Small heterodimer partner negatively regulates C-X-C motif chemokine ligand 2 in hepatocytes during liver inflammation

Recently, we reported that orphan nuclear receptor small heterodimer partner (SHP) is involved in neutrophil recruitment through the regulation of C-X-C motif chemokine ligand 2 (CXCL2) expression in a concanavalin A (ConA)-induced hepatitis model. In the present study, we examined the mechanisms underlying CXCL2 regulation by SHP and the cell types involved in liver inflammation. To this end, either Shp knockout (KO) or wild-type (WT) bone marrow cells were transferred into sublethally-irradiated WT (KO → WT or WT → WT) or Shp KO (KO → KO or WT → KO) recipients, followed by intravenous injection of ConA (20–30 mg/kg) 8 weeks later. The KO recipient groups showed higher ConA-induced lethality than the WT recipient groups. Accordingly, plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels, and inflammatory cytokine expressions were significantly higher in the KO recipients than in the WT recipients regardless of donor genotype. Massively increased hepatocyte death in KO recipients, as determined by H&E and TUNEL staining, was observed after ConA challenge. Bone marrow chimera experiments and in vitro chemotaxis assay also showed that SHP-deficient hepatocytes have an enhanced ability to recruit neutrophils to the injured liver. In vitro promoter assays showed that SHP is a negative regulator of Cxcl2 transcription by interfering with c-Jun binding to the AP-1 site within the Cxcl2 promoter. Collectively, SHP regulates Cxcl2 transcription in hepatocytes, playing a pivotal role in the recruitment of neutrophils. SHP-targeting strategies may represent alternative approaches to control fulminant hepatitis.


Isolation of intrahepatic immune cells.
Liver samples were collected from euthanized chimeric mice and were forced gently through a 70-μm cell strainer (BD Falcon, San Jose, CA, USA) using a sterile syringe plunger with addition of ice-cold phosphate-buffered saline (PBS). The preparation was centrifuged at 40 × g for 5 min at 4 °C. The supernatant was transferred to a new tube and centrifuged at 430 × g for 5 min at 4 °C. The pellet was re-suspended in 40% Percoll (GE Healthcare, Buckinghamshire, UK) in PBS and centrifuged at 1,265 × g with the no-brake setting for 30 min at 4 °C. The supernatant was discarded and erythrocyte lysis buffer (BioLegend, San Diego, CA, USA) was added to the pellet containing the intrahepatic immune cells. The cells were washed once with fluorescence-activated cell sorting (FACS) buffer prior to analysis by flow cytometry.
Isolation of blood immune cells. Peripheral blood were collected from euthanized WT or Shp KO mice 3 h after PBS or 25 mg/kg of ConA administration. The blood were re-suspended in ice-cold PBS, and centrifuged at 430 × g for 5 min at 4 °C. The supernatant was discarded and erythrocyte lysis buffer (BioLegend) was added to the pellet containing the blood immune cells. The cells were washed once with FACS buffer prior to analysis by flow cytometry.
Flow-cytometric analysis of intrahepatic or blood immune cells. Cells  Bone marrow-derived neutrophils. Mice were euthanized, and bone marrow was collected from femur, tibia, pelvis, scapula, and humerus. Bone marrow neutrophils were isolated by Percoll density gradient centrifugation and hypotonic lysis of red blood cells as previously described. Neutrophils were suspended in RPMI 1640 supplemented with 5% fetal bovine serum (FBS) and held on ice prior to testing. Neutrophil preparations were routinely >75% pure as determined by a Gallios ™ Flow Cytometer (Beckman Coulter) ( Supplementary Fig. 2).
Neutrophil chemotaxis assay. WT and Shp KO mice were injected intravenously with 15 mg/kg of ConA and the livers were harvested 1 h later. The livers were soaked in ice-cold sterile Hank's balanced salt solution supplemented with 0.5% fetal calf serum and subjected to two cycles of bead-beating using a TissueLyser (Qiagen, Valencia, CA, USA) followed by centrifugation. Bone marrow-derived neutrophil migration was determined using the TX ChemoTx ® System (101-3; NeuroProbe, Gaithersburg, MD, USA) consisting of a 3-μm-pore filter in a 96-well plate.
Neutrophils were pre-labeled with 5 μM CellTracker Green (Molecular Probes, Eugene, OR, USA). Thirty microliters of liver lysates were placed in the wells and 20-μl droplets of 1.7 × 10 6 /ml pre-labeled neutrophils were placed in hydrophobic rings on the filter, which allowed contact with the liver lysates below. Additionally, a standard curve of stained cells was prepared from cell preparations, of which 30 μl was placed in separate wells and 25 μl RPMI 1640 supplemented with 5% FBS was added onto the filter top. The plate was incubated at 37 °C with 5% CO 2 for 3 h. Following incubation, the medium in the top wells was absorbed, and any remaining cells were wiped off the top of the filter. The bottom plate (with the filter still on top) was centrifuged 430 × g for 5 min. The filter was removed and 100 nM phorbol myristate acetate (Sigma-Aldrich Chemical) was added to the wells and left for 20 min to facilitate permanent adhesion of transmigrated neutrophils; the cells were then washed with PBS. The plate was read on a Victor3 1420 Multilabel Counter (Perkin Elmer, Wellesley, MA, USA) at 485 nm excitation and 530 nm emission wavelengths. The numbers of migrated cells were calculated from the standard curves, based on relative fluorescence units.

Isolation of primary mouse hepatocytes and in vitro TNFα treatment.
Primary mouse hepatocytes were isolated from mice by perfusion with collagenase type I, as previously described 18 . The hepatocytes were cultured overnight, followed by treatment with recombinant TNFα (GIBCO BRL, Gaithersburg, MD, USA).
Enzyme-linked immunosorbent assay for CXCL1 and CXCL2. The levels of CXCL1 or CXCL2 in the medium of TNFα-stimulated primary hepatocytes or liver lysates from ConA-treated mice were determined using the OptEIA mouse CXCL1 or CXCL2 enzyme-linked immunosorbent assay kit (ELISA) (BD Biosciences, San Diego, CA, USA) according to the manufacturer's instructions.
Vector constructs. Mouse CXCL2 promoter (−900/+120) was PCR-amplified from mouse genomic DNA (Promega, Madison, WI, USA) and inserted into the pGL3 basic vector (Promega) using MluI and XhoI restriction enzyme sites. The pGL3-mCXCL2 AP1 mutant was generated by site-directed mutagenesis. All constructs were confirmed by DNA sequencing. pcDNA3-FLAG-SHP and pcDNA3-c-Jun were described previously 19 . Ad-GFP and Ad-FLAG-SHP were generated using the pAd-easy system (Clontech, Palo Alto, CA, USA). All viruses were purified using a CsCl gradient.

Cell culture and transient transfection. HepG2 human hepatoma cells were maintained in
Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and antibiotics. AML12 (alpha mouse liver 12, a non-transformed mouse cell line) cells were cultured in DMEM/F-12 supplemented with insulin-transferrin-selenium, dexamethasone (40 ng/ml), and antibiotics. After overnight incubation in 5% CO 2 at 37 °C, the medium was changed to serum-free culture medium. Vehicle control and TNF-α were added to the serum-free medium for the indicated times and at indicated concentrations. The cells were transiently transfected with Lipofectamine 2000 (Invitrogen, Gaithersburg, MD, USA) according to the manufacturer's instructions. Total plasmid DNA for transfection was adjusted to 1 µg/well by adding an appropriate amount of empty vector, and pNL1.1. TK (Nluc/TK) plasmid (Promega) was used as an internal control. The cells were harvested 24-36 h post-transfection for firefly and Nano-Glo luciferase assays. The firefly luciferase activity was normalized to Nano-Glo luciferase activity.
ChIP assay. The ChIP assay was performed according to the manufacturer's instructions (EMD Millipore, Temecula, CA, USA). Briefly, AML12 cells were fixed with 1% formaldehyde and then harvested. Soluble chromatin was immunoprecipitated with an antibody against c-Jun (Cell Signaling Technology). After recovering of the DNA, PCR was carried out using primers encompassing the mouse CXCL2 promoter (forward 5′-GAGGATTTGGGGAAGGACAT-3′, reverse 5′-CTCATCAGGAAGCACAGAGC-3′).
Statistical analysis. Numerical data are presented as the mean ± SEM. All statistical analyses were performed using JMP software (SAS International Inc., Cary, NC). Student's t-test was used for comparing two experimental conditions. Comparisons of multiple groups were performed using Tukey-Kramer HSD test after the one-way ANOVA. The survival ratio was analyzed using the log-rank test. The threshold of significance was set at P < 0.05.

SHP deficiency in liver parenchymal cells is associated with increased susceptibility to ConAinduced hepatitis.
Previously, it has been reported that SHP is involved in neutrophil recruitment through the regulation of CXCL2 expression in the ConA-induced liver injury model 17  transplantation, as shown in (Fig. 1a). KO-recipient groups (WT → KO or KO → KO) displayed a higher ConAinduced lethality than WT recipient groups (WT → WT/KO → WT), regardless of whether reconstitution was performed with WT or KO BM cells (Fig. 1b). Accordingly, plasma ALT and AST levels were significantly higher 9 h after ConA injection in KO recipients, regardless of the donor genotype (Fig. 2a). Furthermore, H&E and  TUNEL staining showed massively increased hepatocyte death in KO recipients compared with WT recipients 24 h after ConA challenge (Fig. 2b,c). We measured the levels of inflammatory cytokines in liver collected from chimeric mice at 24 h after ConA treatment ( Fig. 2d-g). Consistent with above described findings, Tnfα, Il-1β, Il-6, and Il-10 gene expressions in the liver were substantially increased in KO recipients compared with WT recipients. In contrast to the obvious detrimental effect of SHP ablation in liver parenchymal cells (KO recipients), we observed a modest contribution of SHP in BM cells during ConA-induced hepatitis. From these results, we conclude that SHP deficiency in liver parenchymal cells is critical for enhanced liver damage in the ConAinduced hepatitis mouse model.

SHP deficiency in liver parenchymal cells augments neutrophil recruitment in ConA-induced hepatitis.
To determine if parenchymal or non-parenchymal SHP affects immune cell recruitment to the injured liver in the ConA model, we purified infiltrating immune cells at 3 h after ConA challenge. FACS analyses revealed that numbers of infiltrating immune cells were significantly decreased in the liver samples of WT recipients that received BM from KO mice (KO → WT) (Fig. 3a). Additionally, absolute numbers of infiltrating T cells to the injured liver were significantly decreased in the KO → WT group. This decreased T cell population may influence the reduction in total cell numbers in this group. No significant monocyte/macrophage changes were observed among the different groups after ConA administration (Fig. 3b-d). Interestingly, neutrophil populations showed significantly different patterns based on the genotype of recipients. Notably, all KO-recipient groups displayed higher neutrophil infiltration than WT recipient groups. This increase in the ConA-induced neutrophil population was independent of the transplanted BM genotype (Fig. 4a-c). These findings were maintained 24 h after ConA treatment which were confirmed by immunohistochemical staining and quantification of neutrophils in liver (Fig. 4d,e). These data suggest that the extent of neutrophil migration to the injured liver and the severity of liver damage depend on the level of SHP in hepatic parenchymal cells.
SHP affects neutrophil chemotaxis and controls TNFα-induced CXCL2, but not CXCL1 expression in mouse hepatocytes. To evaluate whether SHP deficiency in the liver affects neutrophil migration, we conducted in vitro chemotaxis assays (Fig. 5a). Neutrophil migration toward liver lysates from ConA treated-Shp KO mice was higher than toward liver lysates from WT mice (Fig. 5b-e), indicating that intrinsic factors in the Shp KO mouse liver promote neutrophil migration in ConA-induced hepatitis. Recent reports have suggested that neutrophil migration to the injured liver is critically dependent on a hepatic chemokine network 20 . Therefore, we hypothesized that increased neutrophil migration toward the liver lysates of ConA-treated Shp KO mice may be influenced by C-X-C chemokines, which are major chemoattractants for neutrophils in various inflammatory and autoimmune diseases 21,22 . To test this hypothesis, we measured CXCL2 secretion first in liver lysates isolated from WT and Shp KO mice 1 h after ConA administration. Upon ConA stimulation, Shp-deficient liver was able to produce high level of CXCL2 than that of WT liver lysates (Supplementary Fig. 3). Next, we confirmed the levels of Cxcl1 and Cxcl2 expression in WT and Shp KO recipient mice liver following 24 h after ConA treatment. Cxcl1 expression showed similar pattern among groups, but Cxcl2 expression were significantly higher in KO recipients, regardless of the donor genotype (Fig. 6a,b). Under these situation, Shp expression was significantly suppressed in KO recipient liver, indicating that Shp is inversely correlated with Cxcl2 expression (Fig. 6c). More specifically, we isolated primary hepatocytes, representative of hepatic parenchymal cells, from WT and Shp KO mice livers. It is well established that TNFα is a major proinflammatory cytokine involved in early inflammatory events and that TNFα stimulates CXCL2 expression in ConA-induced hepatitis 23 . We measured C-X-C chemokine levels after TNFα treatment in primary hepatocytes isolated from WT and Shp KO mice (Fig. 6d). CXCL1 secretion was significantly increased by TNFα stimulation; however, there were no significant differences between both groups. In contrast, CXCL2 secretion was largely up-regulated in SHP-deficient hepatocytes 60 min after TNFα treatment. Finally, we checked TNFα-induced CXCL2 secretion in primary hepatocytes isolated from Shp KO mice under the condition of SHP overexpression by adding Ad-SHP ( Supplementary  Fig. 4). TNFα-evoked CXCL2 secretion was significantly suppressed by SHP overexpression. These data indicate that hepatocytes are a major source for increased CXCL2 production and secretion in Shp KO mice, and SHP acts as a negative downstream regulator of TNFα-mediated CXCL2 in hepatocytes.

SHP regulates CXCL2 transcription through interaction with c-Jun.
Finally, we investigated the mechanism of Cxcl2 expression by SHP. Firstly, we tested whether SHP affects TNFα-induced Cxcl2 promoter activity through a reporter luciferase assay in AML12 and HepG2 cells. In both cell lines, overexpression of SHP drastically suppressed TNFα-induced Cxcl2 promoter activity (Fig. 7a). A previous study reported that Cxcl2 is regulated by NF-κB and c-Jun/Activator Protein-1 (AP-1). In our previous study, we showed that the effect of SHP on Cxcl2 transcription is not influenced by the NF-κB pathway 17 . Therefore, we used reporter assays with AP-1-mutant Cxcl2 promoter constructs to examine the participation of the AP-1 site in the SHP regulatory effect. Our data show that c-Jun/TNFα-mediated WT but not mutant Cxcl2 promoter activity was remarkably suppressed by SHP overexpression (Fig. 7b). In addition, ChIP assays showed that TNFα-induced binding of c-Jun to the AP-1 site of the Cxcl2 promoter is almost entirely abrogated by the overexpression of SHP (Fig. 7c). Taken together, these results demonstrate that the inhibitory effect of SHP on TNFα-mediated CXCL2 expression is mainly through the inhibition of c-Jun-DNA binding.

Discussion
Previously, we observed that Shp KO mice show higher neutrophil trafficking to the injured liver following ConA treatment, accompanied by enhanced CXCL2 expression 17 . Here, using bone marrow chimeric mice, we verified that SHP is a previously unrecognized transcriptional regulator of hepatic CXCL2 that contributes to the pathogenesis of fulminant hepatitis. Shp KO recipient mice, regardless of the donor genotype, were highly susceptible to ConA-induced liver injury and prone to death upon ConA administration. Infiltrating neutrophils to the injured liver, driven by CXCL2, were also significantly increased in KO recipient mice regardless of WT or Shp KO donor, suggesting that SHP deficiency in liver parenchymal cells is critical for ConA-induced hepatitis. Intriguingly, SHP in hepatocytes differentially regulates CXCL1 and CXCL2, and negatively controls TNFα-mediated CXCL2 expression through inhibition of c-Jun-DNA binding.
In ConA-induced hepatitis, TNFα and IFNγ are rapidly produced in response to ConA and induce the expression of pro-inflammatory proteins in both parenchymal and non-parenchymal liver cells 24 . Several studies have reported that CXCL2 is secreted by various hepatic cells under different pathological conditions 25,26 , and mouse CXCL2 is mostly induced by TNFα. Moreover, CXCL2 production stimulated by TNFα partially contributes to liver injury mediated through neutrophil recruitment 23 . We observed that SHP deficiency in hepatocytes led to the increased secretion of CXCL2, but not CXCL1, upon TNFα stimulation, emphasizing that hepatocytes contribute to increased levels of circulating CXCL2 in Shp KO mice during ConA-induced hepatitis. This is consistent with our previous study that demonstrates increased CXCL2 expression, but not CXCL1, in both the liver and plasma of Shp KO mice after ConA administration. Moreover, it has been demonstrated that Shp expression was inversely correlated with Cxcl2 expression in ConA-treated livers 17 , implying that SHP is a possible transcriptional regulator of CXCL2.
Here, to solve the remaining question whether Shp KO neutrophils have an enhanced ability to chemotactically migrate to the injured liver, we confirmed chemokine receptor CXCR2 expression on cell membrane of circulating neutrophils as well as other immune cells after ConA treatment using FACS ( Supplementary Fig. 5). The CXCR2 is expressed at high levels on cell membrane of circulating neutrophils, and is critical for directing their migration to sites of inflammation 27 . In this study, basal levels of CXCR2 membrane expression on neutrophils were not different between WT and Shp KO mice. Activation of Toll-like receptors (TLRs) in neutrophil downregulates CXCR2 membrane expression, and impairs neutrophil migration 28 . It has been reported that ConA induces expression of TLRs in murine macrophages, and renders them more susceptible to subsequent activation by TLR ligands 29 . In these regards, it might be assumed that ConA treatment could regulate CXCR2 expression on cell membrane through modulation of TLR signalling. In the present study, ConA administration lowered CXCR2 membrane expression on neutrophils, which would be regulated by ConA treatment-mediated regulation of TLR signalling. However, these decreased CXCR2 membrane expressions in neutrophil were comparable between WT and Shp KO mice. In regarding to monocytes, CD8 + T cells, and CD4 + T cells, CXCR2 membrane expressions were similar between WT and Shp KO mice, and not influenced by ConA treatment. These results demonstrated that hepatic CXCL2 is associated with augmented neutrophil recruitment to the injured liver of Shp KO mice in a neutrophil CXCR2 expression modulation independent manner. Furthermore, we addressed the question of why CXCL1 and CXCL2 are differentially regulated in the absence of SHP during the ConA-induced inflammatory response. Cxcl1 and Cxcl2 share an NF-κB-binding site in their promoter regions 30 . Intriguingly, NF-κB inhibition effectively prevented the up-regulation of Cxcl1 but not Cxcl2. Complete suppression of Cxcl2 required the dual inhibition of NF-κB and AP-1 31 . Here, we report that SHP specifically controls the transcription of Cxcl2, but not Cxcl1, by directly inhibiting the binding of c-Jun to the AP-1 site within the Cxcl2 promoter. We targeted the AP-1 site to exclude the possibility of other transcription factors sharing a promoter region with Cxcl1. Additionally, human and mouse Cxcl2 share NF-κB-and AP-1-binding sites in their promoter regions 32 , suggesting evolutionary conservation of the identified mechanisms. Hence, we propose that the targeting of SHP might represent a novel therapeutic option in various human inflammatory diseases, including hepatitis.
SHP has been previously reported to negatively regulate Toll-like receptor-induced inflammatory responses in innate immune cells through a biphasic interaction with its cytoplasmic partners TRAF6 and NF-κB p65 15 . We now found that SHP deficiency in liver parenchymal cells aggravates neutrophil-mediated liver injury via modulation of CXCL2 expression in an NF-κB independent manner, suggesting that SHP may have cell type-specific roles in the control of inflammatory processes.
Collectively, we demonstrated that CXCL2-derived neutrophils are critical mediators of experimental hepatitis. We have revealed a novel function of SHP as a negative transcriptional regulator of hepatocyte-derived CXCL2. Therefore, the modulation of SHP may have therapeutic benefits in the treatment of inflammatory liver diseases. . The cells were co-transfected with the indicated plasmids and/or treated with TNFα (30 ng/ml). Firefly luciferase activity was normalized to Nano-Glo luciferase activity. (c) AML12 cells were transfected with vector expressing mCXCL2 promoter and then treated with TNFα (30 ng/ml) in the presence of Ad-GFP or Ad-SHP. Input represents 10% of purified DNA in each sample. Soluble chromatin was immunoprecipitated with antibody against c-Jun and purified DNA samples were analyzed by PCR with primers encompassing the AP-1 site of the mCxcl2 gene promoter. Data are means ± SEMs of 3 individual experiments. *P < 0.05 (Tukey-Kramer HSD test after the one-way ANOVA).