Research ArticleFibrosis

iRhom2 inhibits bile duct obstruction–induced liver fibrosis

See allHide authors and affiliations

Science Signaling  29 Oct 2019:
Vol. 12, Issue 605, eaax1194
DOI: 10.1126/scisignal.aax1194

iRhom2 protects against liver fibrosis

Injury or chronic inflammation in the liver leads to the activation of hepatic stellate cells, which transdifferentiate into matrix-secreting myofibroblasts that promote fibrosis. Sundaram et al. found that mice lacking the rhomboid family pseudoprotease iRhom2, which is required for the proper trafficking and activation of the metalloprotease ADAM17 (see the Focus by Badenes and Adrain), showed increased stellate cell activation and susceptibility to liver fibrosis induced by bile duct ligation (BDL). iRhom2-dependent activation of ADAM17 promoted shedding of tumor necrosis factor receptors (TNFRs) from hepatic stellate cells. Treating iRhom2-deficient mice with the TNF-α inhibitor etanercept reduced BDL-induced stellate cell activation and liver fibrosis. Data from patients with liver cirrhosis were consistent with these observations, suggesting a protective role for iRhom2 in human liver disease.

Abstract

Chronic liver disease can induce prolonged activation of hepatic stellate cells, which may result in liver fibrosis. Inactive rhomboid protein 2 (iRhom2) is required for the maturation of A disintegrin and metalloprotease 17 (ADAM17, also called TACE), which is responsible for the cleavage of membrane-bound tumor necrosis factor–α (TNF-α) and its receptors (TNFRs). Here, using the murine bile duct ligation (BDL) model, we showed that the abundance of iRhom2 and activation of ADAM17 increased during liver fibrosis. Consistent with this, concentrations of ADAM17 substrates were increased in plasma samples from mice after BDL and in patients suffering from liver cirrhosis. We observed increased liver fibrosis, accelerated disease progression, and an increase in activated stellate cells after BDL in mice lacking iRhom2 (Rhbdf2−/−) compared to that in controls. In vitro primary mouse hepatic stellate cells exhibited iRhom2-dependent shedding of the ADAM17 substrates TNFR1 and TNFR2. In vivo TNFR shedding after BDL also depended on iRhom2. Treatment of Rhbdf2−/− mice with the TNF-α inhibitor etanercept reduced the presence of activated stellate cells and alleviated liver fibrosis after BDL. Together, these data suggest that iRhom2-mediated inhibition of TNFR signaling protects against liver fibrosis.

INTRODUCTION

Chronic liver disease resulting in hepatic inflammation can lead to liver fibrosis, cirrhosis, and hepatocellular carcinoma (1). Its underlying causes include alcoholic liver disease, nonalcoholic liver disease, viral-induced hepatitis, and autoimmune diseases targeting the liver or bile duct (1). Together, chronic liver disease accounted for more than 1 million deaths worldwide per year in 2010 (2, 3). Fibrotic changes in liver disease are mediated by increased production of collagens and the transdifferentiation of hepatic stellate cells (HSCs) into myofibroblasts (46). HSCs are progenitor cells that are normally quiescent but proliferate in response to liver injury and contribute to liver regeneration (7). Under normal conditions, quiescent HSCs reside in the space of Disse, store vitamin A, and produce a specific marker called glial fibrillary acid protein (GFAP). After stimulation with CD95L or other death receptor ligands, stellate cells have the paradoxical tendency to proliferate (8). In response to liver injury, HSCs reduce their vitamin A storage, migrate to pericentral areas, and transdifferentiate into collagen type I– and α-smooth muscle actin (α-SMA)–producing myofibroblasts (9, 10). Immune cells, such as Kupffer cells and monocytes, are recruited and/or activated after injury, resulting in increased proinflammatory cytokines, which can contribute to liver fibrosis (11, 12). Another factor known to promote liver fibrosis is transforming growth factor–β (TGF-β), which is secreted by activated HSCs and has been shown to induce them to produce collagen, a key characteristic of myofibroblasts (13, 14). Furthermore, nuclear factor κB (NF-κB) activation in HSCs is critical for their survival and for the establishment of liver fibrosis (15). Tumor necrosis factor–α (TNF-α) and interleukin-1β (IL-1β) can induce NF-κB signaling in HSCs, promoting their survival and differentiation into myofibroblasts (11). Consistent with this, TNF receptor 1 (TNFR1)–deficient mice exhibit decreased liver fibrosis in response to bile duct ligation (BDL) (16). Thus, HSCs and the factors regulating their proliferation, survival, and transformation into myofibroblasts contribute to liver fibrosis.

TNF-α is a proinflammatory cytokine that can induce liver cell death during liver disease (17). TNF-α is synthesized as a membrane-bound protein that is proteolytically cleaved by the metalloprotease ADAM17 (A disintegrin and metalloprotease 17), also known as TACE (TNF–α converting enzyme) (18, 19). The inactive member of the rhomboid intramembrane protease family, iRhom2, which is encoded by the gene Rhbdf2, interacts with ADAM17 (20, 21). Both ADAM17 and iRhom2 are localized to the endoplasmic reticulum, where iRhom2 interacts with ADAM17 and contributes to its translocation to the plasma membrane (20). Lack of iRhom2 results in reduced shedding of ADAM17 substrates including TNF-α, L-selectin, heparin-binding epidermal growth factor (HB-EGF), TNFR1, TNFR2, amphiregulin, epiregulin, EphB4, KitL2, and Tie2 (2123). Because Rhbdf2−/− mice exhibit considerably reduced amounts of soluble TNF-α, the animals are highly susceptible to Listeria monocytogenes infection but are protected against lipopolysaccharide (LPS)–induced septic shock (21) and rheumatoid arthritis (24). iRhom2-deficient mice exhibit alleviated autoimmune-induced nephritis, an observation partly attributed to reduced TNF-α and epidermal growth factor receptor (EGFR) signaling (25). Gain-of-function mutations that affect the cytoplasmic tail of iRhom2 in humans can trigger Tylosis syndrome, which is characterized by hyperproliferation of keratinocytes, causing palmoplantar hyperkeratosis and esophageal cancer (26, 27). Deletion of the cytoplasmic tail results in increased ADAM17 activity, increased TNFR shedding, and concomitant reductions in apoptosis after exposure to TNF-α (22). Posttranslationally, phosphorylation of the cytoplasmic tail of iRhom2 results in dissociation of ADAM17, followed by an increase in ADAM17 catalytic activity (28, 29). Rhbdf2 has also been shown to be transcribed by the transcription factor p63 during cellular stress (30). Together, these findings establish that iRhom2 is a critical regulator of several aspects of ADAM17 biology, including its maturation and catalytic activity. The resulting effect of iRhom2 on shedding of ADAM17 ligands—specifically, TNF-α versus TNFR shedding—can have opposing consequences. Which is dominant would depend on the specific cellular and tissue context.

In the context of liver fibrosis, reduced TNF-α shedding from macrophages should alleviate fibrosis, whereas reduced TNFR shedding from liver cells would have opposite effects. In this study, we therefore chose whole-body Rhbdf2−/− mice to determine which plays a dominant role in liver fibrosis. Because Rhbdf2−/− mice have less soluble TNF-α than wild-type mice and given that iRhom2 can affect ADAM17’s ability to cleave membrane-bound TNF-α as well as its receptors TNFR1 and TNFR2 from the cell surface, we wondered whether iRhom2-deficient mice would have less fibrosis after BDL. Unexpectedly, we found that the absence of iRhom2 resulted in an increased presence of activated HSCs along with increased expression of liver fibrotic markers. Accordingly, treatment of Rhbdf2−/− mice with the TNF-α inhibitor etanercept alleviated liver fibrosis. We therefore propose iRhom2 to be a critical regulator during BDL-induced liver fibrosis.

RESULTS

Lack of iRhom2 results in increased liver fibrosis after BDL

Liver fibrosis can be induced in mice after BDL, which results in increased production of α-SMA and expression of Col1a1, which encodes the α1 chain of collagen type 1, in the liver (fig. S1, A and B). During BDL-induced liver fibrosis, we observed increases in soluble TNFRs and IL-6 receptor (IL-6R), both of which are cleaved by ADAM17 (Fig. 1, A to C). Other ADAM17 substrates we tested were not detectable or increased in the sera of mice suffering from liver fibrosis (fig. S2, A to C). As expected, we identified an increase in mature ADAM17 during liver fibrosis (Fig. 1, D and E). ADAM17 activation can be triggered by iRhom2 (20, 21), transcripts for which were expressed in hepatocytes, Kupffer cells, and stellate cells isolated from untreated mice (fig. S2D). Hepatic Rhbdf2 expression increased during liver fibrosis after BDL in mice (Fig. 1F), leading us to further investigate whether iRhom2 influenced the progression of liver fibrosis. Moreover, we determined TNFR concentrations in plasma samples from various patients with or without liver cirrhosis and healthy volunteers (table S1, cohort A). The cirrhotic patient group consisted of patients with liver cirrhosis who were infected with hepatitis B virus (HBV), HCV, or the HBV isolate HC-C2. The noncirrhotic group included healthy volunteers and noncirrhotic patients who were infected with HBV or who were infected with HBV or HCV and also suffered from nonalcoholic steatohepatitis. We detected increased serum concentrations of TNFRs in cirrhotic patients as compared to noncirrhotic patients and healthy volunteers (fig. S3A). Furthermore, RHBDF2 mRNA was increased in human cirrhotic liver tissue compared to noncirrhotic liver tissue in a different cohort (fig. S3B and table S1, cohort B).

Fig. 1 ADAM17 activity increases during liver fibrosis.

(A to C) Quantification of soluble (s) TNFR1 (A), TNFR2 (B), and IL-6R (C) in serum samples from C57BL/6 mice subjected to sham or BDL operation. Sera were taken at 14 or 20 days after surgery. (D to E) Immunoblotting and quantification of ADAM17 in total liver tissue of sham- and BDL-operated mice at day 14 (D14) and day 20 (D20) after surgery. Pro and mature (mat) forms of ADAM17 are noted on the immunoblot (D). The mature (M) form of ADAM17 was quantified by densitometry (E). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is a loading control. (F) Quantification of Rhbdf2 mRNA in total liver tissue of sham- and BDL-operated mice at day 14 or day 20 after surgery. ns, not significant. For all experiments, n = 5 mice per condition. Data are shown as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA with Bonferroni’s multiple comparisons test (A, B, C, E, and F).

As expected, ADAM17 maturation was decreased in iRhom2-deficient (Rhbdf2−/−) mice during liver fibrosis after BDL compared to control mice subjected to BDL (Fig. 2A and fig. S4A), which is consistent with the reduced presence of circulating TNFRs in the sera of Rhbdf2−/− mice compared to control animals after BDL (Fig. 2B and fig. S4, B to E). The abundance of soluble IL-6R, which is released by ADAM17-mediated processing of IL-6R, was also decreased in Rhbdf2−/− mice compared to control animals after BDL, but EGFR ligands targeted by ADAM17 were not (fig. S4, F to I). The abundance of mRNAs encoding ADAM17 substrates were not significantly different between Rhbdf2−/− and control mice 20 days after BDL (fig. S5). Liver tissue harvested from Rhbdf2−/− mice after BDL, however, exhibited significantly increased areas of fibrosis, indicated by increased Picro Sirius Red and Masson’s trichrome staining compared to liver tissue harvested from control animals (Fig. 2C). Moreover, Rhbdf2−/− BDL mice showed increased abundance of hepatic fibrosis markers such as collagen and α-SMA when compared to wild-type BDL controls (Fig. 2D). Together, these data indicate that the activity of ADAM17 increases during liver fibrosis and that the absence of iRhom2, despite reducing ADAM17 maturation, exacerbates liver fibrosis compared to control animals.

Fig. 2 iRhom2 deficiency enhances liver fibrosis after BDL.

(A) Representative immunoblot showing ADAM17 in total liver tissue from Rhbdf2+/− and Rhbdf2−/− mice at day 0 and day 20 after BDL. β-Actin is a loading control. n = 6 animals per genotype for each time point. (B) Quantification of soluble TNFR1 and TNFR2 in sera collected from Rhbdf2+/− and Rhbdf2−/− mice at day 0 and day 20 after BDL. n = 5 to 6 animals per genotype for each time point. (C) Imaging and quantification of Picro Sirius Red and Masson’s trichrome staining in sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 14 after BDL. n = 4 to 6 mice per genotype. Scale bars, 50 μm. (D) Distribution and quantification of Col1A1 and α-SMA in sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 20 after BDL. Nuclei are labeled with DAPI (blue). Scale bars, 100 μm. Quantification of Col1A1 and α-SMA by mean fluorescence intensity (MFI) is shown in arbitrary units (AU). n = 6 mice per genotype. Data are shown as means ± SEM. *P < 0.05, **P < 0.01, two-way ANOVA with Bonferroni’s multiple comparisons test (B) and Mann-Whitney U test (C and D).

Rhbdf2 expression is induced early in the liver after BDL

To determine whether ADAM17 activation occurred after the initiation of liver fibrosis or was an early event before the occurrence of fibrosis, we assayed earlier time points after BDL. We observed increased expression of Rhbdf2 as early as 6 hours after the BDL operation (Fig. 3A). Increased Rhbdf2 expression was observed in liver tissue but not in other organs tested (fig. S6A). Because we observed increased Rhbdf2 expression, we speculated that ADAM17 activation should also be increased at early time points after the BDL operation. We detected increased expression of mRNAs encoding ADAM17 substrates within 48 hours after BDL in liver tissue compared to sham-operated animals (Fig. 3B) and increased abundance of ADAM17 substrates in the circulation, including soluble forms of TNFRs, IL-6R, and HB-EGF (Fig. 3, C to F). Other ADAM17 substrates we tested were not detected in the sera of BDL or sham-operated mice (fig. S6B). Activation of ADAM17 depended on the presence of iRhom2 because Rhbdf2−/− animals showed reduced abundance of mature ADAM17 (Fig. 3G and fig. S6C).

Fig. 3 Rhbdf2 expression and shedding of ADAM17 substrates increase early after BDL.

(A) Quantification of Rhbdf2 mRNA in total liver tissue of sham and BDL wild-type mice at the indicated early time points after surgery. n = 4 animals per condition for each time point. (B) Quantification of Tnf, Tnfrsf1a, Tnfrsf1b, Hbegf, Areg, and Tgfa mRNAs in liver tissue of sham and BDL mice at the indicated time points after surgery was analyzed by reverse transcription polymerase chain reaction (RT-PCR). The highest value for each transcript, relative to the amount of GAPDH, compared to their expression in sham mice is noted in the heatmap. n = 4 animals per condition for each time point. (C to F) Quantification of soluble TNFR1 (C), TNFR2 (D), IL-6R (E), and HB-EGF (F) in serum from sham and BDL mice at the indicated time points after surgery. n = 3 to 6 animals per condition for each time point. (G) Representative immunoblot showing ADAM17 in total liver tissue of Rhbdf2+/− and Rhbdf2−/− mice at early time points after BDL. β-Actin is a loading control. n = 6 to 9 animals per genotype for each time point. Data are shown as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA with Bonferroni’s multiple comparisons test (A, C, D, E and F).

Lack of iRhom2 triggers increased presence of activated HSCs after BDL

BDL in mice is accompanied by liver damage, resulting in increased amounts of liver enzymes in the blood. However, when we analyzed hepatic tissue damage by measuring liver enzyme activity after BDL, we did not see significant differences during early or late time points between wild-type and iRhom2-deficient mice (Fig. 4, A and B, and fig. S7, A and B). Consistent with this, we did not find a significant difference in the abundance of active caspase-3 between Rhbdf2+/− and Rhbdf2−/− liver tissue (fig. S7C). In addition, when we treated Rhbdf2+/− and Rhbdf2−/− primary hepatocytes with TNF-α plus cycloheximide (CHX), we detected no significant difference in cleaved caspase-8 between the two groups (fig. S7D). Furthermore, although conjugated bile acids were reduced in the sera and liver tissue of Rhbdf2−/− mice compared to controls at day 4 after BDL, we did not detect any difference at later stages (fig. S8, A and B). Together, this suggests that the phenotypes seen in Rhbdf2−/− mice at later time points are not due to defects in hepatocytes but another cell population within the liver. We found an increase in areas densely populated with cells, consistent in appearance with fibrotic lesions (31, 32), in Rhbdf2−/− hematoxylin and eosin–stained liver tissue sections compared to Rhbdf2+/− tissue (Fig. 4C). Because stellate cells contribute to liver fibrosis by differentiating into myofibroblasts and producing collagens (33), we speculated that absence of iRhom2 could induce the proliferation of myofibroblasts derived from HSCs. Activated stellate cells and myofibroblasts can be visualized by staining for platelet-derived growth factor receptor β (PDGFRβ) and the intermediate filament protein Desmin (3436). We found significant increases in PDGFRβ- and Desmin-producing cells in liver tissue harvested from Rhbdf2−/− mice compared to control animals after BDL (Fig. 4, D and E). We also identified decreased abundance of GFAP, an HSC marker that is absent in myofibroblasts (9), in Rhbdf2−/− liver tissue when compared to control liver tissue (fig. S8C). Together, these data indicate that early production of iRhom2 after BDL facilitates ADAM17 activation and reduces the presence of activated HSCs.

Fig. 4 iRhom2 protects against liver fibrosis early after BDL.

(A to B) Quantification of liver enzymes AST (A) and ALT (B) in serum samples from Rhbdf2+/− and Rhbdf2−/− mice at the indicated time points after BDL. n = 6 to 9 animals per genotype for each time. (C) Sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 0 or day 4 after BDL were stained with hematoxylin and eosin. Gray dashed lines indicate boundaries of fibrotic lesions. The areas of fibrotic lesions were measured and expressed as square micrometers. n = 6 to 9 animals per genotype for each time point. Scale bar, 50 μm. (D to E) Sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 4 after BDL were stained for PDGFRβ (D) or Desmin (E). Dashed boxes indicate the areas magnified in the images on the right. PDGFRβ- or Desmin-positive cells were quantified as a percentage of all cells. n = 7 to 8 animals per genotype. Scale bars, 100 μm (left) and 50 μm (right) for (D) and 50 μm (left) and 20 μm (right) for (E). Data are shown as means ± SEM. *P < 0.05, ****P < 0.0001, Mann-Whitney U test (C to E).

Lack of iRhom2 has no major effect on the inflammatory milieu in liver

Liver tissue from Rhbdf2−/− mice had slightly more cells positive for the infiltrating macrophage marker CD68 compared to controls at later stages after BDL (fig. S9A). However, we did not find significant differences in F4/80, which marks Kupffer cells, and LY6G, a marker for monocytes, neutrophils, and granulocytes, between Rhbdf2+/− and Rhbdf2−/− liver tissue after BDL (fig. S9A). These data were supported by flow cytometry measurements at an early fibrotic stage (fig. S9B). In agreement with this, we did not observe a significant difference in the expression of genes encoding cytokines or chemokines between liver tissue from iRhom2-deficient and control mice after BDL surgery (fig. S10, A to C). Furthermore, we did not detect significant differences between concentrations of IL-6, IL-1β, or TGF-β in serum samples of control and Rhbdf2−/− mice (fig. S10D). Previous studies have shown that BDL can result in bacterial translocation (37) and that iRhom2-deficient animals are highly susceptible toward L. monocytogenes infection (21). However, we did not observe any significant differences in bacterial titers between wild-type and iRhom2-deficient animals on day 3 after BDL (fig. S11A). Moreover, previous reports have shown an induction in type I interferon (IFN) signaling in BDL-operated animals (38) and demonstrated that iRhom2 is essential for stimulator of IFN genes (STING) activity to induce type I IFN signaling after infection with a DNA virus (39). However, during early time points after BDL, we did not observe strong induction of IFN-1 or IFN-regulated genes in our experiments (fig. S11, B to E).

Lack of iRhom2 leads to decreased TNFR shedding and increased fibrotic markers in stellate cells

Mouse primary HSCs express fibrotic markers when cultured in vitro (fig. S12A) (40). We confirmed the presence of both the pro and mature forms of ADAM17 in primary HSCs isolated from C57BL/6 mice (fig. S12B). We observed increased shedding of TNFRs, but not other ADAM17 substrates, in primary HSC cultures (Fig. 5A and fig. S12, C and D). As expected, ADAM17 maturation was significantly reduced in stellate cells from Rhbdf2−/− mice compared to those from control animals (Fig. 5, B and C, and fig. S12E). Furthermore, we found increased abundance of mature ADAM17 when we exposed Rhbdf2+/− but not Rhbdf2−/− stellate cells to TNF-α (Fig. 5, B and C, and fig. S12E). Shedding of TNFR1 in Rhbdf2+/− stellate cells was reduced after exposure to TNF-α, whereas shedding of TNFR2 was increased (Fig. 5D). Moreover, shedding of both TNFR1 and TNFR2 depended on iRhom2 because Rhbdf2−/− stellate cells showed reduced TNFRs in the culture supernatant (Fig. 5D). Similar to the data from liver tissue (Fig. 2D), we found an increase in fibrotic markers in primary HSCs from Rhbdf2−/− animals compared to Rhbdf2+/− control cells (Fig. 5, E and F). Together, these data further demonstrate that shedding of TNFRs depends on iRhom2 and that the absence of iRhom2 can increase the expression of fibrotic markers in primary HSCs.

Fig. 5 Increased activation of HSCs in iRhom2-deficient mice.

(A) Quantification of soluble TNFR1 and TNFR2 in supernatants from primary HSCs from C57BL/6 mice at the indicated time points after isolation. n = 4 mice for each time point. (B) Representative immunoblot showing pro and mature forms of ADAM17 in primary HSCs isolated from Rhbdf2+/− and Rhbdf2−/− mice and left untreated (U) or treated (T) with TNF-α for 24 hours. α-Tubulin is a loading control. n = 4 mice per genotype for each condition. (C) Densitometric quantification of mature ADAM17 from (B). n = 4 mice per genotype for each condition. (D) Quantification of soluble TNFR1 and TNFR2 in supernatants from untreated and TNF-α–treated primary HSCs from Rhbdf2+/− and Rhbdf2−/− mice. n = 4 mice per genotype for each condition. OD450, optical density at 450 nm. (E) Quantification of Col1a1, Col3a1, and Acta2 transcripts in 3-day cultures of primary HSCs from Rhbdf2+/− and Rhbdf2−/− mice. n = 4 mice per genotype. (F) Representative immunoblot and quantification of α-SMA in 3-day cultures of primary HSCs from Rhbdf2+/− and Rhbdf2−/− mice. n = 4 mice per genotype. Data are shown as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, Mann-Whitney U test (A and F) and two-way ANOVA with Bonferroni’s multiple comparisons test (C to E).

Increased TNFR signaling in the absence of iRhom2 triggers stellate cell proliferation and liver fibrosis after BDL

ADAM17 is involved in the shedding of TNF-α and its receptors, EGFR ligands and IL-6R. Ligand and receptor shedding affects NF-κB activation, mitogen-activated kinases, extracellular signaling kinase 1/2 (ERK1/2), and signal transducer and activator of transcription 3 (STAT3) signaling pathways (22, 41, 42). Previous studies have demonstrated the importance of these signaling pathways during liver damage and liver regeneration. Mice lacking the p65 (also called RelA) subunit of the transcription factor NF-κB show an embryonic lethal phenotype and massive liver degeneration due to cell death (43). Mice lacking EGFR specifically in hepatocytes show decreased hepatocyte proliferation in the initial phase of liver regeneration (44). IL-6 trans-signaling can be activated by binding of IL-6 to soluble IL-6R, and earlier studies have described the importance of IL-6 trans-signaling for the protection of the liver during acute damage to the organ (45, 46).

We wanted to uncover which of these hepatic signaling pathways were affected by iRhom2. Although we found increased phosphorylation of ERK1/2 (p-ERK1/2) after BDL in both Rhbdf2+/− and Rhbdf2−/− liver tissue (fig. S13A), we did not observe a significant difference between Rhbdf2+/− and Rhbdf2−/− mice. This finding was expected in light of the lack of differences in EGFR ligand concentration between the sera of Rhbdf2+/− and Rhbdf2−/− mice. We detected a significant increase in phosphorylation of the transcription factor STAT3 in liver tissue from naive iRhom2-deficient mice compared to naive Rhbdf2+/− animals (fig. S13B). However, after BDL, there was no significant difference in phosphorylated STAT3 (p-STAT3) between Rhbdf2−/− and Rhbdf2+/− animals (fig. S13B). The transcription factor NF-κB is a heterodimeric protein that consists of two subunits, p50 and p65; the p65 domain contains the transcriptional activation domain (47). Phosphorylation of p65 (p-p65) was increased after BDL in liver tissue harvested from Rhbdf2−/− mice when compared to Rhbdf2+/− controls (fig. S13C), but we did not observe a difference in the abundance of inhibitor of NF-κBα (IκBα) (fig. S13D). However, we observed increased nuclear p65 in Rhbdf2−/− liver tissue compared to controls after BDL at day 20 (Fig. 6A). In addition, we also detected increased expression of the proliferation markers Ki67 and Cyclin-A2 in liver tissue harvested from Rhbdf2−/− mice compared to Rhbdf2+/− controls at early time points after BDL (Fig. 6B and fig. S13E), and more Ki67-positive cells were present in Rhbdf2−/− liver tissue compared to control liver tissue at day 20 after BDL (Fig. 6C). Furthermore, we observed that most of the Ki67-positive cells also stained for α-SMA in the absence of iRhom2, which suggests increased proliferation of HSCs in Rhbdf2−/− compared to Rhbdf2+/− (Fig. 6D).

Fig. 6 iRhom2 deficiency increases p65 phosphorylation and proliferation markers in the liver.

(A) Immunostaining and quantification of NF-κB p65 in sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 20 after BDL. Nuclei are labeled with DAPI (blue). Dashed boxes indicate the areas magnified in the images on the right. Quantification of p65 is shown as both the percentage of cells that were positive for p65 and the MFI of p65-positive cells. n = 6 mice per genotype. Scale bars, 50 μm (left) and 20 μm (right). (B) Quantification of MKI67 mRNA in total liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at the indicated time points after BDL surgery. n = 7 to 11 animals per genotype for each time point. (C) Immunostaining and quantification of the proliferation marker Ki67 in sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 20 after BDL. Ki67-positive cells were quantified as a percentage of all cells. n = 6 animals per genotype. Scale bars, 100 μm (left) and 50 μm (right). (D) Immunostaining for Ki67 and α-SMA in sections of liver tissue harvested from Rhbdf2+/− and Rhbdf2−/− mice at day 20 after BDL. α-SMA–positive cells that were also Ki67 positive were quantified as a percentage and MFI. n = 6 animals per genotype. Scale bars, 50 μm (left) and 20 μm (right). Data are shown as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, Mann-Whitney U test (A, C, and D) and two-way ANOVA with Bonferroni’s multiple comparisons test (B).

It has been previously shown that TNFR2 knockout mice have no major phenotype with regard to BDL-induced fibrosis (16). However, in our setting, soluble TNFR2 shed from ADAM17-positive cells could play a role in competitively inhibiting soluble TNF-α, which would further explain the differences in BDL-induced fibrosis between Rhbdf2+/− and Rhbdf2−/− mice. The unshed TNFRs on HSCs from Rhbdf2−/− mice as well as the increases in nuclear translocation of p65 translocation and proliferation markers in liver tissue of Rhbdf2−/− mice subjected to BDL led us to hypothesize that increased TNFR signaling in HSCs of iRhom2-deficient mice contributes to their increased proliferation and the subsequent progression of liver fibrosis. To test this hypothesis, we inhibited TNFR signaling by treating iRhom2-deficient mice with recombinant TNFR2-Fc (etanercept), which binds to TNF-α and prevents TNF-α–induced signaling (24), starting on the day before BDL surgery. We found reduced nuclear p65 translocation in etanercept-treated Rhbdf2−/− mice compared to phosphate-buffered saline (PBS)–treated mice (Fig. 7A) and reduced proliferation, as determined by Ki67 staining (Fig. 7B). Etanercept-treated mice showed a significant reduction in Ki67 staining in areas positive for α-SMA (Fig. 7C), suggesting that blocking TNF-α signaling in iRhom2-deficient liver inhibited HSC proliferation within the fibrotic areas. We detected no changes in macrophage infiltration (fig. S14A) or in expression of genes encoding inflammatory cytokines (fig. S14, B and C), with the exception of IL-10, which is induced during inflammation and fibrosis (48, 49) but was reduced with etanercept treatment (fig. S14B). Furthermore, we found increased abundance of GFAP in the livers of etanercept-treated Rhbdf2−/− BDL mice (Fig. 8A) compared to PBS-treated BDL mice, indicating the presence of quiescent HSCs rather than myofibroblasts. This is consistent with the observed reduction in liver fibrotic lesions of etanercept-treated Rhbdf2−/− BDL mice (Fig. 8B). PDGFRβ-positive cells were also reduced in etanercept-treated Rhbdf2−/− BDL mice compared to PBS-treated controls (Fig. 8B). These phenotypes were accompanied by reductions in α-SMA and collagen after etanercept treatment compared to PBS treatment (Fig. 8C). Together, these findings demonstrate that the absence of iRhom2 results in increased liver fibrosis, which can be alleviated through anti–TNF-α therapy.

Fig. 7 Etanercept treatment rescues inflammation and cell proliferation in the livers of Rhbdf2−/− mice.

(A) Immunostaining and quantification of NF-κB p65 in sections of liver harvested from Rhbdf2−/− PBS- or etanercept-treated mice 14 days after BDL. Nuclei are labeled with DAPI (blue). Dashed boxes indicate the areas magnified in the images on the right. Quantification of cells with nuclear p65 is shown as both the percentage of positive cells and the MFI. Scale bars, 50 μm (left) and 10 μm (right). (B) Immunostaining and quantification of Ki67 in sections of liver tissue harvested from Rhbdf2−/− PBS- or etanercept-treated mice 14 days after BDL. Ki67-positive cells were quantified as a percentage of all cells. Scale bars, 100 μm (left) and 50 μm (right). (C) Immunostaining and quantification of Ki67 and α-SMA in sections of liver tissue harvested from Rhbdf2−/− PBS- or etanercept-treated mice 14 days after BDL. α-SMA–positive cells that were also Ki67 positive were quantified as a percentage (%) and MFI. Scale bars, 50 μm (left) and 20 μm (right). For all panels, n = 7 mice for each condition. Data are shown as means ± SEM. *P < 0.05, ****P < 0.0001, Mann-Whitney U test.

Fig. 8 Etanercept treatment rescues liver fibrosis in of Rhbdf2−/− mice.

(A) Immunostaining and quantification of GFAP, which marks HSCs that have not transdifferentiated into myofibroblasts, in sections of liver harvested from Rhbdf2−/− PBS- or etanercept-treated mice 14 days after BDL. Nuclei are stained with DAPI (blue). Dashed boxes indicate the areas magnified in the images on the right. GFAP was quantified as MFI. Scale bars, 50 μm (left) and 20 μm (right). (B) Sections of liver tissue harvested from Rhbdf2−/− PBS- or etanercept-treated mice 14 days after BDL were stained with hematoxylin and eosin (H&E) or with antibodies recognizing PDGFRβ. Gray dashed lines indicate the boundaries of fibrotic lesions boundaries. The sizes of fibrotic lesions were quantified in square micrometers, and PDGFRβ-positive cells were quantified as a percentage of all cells in each section. Scale bars, 50 μm. (C) Immunostaining and quantification of α-SMA and Col1A1 in sections of liver tissue harvested from Rhbdf2−/− PBS- or etanercept-treated mice 14 days after BDL. Cells positive for α-SMA or Col1A1 were quantified as MFI. Scale bars, 100 μm. For all panels, n = 7 mice for each condition. Data are shown as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, using Mann-Whitney U test.

DISCUSSION

In this study, we found that iRhom2 protected against liver fibrosis after BDL by promoting ADAM17 maturation, leading to increased shedding of TNFRs and therefore less TNF-α signaling, in HSCs. The absence of iRhom2 caused an increase in the number of activated stellate cells along with increased expression of fibrotic markers after BDL, both of which were alleviated by anti–TNF-α therapy.

The release of TNF-α into the circulation is triggered by ADAM17-mediated cleavage of membrane-bound TNF-α from the cell surface (18, 19). Because iRhom2 stimulates ADAM17 maturation, it consequently promotes shedding of soluble TNF-α (20, 21). Thus, iRhom2-deficient mice are protected against LPS-induced septic shock but susceptible to bacterial infection (21). From these data and considering that iRhom2 is produced in Kupffer cells, it might be logical to speculate that the absence of iRhom2 would alleviate liver fibrosis induced by BDL. However, our data indicate that the absence of iRhom2 aggravated liver fibrosis. ADAM17 cleaves not only membrane-bound TNF-α (18, 19) but also TNFR1 and TNFR2 (50). Compared to iRhom2-deficient animals, we found increased circulating TNFRs in wild-type mice after BDL, suggesting that there were more unshed TNFRs on the surface of HSCs or hepatocytes in the livers of the mutant mice. This would result in a corresponding increase in TNF-α–induced signaling, which would drive liver fibrosis.

Although iRhom2 triggers ADAM17 maturation, its absence does not completely abolish ADAM17 maturation or activity and thus may not block all shedding of TNF-α after BDL (51). In addition, unshed membrane-bound TNF-α can also bind to and activate both TNFR1 and TNFR2, albeit with different efficiencies (52, 53), suggesting a potential iRhom2-independent mechanism for initiating profibrotic TNF-α signaling after BDL. We expect that if the latter is the case, then it would be more likely to occur in HSCs than in hepatocytes, given that our data show differences in NF-κB signaling, which is downstream of TNF-α, and subsequent proliferation in HSCs with no difference in TNF-α–induced apoptosis in primary hepatocytes. Consistent with this, we did not find a statistically significant difference in circulating liver enzymes and cleaved caspase-3 between Rhbdf2+/− and Rhbdf2−/− mice after BDL, which led us to speculate that TNF-α–TNFR signaling in hepatocytes of Rhbdf2+/− and Rhbdf2−/− mice is comparable during BDL. Our experiments were performed in a whole-body iRhom2-deficient mouse model, and although we analyzed the overall effects of iRhom2 deficiency during BDL, we cannot make definitive conclusions regarding the function of iRhom2 in specific cell types. We can only conclude that iRhom2 reduces the numbers of stellate cells and alleviates liver fibrosis after BDL, but this phenotype might not be uniquely linked to a specific iRhom2-producing cell type and could result from the actions of iRhom2 in multiple cell types.

Sequence alterations in the cytoplasmic tail of iRhom2 can increase ADAM17 activity and consequently increase TNFR shedding (22, 26, 27). It is possible that posttranslational modifications of iRhom2 might trigger ADAM17 maturation and TNFR shedding after cholestasis and therefore have a protective effect against BDL-induced liver fibrosis. Mutations that affect the cytoplasmic tail of iRhom2 are also associated with the establishment of esophageal cancer (26). Future studies could investigate whether inactivating mutations or reduced abundance of iRhom2 correlate with liver fibrosis or hepatocellular carcinoma in patients.

iRhom2 has multiple biological functions apart from its role in ADAM17 maturation. Specifically, iRhom2 can bind to STING and prevent its degradation (39). Accordingly, iRhom2 promotes IFN regulatory factor 3 (IRF3) phosphorylation and translocation into the nucleus and consequently contributes to IFN-I production during viral infection (39, 54). Furthermore, STING and IRF3 are able to promote hepatocyte death and liver fibrosis during CCl4 treatment, albeit IFN-I is beneficial after BDL (5557). However, our data in iRhom2-deficient animals suggest that iRhom2 is rather beneficial during the establishment of liver fibrosis. Hence, we speculate that, in our study, the role of iRhom2 in liver fibrosis is independent from its stabilizing effect on STING. Furthermore, iRhom2 can regulate the cytoskeletal scaffolding protein Keratin 16 (K16) (58). In the liver, keratins protect hepatocytes from apoptosis and necrosis (59), and keratins such as K8 and K18 variants and K19 are associated with liver disease (5962). Accordingly, mice expressing variants of K18 exhibit an increase in Fas-induced liver damage compared to TNF-α–induced liver damage (63). In addition, iRhom2 influences K16 abundance with effects extending to its binding partner K6 (58). One could speculate that iRhom2 affects the equilibrium of keratins in liver tissue, thereby affecting the establishment of liver fibrosis. However, when we analyzed tissue damage after BDL in iRhom2-deficient mice and controls, we did not observe any statistically significant differences, making keratins unlikely contributors to liver fibrosis in our model system. We observed an increased presence of cells positive for CD68, a well-established marker for activated macrophages (64), in liver tissue from iRhom2-deficient mice, which could contribute to increased TNF-α–TNFR signaling in stellate cells.

Other ADAM17 substrates have also been reported to play roles in liver fibrosis. Ectodomain shedding of EGFR ligands and TNFR1 can critically regulate acute liver damage (65), and conditional deletion of HB-EGF results in increased liver injury after acute toxic hepatitis (66). Furthermore, overexpression of HB-EGF results in aggravated liver fibrosis after chronic liver injury (67), whereas another report has shown that treatment with the EGFR inhibitor erlotinib alleviated establishment of liver fibrosis (68). Our findings showed no significant difference in HB-EGF in the sera of control or iRhom2-deficient mice. Moreover, decreased HB-EGF concentrations, which would be expected in the absence of iRhom2, would rather alleviate liver fibrosis. Moreover, we observed no significant differences in EGFR signaling. Hence, we speculate that increased liver fibrosis in the absence of iRhom2 cannot be explained mechanistically by changes in EGFR signaling.

The absence of gp130, a critical factor essential for IL-6R signaling, increases liver damage after BDL, which is associated with increased bacterial burden (37). It has also been shown that lack of IL-6, gp130, and STAT3 signaling in the liver promotes the establishment of steatohepatitis (69). Bacterial translocation after BDL triggers an IFN-I signature resulting in immunosuppression (38). In our investigation, we also observed reduced soluble IL-6R in the sera of iRhom2-deficient animals after BDL. However, p-STAT3 after BDL was not affected by iRhom2. In addition, we did not see any differences in bacterial titers or IFN-I–regulated genes in iRhom2-deficient mice compared to control animals. Although other signaling pathways might be affected by iRhom2 and contribute to the establishment of liver fibrosis, in our model, TNF-α blockade reduced the presence of stellate cells in the liver and alleviated fibrosis in iRhom2-deficient mice.

MATERIALS AND METHODS

Animal experiments

Rhbdf2−/− whole-body knockout mice were bred in house on a C57BL/6 background, as previously described, and are available at Mutant Mouse Resource and Research Centers (www.mmrrc.org/) (21). Experiments were performed in 10- to 12-week-old male mice with littermate controls. All animal experiments were conducted according to the German law for the welfare of animals and were approved by local authorities. For BDL, animals were anesthetized by isoflurane and placed on a heating pad. After intubation and ventilation (Harvard MiniVent, Harvard Apparatus), the animals were shaved, and the skin was disinfected with 70% ethanol and povidone iodine. A midline incision in the upper abdomen was made, and the common bile duct and the gallbladder were identified, isolated, and ligated with silk. The fascia and skin of the midline abdominal incision were closed with silk. Sham treatment was performed similarly but without ligation of the bile duct and gallbladder. Animals were monitored during recovery and treated with carprofen (0.05 mg/kg body weight) after surgical intervention. Etanercept (Enbrel, Pfizer) was reconstituted in PBS to a final concentration of 10 mg/ml. Mice were subcutaneously injected with etanercept (10 mg/kg) 24 hours before surgery and every other day after surgery until day 14.

Analysis of human material

Serum samples from patients suffering from liver disease and healthy volunteers were collected (table S1, cohort A). Patients gave informed consent, and analysis was approved by the Ethics Committee of the Faculty of Medicine at the Heinrich Heine University of Düsseldorf under the study no. 5350. Expression levels of RHBDF2 were determined in cohort B (table S1) under approval of the Ethics Committee of the Medical Faculty of University of Heidelberg under the study no. 206/05.

Histology

Briefly, snap-frozen tissue sections were cut to 7-μm sections, air-dried, and fixed with acetone for 10 min. Sections were blocked with 2% fetal calf serum (FCS) in PBS for 1 hour. Sections were stained with Col1A1, α-SMA, PDGFRβ, Desmin, p65, CD68, F4/80, and LY6G for 1 hour, washed with PBS containing 0.05% Tween 20 (Sigma-Aldrich), and incubated with secondary antibody, anti-rabbit–Cy (1:200), anti-rat allophycocyanin for CD68 (1:200), and anti-rat phycoerythrin for Ki67 together with 4′,6-diamidino-2-phenylindole (DAPI) (1:1000) for 1 hour. Then, sections were washed and mounted using fluorescence mounting medium (Dako). Images were taken by the Axio Observer Z1 fluorescence microscope (ZEISS). Analysis of the fluorescence images was performed using ImageJ software. Picro Sirius Red and Masson’s trichrome staining kits were purchased from Polysciences Inc., and staining was performed according to the manufacturer’s instructions.

Reverse transcription polymerase chain reaction analyses

RNA purification was performed according to the manufacturer’s instructions (QIAGEN RNeasy Kit or TRIzol). Gene expression of Rhbdf2, Tnfrsf1a, Tnfrsf1b, Tnf, Hbegf, Areg, Tgfa, Il6r, Il6, Col1a1, Col3a1, and Acta2 was performed using FAM/VIC probes (Applied Biosystems) and iTAQ One-Step PCR (polymerase chain reaction) kit (Bio-Rad). Gene expression of Il1b, Il10, Cd80, Cd86, Ccl2, Ccl3, Ccl4, Ccl5, Ccl8, Ccl9, Ccl12, Ccl17, Ccl20, Cxcl2, Cxcl9, Cxcl10, Cxcl11, and Cxcl13 was performed using SYBR Green probes. For analysis, the expression levels of all target genes were normalized to Gapdh expression (∆Ct). Gene expression values were then calculated on the basis of the ∆∆Ct method, using naive wild-type mice as a control to which all other samples were compared. Relative quantities (RQs) were determined using the equation: RQ = 2−∆∆Ct.

Immunoblotting

Briefly, liver tissue was lysed in PBS containing 1% Triton X-100 (Sigma-Aldrich), EDTA-free protease inhibitor cocktail (Roche), PhosSTOP (1 tablet/10 ml), and the inhibitors BB-2516 (20 μM; Tocris Bioscience) and 1,10-phenanthroline (10 mM; Sigma-Aldrich). After lysis, the sample was used for immunoblotting. Antibodies used for blotting were as follows: anti-ADAM17 (Abcam), anti–p-STAT3, anti-total STAT3, anti–p-ERK1/2, anti-total ERK1/2, anti-total IκBα, anti–p-p65, anti-total p65, anti–α-tubulin, and anti–β-actin (Cell Signaling Technology).

Enzyme-linked immunosorbent assay

The following enzyme-linked immunosorbent assay (ELISA) kits were used: TNFR1, TNFR2, IL-6Rα, HB-EGF, and amphiregulin (R&D Systems); TNF-α (eBioscience); TGF-α (Antibodies Online); and TGF-β1, IL-6, and IL-1β (Invitrogen). All ELISAs were performed according to the manufacturers’ instructions.

Serum biochemistry

Aspartate aminotransferase (AST or serum glutamic-oxaloacetic transaminase), alanine aminotransferase (ALT or serum glutamic-pyruvic transaminase), total bilirubin, and lactate dehydrogenase were measured using the automated biochemical analyzer SPOTCHEM EZ SP-4430 (ARKRAY, Amstelveen, the Netherlands) and the SPOTCHEM EZ Reagent Strips Liver-1.

Bile acid analysis

Bile acids and their glycine and taurine derivatives were analyzed by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC-MS/MS) (70). The raw MS/MS data are provided in data files S1 to S3. The system consists of an ACQUITY UPLC I-Class (Waters, UK) coupled to a Waters Xevo-TQS MS/MS equipped with an electrospray ionization source in the negative ion mode. Data were collected in the multiple reaction monitoring mode.

Bacterial titer estimation

Mesenteric lymph nodes were harvested from Rhbdf2+/− and Rhbdf2−/− mice after BDL, and bacterial growth was measured.

HSC isolation

Primary mouse HSCs were isolated using sequential pronase-collagenase digestion. Briefly, livers were perfused in situ with Hanks’ balanced salt solution (HBSS) buffer without Ca2+ and Mg2+ (Thermo Fisher Scientific) supplemented with 0.5 mM EGTA for 5 min. Then, liver tissue was perfused with pronase E (0.7 mg/ml; Roche) for 5 min and collagenase P (0.25 mg/ml; Roche) for 6 to 8 min, respectively, at a flow rate of 5 ml/min in HBSS buffer containing Ca2+. After excision of the liver, the liver was digested in vitro for 15 min in HBSS containing 1% deoxyribonuclease I (DNase I; Roche). HSCs were purified from the remainder of nonparenchymal cells and hepatocyte-derived debris by floatation through 9% (w/v) OptiPrep (Axis-Shield PoC AS, Oslo, Norway) in HBSS buffer. The isolated HSCs were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F-12 (Thermo Fisher Scientific) supplemented with 10% FCS. For TNF-α treatment, primary HSCs were isolated from Rhbdf2+/− and Rhbdf2−/−, the isolated HSCs were cultured in DMEM/F-12 supplemented with 10% FCS for 3 days, and then, the cells were starved in serum-free DMEM/F-12 overnight. Next day, the cells were treated with and without TNF-α (50 ng/ml) in DMEM/F-12 supplemented with 10% FCS for 24 hours.

Hepatocyte and Kupffer cell isolation

Primary mouse hepatocytes and Kupffer cells were isolated using collagenase digestion. Briefly, livers were perfused in situ with HBSS buffer without Ca2+ and Mg2+ (Thermo Fisher Scientific) supplemented with 0.5 mM EGTA for 5 min. Then, liver tissue was perfused with collagenase P (0.25 mg/ml; Roche) for 6 to 8 min, respectively, at a flow rate of 5 ml/min in HBSS buffer containing Ca2+. After excision of the liver, the liver was digested in vitro for 15 min in HBSS containing 1% DNase I (Roche). The digested liver was centrifuged for 3 min at a speed of 30g. The supernatant was used for Kupffer cell isolation, whereas hepatocytes were isolated from the pellet. After the first centrifugation, supernatant was slowly layered on the Percoll gradient and centrifuged for 30 min at 1200g. The middle interphase was collected and stained for F4/80, and Kupffer cells were sorted (BD FACSAria). Primary hepatocytes isolated from Rhbdf2+/− and Rhbdf2−/− animals were treated with CHX (10 μg/ml) and TNF-α (40 ng/ml) for 8 hours in Williams’ medium.

Statistical analyses

Data are expressed as means ± SEM. Statistically significant differences between two groups were determined with Mann-Whitney U test. Statistically significant differences between several groups were determined with a one-way analysis of variance (ANOVA). Statistically significant differences between groups in experiments involving more than one time point were determined using a two-way ANOVA.

SUPPLEMENTARY MATERIALS

stke.sciencemag.org/cgi/content/full/12/605/eaax1194/DC1

Fig. S1. Activation of fibrotic markers in BDL animals.

Fig. S2. Shedding of ADAM17 substrates in BDL mice.

Fig. S3. Shedding of TNFRs in cirrhotic patients.

Fig. S4. Shedding of ADAM17 substrates in Rhbdf2+/− and Rhbdf2−/− BDL mice.

Fig. S5. Expression of transcripts encoding ADAM17 ligands in BDL mice.

Fig. S6. ADAM17 maturation is inhibited in the absence of iRhom2 after BDL.

Fig. S7. No difference in liver parameters or apoptosis between Rhbdf2+/− and Rhbdf2−/− mice.

Fig. S8. Decreased conjugation of bile acids and reduction in the HSC marker GFAP aftet BDL in the absence of iRhom2.

Fig. S9. Changes in immune cell infiltrates after BDL in the absence of iRhom2.

Fig. S10. Expression of cytokines and chemokines in Rhbdf2+/− and Rhbdf2−/− mice after BDL.

Fig. S11. IFN responses during BDL do not depend on iRhom2.

Fig. S12. Primary HSCs have increased expression of fibrotic markers.

Fig. S13. Abundance of p-ERK1/2, p-STAT3, IκBα, and cyclin-A2 in Rhbdf2+/−, and Rhbdf2−/− mice after BDL.

Fig. S14. Expression of cytokines and chemokines in etanercept-treated Rhbdf2−/− BDL mice.

Table S1. Clinical parameters of cirrhotic and noncirrhotic patients and healthy volunteers.

Data File S1. Measurement of serum bile acid profile at early time points after BDL.

Data File S2. Measurement of liver bile acid profile at early time points after BDL.

Data File S3. Measurement of serum bile acid profile at day 20 after BDL.

REFERENCES AND NOTES

Funding: This work was supported by the German Research Council (SFB974, KFO217, and LA-2558/5-1). Furthermore, this study was supported by the Jürgen Manchot Graduate School MOI III. Author contributions: B.S. designed and conducted experiments and analyzed data. K.B., A.B., M.A.A.-S., and Y.T. performed experiments and analyzed data. R. Polz, R. Pellegrino, Y.Z., P.V.S., H.C.X., J.V., and D. Herebian performed experiments. N.A. analyzed the statistical tests performed in this study. T.L., E.M., H.H.B., P.M., C.K., T.W.M., V.K., D. Häussinger, and J.S. analyzed and discussed data. A.A.P., K.S.L., and P.A.L. designed studies, discussed data, and wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.
View Abstract

Stay Connected to Science Signaling

Navigate This Article