Survival and Inflammation Promotion Effect of PTPRO in Fulminant Hepatitis Is Associated with NF-κB Activation

Fulminant hepatitis (FH) is a rare, but potentially fatal, disease in which the patient’s condition rapidly deteriorates; it can cause hepatic encephalopathy, necrosis of the hepatic parenchyma, coagulopathy, renal failure, and coma. Without supportive management and/or liver transplantation, the mortality is ∼70% (1). T cell–mediated liver damage is usually considered one of the principal pathological changes of FH. Hepatitis B or C viruses and autoimmune factors are the most common causes of lethal T cell–mediated liver damage in humans (2, 3). It is difficult to study FH using human samples because of its short duration and high mortality. Therefore, various animal models of T cell–mediated liver injury have been developed, including an FH model in mice that is induced by i.v. injection of the T cell–stimulatory plant lectin Con A (3).

Protein tyrosine phosphatase, receptor type, O (PTPRO) is an integral membrane protein that contains a single intracellular catalytic domain with a characteristic signature motif. There are several alternatively spliced variants that make up the tissue-specific isoforms of PTPRO (4). For example, hepatocytes contain the full-length PTPRO protein, whereas it appears as a truncated form (PTPROt) in T cells, which are some of the main infiltrating inflammatory cells during FH. Most studies of PTPRO’s role focused on cancer research; a common goal has been to define PTPRO as a tumor suppressor candidate (5–8). Very recently, we (5) reported that PTPRO inhibits development of hepatocellular carcinoma by suppression of JAK2/STAT3 signaling; however, its role in liver inflammation is largely unknown. In this study, we investigated the effect of PTPRO in both mouse and human FH. Our findings demonstrate that PTPRO is associated with the promotion of hepatocyte survival and proinflammatory effects by NF-κB activation via PTPRO/ErbB2/Akt/GSK-3β/β-catenin/NF-κB signaling.

Twenty-four patients who underwent liver transplantation because of FH at the First Affiliated Hospital of Nanjing Medical University from January 2009 to December 2012 were included in the study. Normal liver parenchymal tissues, used as controls, were obtained from 24 Chinese patients with benign disease, such as hemangioma. Informed consent for gene expression analysis was obtained from all patients prior to surgery. The study was approved by the institutional ethics committee of the First Affiliated Hospital of Nanjing Medical University in accordance with the Declaration of Helsinki.

Eight- to ten-week-old male C57BL/6 mice were purchased from the animal center of Nanjing Medical University. The PTPRO-knockout (KO) C57BL/6 mice were donated by Dr. John L. Bixby (University of Miami, Miami, FL) and were maintained and bred in the animal center of Nanjing Medical University. All mice were maintained under specific pathogen–free conditions in the animal facilities.

All animals received care in compliance with the guidelines outlined in the Guide for the Care and Use of Laboratory Animals. In liver injury studies, mice received a Con A injection (20 μg/kg body weight) via the tail vein, as described previously (9). Detailed indices of liver injury were analyzed at 6, 9, 12, and 24 h after Con A injection.

Serum samples from mice were obtained at different times. Serum alanine aminotransferase (ALT) activities were determined using the serum transaminase test kit (Rong Sheng, Shanghai, China), based on methods recommended by the International Federation of Clinical Chemistry.

Reverse-transcription reactions were performed using the SuperScript First-Strand Synthesis System (Invitrogen, Carlsbad, CA). RNA templates were treated with DNase to avoid genomic DNA contamination. To quantify cDNA in the reverse-transcribed samples, real-time PCR analysis was performed using a 7300 Detection System (Applied Biosystems, Foster City, CA). Real-time PCR was performed according to the manufacturer’s instructions with a SYBR Premix Ex Taq kit (TaKaRa, Shiga, Japan). The primers used to detect various genes are provided in Table I. Data were normalized to GAPDH levels in the samples.

Liver mononuclear cells (MNCs) were isolated as described (10). Briefly, mouse livers were removed and pressed through a 100-μm cell strainer (BD), suspended in RPMI 1640 medium (Life Technologies, BRL), and centrifuged at 50 × g for 5 min. Supernatants containing hepatic MNCs were collected, washed in PBS, and resuspended in 40% Percoll (Sigma-Aldrich, St. Louis, MO) in RPMI 1640 medium. The cell suspension was gently overlaid onto 70% Percoll and centrifuged for 20 min at 750 × g. MNCs were collected from the interphase and washed twice in PBS.

For EMSAs, liver nuclear extracts were prepared, and the binding reaction was performed using a Light Shift Chemiluminescent EMSA Kit (Thermo Scientific, Waltham, IL), according to the manufacturer’s instructions. Free DNA and DNA protein complexes were resolved on a 6% polyacrylamide gel. A biotin-labeled oligonucleotide representing an NF-κB consensus site (5′-TAGTTGAGGGGACTTTCCCAGGCA-3′) was used as probe.

Mouse livers were perfused in situ with 45 ml Life Technologies Liver Perfusion Media, followed by 45 ml Life Technologies Liver Digestion Media (both from Invitrogen). The liver digests were filtered through a cell strainer and washed with Gey’s Balanced Salt Solution (Sigma-Aldrich) containing DNase I (2 mg/ml; Roche Diagnostics). The homogenate was centrifuged at 25 × g for 5 min at room temperature to obtain the hepatocytes. Isolated hepatocytes were washed thoroughly with PBS. After determination of cell viability (95%), hepatocytes were plated on 6-cm dishes and maintained in DMEM (Sigma-Aldrich) containing 10% FCS and a combination of penicillin and streptomycin for further analysis.

A total of 5 × 10⁶ hepatocytes was isolated from wild-type (WT) and PTPRO mice liver that was treated or not with TNF-α. Six hours later, the cells were lysed by incubation for 20 min at 4°C with RIPA plus a protease inhibitor mixture (Roche, Basel, Switzerland). The IKK complex was immunoprecipitated using Ab for IKKα and IKKβ (5 μg each). An enzymatic reaction was performed using 40 μl IKK complex, 2 μg GST-IκBα (1-45), and 0.5 μmol ATP and incubation at 30°C for 2 h. GST-IκBα was purified by GST Magnetic Beads (Promega, Madison, WI), size separated by SDS-PAGE, and probed by Western blotting with anti–p-IκBα (Abcam, Cambridge, MA).

The TUNEL test was performed with a TUNEL Apoptosis Detection Kit (Millipore), according to the manufacturer’s instructions, using a fluorescence microscope. H&E staining was performed using a standard protocol for cryosections. Immunohistochemistry was performed on paraffin-embedded sections, frozen sections, or isolated hepatocytes fixed on slides using Abs against TNF-α, IFN-γ, IL-6, PTPRO, p-IκBα, p-p65, p-ErbB2, and β-catenin (Santa Cruz Biotechnology, Santa Cruz, CA).

Mouse serum was collected and assessed using mouse epithelial growth factor (EGF) (Abcam) and TNF-α and IFN-γ ELISA kits (eBioscience). The experiments were performed in strict accordance with the instructions provided by the manufacturer.

Six- to ten-week-old WT and PTPRO-KO mice were used as donors. Recipients were gamma irradiated with a cesium source (two 450-rad doses 4–5 h apart). A total of 1 × 10⁶ whole bone marrow cells isolated from the donor mice was transplanted into the recipients, and hematologic parameters were assessed at 8 wk posttransplant. Transplanted mice were maintained on 2.5 mg/100 ml Sulfatrim Pediatric Suspension (Alpharma, Baltimore, MD).

Proteins were extracted from mouse tissues and cells and quantitated using a protein assay (Bio-Rad, CA). Protein samples (30 μg) were fractionated by SDS-PAGE and transferred to a nitrocellulose membrane. Immunoblotting was conducted using Abs against PTPRO, p-p65 (S536), p65, p-IκBα (s32+s36), IκBα, XIAP, p-IKKα/β (S180+S181), total IKKα/β, p-ErbB2 (Y1248), ErbB2, p–β-catenin (Y654), β-catenin, p–GSK-3β (S9), GSK-3β, and TCF-1 (Santa Cruz Biotechnology). The results were visualized using a chemiluminescent detection system (Pierce ECL Western Blotting Substrate detection system; Thermo Scientific) and exposure to autoradiography film (Kodak XAR film).

The isolated hepatocytes from WT or PTPRO-KO mice were treated with TNF-α (100 ng/ml) for various times and collected for apoptosis detection following the instructions for the PI–Annexin V staining kit (Invitrogen).

In order to detect the transcription of inflammatory related molecules, fresh liver tissues were obtained from the mice 12 h after Con A treatment. Reactions were carried out with an ABI 7900 HT thermal cycler using a TaqMan Array Mouse Immune Panel (4367786; both from Applied Biosystems). To screen for the differences in activity in receptor tyrosine kinases (RTKs) within TNF-α–treated hepatocytes from WT and PTPRO-KO mice, protein was extracted and analyzed according to the instructions for the Mouse Phospho-Receptor Tyrosine Kinase Array (cat. no. ARY014; R&D Systems).

Results are expressed as mean ± SD. Comparisons between two groups were performed using the unpaired Student t test. Correlations between parameters were determined using Pearson correlation and linear regression analysis, as appropriate. All statistical analyses were performed using SPSS statistical software version 13.0, and two-tailed tests were applied to all data unless otherwise specified. A p value < 0.05 (95% confidence interval) was considered statistically significant.

Paired PTPRO-KO mice and littermate WT mice were injected with Con A. Significantly elevated serum ALT activity indicated that severe hepatic injury had occurred in PTPRO-KO mice (Fig. 1A). Liver damage was more severe in PTPRO-KO mice compared with WT mice; representative H&E results for both groups 24 h after Con A administration are shown in Fig. 1B1. Multiple necrosis and only a few integrated hepatic lobules were detected in livers of PTPRO-KO mice, and there also were dramatic differences between WT and PTPRO-KO mice at other time points (Supplemental Fig. 1A). Moreover, Suzuki’s criteria (10) were used to systematically evaluate the liver damage, which also indicated that significant serious injury had occurred as a result of depletion of PTPRO (Fig. 1B2). Our findings showed that apoptosis correlated with increased liver injury, as detected by TUNEL assay. Significantly increased apoptosis in PTPRO-KO liver also was seen compared with WT liver at each time point (Fig. 1C1, 1C2). MNCs were isolated from mouse livers at 9 and 24 h following Con A injection. As controls, WT and PTPRO-KO mice were injected with PBS and sacrificed for MNC isolation. Inflammatory cell infiltration increased significantly in the livers of both WT and PTPRO-KO mice 9 and 24 h after Con A injection, but the numbers of MNCs decreased significantly in PTPRO-KO mice compared with WT mice at these same time points (Fig. 1D). Furthermore, the expression levels of a set of inflammatory molecules were analyzed with low-density real-time PCR array using mRNA from total livers of KO and WT mice 12 h after Con A injection. The results were unexpected. As shown in Fig. 1E and Supplemental Fig. 1B, expression of proinflammatory cytokines, such as TNF-α, IFN-γ, IL-1β, and IL-6, decreased in KO mice compared with WT mice, and the expression of chemokines, such as ccl2, ccl3, and cxcl10, also decreased, which may explain the decrease in infiltrating cells in PTPRO-KO mice. Further investigations were performed to verify the differences among the three indices in which we were interested. The differences in TNF-α, IFN-γ, and IL-6 were confirmed by real-time PCR and immunohistochemistry, with both indicating significantly decreased secretion in PTPRO-KO mice (Supplemental Fig. 1C, 1D). Moreover, decreased expression of TNF-α, IFN-γ, and IL-6 in MNCs extracted from PTPRO KO mice compared with WT controls was observed by both real-time PCR and Western blot (Fig. 1F1, 1F2).

Based on the cytokine-expression array results, we hypothesized that cell signaling also may affect the attenuated inflammation and decreased antiapoptotic abilities. Based on established knowledge about inflammation, particularly acute hepatitis, we investigated the status of NF-κB signaling in both groups.

To test this hypothesis, we first examined phosphorylation of NF-κBp65 (p-p65) and IκBα (p-IκBα), as well as expression of XIAP, an antiapoptotic protein that is one of the transcriptional products of NF-κB, in mouse livers. The levels of p-p65, p-IκBα, and XIAP were significantly decreased in PTPRO-deficient mice compared with WT controls (Fig. 2A, 2B, Supplemental Fig. 1E), suggesting that NF-κB activity was impaired in PTPRO-deficient mice. Furthermore, EMSA was used to assess NF-κB activation, and blockage of NF-κB activation in the liver of PTPRO-KO mice was confirmed (Fig. 2C).

To further study the blockage effect of NF-κB activation as a result of PTPRO depletion, hepatocytes were isolated and maintained in vitro. TNF-α–induced NF-κB activation was detected by immunofluorescence assay for p65 nuclear translocation (Fig. 2D1); nuclear translocation of p65 was quite evident, whereas it was blocked or attenuated in PTPRO-KO hepatocytes (Fig. 2D2). Meanwhile, apoptosis was detected by a PI–Annexin V costaining assay. No spontaneous apoptosis was detected in WT or KO hepatocytes; however, significantly increased apoptosis was observed in PTPRO-KO hepatocytes (Fig. 2E, Supplemental Fig. 2A). Protein expression analysis by Western blot gave similar results; decreased p65 was detected in the nucleus, and decreased p-IκBα and XIAP were found in PTPRO-KO hepatocytes that underwent 6 or 24 h of TNF-α treatment. Surprisingly, the activation of IKKα/IKKβ, which are upstream kinases for NF-κB, was nearly intact in PTPRO hepatocytes compared with WT controls (Fig. 2F, Supplemental Fig. 2B), suggesting that an alternative pathway regulates NF-κB activation in PTPRO-KO hepatocytes.

Con A–induced hepatitis is characterized by massive hepatocellular degeneration and lymphoid infiltration of the liver; these depend primarily upon T and NK cell activation, which increases plasma levels of various cytokines, including IFN-γ and TNF-α. Liver MNCs were isolated from WT and KO mice, and cytokine secretion was determined by flow cytometry. MNCs were maintained in vitro and stimulated by Con A (10 μg/ml) for 24 h; decreased secretion of both IFN-γ and TNF-α was observed in T cells (Fig. 3A1, Supplemental Fig. 2C). Important subsets of hepatic inflammatory cells, NK/NKT cells, also were analyzed; significantly decreased cytokine secretion was observed, as well (Fig. 3A2, Supplemental Fig. 2D). CD4⁺ and CD8⁺ T cells were isolated from peripheral blood of WT and PTPRO-KO mice, and IFN-γ and TNF-α secretion was determined by ELISA with Con A treatment and controlled by PBS; secretion of both cytokines decreased significantly in PTPRO KO CD4⁺ and CD8⁺ T cells with Con A treatment (Fig. 3B1, 3B2). Next, CD4⁺ T cell subsets within MNCs were analyzed; there was no significant difference in the percentages of Th17 or regulatory T cells within liver MNCs after Con A treatment (Supplemental Fig. 3A). However, the percentage of Th1 cells was significantly decreased within the MNCs as a result of PTPRO depletion (Fig. 3C), which may have resulted from decreased IFN-γ secretion. To further study the mechanism of PTPROt, a truncated form of PTPRO, in hepatic inflammation (11, 12), T cells and NK/NKT cells were isolated from mouse spleen. In addition, the different forms of PTPRO, full-length PTPRO in hepatocytes and PTPROt in inflammatory cells, were confirmed by RT-PCR and Western blot (Table I, Supplemental Fig. 3B). ELISPOT was used to evaluate functional changes in PTPROt-depleted inflammatory cells. Significant decreases in both IFN-γ and TNF-α were found in PTPROt-depleted T and NK/NKT cells compared with WT cells (Fig. 3D1, 3D2); this confirmed the results from in vivo studies and flow cytometry. The protein also was extracted from both T and NK/NKT cells, and the results were similar to hepatocytes; NF-κB signaling also was suppressed in inflammatory cells and decreased phosphorylation of p65 was found in both T and NK/NKT cells extracted from PTPRO-KO spleen treated with Con A. Phosphorylated IKKα/IKKβ increased when T and NK/NKT cells were stimulated with Con A, but there was no marked difference between WT and PTPRO-KO groups, which was similar to hepatocytes (Fig. 3E).

Based on the above results, we hypothesized that PTPRO might be associated with survival of hepatocytes and secretion of inflammatory cells in the mouse model of FH. To test this hypothesis, bone marrow transplantation (BMT) was performed. As indicated in Fig. 4A, bone marrow samples from PTPRO-KO and WT mice were collected and transplanted into each other; the recipient mice were identified as BMT from WT mice to PTPRO-KO mice (W-P) and BMT from PTPRO KO mice to WT mice (P-W). WT BMT to WT was used as a control. First, the expression of PTPROt in peripheral MNCs was detected by Western blot analysis, indicating the success of BMT (Supplemental Fig. 3C). Next, the Con A–induced mouse FH model was set up; the three groups of mice were sacrificed for analysis 12 h following Con A administration. Serum ALT was assessed; the results revealed that W-P mice exhibited the most severe hepatic injury, whereas the livers of P-W mice exhibited the least damage (Fig. 4B). The histological results were in agreement with serum ALT analysis. Serious liver structural damage occurred in W-P mice, whereas there was almost no obvious damage in P-W mice (Fig. 4C); the difference in liver damage was reflected by Suzuki scores (Fig. 4D). The apoptosis of hepatocytes was evaluated by TUNEL assay, and the significance of the differences between W-P and the other two groups was assessed (Fig. 4E, Supplemental Fig. 3D). Without PTPRO, the hepatocytes were not exposed to the impact of cytokines, such as TNF-α; moreover, the function of inflammatory cells was restored after BMT. The degree of liver damage in W-P mice was even more marked in comparison with PTPRO-KO mice. In conclusion, PTPRO is associated with survival and function of inflammatory cells during mouse FH in association with NF-κB activation.

To identify a factor that is responsible for NF-κB suppression in PTPRO-deficient cells, a mouse RTK Ab array was used to detect TNF-α–treated hepatocytes isolated from both WT and PTPRO KO mice. We found that loss of PTPRO led to increased activation of several RTKs, including ErbB2, IGFr-1, Axl, Tie1, and Tie2 (Supplemental Fig. 4A). Among these activated RTKs, ErbB2 might serve as an ideal candidate because it can act either as a direct substrate of PTPRO or as an agonist for Wnt–β-catenin signaling (8, 13) and because EGF/ErbB2 signaling was involved in liver biology (14). Also, recent studies suggested that the accumulation of β-catenin can block activation of NF-κB via direct binding (15, 16). Thus, we reasoned that ErbB2 and β-catenin may mediate the suppression of NF-κB by PTPRO deficiency.

To test this scenario, the expression and activation of both ErbB2 and Wnt–β-catenin signaling were investigated in mouse liver tissues; markedly increased activation of ErbB2 and accumulation of β-catenin were verified in the livers of PTPRO-KO mice that were treated with Con A for 9 or 24 h. Moreover, phosphorylation of Akt and GSK-3β, as well as accumulation of their downstream targets β-catenin and TCF-1, was largely induced in PTPRO-deficient cells (Fig. 5A–C, Supplemental Fig. 4B). Furthermore, the hepatocytes were isolated and treated with TNF-α. Activation of Wnt–β-catenin in WT hepatocytes was blocked, whereas nuclear translocation of p65 was quite obvious; in contrast, activation of Wnt–β-catenin in KO hepatocytes was significantly increased, and nuclear translocation of p65 was suppressed. Moreover, the inhibitors of Wnt–β-catenin and ErbB2 also were involved. When PTPRO-KO hepatocytes were treated with lapatinib, an universal inhibitor for the EGFR family, including ErbB2, decreased activation of Wnt–β-catenin and increased p65 nuclear translocation were observed. However, lapatinib cannot completely compensate for the suppression of NF-κB in PTPRO-KO hepatocytes, which might indicate that ErbB2 was just one of the RTKs promoting Wnt–β-catenin signaling. However, when Wnt–β-catenin was blocked by its inhibitor, FH535, nuclear translocation of p65 increased markedly, indicating that Wnt–β-catenin potentially plays a dominant role in PTPRO-regulated NF-κB suppression (Fig. 5D). Meanwhile, the functions of both T cells and NK/NKT cells were restored to a varying degree when treated with lapatinib and FH535 (Fig. 5E). In addition, the nuclear existence of p65 was analyzed in hepatocytes, T cells, and NK/NKT cells treated with TNF-α or Con A. Lapatinib only partially rescued p65 nuclear translocation, whereas FH535 restored it almost completely (Fig. 5F). Furthermore, coimmunoprecipitation (co-IP) assays were performed to confirm whether β-catenin can inhibit NF-κB by direct binding within hepatocytes. In PTPRO-KO hepatocytes, with or without TNF-α treatment (Fig. 5G1), β-catenin interacts with p65, which is in accordance with the outcome of previous research (15).

To determine whether PTPRO directly regulates ErbB2 activity, we first examined the serum level of EGF in both WT and PTPRO-deficient mice. Although the serum level of EGF was almost equal in WT and KO mice (Supplemental Fig. 4C), activation of ErbB2 by EGF clearly increased in PTPRO-deficient hepatocytes (Supplemental Fig. 4D). A further immunoprecipitation assay revealed that PTPRO was associated with ErbB2 in hepatocytes (Fig. 5G2), and our data confirmed previous results obtained from mammary epithelial cells that PTPRO can inhibit ErbB2 activation by direct binding (8).

Taken together, our data suggest that negative regulation of NF-κB by PTPRO is mediated by the ErbB2/Akt/Gsk-3β/β-catenin axis.

To investigate the role of PTPRO in human FH, 24 patients who suffered from FH and had undergone liver transplantation were studied. Paraffin-embedded liver tissues were used for immunohistochemistry staining. The expression of PTPRO, p–ErbB-2, β-catenin, p-p65, and IFN-γ were detected and quantified by image software. Compared with normal liver tissues, the expression of IFN-γ, p-ErBb2, and β-catenin was significantly increased. Interestingly, there was almost no PTPRO expression in hepatocytes, whereas high expression levels of PTPROt were detected in nonparenchymal cells. Similarly to PTPRO, p-p65 was very low in hepatocytes, whereas strong staining was evident in inflammatory cells (Fig. 6A, 6B). After statistical analysis, we found that PTPRO was negatively correlated with the expression of β-catenin (Fig. 6C) and positively correlated with INF-γ expression (Fig. 6D) in the liver tissues of human FH. Therefore, our data suggested that low expression of PTPRO in hepatocytes and high expression of PTPROt in inflammatory cells may be associated with massive apoptosis and strong inflammation in human FH.

Numerous studies suggested that PTPRO acts as a tumor suppressor because of its characteristic kinase inhibition. Low or aberrant expression of PTPRO has been verified in many kinds of human tumors, including lung cancer, hepatocellular carcinoma, colon cancer, and breast cancer (3, 4, 6, 17). A variety of its targets, which are involved in essential tumor-associated cell signaling, has been explored, and most of these also play an essential role in inflammation (18, 19). However, whether PTPRO is directly involved in regulating inflammatory responses is unclear.

Other protein tyrosine phosphatases were reported to play a regulatory role in inflammation (20). For example, loss of T cell protein tyrosine phosphatase expression results in synovitis, with several hallmarks of inflammatory arthritis, within mouse femurs (21). In the current study, we demonstrated that PTPRO is a critical regulator in both human and mouse FH. Low or depleted expression of PTPROt was associated with impaired cell function in both T and NK/NKT cells; secretion of IFN-γ, TNF-α, and other proinflammatory factors was associated with expression of PTPROt. A novel PTPRO/PTPROt/ErbB2/Akt/Gsk-3β/β-catenin/NF-κB signaling pathway was identified within inflammatory cells and hepatocytes.

The strong increase in transaminase levels and more serious liver damage in PTPRO-KO mice suggested a protective role for PTPRO during Con A–induced acute liver injury. To test the functional relevance of PTPRO in hepatocytes, we stimulated them with TNF-α. Two pathways exist in hepatocytes following treatment with TNF-α: the survival pathway and the death pathway. The survival pathway is TRAF2-mediated, classic NF-κB–signaling activation that triggers transcription of a series of antiapoptosis genes, such as XIAP, Bcl-2, A-20, and c-Flip (22–24). As shown in this simple and feasible TNF-α–treated hepatocyte model, PTPRO is an essential factor that maintains the survival pathway. In contrast to WT control, TNF-α induces massive hepatocyte apoptosis in association with decreased NF-κB activation. As mentioned above, NF-κB is essential to trigger antiapoptotic gene expression in hepatocytes, which explains the lack of protection and increased apoptosis in PTPRO-KO mice. However, WT and PTPRO-deficient hepatocytes exhibited a similar activation of IKKα/IKKβ, whereas NF-κB activation was dramatically decreased in PTPRO-deficient hepatocytes, indicating that other factors are involved in regulating NF-κB activation.

Our results indicated that activation of NF-κB signaling was impaired in hepatocytes, as well as in inflammatory cells, such as T and NKT cells, as a result of PTPRO or PTPROt depletion. Impairment of NF-κB in inflammatory cells could result in decreased secretion of cytokines that reflected as decreased inflammation, whereas impaired NF-κB activation in hepatocytes could decrease their antiapoptotic ability during the inflammation strike, which could result in increased liver damage. Therefore, more severe liver damage was observed in PTPRO-KO mice with Con A treatment, even when the inflammation was less severe compared with WT mice.

Indeed, using a mouse RTK array, we identified that ErbB2, a member of the EGFR/ErbB family, was activated within liver with FH. It was suggested that amplification or overexpression of this gene has been shown to play an important role in the pathogenesis and progression of certain types of cancer (17, 25, 26). In breast cancer, ErbB2 was found to upregulate β-catenin via inactivation of GSK-3β by the Erk pathway (13). Also, in hepatocellular carcinoma, upregulation of ErbB-2 expression increased the existence of β-catenin (13, 27). Further, a study suggested that serum levels of ErbB2 might be associated with liver dysfunction (28). However, a role for ErbB2 in human liver disease is largely unclear. In the current study, we found that p-ErbB2 was mostly positive (75%) in human FH tissues and identified a PTPRO/ErbB2/Akt/β-catenin/NF-κB cascade (Fig. 6E).

β-Catenin is part of a complex of proteins that make up adherens junctions. Recent evidence suggested that β-catenin plays an important role in liver biology, including liver development, liver regeneration, and pathogenesis of liver cancer (29–31). Aside from its role in cancer, several studies focused on its role in acute liver injury; they all indicated that high expression of β-catenin can induce apoptosis through different mechanisms. Li et al. (32) reported that high levels of β-catenin in the HCC cell line can induce more rapid apoptosis compared with low β-catenin cells treated by IFN-γ as a result of p53 accumulation and Bcl-X_L downregulation. Another study using a mouse hepatic ischemia-reperfusion model revealed that activation of β-catenin inhibited the TLR4-driven inflammatory response in dendritic cells by triggering PI3K/Akt signaling (33); the phenotype is similar to our model in which the accumulation of β-catenin in T or NK/NKT cells impairs their secreting function. Moreover, Zou et al. (16) demonstrated that the loss of β-catenin in hepatocytes significantly reduced their apoptosis in GalN/LPS-induced liver damage; its effect was attributed to inhibition of NF-κB signaling. Another study observed that the p65/β-catenin complex in hepatocytes undergoes dynamic changes during TNF-α–induced hepatic injury and plays a critical role in NF-κB activation and cell survival (15). The dynamic changes in p65/β-catenin complexes were impaired as a result of overactivated ErbB2 in PTPRO-depleted hepatocytes, which, in turn, activated Akt, but inactivated GSK-3β, leading to accumulation of β-catenin. Accumulated β-catenin formed a complex with NF-κB and blocked its nuclear translocation (Fig. 6E).

In summary, we identified that PTPRO is critical for FH development and that it negatively regulates NF-κB in FH via activation of the ErbB2/Akt/GSK-3β/β-catenin cascade. Our data suggest that PTPRO may be a therapeutic target for FH.

The authors have no financial conflicts of interest.