Macrophage-derived vascular endothelial growth factor and angiogenesis within the hepatic scar-new pathways unmasked in the resolution of fibrosis.

Abstract

Potential conflict of interest: Nothing to report. See Article on Page 2042 The potential for regression of liver fibrosis and even cirrhosis has been demonstrated in a catalogue of rodent models and in therapeutic studies in humans. Crucially, fibrosis regression in patients with cirrhosis has been shown to correlate with improved clinical outcomes (e.g., survival, reduction in portal hypertension). For this to occur there must be cessation of active scar production and degradation of extracellular matrix (ECM) that has already been deposited in order to restore tissue architecture toward normality. Observations from rodent models of resolving liver fibrosis have determined that loss of profibrogenic myofibroblasts through inactivation, apoptosis, or senescence and modulation of the balance of proteolytic enzymes (matrix metalloproteinases, MMPs) and their inhibitors (tissue inhibitors of metalloproteinases, TIMPs) in the fibrotic microenvironment are central mechanisms in mediating homeostatic liver repair. Furthermore, scar‐associated macrophages (SAMs) have emerged as critical effector cells in mediating both the progression and regression of fibrosis. A new, and perhaps surprising, relationship between fibrosis regression and angiogenesis is revealed by Kantari‐Mimoun et al. in this issue of hepatology.1 Here, the authors show that myeloid cell–derived vascular endothelial growth factor (VEGF) drives revascularization of the receding hepatic scar and promotes ECM degradation through modulation of MMP and TIMP expression in liver sinusoidal endothelial cells (LSECs) (Fig. 1).Figure 1: The dichotomous role of VEGF in liver fibrosis. (A) In advanced experimental murine liver fibrosis and human fibrosis/cirrhosis biopsies, an increase in VEGFR2+ vessels was localized to the perifibrotic regions, whereas the scar area itself was relatively low in VEGF and sinusoids were almost absent. During chronic fibrogenic injury in mice, hepatic VEGF levels increased, then subsequently declined after prolonged (12‐week) CCl4 treatment. Genetic ablation of myeloid cell–derived VEGF in chronic CCl4 or bile duct ligation injury in mice had no effect on the development of liver fibrosis despite reduced angiogenesis, indicating that other VEGF‐mediated effects (e.g., sinusoidal permeability) are inhibited. (B) In the context of resolving CCl4‐induced fibrosis, wild‐type mice exhibited revascularization of the hepatic scar and ECM degradation. In contrast, genetic ablation of myeloid cell–derived VEGF or treatment with a VEGFR2 inhibitor abrogated angiogenesis within the scar and was associated with a failure of ECM degradation. Myeloid cell–derived VEGF promoted a proresolution phenotype in LSECs, characterized by up‐regulation of hepatic MMP2 and MMP14 and downregulation of TIMPs. Links between VEGF expression in resolving fibrosis and downstream effects on liver regeneration and portal hypertension require further investigation. Abbreviations: MF, myofibroblast; HC, hepatocyte.The negative consequences of pathological angiogenesis and sinusoidal remodeling in liver fibrosis are well documented. Angiogenesis and fibrosis occur in parallel in many organs, including the liver; and in human liver specimens as well as in animal models, microvessel density correlates with the degree of liver fibrosis. Furthermore, the pattern of fibrosis (i.e., postnecrotic, biliary, centrilobular, or perisinusoidal) influences the degree of associated angiogenesis, which in turn may increase the rate of disease progression to cirrhosis and limit the potential for fibrosis regression. Angiogenesis and pivotal proangiogenic molecules such as VEGF have been linked to hepatic fibrogenesis, the development of portal hypertension, and, potentially, the evolution of hepatocellular carcinoma. Hypoxia and inflammatory cues activate angiogenic and fibrotic pathways in chronic liver injury. Ehling et al. recently showed that CCL2‐dependent inflammatory monocyte–derived macrophages controlled angiogenesis in a mouse liver fibrosis model, whereby macrophages were abundant in injured livers, colocalized with newly formed vessels in portal tracts, and exhibited a proangiogenic gene signature including up‐regulation of VEGF and MMP9.2 Subsequent inhibition of monocyte infiltration through targeting of CCL2 prevented fibrosis‐associated angiogenesis but not fibrosis progression. Interestingly, genetic ablation of myeloid cell–derived VEGF in chronic carbon tetrachloride (CCl4) or bile duct ligation injury had no effect on the development of liver fibrosis, despite reduced angiogenesis (Fig. 1). Vascular endothelial growth factor regulates angiogenesis and vascular remodeling through effects on endothelial cell proliferation and survival and mediates profibrogenic effects in hepatic stellate cells (HSCs) and LSECs through the VEGFR2 receptor, whereas angiopoietin‐1 promotes vessel stability by activating Tie‐2 receptors on LSECs. Concurrently, LSECs also produce platelet‐derived growth factor, which has strong mitogenic effects on HSCs and regulates their recruitment to the endothelial tubule, thereby stabilizing the vascular architecture of nascent vessels. The strong link between angiogenesis and fibrogenesis is reinforced by the observation that in vivo targeting of the VEGF/VEGFR2 pathway, using either anti‐VEGF antibodies or receptor tyrosine kinase inhibitors (e.g., sorafenib), ameliorates hepatic fibrosis. The role of VEGF in the resolution of hepatic fibrosis has only recently been described. Yang et al.3 showed that VEGF had dual and opposing effects on liver fibrogenesis and fibrosis resolution in mice using reversible models of CCl4 and bile duct ligation. Moreover, treatment with a VEGF neutralizing antibody prevented fibrogenesis but also abrogated the resolution of fibrosis by impeding vascular permeability, monocyte infiltration, and macrophage‐dependent ECM degradation. In contrast, Kantari‐Mimoun et al.1 used targeted deletion of VEGF‐A in myeloid cells, using a loxP‐flanked Vegfa allele crossed to the lysozyme M promoter‐driven Cre recombinase, and showed that in resolving hepatic fibrosis myeloid cell–derived VEGF promoted angiogenesis within the scar and matrix remodeling (Fig. 1). Furthermore, pharmacological inhibition of VEGFR2 signaling using a small molecule inhibitor had similar effects, suggesting involvement of VEGFR2 in the resolution of hepatic fibrosis. These intriguing findings are another example of the crucial importance of SAMs in fibrosis regression. The “restorative” macrophage population is derived from an in situ phenotypic switch of profibrogenic Ly‐6Chi inflammatory monocyte‐macrophages, potentially induced by phagocytosis, resulting in down‐regulation of proinflammatory cytokines and chemokines, increased expression of fibrolytic MMPs (particularly MMP12 and MMP13), other antifibrotic genes such as CX3CR1 and arginase‐1, and proliferative signals for hepatocytes and hepatic progenitor cells.4 However, whereas Yang et al.3 demonstrated that VEGF regulated sinusoidal permeability and monocyte trafficking and that SAM‐mediated fibrosis resolution was dependent upon CXCL9 and MMP13 production, Kantari‐Minoun et al.1 did not observe differences in SAM (or neutrophil) kinetics in the fibrotic livers of CCl4 or bile duct ligation–treated, myeloid cell VEGF‐deficient mice. This indicates, at least in mouse models of resolving fibrosis, that autocrine VEGF production is not required for myeloid cells to extravasate and enter the fibrotic scar. Furthermore, although hepatic CXCL9 and MMP13 transcripts were up‐regulated in the resolution phase of CCl4 fibrosis, there were no differences between wild‐type and mutant mice. The authors attribute these discrepancies to differences in duration of CCl4 treatment (12 weeks versus 6 weeks) and the methods of VEGF ablation used in these two studies, wherein LysMCre/VEGF+f/+f mice exhibit a defect in autocrine VEGF signaling in SAMs but VEGF‐neutralizing antibodies also block VEGF derived from paracrine and endocrine sources. Indeed, different anti‐VEGF strategies may affect discrete VEGF isoforms, which may in turn determine the specific net tissue response. A further twist in the current study was the identification of LSECs as master regulators of fibrolysis through up‐regulation of MMP2 and MMP14 as well as down‐regulation of TIMP1 and TIMP2. The LSECs have not hitherto been considered to be a major source of proteolytic enzymes in the resolution of fibrosis. Importantly, adoptive transfer of wild‐type bone marrow cells into myeloid cell VEGF‐deficient mutant mice prior to the recovery phase of the CCl4 model restored the proresolution LSEC phenotype, collagenolytic activity, and resolution of fibrosis. The VEGF/VEGFR2 pathway is also essential for effective liver regeneration. Inducible VEGFR2‐deficient mice have impaired liver regeneration after partial hepatectomy.5 Increased hepatic VEGF levels after chronic injury or partial hepatectomy leads to the recruitment of bone marrow–derived sinusoidal endothelial cell progenitor cells to the liver.6 Also, VEGF stimulates hepatocyte growth factor expression by LSECs. However, in the current study there was no difference in the number of proliferating cell nuclear antigen–positive proliferating hepatocytes in the livers of mutant compared to wild‐type mice. This may be due to the involvement of other important factors that link angiogenesis and liver regeneration, such as LSEC‐derived angiopoietin‐2. What are the therapeutic implications of these findings? It is tempting to speculate that augmentation of proangiogenic effects in resolving fibrosis (in the absence of ongoing injury) might accelerate liver repair and regeneration. Infusion of bone marrow–derived macrophages in CCl4‐treated mice promoted fibrolysis and liver regeneration through up‐regulation of MMPs, cytokines, and growth factors including VEGF.7 Additionally, transplantation of endothelial progenitor cells reversed established fibrosis in rats, increased hepatic VEGF and endothelial nitric oxide synthase levels, reduced portal hypertension, augmented hepatic blood flow, increased vascular density in the fibrotic liver, and improved hepatic regeneration.8 In practice, there would inevitably be concerns about exacerbating portal hypertension or promoting the development of hepatocellular carcinoma in patients with advanced‐stage disease. In summary, Kantari‐Mimoun et al.1 have accentuated the proresolution effects of VEGF and SAMs in liver fibrosis but, more importantly, have unmasked a new and unanticipated link between angiogenesis within the scar and the resolution of fibrosis. Time will tell whether these new concepts can be disentangled and exploited in the clinic.