Tax1BP1 limits hepatic inflammation and reduces experimental hepatocarcinogenesis

The nuclear factor kappa beta (NFκB) signaling pathway plays an important role in liver homeostasis and cancer development. Tax1-binding protein 1 (Tax1BP1) is a regulator of the NFκB signaling pathway, but its role in the liver and hepatocellular carcinoma (HCC) is presently unknown. Here we investigated the role of Tax1BP1 in liver cells and murine models of HCC and liver fibrosis. We applied the diethylnitrosamine (DEN) model of experimental hepatocarcinogenesis in Tax1BP1+/+ and Tax1BP1−/− mice. The amount and subsets of non-parenchymal liver cells in in Tax1BP1+/+ and Tax1BP1−/− mice were determined and activation of NFκB and stress induced signaling pathways were assessed. Differential expression of mRNA and miRNA was determined. Tax1BP1−/− mice showed increased numbers of inflammatory cells in the liver. Furthermore, a sustained activation of the NFκB signaling pathway was found in hepatocytes as well as increased transcription of proinflammatory cytokines in isolated Kupffer cells from Tax1BP1−/− mice. Several differentially expressed mRNAs and miRNAs in livers of Tax1BP1−/− mice were found, which are regulators of inflammation or are involved in cancer development or progression. Furthermore, Tax1BP1−/− mice developed more HCCs than their Tax1BP1+/+ littermates. We conclude that Tax1BP1 protects from liver cancer development by limiting proinflammatory signaling.

The nuclear factor kappa beta (NFκB) signaling pathway plays an important role in liver homeostasis and cancer development. Tax1-binding protein 1 (Tax1BP1) is a regulator of the NFκB signaling pathway, but its role in the liver and hepatocellular carcinoma (HCC) is presently unknown. Here we investigated the role of Tax1BP1 in liver cells and murine models of HCC and liver fibrosis. We applied the diethylnitrosamine (DEN) model of experimental hepatocarcinogenesis in Tax1BP1 +/+ and Tax1BP1 −/− mice. The amount and subsets of non-parenchymal liver cells in in Tax1BP1 +/+ and Tax1BP1 −/− mice were determined and activation of NFκB and stress induced signaling pathways were assessed. Differential expression of mRNA and miRNA was determined. Tax1BP1 −/− mice showed increased numbers of inflammatory cells in the liver. Furthermore, a sustained activation of the NFκB signaling pathway was found in hepatocytes as well as increased transcription of proinflammatory cytokines in isolated Kupffer cells from Tax1BP1 −/− mice. Several differentially expressed mRNAs and miRNAs in livers of Tax1BP1 −/− mice were found, which are regulators of inflammation or are involved in cancer development or progression. Furthermore, Tax1BP1 −/− mice developed more HCCs than their Tax1BP1 +/+ littermates. We conclude that Tax1BP1 protects from liver cancer development by limiting proinflammatory signaling.

NFκB
Nuclear factor kappa beta Tax1BP1 Tax1 www.nature.com/scientificreports/ 50 ml with medium and a centrifugation step of 5 min with 50×g at 4 °C was performed. The supernatant was discarded and the cells were resuspended in 50 ml of fresh medium, followed by an additional centrifugation step at 4 °C for 5 min. The supernatant was again discarded and the cells were once again resuspended in 10 ml medium. The number of viable cells was assessed with trypan blue staining. 10 6 viable cells were seeded in 2 ml medium in each well of collagen coated six well plates. After 2 h the supernatant was removed and the attached cells were washed with phosphate buffered saline (PBS) and covered with 2 ml fresh medium containing 10% FCS. On next day the hepatocytes were stimulated with murine TNFα (10 ng/ml, Peprotech GmbH, Hamburg, Germany) for the indicated time.
Isolation and culture of mouse Kupffer cells. Kupffer cells were isolated according to Froh et al. 20 with slight modifications. For isolation of Kupffer cells, the livers of anesthetized mice were perfused as described above. After resuspension of the liver homogenate in 10 ml Williams Medium E containing 10% FCS, the suspension was centrifuged 50×g at 4 °C for 5 min. The supernatant was collected and transferred to a new 50 ml tube and a second centrifugation step with 50×g at 4 °C for 5 min was performed. The supernatant obtained at the second step was centrifuged for at 650×g at 4 °C for 7 min. The pellet was resuspended in 1 ml HBSS and added to a 15 ml tube, which was overlayed with 7 ml layers of 50% and 25% Percoll, respectively. The tube was centrifuged at 1800×g and at 4 °C for 15 min. The medium layer of the newly formed solution was transferred to a new 15 ml tube and filled up to 15 ml with HBSS. After an additional centrifugation step at 650×g and at 4 °C for 7 min, the pellet was resuspended in 1 ml HBSS. After cell counting, the macrophages were isolated with anti F80/4-biotin labelled antibodies using magnetic activated cell sorting (MACS) (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) according to the recommendation of the manufacturer. Isolated Kupffer cells were resuspended in RPMI 1640 medium containing 10% FCS and 1% penicillin/streptomycin and were plated on 6 well plates. After 1 h the medium was changed. On the next day the Kupffer cells were stimulated with LPS (300 ng/ ml, InvivoGen, Toulouse, France) for the indicated time points.
Isolation and characterization of hepatic leucocytes by flow cytometry. Characterization of immune cell subsets in the liver was performed essentially as described previously 21 . The individual samples were analyzed with a LSRII/Fortessa flow cytometer (BD Biosciences) and the FlowJo software Vx (Treestar). All indicated antibodies and reagents were titrated to determine optimal concentrations. CompBeads (BD) were used for single-color compensation to create multi-color compensation matrices. For gating, fluorescence minus one controls were used. The instrument calibration was controlled daily using cytometer setup and tracking beads  8 . After eight months the mice were sacrificed, the livers were removed, fixed in formalin and subjected to pathological examination. In detail, the right medial lobe was cut in sections of two millimeter and two hematoxylin and eosin stained slides from every section were used for determination of the number of nodules for every mouse, respectively.

Fibrosis induction with CCl 4 .
Female and male mice aged 7-8 weeks received intraperitoneal injections of 0.375 µg/g CCl 4 (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) dissolved in olive oil twice weekly for six following weeks. One week after last injection, the mice were sacrificed. Liver were excised and were fixed in buffered formalin. For fibrosis evaluation four different sections at different depths through the left liver lobe were performed and were evaluated for fibrosis development.
Histochemistry and immunohistochemistry. Murine hepatic tissue samples were formalin fixed and paraffin embedded and cut at a thickness of 4 µm. After dewaxing and rehydration the slides were stained with hematoxylin and eosin for standard histochemistry and/or Sirius red staining for fibrosis staining. For immunhistochemistry the slides were deparaffinized and rehydrated. Then slides were incubated with PBS containing 3% BSA for 60 min. Primary antibodies were added for 60 min at room temperature and after washing several www.nature.com/scientificreports/ times the biotinylated secondary antibodies were added for 30 min. After several additional washing steps with PBS, the tissue slides were incubated for 30 min with streptavidin peroxidase and the development reaction with the DAB chromogen was performed for 5-15 min. mRNA extraction and real-time PCR. Total RNA from mouse livers was extracted with the high pure mRNA extraction kit from Roche. cDNA was generated with random primers. Real-time PCR was performed on an Applied Biosystems StepOnePlus system with commercially available TaqMan Copy Number Assays for  Tax1BP1, IL-1β, IL-6, TNFα, and Glycerol-3-phosphate dehydrogenase (GPDH) (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). Expression of Tax1BP1, IL-1β, IL-6, and TNFα were normalized to GPDH expression.
Western blotting. Western blotting was performed as described previously 22 . The membranes were incubated with the SuperSignal West Pico Chemiluminescent Substrate from Thermo Scientific (Rockford, IL) and the signals were detected by the Fuji LAS-4000 detection system. The following primary antibodies were used: mouse (1:500) and rabbit (1:500) Tax1BP1 and mouse β-actin antibodies (1:5000) were from Sigma-Aldrich Chemie GmbH, Taufkirchen Gene and miRNA arrays. Materials

Results
Tax1BP1 expression in murine liver. Female and male Tax1BP1 +/− mice were crossed to obtain Tax1BP1 +/+ , Tax1BP1 +/− and Tax1BP1 −/− mice. The mice were genotyped with primer pairs as described previously 15 . Tax1BP1 is widely expressed in different organs including liver tissue 23 . Western blotting analysis of the livers from Tax1BP1 +/+ and Tax1BP1 −/− mice revealed Tax1BP1 expression in Tax1BP1 +/+ and absence of its expression in Tax1BP1 −/− mice (Fig. 1A). To assess the expression pattern of Tax1BP1 in parenchymal and non-parenchymal cells, confocal microscopy of murine liver tissue sections from Tax1BP1 +/+ mice was performed. Tax1BP1 expression could be detected in non-parenchymal cells as well as in hepatocytes (Fig. 1B).
To further analyze Tax1BP1 expression in different hepatic cell compartments, quantitative PCR of Tax1BP1 mRNA was performed in the fraction of isolated hepatocytes as well as in the fraction of non-parenchymal cells.
Additionally, quantitative PCR of Tax1BP1 mRNA was performed in isolated Kupffer cells. As shown in Fig. 1C non-parenchymal cells, and particular Kupffer cells, showed a higher Tax1BP1 expression than hepatocytes.

Increased number of inflammatory cells and sustained inflammatory responses of macrophages in livers of Tax1BP1 −/− mice.
Increased numbers of macrophages in livers of Tax1BP1 −/− mice were described previously 17 . However, a more detailed characterization of the innate and adaptive immune cells has not been performed in Tax1BP1 −/− mice. Therefore, hepatic leukocyte patterns from terminally perfused mice were characterized. FACS analysis of isolated leukocytes of Tax1BP1 +/+ and Tax1BP1 −/− mice revealed elevated numbers of leukocytes in livers of Tax1BP1 −/− mice (P = 0.028) ( Fig. 2A). Further FACS analysis of suspended liver cells revealed increased levels of immune cells of the innate as well as adaptive immune system in Tax1BP1 −/− mice (Fig. 2B). Especially, increases in regulatory T cells, CD8 positive T cells, neutrophils and monocytes were statistically significant (Fig. 2B). To further characterize innate inflammatory responses in Tax1BP1 −/− mice compared to Tax1BP1 +/+ mice, Kupffer cells from the indicated mice were isolated and stimulated with LPS. As shown in Fig. 2C-E Kupffer cells from Tax1BP1 −/− mice showed a significantly higher transcription mRNA of TNFα and IL1β compared to their wildtype littermates. The transcription of IL-6 mRNA was also increased. However, the increase was not statistically significant.
Loss of Tax1BP1 leads to sustained NFκB activation in hepatocytes. As Tax1BP1 expression was found in parenchymal as well as in non-parenchymal cells and Tax1BP1 −/− mice showed an increase in hepatic leukocytes and increased production of proinflammatory cytokines following stimulation with LPS, we investigated if hepatocytes from Tax1BP1 −/− mice show increased NFκB activation. Therefore, hepatocytes were isolated from Tax1BP1 +/+ as well as Tax1BP1 −/− mice. Tax1BP1 −/− mice showed a reduced hepatic expression of IκBα at basal conditions compared to Tax1BP1 +/+ mice (Fig. 3A). Moreover, hepatocytes from Tax1BP1 −/− mice showed higher IκBα degradation following TNFα stimulation compared to hepatocytes from Tax1BP1 +/+ mice at the indicated time points (Fig. 3A). TNFα led to increased NFκB-p65 phosphorylation and to a more profound decrease in p65 levels from to the cytoplasmatic fraction upon stimulation in Tax1BP1  www.nature.com/scientificreports/ www.nature.com/scientificreports/ Tax1BP1 +/+ mice, respectively (Fig. 3B). Furthermore, hepatocytes from Tax1BP1 −/− mice tended to show a faint increase in JNK activation following TNFα challenge compared to hepatocytes from Tax1BP1 +/+ mice (Fig. 3A).
CCl 4 induced liver fibrosis. As Tax1BP1 −/− mice showed an inflammatory phenotype, we deduced that experimental fibrosis development might be affected by Tax1BP1. Therefore, fibrosis was induced in Tax1BP1 +/+ mice as well as in their Tax1BP1 −/− littermates by CCl 4 treatment. In each group eight mice were used with 50% of female mice. Both mouse strains developed fibrosis after CCl 4 treatment. There was no difference in the stage of fibrosis between Tax1BP1 −/− mice and the Tax1BP1 +/+ mice seven weeks after start of fibrosis induction, respectively ( Fig. 4C-F). Exemplary untreated mouse tissue sections of livers of Tax1BP1 +/+ and Tax1BP1 −/− and mice are displayed in Fig. 4A,B.
DEN induced hepatocellular carcinoma. The NFκB signaling pathway plays an important role in liver homeostasis and inflammation. Dependent on its activity in the different hepatic cell compartments, it may limit tumor formation but it can also promote carcinogenesis. As Tax1BP1 was expressed in the different hepatic cell www.nature.com/scientificreports/ types, we concluded that it could be an important factor in hepatic carcinogenesis. To investigate if Tax1BP1 might modulate liver cancer development, hepatocarcinogenesis was induced in Tax1BP1 +/+ , Tax1BP1 +/− and Tax1BP1 −/− mice using DEN treatment. After eight months, the livers were removed and the number of HCCs and preneoplastic tumors including dysplastic foci and nodules was assessed by a pathologist in a blinded manner. Tax1BP1 +/+ , Tax1BP1 +/− as well as Tax1BP1 −/− developed pre-neoplastic tumors and HCCs. There were significant differences among Tax1BP1 +/+ , Tax1BP1 +/− and Tax1BP1 −/− mice in the number of HCCs with Tax1BP1 −/− mice showing significantly more HCCs (Fig. 5A). Tax1BP1 −/− mice also developed more preneoplastic lesions compared to their Tax1BP1 +/+ and Tax1BP1 +/− littermates. However, this difference was not statistically significant (Fig. 5B). Representative macroscopic and microscopic images for Tax1BP1 +/+ and Tax1BP1 −/− mice are shown (Fig. 5C,D).

Discussion
The equilibrium between cell death and survival is decisive for organ homeostasis. Dysregulations of both ways may be fatal leading either to massive cell death, cancer development or both 6,9 . The NFκB signaling pathway is a central regulator of cell death and survival and its effects on cells and tissues are highly dependent on the cell type 24,25 . In the DEN model of experimental HCC, NFκB activity in hepatocytes protects from cancer development, whereas the activation of this pathway in the bone marrow derived cells promotes cancer development and progression 6-8 . Tax1BP1 is highly expressed in myeloid derived cells and the expression of Tax1BP1 in bone marrow derived cells determines its phenotype as NFκB inhibiting protein in an experimental model mimicking sepsis 15 , and Tax1BP1 −/− mice show a spontaneous hepatic inflammatory phenotype 17 . Here we found an increased number of cells of the innate and adaptive immune system in the livers of Tax1BP1 −/− mice and an enhanced expression of proinflammatory cytokines such IL1β and TNFα in isolated Kupffer cells derived from Tax1BP1 −/− mice after LPS stimulation. This supports the observation that the livers of Tax1BP1 −/− mice show an inflammatory phenotype. Nevertheless, we further show that Tax1BP1 is not only expressed in non-parenchymal hepatic cells but also in hepatocytes and that loss of Tax1BP1 expression was associated with a stronger activation of the NFκB signaling in isolated hepatocytes. Interestingly, we also found a slightly increased SAPK/JNK activity in hepatocytes from Tax1BP1 −/− mice in comparison to their wildtype littermates after TNFα challenge. The JNK pathway plays a critical role in liver disease, and hyperactivation or loss of the pathway may influence inflammation and cancer development 26 . Two different Tax1BP1 knock-out mouse models revealed differences concerning the activation of the SAPK/JNK pathway 15,27 . In our current mouse model no relevant increase in JNK activation was observed in macrophages in contrast to another Tax1BP1 knock-out model 15,27 . Namely, we found a slight increase in JNK phosphorylation in hepatocytes from Tax1BP1 −/− mice, indicating that the extent of affection of this pathway by Tax1BP1 might be cell-type dependent. www.nature.com/scientificreports/ Nonetheless, Tax1BP1 seems to play its major role in immune cells, as we found an increase of cells of the innate as well as adaptive immune systems in Tax1BP1 −/− mice supporting the role of Tax1BP1 as an important regulator of inflammation. Furthermore, it has recently been shown that Tax1BP1 is not only an adaptor for deubiquitinating enzymes such as A20 15 , but it also affects adaptive immunity including B and T cells by regulating autophagy 28,29 . Tax1BP1 −/− mice exhibited increased levels of lymphocytes and especially CD8 positive T cells and regulatory T cells were significantly augmented. The increase in the number of cells of the innate and adaptive immunity in livers Tax1BP1 −/− mice of came along increased expression of regulators of inflammation and immune cell differentiation such as BCL6, CXCL13, KLF10, and Lysozyme 2 in Tax1BP1 −/− mice in array based analyses from liver tissue. miRNAs are epigenetic regulators of gene expression. In our array based miRNA analyses we found several miRNAs up-or downregulated in Tax1BP1 −/− mice. Some of these are well known as regulators of inflammation or are involved in cancer development or progression. Namely, miRNAs such as www.nature.com/scientificreports/ miR-31 and miR-192 which limit inflammation and cancer development [30][31][32] , were downregulated and miR-92, which increases liver tumor development 33 was overexpressed in Tax1BP1 −/− mice. Interestingly, we did not observe increase fibrosis development in Tax1BP1 −/− mice with the CCL 4 model we used. This was unexpected as NFκB activiation is associated with inflammation and fibrosis development 24 . Therefore, additional models of fibrosis such as bile duct ligation and other concentrations of CCl 4 might be used to further decipher the effect of Tax1BP1 in fibrosis development. Increased NFκB activation is an early event in HCC development 34 . However, the role of NFκB as a promotor or an inhibitor seems to be rather cell type specific 24 . In transformed hepatocytes, inhibition of NFκB signaling attenuates tumor progression 6 , whereas ablation of central NFκB signaling pathways leads to spontaneous HCC development 7 , or may trigger JNK activation, which promotes carcinogenesis 35 . Our Tax1BP1 −/− mice showed an increased activation of the NFκB signaling pathway in hepatocytes as well as increased numbers of inflammatory cells leading to a spontaneous inflammatory phenotype in the liver. Inhibition of NFκB signaling by ablation of IKKβ in myeloid derived cells reduces HCC development in mice 8 . Therefore, our observation of an increased development of HCCs in Tax1BP1 −/− mice with increased numbers of non-parenchymal liver cells and increased levels of proinflammatory cytokines upon stimulation of macrophages fits in the theory of a tumor promoting effect of NFκB in myeloid derived cells by enhancing anti-apoptotic signaling in hepatocytes triggered by Kupffer and hepatic stellate cells predisposes transformed cells to cancer development [3][4][5] .
Another interesting aspect of Tax1BP1 is that Tax1BP1 has been described as a regulator of antiviral signaling. Together with A20, it targets the TBK1-IKKi kinases and restricts INFβ production 36 . IKKi/IKKepsilon has an important antiviral function in chronic hepatitis C virus infection, as it limits viral replication and itself is down regulated in infected cells 37 . Furthermore enhanced genomic amplification of Tax1BP1 has been described in a subset of hepatocellular carcinomas, which developed in patients with hepatitis C virus infection 18 . Therefore, further investigation of Tax1BP1 in chronic liver disease may be of great interest and may open new vistas on this protein and its role in diseases.