Janus kinase 2 inhibition by pacritinib as potential therapeutic target for liver fibrosis

Background and Aims:

Janus kinase 2 (JAK2) signaling is increased in human and experimental liver fibrosis with portal hypertension. JAK2 inhibitors, such as pacritinib, are already in advanced clinical development for other indications and might also be effective in liver fibrosis. Here, we investigated the antifibrotic role of the JAK2 inhibitor pacritinib on activated hepatic stellate cells (HSCs) in vitro and in two animal models of liver fibrosis in vivo.

Approach and Results:

Transcriptome analyses of JAK2 in human livers and other targets of pacritinib have been shown to correlate with profibrotic factors. Although transcription of JAK2 correlated significantly with type I collagen expression and other profibrotic genes, no correlation was observed for interleukin‐1 receptor‐associated kinase and colony‐stimulating factor 1 receptor. Pacritinib decreased gene expression of fibrosis markers in mouse primary and human‐derived HSCs in vitro. Moreover, pacritinib decreased the proliferation, contraction, and migration of HSCs. C₅₇BL/6J mice received ethanol in drinking water (16%) or Western diet in combination with carbon tetrachloride intoxication for 7 weeks to induce alcoholic or nonalcoholic fatty liver disease. Pacritinib significantly reduced liver fibrosis assessed by gene expression and Sirius red staining, as well as HSC activation assessed by alpha‐smooth muscle actin immunostaining in fibrotic mice. Furthermore, pacritinib decreased the gene expression of hepatic steatosis markers in experimental alcoholic liver disease. Additionally, pacritinib protected against liver injury as assessed by aminotransferase levels.

Conclusions:

This study demonstrates that the JAK2 inhibitor pacritinib may be promising for the treatment of alcoholic and nonalcoholic liver fibrosis and may be therefore relevant for human pathology.

INTRODUCTION

Chronic liver injury, such as chronic viral infection, alcoholic steatohepatitis (ASH), and non‐ASH (NASH), induces progressive fibrosis and may result in cirrhosis with portal hypertension.1,2 The main profibrotic and procontractile cell types in the liver are the hepatic stellate cells (HSCs), which, once activated, produce excessive amounts of extracellular matrix, especially collagen.3,4 Recently, we and others have shown that Janus kinase 2 (JAK2) is upregulated in human and murine HSCs and that it mediates fibrogenesis and contraction of HSCs downstream of the angiotensin‐II‐type‐I receptor (AT1R).5–9 Previously, we could show that Jak2 inhibition and overexpression have, respectively, blunted or enhanced liver fibrosis in vitro and in vivo.5,6 The AT1R is coupled to JAK2, which in turn activates the Rho guanine nucleotide exchange factor 1 (Arhgef1)/Ras homolog family member A (RhoA)/Rho‐kinase (ROCK) pathway, which is upregulated in human cirrhosis.5,6 Pharmacological inhibition and genetic JAK2 deficiency blunted fibrogenesis and decreased portal pressure in different mouse models and decreased the contractility of HSCs in vitro.5–7

Interestingly, Janus kinase (JAK) inhibitors are already in clinical use for myelofibrosis, a bone marrow blood cancer characterized by increased inflammation and fibrosis.10 Although the US Food and Drug Administration has approved two JAK inhibitors (ruxolitinib, fedratinib),11 pacritinib is in advanced clinical development for myelofibrosis.12–14 It is more specific for JAK2 than existing JAK inhibitors and does not induce myelosuppression.15 Although pacritinib specifically inhibits JAK2, several studies indicate that it also can inhibit the signaling of the nontyrosine kinases interleukin‐1 receptor‐associated kinase (IRAK1) and colony‐stimulating factor 1 receptor (CSF1R), which are involved in microenvironmental tumor interactions, as well as in tumor progression.16–18

In this study, we investigate the therapeutic role of the JAK2 inhibitor pacritinib in chronic liver disease with fibrosis. We show that pacritinib decreases the activation of HSC in vitro and that oral gavage of pacritinib ameliorates liver fibrosis progression in two experimental models of liver fibrosis with different etiologies in vivo.

MATERIALS AND METHODS

Human cell lines

Snap‐frozen LX‐2 cells (provided by V.S., Mayo Clinic) and TWNT‐4 cells (provided by M.‐L.B., RWTH Aachen),19–21 both immortalized human‐derived HSCs, were incubated with cell culture medium (Dulbecco's Modified Eagle Medium [DMEM] +10% fetal bovine serum +penicillin/streptomycin) in plastic flasks at 37°C. After reaching 80% confluency, cells were passaged with a 1:2 split ratio.

Isolation of primary HSCs

HSCs of C₅₇BL/6J wild‐type mice and floxed‐Jak2 knockout mice with smooth muscle protein 22 Cre‐promotor (SM22^Cre+‐Jak2^f/f) were isolated as described previously.5,22 Briefly, primary HSCs were isolated in a two‐step pronase–collagenase liver perfusion from healthy mice by density gradient centrifugation. Viability and purity were systematically above 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.

Pacritinib treatment in vitro

Human‐derived (LX‐2 and TWNT‐4) and primary murine HSCs were incubated with 1 μm of pacritinib (SB1518, Cayman Chemical) diluted in DMEM +penicillin/streptomycin for 24 h and harvested for protein and gene expression analysis afterwards.

Mice

All experiments were approved and performed in accordance with the German animal protection law of the Animal Care and Use Committee Hesse, Germany, in the animal care facility (application number: V54‐19c20/15‐FK/2005). All animals received human care according to the criteria outlined in the EU regulations on animal research (2010/63/EU). Mice were purchased by Charles River Laboratories. SM22^Cre+‐Jak2^f/f breeding pairs were kindly provided by Peter Sayeski, Department of Physiology and Functional Genomics, University of Florida, College of Medicine. At the end of the experiments, sera and livers were collected for histological and molecular readouts of fibrosis, inflammation, and steatosis.

Toxic models of liver fibrosis in mice

ASH model

Male 10–12‐week‐old wild‐type C₅₇BL/6J male mice were fed ad libitum for 7 weeks with ethanol in the drinking water (4% during week 1, 8% during week 2, and 16% until euthanized) and with normal chow (Sniff Spezialdäten GmbH). Mice were randomly divided into two groups: corn oil + ethanol group and carbon tetrachloride (CCl₄)/ethanol group. The intraperitoneal injection dose was 0.75 μl per g of body weight corn oil or 2 μl per g of body weight CCl₄ (CCl₄:corn oil = 1:3; volume:volume [v:v]) for 7 weeks (twice a week) to induce liver fibrosis.23

NASH model

Male 10–12‐week‐old wild‐type C₅₇BL/6J mice were fed ad libitum for 7 weeks with high fat and cholesterol rich Western diet (WD) (diet #S0279‐S011, 1.25% cholesterol, Sniff Spezialdäten GmbH). Mice were randomly divided into two groups: corn oil/WD group and CCl₄/WD group. The intraperitoneal injection dose was 0.75 μl per g of body weight corn oil or 2 μl per g of body weight CCl₄ (CCl₄:corn oil = 1:3; v:v) for 7 weeks (twice per week) to induce liver fibrosis.23

Pacritinib treatment in vivo

After 3 weeks of ASH‐ or NASH‐induced liver injury, mice received pacritinib (SB1518, Cayman Chemical) either orally by gavage or by intravenous (i.v.) injection for an additional 4 weeks. Mice were gavaged with pacritinib (300mg/kg body weight) or vehicle three times per week for 3 more weeks. Another group of mice with CCl₄/ethanol‐induced liver injury was injected with pacritinib (65 μg/kg body weight) or vehicle intravenously twice per week.

Transcriptome analysis

Transcriptome analyses were performed on whole liver samples of patients with cirrhosis (n = 23) and healthy individuals (n = 7) using the Illumina NovaSeq 6000 S4 reagent kit (200 cycles) with the NovaSeq Xp 4‐Lane kit. The count matrix was generated using BioJupies,24 data were normalized with the R package edgeR (version 3.30.0),25 and differential gene expression analysis was performed with the R package limma (version 3.44.1).26 Correlation analysis of whole liver samples was performed and visualized with the R package corrplot using the Spearman correlation method.27 The local ethics committee of the University of Bonn approved the study (029/13), and all patients agreed and signed an informed written consent in accordance with the Declaration of Helsinki for the procedures they underwent.

Statistical analysis

Statistical analyses among groups were performed using Prism, version 5.0 (GraphPad). Data were expressed as mean ± SD for in vitro experiments and ± SEM for in vivo experiments. Comparisons between two groups were done by nonparametric Mann–Whitney U t tests. Comparisons among three and four groups were done with analysis of variance (ANOVA), followed by Tukey's multiple comparison posttest. A p < 0.05 was considered as significant (^# p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). All experiments were performed in triplicate at least three times, and representative images or blots are shown in the manuscript.

RESULTS

JAK2, the main target of pacritinib, correlates with fibrosis and portal hypertension in patients with cirrhosis

Although JAK2 is the main target of pacritinib, it might also target IRAK1 and CSF1R (Figure 1A). Therefore, we correlated the gene expression of these three kinases with the main gene markers of liver fibrosis and portal hypertension in patients with cirrhosis to identify their individual association with these processes. The clinical characteristics of patients with cirrhosis (n = 23) and noncirrhotic controls (n = 7) are presented in Table 1. The heat maps showed a strong correlation of Jak2 with the genes associated with HSC activation and fibrosis, based upon Spearman's correlation coefficient for all analyzed genes (log2 fold) (Figure 1B), especially collagen‐1 (Figure 1C), in patients with cirrhosis. Interestingly, Irak1 and Csf1r showed negative correlations to those genes (Figure 1B) and were not correlated with type I collagen (Col1a1) (Figure 1C).

FIGURE 1.

JAK2, the main target of pacritinib, correlates with fibrosis and portal hypertension in patients with cirrhosis. (A) Diagram of the pacritinib targets. (B) Heat map of the correlation among Jak2, Irak1, and Csf1r with genes associated to HSC activation and fibrosis. (C) Correlation of hepatic transcriptional levels of Col1a1 with Jak2, Irak1, and Csf1r. (D) Heat map of correlation among Jak2, Irak1, and Csf1r with surrogates of portal hypertension. (E) Hepatic transcriptional levels of Infgr and Il12. (F) Heat map of correlation among Jak2, Irak1, and Csf1r with Infgr and Il12. Data are shown as Spearman rank coefficient (Rs) and p value. APRI, platelet ratio index; Csf1r, colony‐stimulating factor 1 receptor; Col1a1, type I collagen; FIB‐4, fibrosis‐4 score; Gα, G protein alpha subunit; HSC, hepatic stellate cell; Ifngr, interferon‐gamma receptor; Il12r, interleukin‐12 receptor; Irak1, interleukin‐1 receptor‐associated kinase; JAK2, Janus kinase 2; TIPS, transjugular intrahepatic portosystemic shunt.

TABLE 1. General characteristics of patients.

Parameters	Value
Patients (n)	30
Sex (female/male/no data)	7/16/7
Age (in years) as median (range)	46 (32–64)
Etiology (alcoholic/chronic hepatitis)	7/16
MELD score as median (range)	15 (6–23)
Child score as median (range)	9 (5–12)
Child category (A/B/C)	4/9/10
Ascites (absent/mild/severe)	19/6/5
Esophageal varices (absent/grade I–II/grade III–IV)	15/12/3

Abbreviation: MELD, model for end‐stage liver disease.

Portal hypertension is associated with the development of ascites and variceal bleeding, which could be reduced with the implantation of a transjugular intrahepatic portosystemic shunt (TIPS). Because we could not have a direct portal pressure measurement in these patients, we considered the presence of esophageal varices, ascites, variceal bleeding, Child–Pugh score, fibrosis‐4 score, platelet ratio index score, and TIPS insertion in patients with cirrhosis as a surrogates of portal hypertension. To assess the correlation of portal hypertension with the expression of Jak2, Irak1, and Csf1r, we analyzed the correlation of the surrogate's grade with the expression of the three targets of pacritinib, and interestingly, Jak2 correlates highly using portal hypertension surrogates in relation to Irak1 and Csf1r (Figure 1D).

Because JAK2 is activated in the liver by, inter alia, cytokines, we analyzed the expression of interferon‐gamma receptor (Ifngr) and interleukin‐12 receptor (Il12r), which were significantly increased in livers of patients with cirrhosis (Figure 1E).

We also analyzed their correlation with Jak2, Irak1, and Csf1r in livers of patients with cirrhosis (Figure 1F) and found that only Jak2 was strongly correlated to Il12r and weakly to Ifngr in livers of patients with cirrhosis (Figure 1F). These results confirm our previous data,5,6 namely, that the expression of Jak2 is strongly associated with fibrosis progression and especially collagen in humans. Therefore, we used the JAK2 inhibitor pacritinib to assess its effects in vitro and in vivo experimental liver fibrotic models.

Pacritinib treatment JAK2‐dependently blunts fibrogenesis and proliferation in rodent and immortalized human‐derived HSCs

Because the main contributors to liver fibrosis are activated HSCs, isolated primary HSCs from wild‐type and specific floxed‐Jak2 knockout mice (SM22^Cre+‐Jak2^f/f) were used to assess the effect of pacritinib. Therefore, HSCs with clearly activated phenotype after 7 days of culture growth were incubated with pacritinib.

First, we assessed protein levels and mRNA expressions of collagen‐1, the main profibrotic marker, which were significantly reduced by pacritinib treatment and blunted in SM22^Cre+‐Jak2^f/f mouse HSCs (Figure 2A). We also assessed phospho‐moesin (p‐moesin) protein levels, which showed the same effect as in collagen‐1 expression (Figure 2A). Moreover, we analyzed the related profibrotic genes, Jak2, Rock, alpha‐smooth muscle actin (Acta2), and transforming growth factor beta (Tgfb), which showed a clear reduction after pacritinib incubation in wild‐type and SM22^Cre+‐Jak2^f/f mouse HSCs (Figure 2B).

FIGURE 2.

Pacritinib treatment JAK2‐dependently blunts fibrogenesis and proliferation in rodent and immortalized human‐derived HSCs. Isolated wild‐type and SM22^Cre+‐Jak2^f/f mice HSCs were treated with pacritinib at 1 μm during 24 h to measure Col1a1 and p‐moesin protein levels and Col1a1 gene expression (A) and mRNA expression levels of Jak2, Rock, Acta2, and Tgfb (B). TWNT‐4 and LX‐2 human HSCs were treated with pacritinib at 1 μm for 24 h to measure Col1a1 and p‐moesin protein levels and Col1a1 gene expression (C) and mRNA expression of Tgfb, Pdgfrb, Jak2, Rock, and Acta2 in TWNT‐4 and LX2 and Timp1, Col4a1, Pprg, and Lrat in TWNT‐4 (D). BrdU proliferation assay (E); Pcna mRNA expression (F). Results are expressed as the mean ± standard deviation; ^# p < 0.1. *p < 0.05, **p < 0.01, and ***p < 0.001. Acta2, alpha‐smooth muscle actin; BrdU, bromodeoxyuridine; Col1a1, type I collagen; Col4a1, collagen‐IV; GAPDH, glyceraldehyde‐3‐phosphate dehydrogenase; HSC, hepatic stellate cell; Jak2, Janus kinase 2; Lrat, lecithin‐retinol acyltransferase; NT, nontreated; Pac, pacritinib; Pdgfrb, platelet‐derived growth factor receptor beta; p‐moesin, phospho‐moesin; Pprg, peroxisome proliferator activated receptor gamma; Rock, RhoA/Rho‐kinase; SM22^Cre+‐Jak2^f/f, floxed‐Jak2 knockout mice with smooth muscle protein 22 Cre‐promotor; Tgfb, transforming growth factor beta; Timp1, tissue inhibitor of metalloproteinases 1; WT, wild type.

To confirm the results from rodent HSCs, we used two well‐characterized human HSC lines (LX‐2 and TWNT‐4) to investigate the role of pacritinib in vitro. First, we analyzed mRNA expression and protein levels of collagen‐1, and the results showed a significant decrease after pacritinib treatment in both human HSC lines (Figure 2C). The p‐moesin protein levels showed the same outcome (Figure 2C). Furthermore, pacritinib reduced the expression of other genes related to HSC activation, such as Tgfb and platelet‐derived growth factor receptor beta, and components of the AT1R pathway, such as JAK2 and Rock, in both cell lines (Figure 2D). Acta2 was only significantly reduced in LX‐2 HSCs (Figure 2D). Moreover, we analyzed other fibrogenic targets, such as tissue inhibitor of metalloproteinases 1, collagen‐IV, peroxisome proliferator activated receptor gamma, and lecithin‐retinol acyltransferase (Lrat) in TWNT‐4 cells (Figure 2D), which, except for Lrat, showed a significant reduction after pacritinib treatment.

Next, we evaluated the proliferation capacity in rodent HSCs and in both human cell lines with or without pacritinib incubation. Our findings show that pacritinib reduced bromodeoxyuridine (BrdU) proliferation assay and the gene expression levels of the proliferation marker proliferating cell nuclear antigen (Pcna) (Figure 2E,F). Similar effects were observed in SM22^Cre+‐Jak2^f/f mouse HSCs with minor but significant changes after pacritinib treatment (Figure 2E,F).

Jak2 inhibition with pacritinib treatment reduces contraction and migration of rodent and immortalized human‐derived HSCs.

To further underline the effects of pacritinib on HSC activation, we assessed the contraction capacity in rodent and both human cell lines, which resulted in a significant relaxation of rodent HSCs treated with pacritinib and also in SM22^Cre+‐Jak2^f/f HSC (Figure 3A). Pacritinib‐treated activated TWNT‐4 and LX‐2 cells were less contractile than nontreated cells (Figure 3A).

FIGURE 3.

JAK2 inhibition with pacritinib treatment reduces contraction and migration of rodent and immortalized human‐derived HSCs. (A) Quantification and representative images of contraction assay at 72 h, and (B) scratch assay quantification and representative pictures (B) at 24 h. Results are expressed as the mean ± standard deviation; ^# p < 0.1. *p < 0.05, **p < 0.01, and ***p < 0.001. HSC, hepatic stellate cell; JAK2, Janus kinase 2; SM22^Cre+‐Jak2^f/f, floxed‐Jak2 knockout mice with smooth muscle protein 22 Cre‐promotor; NT, nontreated; Pac, pacritinib; WT, wild type.

After showing the effect of JAK2 inhibition on fibrogenesis, proliferation, and contraction, we evaluated their migration ability. To measure the migration ability, we performed the wound‐healing assay. Although the nontreated cells closed the gap, pacritinib‐treated cells showed decreased migration, and a larger gap remained (Figure 3B). Migration ability was significantly reduced in nontreated and treated SM22^Cre+‐Jak2^f/f mice HSCs with pacritinib (Figure 3B), indicating no additional effect. Collectively, these findings indicated that pacritinib in SM22^Cre+‐Jak2^f/f mice significantly reduces the activation of HSC, demonstrated by reduced fibrogenesis, proliferation, contraction, and migration. This underlines the potential therapeutic effect of pacritinib on the key hepatic cell type that contributes to liver fibrosis progression.

Intravenous pacritinib administration ameliorates CCl₄/ethanol‐induced liver fibrosis in mice

To better understand the observed effects of pacritinib in vitro, we assessed the effect of JAK2 inhibition on hepatic fibrogenesis in an ASH model with liver fibrosis, which mimics chronic alcohol abuse, one of the main causes of chronic liver disease that can further progress to liver fibrosis. Wild‐type mice were treated with ethanol in their drinking water or in combination with CCl₄ intoxications for 7 weeks. We evaluated the effect of pacritinib administered by i.v. injection (65 μg/kg body weight [bw], twice per week) starting 3 weeks after the first CCl₄ injection for a further 4 weeks (Figure S1A}.

After induction of liver fibrosis by CCl₄/ethanol, pacritinib i.v. treatment showed a significant reduction in serum alanine aminotransferase (ALT) levels (Figure S1B) and aspartate aminotransferase (AST) levels (Figure S1C). Hematoxylin–eosin (H&E) stainings showed an improvement in the liver parenchyma after pacritinib treatment (Figure S1D, upper panel).

Next, we examined hepatic fibrosis by Sirius red (SR) stainings of collagen fibers and the activation of HSCs by alpha‐smooth muscle actin (α‐SMA) immunohistochemistry stainings. The results showed a significant decrease in SR stainings with pacritinib treatment (Figure S1D, middle panel) and a trend toward reduction in α‐SMA immunostaining (Figure S1D, lower panel). Furthermore, pacritinib significantly reduced collagen‐1 and p‐moesin, which is a surrogate for ROCK activity (Figure S1 E). Additionally, a trend toward reduction was shown in Col1a1 mRNA expression with pacritinib treatment, whereas no significant changes in Acta2 and Tgfb were observed (Figure S1F).

Given that alcohol consumption increases liver triglyceride and fat droplet accumulation, we assessed hepatic steatosis by analysis of mRNA expression levels of central regulators of lipid homeostasis, sterol regulatory element binding factor 1c (Srebp1c), and fatty acid synthase (Fasn) and the presence of small lipid droplets by oil red O staining, quantitative hepatic triglyceride (TGG) content, and liver‐to‐body weight ratio. The results revealed no changes with the pacritinib i.v. treatment (Figures S1G and S2A–C). These results suggest that i.v. pacritinib treatment reduces liver fibrosis but does not cause changes in steatosis.

Therapeutic approach of pacritinib in CCl₄/ethanol‐induced liver fibrosis in mice

Next, we examined whether oral administration of pacritinib (300mg/kg bw, three times per week) is as effective as i.v. administration because this is the route of administration in clinics (Figure 4A).

FIGURE 4.

Therapeutic approach of pacritinib in CCl₄/ethanol‐induced liver fibrosis in mice. Experimental design (A), serum aminotransferases levels ALT (B) and AST (C), hematoxylin and eosin (H&E), Sirius red staining and α‐SMA immunohistochemistry representative images and quantifications (D), protein levels (E) and mRNA expression (F) of fibrotic markers. Srebp1c and Fasn gene expression (G). Results are expressed as the mean ± SEM; ^# p < 0.1, *p < 0.05, **p < 0.01, and ***p < 0.001. Representative photomicrographs were captured at 100× (scale bars = 200 μm) and 200× magnification (scale bars = 100 μm). α‐SMA, alpha‐smooth muscle actin; Acta2, alpha‐smooth muscle actin; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CCl₄, carbon tetrachloride; Col1a1, type I collagen; ETOH, ethanol; Fasn, fatty acid synthase; GAPDH, glyceraldehyde‐3‐phosphate dehydrogenase; Jak2, Janus kinase 2; p‐moesin, phospho‐moesin; Srebp1c, sterol regulatory element binding factor 1; Tgfb, transforming growth factor beta; v:v, volume:volume.

Oral pacritinib treatment reduced ALT and AST levels (Figure 4B,C), which paralleled the reduced liver injury observed by H&E after pacritinib treatment (Figure 4D, upper panel). In addition, we analyzed the levels of inflammatory (Ccl2) and proliferative markers (Pcna), and the results showed a significative reduction after pacritinib treatment in Ccl2 gene expression (Figures S3 and S4D). F4/80 and Terminal deoxynucleotidyl transferase (TdT) dUTP Nick‐End Labeling (TUNEL) staining showed a significant reduction with oral pacritinib treatment in fibrotic mice (Figure S4E,F). SR stainings and α‐SMA immunohistochemistry quantifications revealed a significant reduction after pacritinib treatment (Figure 4D, middle and lower panel). These findings were accompanied by decreased protein levels of collagen‐1 and p‐moesin and with a trend toward reduction in α‐SMA and Jak2 (Figure 4E). Also, mRNA expression of Col1a1 and Acta2 showed a significant reduction (Figure 4F). Liver fibrosis was further evaluated by extracellular volume fraction (ECV) using quantitative magnetic resonance imaging mapping. Representative images of ECV showed a significant increase in fibrotic animals compared with the control group, and this effect was prevented with pacritinib treatment (Figure S4A). Portal pressure was measured and showed a significant reduction with oral pacritinib treatment in fibrotic mice (Figure S4C).

Additionally, significant changes were observed in mRNA expression of Srebp1c and Fasn steatosis markers (Figure 4G). Although no changes were observed in TGG hepatic content, a significant reduction was shown in oil red O staining and liver‐to‐body weight ratio after oral pacritinib treatment (Figure S2A–C), indicating that oral pacritinib has an effect on hepatic steatosis in alcoholic liver fibrosis in mice. Finally, regarding the effect of pacritinib on JAK/signal transducer and activator of transcription (STAT) signaling, we tested phospho‐STAT 3 (pSTAT3) and pSTAT5 in this experimental model, but no major impact was observed in STAT phosphorylation proteins (Figure S5). Thus, our data indicate that pacritinib treatment has beneficial effects on liver fibrosis in ASH murine models.

Oral pacritinib treatment reduces the progression of CCl₄/WD‐induced liver fibrosis in mice

Finally, we studied the effects of pacritinib in a NASH model, using a combination of CCl₄/WD, which induces liver fibrosis within 7 weeks. During the last 4 weeks, NASH mice were treated with pacritinib by oral gavage (300mg/kg bw, three times per week), based on the results observed in the ASH model (Figure 5A).

FIGURE 5.

Oral pacritinib treatment reduces the progression of CCl₄/WD‐induced liver fibrosis in mice. Experimental design (A), serum aminotransferases levels ALT (B) and AST (C), hematoxylin and eosin (H&E), Sirius red staining and α‐SMA immunohistochemistry representative images and quantifications (D), protein levels (E), and mRNA expression (F) of fibrotic markers. Srebp1c and Fasn gene expression (G). Results are expressed as the mean ± SEM; ^# p < 0.1, ⁺ p < 0.05, **p < 0.01, and ***p < 0.001. Representative photomicrographs were captured at 100× (scale bars = 200 μm) and 200× magnification (scale bars = 100 μm). α‐SMA, alpha‐smooth muscle actin; Acta2, alpha‐smooth muscle actin; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CCl₄, carbon tetrachloride; Col1a1, type I collagen; Fasn, fatty acid synthase; GAPDH, glyceraldehyde‐3‐phosphate dehydrogenase; Jak2, Janus kinase 2; Srebp1c, sterol regulatory element binding factor 1; Tgfb, transforming growth factor beta; v:v, volume:volume; WD, Western diet.

Elevations of serum ALT and AST due to CCl₄/WD decreased in groups treated with pacritinib (Figure 5B,C). Pacritinib protection against the NASH model was also shown in H&E staining (Figure 5D, upper panel). At the same time, we analyzed the gene expression of inflammation (Ccl2) and proliferative (Pcna) markers, whereby only the inflammation marker was decreased after pacritinib treatment (Figures S3 and S4D). In line with these results, F4/80 and TUNEL staining showed the same outcome (Figure 4E,F). SR staining showed a trend toward decreased collagen‐1 deposition after pacritinib treatment (Figure 4D, middle panel). Nevertheless, representative images of α‐SMA immunohistochemistry and quantification of the α‐SMA‐positive tissue area (Figure 4D, lower panel) showed protection after pacritinib treatment compared with CCl₄/WD‐treated mice. These data were accompanied by collagen‐1, α‐SMA, and p‐moesin protein expression, which revealed a tendency toward decrease after pacritinib treatment (Figure 4E). Quantitative PCR of the fibrotic genes Col1a1, Acta2, and Tgfb confirmed the protein results of reduced expression after pacritinib treatment (Figure 4F).

However, no changes in mRNA expression of steatosis markers were observed in the NASH model treated with pacritinib (Figure 4G). In addition to these findings, no changes were shown in oil red O stainings, TGG content, and liver‐to‐body weight ratios (Figure S2D–F). Finally, regarding the effect of pacritinib on JAK/STAT signaling, the same results were observed in STAT phosphorylation proteins in the NASH model (Figure S5). Overall, these results suggest that pacritinib treatment reduces fibrosis development in the NASH mouse model without changes in lipid accumulation.

DISCUSSION

n this study, we assessed the antifibrotic effect of the JAK2 inhibitor pacritinib in vitro in murine and human HSCs and in vivo in ASH and NASH murine models. Pacritinib is a type 1 inhibitor of JAKs, which targets the ATP binding site of the JAK2 under active conformation of the kinase domain.16,28 To date, three clinical trials to test pacritinib for the treatment of myelofibrosis (phase 3 PERSIST‐1 and PERSIST‐2 trials and phase 2 PAC203 trials) have been completed.29–31 It is known that pacritinib does not induce myelosuppression and exerts less immunosuppressive properties than other JAK inhibitors, therefore having the potential to be used in combination with established therapies for the treatment of other inflammatory diseases.16

Our previous findings demonstrated that the AT1R/JAK2/ROCK pathway is induced in patients with portal hypertension and decompensation in cirrhosis.5,6 In line with other findings,9 we confirm the correlation of Jak2 and Col1a1 in patients with cirrhosis in relation to liver fibrosis and portal hypertension. Interestingly, among the pacritinib targets, only JAK2 shows correlation with liver fibrosis and activation of HSC‐related genes in patients with cirrhosis. These results confirm the role of JAK2 in activated HSCs, as well as its effect on fibrosis progression.5,6 Importantly, Jak2 expression and phosphorylation were particularly common in HSCs.5 Additionally, our previous finding showed that inhibition of JAK2 using AG490 decreased hepatic resistance and thereby portal pressure in these animals, indicating that JAK2 plays an important role in portal hypertension.7 Another study demonstrated that selective inhibition of JAK2, with the antagonist TG101348 (fedratinib), not only blunted HSC activation but also macrophage‐driven inflammation in NASH, both in vitro and in vivo.9 Moreover, mice with hepatocyte‐specific deletion of JAK2 were protected from diet‐induced steatohepatitis, insulin resistance, and liver carcinogenesis.32,33

In this study, we report that the JAK2 inhibitor pacritinib reduces proliferation, migration, contraction, and gene and protein expression from fibrosis markers in two human‐derived HSCs and primary murine HSCs. In line with these in vitro findings, we treated the following two mouse models of chronic liver disease with pacritinib: the CCl₄/ethanol model to mimic ASH, which induces fibrosis, inflammation, and steatosis, and the CCl₄/WD model to mimic NASH, with a rapid progression of steatohepatitis, obesity, slight inflammation, and fibrosis. Pacritinib treatment was able to blunt or interrupt liver fibrosis development in ASH and NASH mouse models, assessed by SR stainings. HSC activation was assessed by α‐SMA immunostaining and the expression of fibrotic markers. Furthermore, pacritinib decreased hepatic steatosis markers in the ASH model, which may be explained either by reduction in ROS‐induced lipid peroxidation, protein oxidation, and consequently a reduction in DNA damage34,35 or by the anti‐inflammatory effect. Consequently, pacritinib decreased liver injury–induced aminotransferase levels in both experimental fibrotic models, suggesting that pacritinib significantly improved liver function and that it is not toxic in liver disease. Curiously, we observed that the levels of Jak2 mRNA and protein decreased with pacritinib; such as in other studies, it could be a feedforward mechanism36 or an autoinhibition mechanism.37 Moreover, other studies suggest that the inhibitory functions of Tyr kinase domain JAK homogy 2 (JH2) kinase domain of Jak kinases are important for maintaining Jak2 in a low‐activity state in the absence of ligand stimulation; autoinhibitory domains also have been proposed.38 However, the complexity of the three‐dimensional structure of Jak kinases and the different interactions of the two kinase domains (JH1 and JH2)38 have not been clearly defined. In conclusion, we provide evidence that pacritinib could be a potential for therapy in patients with liver fibrosis.

Supplementary Material

hep-77-1228-s001.tif ^{(1.2MB, tif)}

hep-77-1228-s002.tif ^{(3MB, tif)}

hep-77-1228-s003.tif ^{(110.9KB, tif)}

hep-77-1228-s004.tif ^{(1.2MB, tif)}

hep-77-1228-s005.tif ^{(148.5KB, tif)}

hep-77-1228-s006.docx ^{(44.4KB, docx)}