Department of Internal Medicine I, University of Bonn, Bonn, Germany
Address reprint requests to: Jonel Trebicka, M.D., Department of Internal Medicine I, University of Bonn, Sigmund-Freud Straße 25, D-53105 Bonn, Germany. E-mail: email@example.com; Fax: +49 228 287 19718.
Potential conflict of interest: Nothing to report.
The study was supported by grants from Deutsche Forschungsgemeinschaft (SFB TRR57 projects 1, 10, 15, 18, and Q1).
Activation of the renin angiotensin system resulting in stimulation of angiotensin-II (AngII) type I receptor (AT1R) is an important factor in the development of liver fibrosis. Here, we investigated the role of Janus kinase 2 (JAK2) as a newly described intracellular effector of AT1R in mediating liver fibrosis. Fibrotic liver samples from rodents and humans were compared to respective controls. Transcription, protein expression, activation, and localization of JAK2 and downstream effectors were analyzed by real-time polymerase chain reaction, western blotting, immunohistochemistry, and confocal microscopy. Experimental fibrosis was induced by bile duct ligation (BDL), CCl4 intoxication, thioacetamide intoxication or continuous AngII infusion. JAK2 was inhibited by AG490. In vitro experiments were performed with primary rodent hepatic stellate cells (HSCs), Kupffer cells (KCs), and hepatocytes as well as primary human and human-derived LX2 cells. JAK2 expression and activity were increased in experimental rodent and human liver fibrosis, specifically in myofibroblastic HSCs. AT1R stimulation in wild-type animals led to activation of HSCs and fibrosis in vivo through phosphorylation of JAK2 and subsequent RhoA/Rho-kinase activation. These effects were prevented in AT1R−/− mice. Pharmacological inhibition of JAK2 attenuated liver fibrosis in rodent fibrosis models. In vitro, JAK2 and downstream effectors showed increased expression and activation in activated HSCs, when compared to quiescent HSCs, KCs, and hepatocytes isolated from rodents. In primary human and LX2 cells, AG490 blocked AngII-induced profibrotic gene expression. Overexpression of JAK2 led to increased profibrotic gene expression in LX2 cells, which was blocked by AG490. Conclusion: Our study substantiates the important cell-intrinsic role of JAK2 in HSCs for development of liver fibrosis. Inhibition of JAK2 might therefore offer a promising therapy for liver fibrosis. (Hepatology 2014;60:334–348)
sodium dodecyl sulfate polyacrylamide gel electrophoresis
standard error of the mean
alpha-smooth muscle actin
signal transducers and activators of transcription
Chronic liver diseases represent a major global health problem with annually approximately 800,000 deaths worldwide.[1, 2] Persistent liver injury induces hepatic fibrosis defined as excessive hepatic production and deposition of extracellular matrix (ECM) by myofibroblasts. In liver fibrosis, the main source of these myofibroblasts are activated hepatic stellate cells (HSCs) located in the space of Disse. These myofibroblastic HSCs express alpha-smooth muscle actin (α-SMA) as a marker of their activation.[3-5] Stimulation of the angiotensin-II (AngII) type 1 receptor (AT1R) by local or systemic activation of the renin angiotensin system (RAS) plays a crucial role in HSC activation and fibrogenesis.[6-13] Subsequent accumulation of ECM in the liver disturbs the intrahepatic angioarchitecture and therefore leads to further complications. Activated HSCs play a pivotal role in the progression of fibrosis to decompensated cirrhosis with high morbidity and mortality.[14, 15]
AT1R is coupled to heterotrimeric G proteins (Gaq/11 and Ga12/13), allowing stimulation and activation of several signal pathways involved in cell contraction and ECM production. One of these pathways is the RhoA/Rho-kinase pathway, which is crucially involved in fibrosis and portal hypertension as previously shown by our group.[17-20] Recently, a link between AT1 receptor and RhoA/Rho-kinase pathway was established in smooth muscle cells showing the involvement of the tyrosine kinase, Janus kinase 2 (JAK2). AT1R stimulation activates JAK2, which, in turn, induces Arhgef1, the nucleotide exchange factor responsible for activation of RhoA, which subsequently activates Rho-kinase.
JAK2 is involved in the intracellular signaling of many other receptors, for example, for hormones and cytokines toward transcription regulators of the signal transducers and activators of transcription (STAT) family. In HSCs, JAK2 acts through the STAT pathway or independent of it.[22, 23] The role of these JAK2-induced pathways in hepatocytes has been investigated for hepatic steatosis, ischemia-reperfusion (I/R) injury and cancer, whereas the role of JAK2 in fibrosis, especially mediated by AT1R stimulation, has not been investigated to date.
The present study showed that AT1 receptor-mediated JAK2 activation induces liver fibrosis. Consequently, inhibition of JAK2 blunts fibrosis. This activation of the JAK2/Rho-kinase pathway—dependent on stimulation of AT1R and leading to activation of HSC—was shown in different animal models and cell culture experiments. Furthermore, we confirmed the up-regulation of JAK2/Rho-kinase expression in human fibrosis.
Materials and Methods
We used 190 Sprague-Dawley wild-type (WT) rats and 155 mice (95 C57BL/6J WT and 60 AT1aR−/− mice) for our experiments. AT1aR-deficient mice were kindly provided by Nikos Werner (Department of Internal Medicine II, University of Bonn, Bonn, Germany). The responsible committee for animal studies in North Rhine-Westphalia approved the study (LANUV 8.87-50.10.31.08.28).
Cholestatic Model of Fibrosis
Bile duct ligation (BDL) was performed in rats with an initial body weight of 180-200 g, as described previously.[17, 18] Experiments were carried out 2 weeks after BDL in 19 rats, whereas 28 sham-operated rats served as controls, respectively. Ten rats undergoing BDL for 2 weeks received AG490 (1 mg/kg/day, intraperitoneally [IP]) on the last 7 days before sacrifice. BDL and sham operation was performed in 5 20-25 g WT mice, which were sacrificed after 2 weeks.
Toxic Models of Fibrosis
Twenty-three rats with an initial body weight of 100-120 g underwent twice-weekly inhalation of 1 L/min CCl4 for 14-16 weeks until ascites were present, as described previously.[17, 18] Twenty-one age-matched control rats did not receive CCl4. Periodic CCl4 inhalation of 2 L/min was performed in 25 mice (15 WT and 10 AT1aR−/−) for 4 weeks, as described previously, whereas 38 mice served as controls (11 WT and 27 AT1aR−/−). Additionally, 24 rats with an initial body weight of 200-250 g underwent weekly thioacetamide (TAA) administration with adjusted dosing in their drinking water for 18 weeks, as described previously.
For continuous release of AngII, osmotic pumps (2ML2 for rats, 2002 for mice, Alzet; Charles River Laboratories, Sulzfeld, Germany) were subcutaneously implanted in vivo in 18 rats, 15 AT1aR−/−, and 19 WT littermates, as described previously. Each pump released 0.7 mg/kg/day of AngII in rats as well as mice for 14 days, which has been shown to induce hepatic fibrosis. Pumps releasing saline were implanted in 5 rats and 24 mice as controls. Additionally, 9 of the rats received AG490 (1 mg/kg/day, IP) for the last 7 days before sacrifice.
Human Liver Samples
The human ethics committee of the University of Bonn (202/01) approved the use of human liver samples, obtained during liver transplantation from patients with alcohol-induced cirrhosis (n = 16). Liver samples from patients without cirrhosis undergoing liver resection served as controls (n = 10). None of the patients or donors received catecholamines, angiotensin converting enzyme (ACE) inhibitors, or angiotensin receptor antagonists before transplantation. Samples were snap-frozen after excision.
Snap-frozen cells and liver samples were processed as previously described using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels and nitrocellulose membranes.[17, 18] Ponceau S staining assured equal protein loading. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or β-actin served as endogenous controls. Membranes were incubated with the respective primary antibody (Ab; Supporting Table 1) and corresponding secondary peroxidase-coupled Ab (Santa-Cruz-Biotechnology, Santa Cruz, CA). After enhanced chemiluminescence (ECL; Amersham, Bucks, UK), digital detection was evaluated using Chemi-Smart (PeqLab Biotechnologies GmbH, Erlangen, Germany).
Western blotting analysis was quantified by densitometry of all experiments (means ± standard error of the mean [SEM]) with values of controls set to 100 densitometric units (representative western blottings are shown in the Supporting Figures). Expression of phosphorylated JAK2 (pJAK2) at Tyr1007/1008 served as a marker of JAK2 activation, and Rho-kinase activity was analyzed as phosphorylation of its substrate moesin at Thr558 detected by phospho- and site-specific Abs.
Quantitative Real-Time Polymerase Chain Reaction
RNA isolation, reverse transcription, and detection by real-time polymerase chain reaction (RT-PCR) were performed as described previously.[17, 18] Assays provided by Applied Biosystems (Foster City, CA) are listed in Supporting Table 2. 18S ribosomal RNA served as an endogenous control. The results of HSC and liver samples were expressed as 2−ΔΔCt, and the data are reported as relative gene expression compared to the control group.
Hepatic Hydroxyproline Content
Hepatic hydroxyproline content was determined photometrically, as described previously.
Sirius Red Staining
For the detection of collagen fibers, paraffin-embedded liver sections were stained with Sirius red using standard methods described previously.
Immunohistochemical Staining for JAK2, pJAK2, and α-SMA
Staining for JAK2, pJAK2, and α-SMA was performed in cryosections from liver tissue (3 and 7 µm). The detailed method is described in the Supporting Information. The amount of staining was evaluated by computational analysis (Histoquant; 3DHistech, Budapest, Hungary), as described previously.[27, 28]
Quantification (% of stained area) of immunohistochemical (IHC) staining is expressed as mean ± SEM of all experiments. For representative sections, please see the Supporting Figures.
Coimmunofluoresence Stainings of pJAK2 and α-SMA
Colocalization of pJAK2 and α-SMA was analyzed by immunofluorescent staining of 7-µm cryosections, as described in the Supporting Information.
Isolation of Primary HSCs
Rat and mouse HSCs were isolated as described previously.[18, 29] Briefly, primary HSCs were isolated in a two-step pronase-collagenase perfusion from livers of healthy rats (n = 36), as well as WT (n = 8) or AT1aR−/− mice (n = 8), and fractionated by density-gradient centrifugation. Viability and purity were systematically over 95%. Cells were seeded on uncoated plastic culture dishes. Experiments were performed 7 days after isolation or after the first passage (10 days) when HSCs were fully activated.
Primary human HSCs were obtained from ScienCell (San Diego, CA) and were cultured and harvested as described previously.[30-32]
Isolation of Hepatocytes
Primary hepatocytes were isolated from male WT rats (n = 5) or WT mice (n = 8), as described previously.
Isolation of Kupffer Cells
Primary Kupffer cells (KCs) were isolated from WT mice (n = 10), as described previously.
Incubation with AngII, AG490, and Losartan
LX2 cells were provided by Vijay H. Shah (Mayo Clinic, Rochester, NY), originally established by Scott Friedman. AngII (10 µM) and/or AG490 (1.5, 5, and 25 µM) and/or Losartan (10 µM) was added to the culture medium of these cells, as indicated, for 3 days, or cells remained untreated.
Transfection With JAK2 Plasmid
Twenty-four hours before transfection, 6 x 105 LX2 cells were incubated with transfection media (Dulbecco's modified Eagle's medium [DMEM] with 10% fetal calf serum [FCS] without penicillin/streptomycin). The JAK2 plasmid and the respective empty plasmid control were isolated according to manufacturer instructions (NucleoBond Xtra Maxi kit; Machery, Nagel, Germany). Plasmid (15 µL) and 37.5 µL of lipofectamine were incubated for 20 minutes with a total volume of 3.6 mL of media. This plasmid/lipofectamine mix was added drop-wise to cells after removal of media. After 3-4 hours, cells were again incubated with media, containing 10% FCS, and harvested after 3 days. Efficacy of transfection was tested by RT-PCR and western blotting.
Transduction With AdJAK2
The recombinant replication-deficient adenoviral vector, AdJAK2, was generated using the adenoviral backbone vector, pAdEasy-1, containing the sequence of human adenovirus 5 with deletion of E1- and E3-genes with the transfer vector, pShuttleCMV (Stratagene, La Jolla, CA). JAK2 transgene was obtained from the plasmid, pUNO1-mJAK2a (Invivogen, Toulouse, France).
LX2 cells (6 × 105) were cultured in 10-cm plates. After 24 hours, cells were transduced with AdJAK2 and AdLacZ as a control at a transfection multiplicity of infection of 250 in DMEM supplemented with 2% FCS and 1% penicillin/streptomycin for 2 hours. Cells were washed twice with media, and AdJAK2-transduced cells were treated with 10% FCS and 1% penicillin/streptomycin alone, or additionally with 5 µM of AngII, or with 5 µM of AngII and AG490. AdLacZ-transduced cells were prepared in the same way as AdJAK2-transduced cells. Efficacy of transfection was tested by RT-PCR and western blotting.
Detection of Reactive Oxygen Species
Serum-starved LX2 cells (1.8 × 104 cells) plated in six-well plates were loaded with the reagent, 2',7'-dichlorofluorescein diacetate (DCFDA), a fluorogenic dye that measures hydroxyl, peroxyl, and other reactive oxygen species (ROS) activity within the cell (DCFDA-Cellular Reactive Oxygen Species Detection Assay Kit, catalog no. ab113851; Abcam, Cambridge, MA). After 20 minutes at 37°C incubation with AngII (10−5 M) with or without AG490 (5 µM), cells were washed and measured for the indicated time in a multiwell fluorescence plate reader using excitation and emission filters of 485 and 535 nm, respectively.
Apoptosis and Cycle Analysis
Analysis of apoptosis (Annexin V Apoptosis Detection Kit; BD Biosciences, Heidelberg, Germany) and cell-cycle analysis was performed as previously described.
Data are presented as mean ± SEM. The Student t test was used for comparison, where appropriate. Mann-Whitney's U test or analysis of variance were used for comparison between groups (minimum n = 5/group). P values <0.05 were considered statistically significant.
Increased Expression and Phosphorylation of JAK2 in Experimental and Human Liver Fibrosis is Found Predominantly in Activated HSCs and Myofibroblasts
In human and experimental fibrosis, hepatic expression of the components of RAS was increased (Fig. 1A), suggesting that RAS is activated in liver fibrosis and cirrhosis. Accordingly, hepatic expression of AT1R and its downstream effectors (JAK2, Arhgef1, RhoA, and Rho-kinase) were increased in human liver fibrosis (Fig. 1B,D and Supporting Fig. S1A). We also observed that activation of JAK2, analyzed as JAK2 phosphorylation at Tyr1007/1008 (pJAK2), and of Rho-kinase, analyzed by the phosphorylation of its substrate, moesin, at Thr558, were increased in human cirrhosis, when compared to nonfibrotic controls (Fig. 1B and Supporting Fig. S1A). This was confirmed by IHC staining. Similarly, using two different models of fibrosis in mice and three different models in rats, we found that AT1R-mediated JAK2 phosphorylation at Tyr1007/100821 was increased in fibrotic livers, when compared to control livers (Fig. 2A,B and Supporting Fig. S1B,C). Furthermore, the JAK2 downstream effector, Arhgef1, its target, RhoA, and the downstream effector, Rho-kinase, were strongly expressed and activated in liver fibrosis (Fig. 2A,B).
Interestingly, although faint JAK2 staining was found in most liver cells, it predominated in fibrotic septa as well as in perivascular and -sinusoidal regions (Supporting Figs. S3A-S5A). In contrast, pJAK2 was only found in fibrotic septa and around vessels and small sinusoids (Figs. 1C and 2C and Supporting Figs. S3B-S5B). Immunofluorescence costaining of pJAK2 and α-SMA in fibrotic livers showed the expression of pJAK2 in α-SMA-positive cells in fibrotic septa as well as in perivascular and -sinusoidal areas (Fig. 1E and Supporting Fig. S3C). These data show, for the first time, that AT1R-dependent JAK2 phosphorylation at Tyr1007/1008 occurs in hepatic myofibroblasts, which derive mainly from activated HSCs.
Protein expression of JAK2-dependent STAT3/suppressor of cytokine signalling 3 signaling was also activated in different fibrosis models of mice and rats (Supporting Fig. S2A,B), suggesting an increased activity of JAK2.
AT1R Stimulation Induced Fibrosis Through the JAK2 PathwayIn Vivo
We assessed the effect of JAK2 inhibition on hepatic fibrogenesis in rats and mice. AG490, injected daily for 1 week before killing of BDL rats (14 days of BDL in total) attenuated fibrogenesis, as assessed by hydroxyproline content and Sirius red staining. This treatment attenuated the activation of HSCs, as mirrored by α-SMA immunostaining as a result of inhibition of JAK2 phosphorylation at Tyr 1007/1008, as assessed by IHC (Fig. 3A,B and Supporting Fig. S6).
Transcription of the JAK2 downstream proteins, as well as HSC activation documented by elevated α-SMA levels and collagen production, was significantly up-regulated at the messenger RNA (mRNA) level after BDL. This effect was attenuated by JAK2 inhibition with AG490 (Fig. 3C).
To further underline the crucial role of AT1R, continuous stimulation of AT1R was performed using AngII infusion in rats and mice for 14 days. Indeed, this treatment induced expression and activation of JAK2 pathway components (Fig. 3D), resulting in mild fibrosis.
Interestingly, concomitant JAK2 inhibition during the last 7 days before sacrifice of rats receiving AngII infusion significantly decreased fibrosis, as assessed by hydroxyproline content and Sirius red staining, as well as inhibition of JAK2 phosphorylation at Tyr 1007/1008 and JAK2 downstream signaling (Fig. 4 and Supporting Fig. S7). These findings demonstrate that AngII injection induced fibrosis as a result of AT1R-mediated activation of JAK2 and its downstream effectors.
In these models, we could not detect any difference regarding thrombogenic or angiogenic effects of the JAK2 inhibitor, as assessed by histology.
Inhibition of JAK2 Pathway or Absence of AT1R With Consecutive Down-Regulation of JAK2 PathwayIn VivoAttenuated Fibrosis
To further assess the role of AT1R in vivo, we induced another form of chronic liver injury by means of CCl4 inhalation in AT1aR−/− mice and their respective WT littermates. (AT1aR is the subtype of AT1R in mice responsible for contraction of smooth muscle cells and important for development of liver fibrosis,[10, 16] further abbreviated as AT1R). Expression of JAK2 and its downstream effectors was significantly down-regulated in these mice (Fig. 5 and Supporting Fig. S8). This was associated with reduced hepatic fibrosis after challenge (Fig. 5 and Supporting Fig. S8). AngII induces fibrosis and activation of JAK2 and its downstream effectors in wt mice, but not in AT1R−/− mice. (Fig. 6A,B and Supporting Fig. S9).
Thus, JAK2 inhibition upon fibrosis induction blunted fibrosis. Similarly, AT1R-deficient animals showed attenuated fibrogenesis and lack of activation of JAK2 and its downstream effectors.
AT1R-Dependent Activation of JAK2 Pathway-Induced Activation of HSC and Collagen Production
In vitro experiments confirmed the results from IHC staining. The components of the JAK2 pathway (JAK2, pJAK2, Arhgef-1, Rho-kinase, and pMoesin) were highly expressed and activated in HSCs after their activation (day 7 in culture), when compared to hepatocytes and KCs (Fig. 6C,D), confirming that pJAK2 was localized in activated HSCs (as shown in Fig. 6C,D).
Activated HSCs showed increased expression and activation of components of the AT1R-mediated JAK2 pathway. To further decipher the role of JAK2, specifically for fibrogenesis, we overexpressed JAK2 in an immortalized human HSC line, namely, LX2 cells, using a JAK2 plasmid (Fig. 7A). We observed elevated transcription of its downstream effectors, RhoA and Rho-kinase, together with higher protein expression of α-SMA and collagen I (Col1) as markers for profibrotic activity of HSCs (Fig. 7B). Efficacy of JAK2 transfection is shown in Supporting Fig. S10.
Additional stimulation of these cells with AngII further increased the transcription of downstream effectors of JAK2 and their profibrotic activity (α-SMA and Col1; Fig. 7B). Again, the JAK2 inhibitor, AG490, blunted the AngII effect (Fig. 7B). In order to exclude a vector-specific effect, we repeated these experiments using adenoviral vectors containing JAK2 or LacZ (as a control) and obtained similar results (Fig. 7B). AT1R dependence was shown in further experiments using HSCs isolated from AT1R-deficient mice. In these mice, neither AngII nor AG490 elicited any significant effect on the transcription of AT1R/JAK2 pathway effectors (Fig. 7C). Efficacy of JAK2 transduction with the adenoviral vector is shown in Supporting Fig. S10.
In VitroJAK2 Inhibition and Blocking Experiments
AT1R stimulation, assessed by incubation of human immortalized HSCs (LX2 cells) with AngII, induced JAK2 phosphorylation at Tyr 1007/1008 (Supporting Fig. S10). The pharmacologic inhibitor of JAK2, AG490, dose dependently blocked phosphorylation of JAK2 and expression of the downstream effectors, RhoA and Rho-kinase (Supporting Fig. S10). Also, the elevated Arhgef1 expression and Rho-kinase activation was blunted after AG490 coincubation (Fig. 8A). Experiments with human primary HSCs confirmed the findings in LX2 cells (Fig. 8B,C). AngII treatment of LX2 cells increased ROS, but after coincubation with AG490, ROS dropped to normal levels (Fig. 8E). Blocking experiments using AG490 (JAK2 inhibitor) or Losartan (AT1R blocker) demonstrated blunting of protein expression of downstream effectors of JAK2 (Arhgef1 and Rho-kinase) and α-SMA (activation marker) in primary HSCs similarly to LX2 cells (Fig. 8D). Interestingly, blockade of AT1R or inhibition of JAK2 had no effect on apoptosis or cell cycle (Supporting Fig. S10).
In summary, our in vitro data demonstrated that AT1R-dependent JAK2 induced activation and profibrotic activity of HSCs though its downstream effectors. Inhibition of JAK2 blunted AT1R-mediated activation, profibrotic properties, migration, and contraction of HSCs.
Our in vivo and in vitro findings show the importance of AT1R-dependent JAK2 activation in hepatic fibrosis in mice, rats, and humans. These findings are uniformly supported by different approaches using several animal models of liver fibrosis as well as by pharmacological and genetic modulations. JAK2 expression and activation was found to occur mainly in activated HSCs, underlining the pivotal role of these cells for collagen production.
RAS is critically involved in the development of progressive fibrosis in different tissues (heart, kidney, and lung) as well as in the liver, mainly as a result of AT1R stimulation.[6, 34-36] In chronic liver injury, activation of RAS is a hallmark of progressive liver disease.[2, 7, 14, 15] RAS activation leads to sodium retention with formation of ascites in later stages of liver disease. But, it also leads to initiation, perpetuation, and augmentation of inflammatory and fibrogenic processes within the liver, where HSCs are key players.
It has been shown only recently that AngII exerts its AT1R-dependent contractile effect in the vascular smooth muscle cells through phosphorylation of JAK2 and the nucleotide exchange factor for RhoA, Arhgef1, which then activates the RhoA cascade toward Rho-kinase, finally resulting in contraction of these cells. Importantly, JAK2-mediated signaling in fibrosis is caused by both increased JAK2 expression and increased JAK2 activation. In this landmark study, the researchers described the missing link between AT1R stimulation and the RhoA/Rho-kinase pathway in arterial smooth muscle cells as a potential mechanism for arterial hypertension. Here, we demonstrate that the same holds true for myofibroblastic-activated HSCs. We show that HSCs—but not hepatocytes or KCs—harbor an activated JAK2/RhoA/Rho-kinase signaling cascade that can be further stimulated by AngII and which is blocked or blunted by an AT1R blocker, AT1aR, deletion in mice (the subtype of AT1R in mice responsible for contraction of smooth muscle cells and important for development of liver fibrosis[10, 16]) or JAK2 inhibition, respectively. This altered signaling clearly leads to functional changes with increased α-SMA and collagen expression of these cells. Of note, no effects on survival or cell-cycle progress were observed in these cells.
In line with these in vitro findings, we show that fibrosis in different animal models and in humans is associated with increased JAK2 expression and activation. Furthermore, fibrosis is attenuated by application of JAK2 inhibitors, whereas continuous exogenous AngII infusion[9, 37] increases collagen formation and HSC activation. In addition, we demonstrated in in vivo experiments that cells showing JAK2 phosphorylation were, in fact, activated myofibroblastic HSCs, because they were located in fibrotic septa and were positive for α-SMA, whereas other α-SMA-negative liver cells (e.g., hepatocytes and inflammatory cells) lacked evidence of JAK2 phosphorylation.
Activation of RAS is a uniform finding in advanced cirrhosis.[14, 15, 38] One possible major reason for this could be a systemic hemodynamic derangement with splanchnic and peripheral vasodilation, leading to “underfilling” in the central intrathoracic compartment and reactive secretion of vasoconstrictors. However, the response from the peripheral and splanchnic vessels to AngII is inadequate, whereas the hepatic vascular bed overreacts. Although this has been shown repeatedly in human and animal models, it is only partially understood.[14, 15, 38] It is questionable whether systemic RAS activation alone accounts for hepatic JAK2 activation, as shown here. Certainly, there must be a systemic effect because we were able to show that AngII infusion enhanced hepatic collagen formation and activated the AT1R/JAK2/Rho-kinase axis in the liver. However, there are probably also paracrine intrahepatic factors involved that lead to up-regulation of AT1R expression on activated HSCs and local formation of AngII with subsequent phosphorylation of JAK2, as shown for oxidative stress and the I kappa B kinase/RelA pathway.[6, 39] Furthermore, JAK2 may also be activated by cytokines or several concomitant pathological conditions in the liver.[40-44]
In our animal models, neither inflammatory infiltrates nor isolated KCs expressed pJAK2. Thombogenic and angiogenic properties of JAK2 were not observed in our models, as assessed by histology. We focused on the AT1R-dependent JAK2/Arhgef1/Rho-kinase pathway. Effects that might be driven by the JAK/STAT pathway, as shown for the transdifferentiation of HSC, I/R, ischemia/reperfusion liver injury, hepatocellular carcinoma, and steatohepatitis, could also play a role in liver cirrhosis.[22, 40, 44] Indeed, these effects were present in our animal models as well.
In summary, in the present work, we showed that JAK2 expression and activation are increased in activated HSCs in fibrosis and inhibition of JAK2 decreased liver fibrosis. Recently, new JAK2 inhibitors have been released for human use in malignant disorders,[45, 46] which, based on these findings, might also be considered for treatment of fibrosis in a clinical setting. In this case, liver-specific targeting is warranted.
The authors thank G. Hack and S. Bellinghausen for their excellent technical assistance, as well as U.B. Kaupp and S. Dentler for their critical reading.