RelA at Ser536 to Promote Myofibroblast Survival and Liver Fibrosis

Angiotensin II Activates IκB Kinase Phosphorylation of RelA at Ser536 to Promote Myofibroblast Survival and Liver Fibrosis



Fiona Oakley, Victoria Teoh‡, Gemma Ching–A–Sue‡, Ramon Bataller§, Jordi Colmenero§, Julie R. Jonsson∥, Aristides G. Eliopoulos¶, Martha R. Watson, Derek Manas, Derek A. MannCorresponding Author Informationemail address

published online 19 March 2009.

Uncorrected Proof




"Fourteen patients with chronic hepatitis C virus (HCV)-induced fibrosis were included in an open-label study investigating antifibrogenic effects of losartan. Inclusion criteria were as follows: (1) age 35–65 years; (2) elevation of serum aminotransferases for >6 months, and positive RNA-HCV; (3) significant histologic fibrosis in liver biopsy (≥F2 in METAVIR score); and (4) no previous response and/or contraindications to antiviral therapy. Exclusion criteria were as follows: (1) existence of other known liver disease; (2) past history of hepatic decompensations or hepatocellular carcinoma; (3) alcohol consumption (>20 g/day); (4) arterial hypertension; (5) serum creatinine >1.5 mg/dL; (6) treatment with AT1 receptor blockers, angiotensin converting enzyme (ACE) inhibitors, or interferon in the preceding 12 months; or (7) contraindications to oral losartan. Patients were treated for 18 months with oral losartan 50 mg/day (Cozaar; MSD, Wilmington, DE). In all patients, 2 liver biopsies were performed: the first biopsy 24 hours prior to initiation of treatment and the second biopsy on the following day of the last dose of losartan. Fibrosis was blindly evaluated by the same pathologist using the METAVIR scoring system. The protocol was approved by the Ethics Committee of the Hospital Clínic of Barcelona and the Agencia Española del Medicamento (ARAHEPC 02-0491) and registered into the protocol registration system (NCT00298714). All patients gave written informed consent. Paired biopsy samples from 11 out of the original 14 patients were available for the current study."



Treatment of human myofibroblasts with ACE inhibitor (enalopril) or ZD7155 provoked apoptosis (Figure 5A and 5B). ZD7155 also suppressed expression of the NF-κB-regulated antiapoptotic factors BCl-2 and Gadd45β (Figure 5C). Apoptosis induced by ZD7155 was prevented by pretreatment with JNK inhibitor SP600125 (Figure 5D), which we previously reported to block apoptosis induced by NF-κB inhibitors.9, 13 Angiotensin II therefore promotes survival of myofibroblasts at least in part by stimulating NF-κB-dependent expression of the naturally occurring JNK inhibitor Gadd45β.



Constitutive P-Ser536-RelA Is a Biomarker for Liver Fibrosis That Is Susceptible to Regression in Response to Blockade of Angiotensin II and IKK Signaling



We next determined whether Ser536 phosphorylation is suppressed upon in vivo treatment with ACE and IKK inhibitors. The ACE inhibitor captopril promotes regression of BDL-induced fibrosis despite progressive injury,18 and the IKK/NF-κB inhibitor sulphasalazine augments spontaneous regression of CCl4-induced fibrosis.9 Both treatments are associated with loss of P-Ser536-RelA in cells associated with fibrotic tissue (Figure 7). Of note, both treatments were originally reported to result in roughly 50% reduced numbers of α-SMA+ cells relative to controls9, 18; however, this would not account for the 90% loss of P-Ser536-RelA observed here and may indicate that diminution of P-Ser536-RelA precedes myofibroblast apoptosis. To translate these discoveries to human disease, we examined P-Ser536-RelA in biopsy slides from the fibrotic livers of patients with chronic hepatitis C prior to and after 18 months of treatment with the AT1 antagonist losartan. As described elsewhere,19 this treatment promoted regression of fibrosis in half of the patients (Supplementary Table 1). Staining biopsy slides for P-Ser536-RelA indicated that patients who underwent regression in response to losartan began (prior to treatment) with more P-Ser536-RelA positive cells than found in nonresponders (Figure 8). Following treatment, numbers of P-Ser536-RelA-positive cells diminished in the losartan-responder group to levels around those measured in nonresponders, which were unchanged between pre- and posttreatment biopsies. Loss of P-Ser536-RelA-positive cells is therefore associated with regression of human liver fibrosis in response to treatment with losartan. Moreover, the level of constitutive P-Ser536-RelA may be predictive for patients who are likely to benefit from losartan treatment.




In Vivo Phosphorylation at Ser536-RelA in Myofibroblasts Is Angiotensin Dependent and Profibrogenic



Myofibroblasts isolated from livers at 24 hours post-CCl4 were positive for P-Ser536-RelA (Figure 6A, left photomicrograph). In vivo induction of P-Ser536-RelA in myofibroblasts in response to acute CCl4-induced liver injury was attenuated by 4-hour treatment with ZD7155 (Figure 6A, right photomicrograph). A criticism of this experiment is that the postisolation culture of myofibroblasts may activate angiotensin II-independent pathways that influence P-Ser536-RelA. However, despite this possibility, our data clearly demonstrate that short-term in vivo treatment with an angiotensin II inhibitor reduces the number of P-Ser536-RelA-positive myofibroblasts isolated from the injured liver and as such makes a strong case for angiotensin II as a major contributor to myofibroblast NF-κB activity and survival in the injured liver. Administration of cell-permeable peptides competing with P-Ser536-RelA (P6) blocked the induction of the fibrogenic genes procollagen I, TIMP1, and α-SMA in response to acute liver injury. Direct targeting of P-Ser536-RelA is therefore a potential antifibrotic strategy. Of note, the control peptide lacking Ser536 promoted an increase in expression of TIMP1 and α-SMA (Figure 6B), an effect currently under investigation.



Discussion



We have described a constitutive positive feedback fibrogenic signaling pathway (angiotensin II/IKKβ/P-Ser536-RelA/active nuclear NF-κB/angiotensinogen gene transcription) that promotes survival of human hepatic myofibroblasts. This pathway explains our previous finding that the IKK inhibitor sulphasalazine can repress myofibroblast NF-κB despite low-level expression of IκBα in these cells.9 Although a significant proportion of RelA:p50 dimers in hepatic myofibroblasts are not subject to regulation by IκBα,12, 15, 16 they are clearly still subject to additional regulatory checkpoints such as subunit phosphorylation. P-Ser536-RelA is required for efficient nuclear transport of RelA-containing NF-κB,22 which in hepatic myofibroblasts is regulated by constitutive IKKβ (a target for sulphasalazine), and can be suppressed by ACE inhibitors and AT1 antagonists.



Angiotensin II has emerged as an important stimulator of fibrogenesis. Patients with chronic HCV infection carrying a genetic polymorphism associated with increased angiotensin II synthesis develop more severe fibrosis.30 Inhibition of angiotensin II synthesis, blockade of AT1 receptor, or genetic ablation of AT1 receptor attenuates fibrosis in animal models of chronic liver disease.26, 27 Direct fibrogenic effects of angiotensin II have been described, including stimulation of proliferation and the expression of fibrogenic genes encoding extracellular matrix proteins, cytokines, growth factors, and fibrogenic receptors.26 One mechanism by which angiotensin II mediates these effects is via the phosphorylation of the p47phox component of NADPH oxidase; induction of reactive oxygen species; and subsequent activation of AKT, MAPKs, and AP-1 DNA binding.31 Here, we describe a second angiotensin II-activated fibrogenic pathway involving activation of IKK, phosphorylation of RelA, and induction of NF-κB-dependent transcription of cell survival genes. Taken together, these observations lend support for angiotensin II functioning as a vital paracrine and autocrine regulator of the survival and fibrogenic actions of hepatic myofibroblasts.



The therapeutic potential for targeting this pathway was demonstrated by a correlation between the loss of scar-associated P-Ser536-RelA-positive cells and regression of fibrosis in diseased rodent livers treated with ACE and IKK inhibitors. Moreover, our observation that fibrotic HCV patients who undergo regression of fibrosis in response to treatment with losartan initially express elevated numbers of P-Ser536-RelA (relative to nonresponders) that diminish following treatment suggests that P-Ser536-RelA may be a useful biomarker for fibrosis that is susceptible to treatment with pharmacologic attenuators of the renin-angiotensin system. Hepatic myofibroblasts are a phenotypically diverse and highly plastic population of cells.2 There are several distinct mechanisms by which myofibroblast can avoid apoptosis.7 It can be speculated that myofibroblasts may modulate their phenotype to predominantly employ a particular survival mechanism as dictated by apoptotic signals present in the microenvironment. P-Ser536-RelA may be a surrogate for a predominantly angiotensin II-driven survival mechanism and fibrosis that will regress in response to drugs that are well tolerated by humans.





ABSTRACT



Background & Aims



The transcription factor nuclear factor (NF)-κB promotes survival of hepatic myofibroblasts and fibrogenesis through poorly defined mechanisms. We investigated the activities of angiotensin II and IκB kinase (IKK) in regulation of NF-κB activity and the role of these processes in liver fibrosis in rodents and humans.



Methods



Phosphorylation of the NF-κB subunit RelA at serine 536 (P-Ser536-RelA) was detected by immunoblot and immunohistochemical analyses. P-Ser536-RelA function was assessed using vectors that expressed mutant forms of RelA, cell-permeable blocking peptides, and assays for RelA nuclear transport and apoptosis. Levels of P-Ser536-RelA were compared with degree of fibrosis in liver sections from chronically injured rats and patients with hepatitis C virus-mediated fibrosis who had been treated with the AT1 antagonist losartan.



Results



Constitutive P-Ser536-RelA is a feature of human hepatic myofibroblasts, both in vitro and in situ in diseased livers. Autocrine angiotensin II stimulated IKK-mediated phosphorylation of RelA at Ser536, which was required for nuclear transport and transcriptional activity of NF-κB. Inhibition of angiotensin II, the angiotensin II receptor type 1 (AT1), or IKK blocked Ser536 phosphorylation and stimulated myofibroblast apoptosis. Treatment of fibrotic rodent liver with the angiotensin converting enzyme (ACE) inhibitor captopril or the IKK inhibitor sulphasalazine resulted in loss of P-Ser536-RelA-positive myofibroblasts and fibrosis regression. In human liver samples, increased numbers of P-Ser536-RelA-positive cells were associated with fibrosis that regressed following exposure to losartan.



Conclusions



An autocrine pathway that includes angiotensin II, IKK, and P-Ser536-RelA regulates myofibroblast survival and can be targeted to stimulate therapeutic regression of liver fibrosis.



Abbreviations used in this paper: ACE, angiotensin converting enzyme, α-SMA, smooth muscle α-actin, AT1, angiotensin type 1 receptor, IκB, inhibitor of NF-κB, IKK, IκB kinase, NF-κB, nuclear factor-κB, P-Ser536-RelA, RelA phosphorylated on serine 536



Liver fibrosis is a dynamic process that can either progress to end-stage cirrhosis or regress and enable regeneration of normal functional hepatic tissue.1, 2 Clinical studies show that effective treatment for the underlying cause of liver injury can provoke regression of fibrosis, although reversion of cirrhosis remains unlikely.1, 2 Unfortunately, long-term effective control of the causes of chronic liver disease remains a significant challenge and, for the majority of patients, is currently not possible. Development of therapeutics that stimulate regression of preestablished fibrosis would have substantial clinical impact.



Transient accumulation of activated smooth muscle α-actin (α-SMA)-positive myofibroblasts is a pivotal event in tissue injury that enables formation of a temporary scar composed of fibril-forming collagens secreted by the myofibroblast.3 In addition, the contractile properties of the myofibroblast promote wound closure, which protects the injured tissue from infection and further damage.4 Resolution of wound healing and remodelling of scar tissue is associated with diminution of the myofibroblast population.5, 6 By contrast, progressive fibrosis is associated with a steadily increasing myofibroblast population because of proliferation and prolonged survival.3, 7 Experimental rodent-based models of liver disease have revealed an association between myofibroblast apoptosis and regression of fibrosis.5, 6, 7 Agents such as gliotoxin, sulphasalazine, and thalidomide promote apoptosis of liver myofibroblasts and augment the rate at which fibrosis spontaneously regresses.8, 9, 10 Selective targeting of gliotoxin to myofibroblasts provokes regression of fibrosis even under conditions of sustained liver injury.11 Myofibroblast apoptosis is therefore not simply a bystander event in fibrosis regression; rather, it is a naturally occurring physiologic process that enables dynamic remodelling of extracellular matrix and can be stimulated to promote therapeutic regression of fibrosis. The next step toward the translation of these discoveries to the clinic is the identification of fibrogenic-signaling pathways that can be effectively and safely targeted in humans.



Human and rodent liver myofibroblasts express constitutive active nuclear factor (NF)-κB (RelA:p50), which promotes survival by stimulating the expression of antiapoptotic proteins Gadd45β and BCl-2.12, 13 Specific inhibition of NF-κB is sufficient to provoke apoptosis of fully mature human myofibroblasts that are relatively resistant to apoptosis.13, 14 These discoveries raise the important questions of how constitutive NF-κB is regulated in myofibroblasts and whether the regulatory checkpoints that control NF-κB activity can provide clinically safe therapeutic targets? A physiologically relevant model for the generation of liver myofibroblasts is the culture-induced transdifferentiation of quiescent hepatic stellate cells (HSC) into activated myofibroblasts.6 This model recapitulates the molecular and cellular events involved in sinusoidal fibrosis that is common to alcoholic and fatty liver diseases. Using this model, we have shown that HSC transdifferentiation is accompanied by a sustained transcriptional repression of inhibitor of NF-κB (IκB) α, the natural occurring inhibitor of NF-κB.15, 16 IκBα is abundantly expressed in the majority of mammalian cells and forms a direct interaction with NF-κB that masks its nuclear localization and DNA-binding motifs, resulting in an inactive cytoplasmic location for the transcription factor. Upon stimulation of cells by a variety of agents, including lipopolysaccharides and tumor necrosis factor-α, 2 serine residues (Ser32 and Ser36) on IκBα are phosphorylated by the I κ kinase (IKK) β component of the IKK complex.17 This results in polyubiquitination and proteasome-mediated degradation of IκBα which releases NF-κB for nuclear transport and interaction with target genes. Depletion of IκBα in hepatic myofibroblasts effectively bypasses this regulatory checkpoint. However, the IKK inhibitor sulphasalazine inhibits myofibroblast NF-κB activity and stimulates apoptosis.9 This suggests existence of additional regulatory IKK-dependent checkpoints that control myofibroblast NF-κB.



In this study, we discover that the NF-κB RelA subunit is constitutively phosphorylated at residue serine 536 (P-Ser536-RelA) both in cultured HSC and myofibroblasts of diseased human liver. We determine that myofibroblasts express constitutive IKKβ activity under the autocrine control of the renin-angiotensin system and is required for maintenance of P-Ser536-RelA, nuclear transport of NF-κB, and myofibroblast survival. The clinical relevance of this autocrine angiotensin II/IKKβ/NF-κB pathway is demonstrated with animal models of liver fibrosis and human liver biopsy samples that collectively provide strong support that it can be targeted to stimulate fibrosis regression with drugs that are safe for use in humans.