lead to vascular remodeling in the liver: On a cellular level, resting HSCs become activated and not only deposit matrix, which leads to formation of a basement membrane around the sinusoids, but also proliferate and enhance their coverage of the sinusoids. This increase in mass of “contractile machinery” transforms the low-resistance vascular bed into a higher-resistance bed, leading to portal hypertension. The development of advanced fibrosis and cirrhosis presents two further models of HSC behavior that show similarities with pericyte behavior in other contexts. In the first, the development of fibrosis and its spontaneous resolution, there is a progressive activation of HSC to become myofibroblast-like. As noted previously, these cells contribute to remodeling of the sinusoid and will ultimately secrete the complex fibrillar collagen-rich matrix that is characteristic for fibrosis and results in the gross disruption of the organ structure. Studies in animal models have indicated that where spontaneous resolution occurs (and this is now a well-documented phenomenon in specific human models of disease also) matrix degradation occurs and is accompanied by apoptosis of the activated HSC.43, 103-105 This process also remodels the sinusoids, returning them to normal or near normal architecture. This process of remodeling, however, may require the presence of macrophages.106 In models of advanced human and rodent fibrosis, good evidence now exists for bone marrow stem cell contribution to the myofibroblast and HSC population.107, 108 Thus, pericytes may be derived from hematopoietic stem cells77 or from mesenchymal progenitor cells,109 depending on the pathobiological setting. Whether the recruited cells will pass through a HSC/pericyte phenotype en route to becoming a myofibroblast or whether cells can be recruited directly as myofibroblasts is unknown. A further form of sinusoidal remodeling occurs in the liver in more advanced cirrhosis. Here, dense fibrils linking vascular structures undergo remodeling, and angiogenesis occurs. These scars contain myofibroblasts, likely derived from activated HSC. Intriguingly, if a comparison is drawn between a mature scar within a cirrhotic liver and a less mature scar, the cells within the mature scar express markers more commonly associated with HSC than with myofibroblasts, such as glial fibrillary acidic protein and desmin in the absence of alpha smooth muscle actin.104 These smooth muscle actin–negative HSC may participate in sinusoidal remodeling in this context. Currently, the pathogenesis of remodeling within fibrotic bands is incompletely defined, although roles for hypoxia, acting as a potent stimulus for VEGF secretion by HSC, have been implicated in some elegant models.110 In summary, HSC share many anatomic and phenotypic similarities with pericytes in other locations. This is paralleled by the important role of HSC in the emerging fields of angiogenesis and vascular remodeling, which are fundamental processes in liver tumor angiogenesis and cirrhosis/portal hypertension.