IFN-γ-Inducible Protein-10 (CXCL10) Is Hepatoprotective During Acute Liver Injury Through the Induction of CXCR2 on Hepatocytes¹

Acetaminophen (APAP)³ is an over-the-counter analgesic that is generally thought to not have deleterious side effects, yet toxicity is the most common cause of acute liver failure (1, 2) in both the adult and the pediatric population (3), sometimes necessitating a liver transplant for the survival of the patient (4). In the adult population overdose generally occurs either intentionally or in the context of altered hepatic metabolism due to alcohol, drugs such as isoniazid (5), and viral infections, yet in the pediatric population the most common cause of overdose is iatrogenic (from both parents and doctors) (3). While the standard of care is N-acetylcysteine, which is effective, it must be administered within 8 h after ingestion of APAP to have a therapeutic effect (6).

Chemokines are chemotactic cytokines that can be divided into four families dependent on the location of the conserved cysteine residues. For example, the CXC chemokines contain two conserved cysteine residues separated by a nonconserved amino acid. The CXC chemokines can be further divided dependent on which amino acids precede the CXC motif. The NH₂ terminus of some of the CXC chemokines contains a 3-aa motif (Glu-Leu-Arg: the ELR motif) immediately adjacent to the CXC motif. The ELR-CXC chemokines, such as IL-8, macrophage-inflammatory protein (MIP)-2, and epithelial cell-derived neutrophil-activating protein-78, which bind to CXCR2, have previously been shown to have a therapeutic role in acute liver injury after APAP challenge, and after APAP challenge in combination with an adenovirus (7, 8, 9). The roles of the non-ELR-CXC chemokines, including IFN-γ-inducible protein-10 (IP-10), monokine induced by IFN-γ, and IFN-inducible T cell α chemoattractant, which bind to CXCR3 and are inducible by the IFNs, are controversial in liver injury. While some studies have shown that IP-10 (also known as CXCL10) levels correspond to liver injury (10, 11), IP-10 has been shown to have hepatoprotective effects (12), and recent studies have shown that its elevation correlates to the peak of DNA synthesis (13). Therefore, the aim of the current study was to elucidate the role of IP-10 during APAP-induced liver injury. To this end, we examined the expression pattern of IP-10 and its receptor CXCR3 during an APAP challenge and found that it was elevated in the serum correlating to the peak of alanine aminotransferase (ALT), a measurement of liver injury. Maintaining the elevated IP-10 levels 10 h after an APAP challenge proved to be quite beneficial by reducing liver injury and promoting the expression of CXCR2, the receptor for MIP-2, which appeared to be absolutely necessary for this protective effect.

Female CD-1 (6–8 wk of age) were purchased from Charles River Breeding Laboratories (Portage, MI) and maintained under specific pathogen-free conditions with free access to water and food before each experiment. Fresh suspensions of APAP (Sigma-Aldrich, St. Louis, MO) were made daily by dissolving the compound in PBS warmed to 50°C. In all experiments, mice were deprived of food, but not water, for 18 h before an APAP challenge. APAP was given to each mouse by i.p. injection at a dose of 400 mg/kg, as described previously in detail (9).

In preliminary experiments, mice were fasted for 18 h and received 400 mg/kg APAP. Liver and serum samples were removed from each mouse at 0 (immediately before APAP challenge), 4, 8, 24, and 48 h after APAP challenge (Fig. 1,A). In a second set of experiments, fasted mice received either PBS (0.5 ml) as control or 1 μg of murine rIP-10 (PeproTech, Rocky Hill, NJ) dissolved in 0.5 ml of PBS via an i.v. injection at 10 h after the APAP challenge. Liver and serum samples were removed from each mouse at 4, 8, 26, 34, 40, and 48 h after PBS or murine rIP-10 therapy (Fig. 1 B).

In a separate set of experiments, mice were fasted for 18 h, subsequently challenged with 400 mg/kg, and then given a therapeutic treatment with 1 μg of murine rIP-10 (i.v.) 10 h later (Fig. 1,B, filled arrows). Immediately after IP-10 treatment, mice received 0.5 mg of either anti-CXCR2 or goat IgG (i.p.), and then received another dose 2 h later (Fig. 1 B, open arrows). Liver and serum samples were removed from the mice 8 h after PBS or IP-10 treatment.

Murine IP-10 was quantified using a modification of a double ligand method, as previously described. Briefly, flat-bottom 96-well microtiter plates (Nunc Immuno-Plate I 96-F; Nunc, Roskilde, Denmark) were coated with 50 ml/well anti-mouse cytokine Ab (1 mg/ml in 0.6 M NaCl, 0.26 M H₃BO₄, and 0.08 M NaOH, pH 9.6) for 16 h at 4°C and then washed with wash buffer (PBS, pH 7.5, 0.05% Tween 20). Nonspecific binding sites in each plate were blocked with 2% BSA in PBS and incubated for 90 min at 37°C. Plates were rinsed four times with wash buffer and diluted (neat and 1/10) serum (50 ml) in duplicate was added to each plate and incubated for 1 h at 37°C. Plates were washed four times, followed by the addition of 50 ml/well biotinylated rabbit Abs against the specific cytokines (3.5 mg/ml in PBS, pH 7.5, 0.05% Tween 20, and 2% FCS), and incubated for 30 min at 37°C. After washing, streptavidin-peroxidase conjugate (Bio-Rad, Richmond, CA) was added and the plates were incubated for 30 min at 37°C. After washing again, chromagen substrate (Bio-Rad) was added. The plates were incubated at room temperature to the desired extinction, and the reaction was terminated with 50 ml/well 3 M H₂SO₄ solution. Plates were read at 490 nm in an ELISA reader. Standards were one-half log dilutions of LPS-free recombinant murine cytokines (R&D Systems, Minneapolis, MN, or PeproTech) from 1 pg/ml to 100 ng/ml. This ELISA method consistently detected murine cytokine concentrations above 25 pg/ml, and ELISA specificity was confirmed for each cytokine and chemokine measured.

Total RNA was isolated from whole liver samples before and at 4 and 8 h after APAP challenge. A total of 0.5 μg of total RNA was reverse transcribed to yield cDNA, and IP-10 gene expression was analyzed by real-time quantitative RT-PCR procedure using an ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). GAPDH was analyzed as an internal control. All primers and probes were purchased from Applied Biosystems. IP-10 gene expression was normalized to GAPDH before the fold change in IP-10 was calculated. The fold increase in IP-10 gene expression in APAP-challenged livers was calculated via the comparison of gene expression of this chemokine after the APAP challenge to that detected before the APAP challenge (i.e., t = 0). IP-10 mRNA levels before APAP were assigned an arbitrary value of 1.

mRNA expression in the liver of APAP-challenged mice was examined using RT-PCR. Briefly, total RNA was isolated from liver homogenates, and 5 μg of total RNA was reverse transcribed to yield cDNA using techniques previously described in detail (14). The following sense and antisense primers, respectively, were used in the PCR reaction: cyclophilin sense 5′-CATCTGCACTGCCAAGACTG-3′ and antisense 5′-CTGCAATCCAGCTAGGCATG-3′, and CXCR3 sense 5′-ATCAGCGCTTCAATGCCAC-3′ and antisense 5′-TGGCTTTCTCGACCACAGTT-3′.

PCR samples were initially incubated for 94°C for 5 min and then cycled 30 times through denaturation at 95°C for 30 s, annealing at 55°C for 45 s, and extension at 72°C for 75 s. PCR products were then separated on 2% agarose gels containing 0.3% ethidium bromide, and the bands corresponding to the intended products were photographed under UV illumination.

Serum levels of ALT were determined at various times after APAP challenge by Clinical Pathology at the University of Michigan Medical School (Ann Arbor, MI) using standardized techniques.

A portion of resected liver from each mouse was immediately fixed in 4% paraformaldehyde for a minimum of 12 h. Fixed liver samples were subsequently processed, embedded in paraffin, thin sectioned, and placed on l-lysine-coated slides. H&E staining was used to reveal morphometric differences among groups of mice. Additional slides containing unstained liver sections were used for immunohistochemical analysis. To reveal the presence of CXCR2 in liver sections, other slides were deparaffinized, microwaved for approximately 20 min in 10 mM citric acid buffer, and then allowed to cool to room temperature. Slide-mounted liver sections were blocked using normal goat serum (blocking solution) for 1 h. Tissue sections were treated with purified polyclonal anti-mouse CXCR2 Ab or rabbit IgG for control. All were diluted at 1/25 with TBS containing blocking solution (1/100) and incubated overnight at 4°C. After incubation, slides were washed twice for 5 min in TBS. A 1/35 dilution of biotinylated goat anti-rabbit Ab (BioGenex, San Ramon, CA) was placed on the slides for 2 h at 37°C in a humidified chamber. Slides were again washed twice in TBS and incubated with a 1/35 dilution of streptavidin conjugated to HRP (BioGenex) for 45 min. Following two washes in TBS with 50 mM levamisole, Fast Red chromagen (BioGenex) was placed on each slide, and staining was visualized at low power until color development was complete. The staining reaction was terminated in sterile water, and each slide was counterstained with Mayer’s hematoxylin (0.1%; Sigma-Aldrich).

Preparation of cytoplasmic extracts from liver was conducted as follows. Briefly, liver samples were rapidly homogenized in PBS containing Complete protease inhibitor (10 mg/ml; Boehringer Mannheim, Indianapolis, IN) on ice and washed with fresh PBS. Homogenates were then suspended in buffer A (10 mM HEPES, 10 mM KCl, 0.5 mM DTT, 1% Nonidet P-40) for 10 min and centrifuged for 10 min at 14,000 × g, and the supernatant containing cytoplasmic components was removed.

After cytoplasmic protein levels were determined using a Bradford assay (Bio-Rad), 50 μg of liver cytoplasmic extracts were electrophoresed on a 12% polyacrylamide gel and then transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad). Equal protein loading was reconfirmed by Coomassie blue staining of the gel after transfer. PVDF membranes were blocked for 1 h at room temperature in 5% dry milk. CXCR2 Abs were diluted to 1/500 and incubated with PVDF membranes overnight at 4°C. HRP-linked secondary Ab (Pierce, Rockford, IL) were then added at a 1/3000 dilution for 2 h at room temperature, and protein bands were visualized by chemiluminescence (Bio-Rad).

Results are expressed as means ± SEM of 5–10 mice per group at each time point after treatment. All statistical calculations were performed using GraphPad Prism 2.0 computer software (GraphPad, San Diego, CA); for all analyses, a Student’s t test was used to test for significance. p ≤ 0.05 was considered statistically significant.

IP-10 protein levels in the serum and IP-10 mRNA levels in the liver of APAP-challenged mice were examined before challenge and at 4, 8, and 24 h after challenge. As shown in Fig. 2,A, immediately prior (i.e., 0 h) to an i.p. challenge of 400 mg/kg APAP and 4 h after this challenge, there were no detectable levels of IP-10 in the serum. In contrast, 8 h after APAP challenge, levels of IP-10 were dramatically increased (p = 0.01). These elevated levels are not sustained throughout recovery from APAP toxicity, and IP-10 levels begin to fall by 24 h after challenge. Because the regulation of IP-10 in the serum after APAP challenge was quite dramatic, we next determined whether there was a quantitative change in IP-10 mRNA in the liver and found that the levels of IP-10 in the liver were significantly increased at 4 and 8 h after APAP challenge (Fig. 2 B). Thus, an APAP challenge in mice was associated with systemic and local increases in IP-10 protein and gene expression.

Considering that the levels of IP-10 were markedly altered following APAP-induced liver injury, we next examined whether there were changes in CXCR3, the receptor for IP-10. As shown in Fig. 3,A, RT-PCR analysis revealed that constitutive CXCR3 mRNA was expressed in liver homogenates before APAP challenge and increased following the challenge (Fig. 3,B). Four hours following the APAP challenge, there was an almost 2-fold increase in CXCR3 mRNA expression as compared with time 0 as determined by densitometry, and at 8 h following the challenge there was almost three times more mRNA expressed. Thus, IP-10 and its receptor CXCR3 were both up-regulated following a sublethal APAP challenge in mice, which correlates to the time when ALTs, a marker of acute hepatocellular injury, are at their peak levels (Fig. 4). Therefore, IP-10 levels were not regulated in the liver, but levels of its receptor were markedly increased at the gene level corresponding to the time of peak liver injury.

Given the marked increase in liver levels of CXCR3 in the APAP-challenged mice, we next assessed the effect of endogenously administered IP-10. Given that the greatest increases in CXCR3 were observed at 8 h after APAP, rIP-10 was administered to mice at 10 h after the APAP challenge. When animals were treated with rIP-10 or control (PBS) 10 h after a 400-mg/kg dose of APAP (see protocol outlined in Fig. 1,B), IP-10 dramatically decreased hepatic injury 2 days after challenge. When control animals were treated with PBS, there is dramatic reduction in centrilobular hepatocyte necrosis and hemorrhagic injury (Fig. 5,A). In contrast, when animals were treated with rIP-10, there is maintenance of hepatic architecture and little evidence of hepatocyte injury (Fig. 5,B). As shown in Fig. 6, IP-10 significantly decreased liver injury at 8 h after administration by ∼3-fold (p = 0.0078). Taken together, these data demonstrate that IP-10 has a dramatic therapeutic effect 10 h after an APAP challenge. Subsequent experiments were designed to address the mechanisms for this therapeutic application.

Previous studies in our laboratory have shown that MIP-2 and its receptor CXCR2 have a tremendous therapeutic effect in acute liver injury due to their hepatoprotective effect (9). To assess whether CXCR2 mediated the therapeutic effect of IP-10, we examined the expression of CXCR2 in the presence and absence of exogenous IP-10. As shown in Fig. 7, IP-10 dramatically enhanced the expression of CXCR2 protein, as assessed by immunohistochemistry. Fig. 7,A depicts low CXCR2 expression on hepatocytes after an i.p. challenge with 400 mg/kg APAP and i.v. injections of PBS 10 h later. In contrast, histological liver sections from mice that received 1 μg of IP-10 exhibited strong expression of CXCR2 on hepatocytes at the same time after treatment (Fig. 7,C). Depicted in Fig. 7, B and D, are the negative controls of Fig. 7, A and C, respectively.

Temporal changes in CXCR2 expression protein were examined by Western blot at 4, 26, 34, 40, and 48 h after IP-10 or saline treatment in APAP-challenged mice (Fig. 8). Although there was similar expression in both groups of mice at 4 h after treatment, by 26 h there was a dramatic divergence in CXCR2 expression. For example, mice treated with exogenous IP-10 expressed much greater levels of CXCR2 protein (as indicated by the 42- and 84-kDa bands) at all subsequent time points examined compared with controls. The monomeric form of CXCR2 is represented by the 42-kDa band, and the 84-kDa band represents a SDS-resistant CXCR2 dimer similar to other G protein-coupled receptors (15, 16). Thus, IP-10 treatment increased the protein levels of CXCR2 in the liver.

Since IP-10 appeared to up-regulate CXCR2 on hepatocytes within the liver, we next examined whether this effect accounted for the therapeutic effect of IP-10. To this end, fasted mice received 400 mg/kg APAP and 1 μg of IP-10 (i.v.) 10 h later. Immediately after the IP-10 treatment, mice were given 0.5 mg of either anti-CXCR2 or goat IgG (i.p.) and then received another dose 2 h later. The IP-10 and goat IgG treatment dramatically reduced ALT levels at 8 h following treatment as compared with control animals that received PBS and goat IgG (Fig. 9). In contrast, CXCR2 immunoneutralization strikingly abrogated the therapeutic effect of IP-10, resulting in 3-fold higher ALT levels than IP-10 treatment alone. These data strongly suggested that IP-10 mediated its hepatoregenerative effects via CXCR2. Further evidence for this was observed in histological samples from mice that received anti-CXCR2 and IP-10 after an APAP challenge (Fig. 10).

While changes in IP-10 levels have been correlated to various types of liver injury, it has yet to be elucidated whether it has a beneficial or detrimental role. Enhanced IP-10 levels have been detected during liver injury due to an adenoviral challenge in mice (17), and elevated hepatic IP-10 levels correlate with more profound histological findings in biliary atresia patients (11). Additionally, lymphocytes infiltrating hepatitis C-infected patients expressed CXCR3, and IP-10 was up-regulated on sinusoidal epithelium (10). In contrast, when IP-10 was elevated in response to IFN-γ treatment, there was less injury in an ischemia-reperfusion model, as shown by a decrease in ALT values (12). Additionally, more recent data suggest that IP-10 expression correlates with a regenerative response in the liver (13). In the present study, we examined the role of IP-10 in an acute liver injury model initiated by APAP overdose. While other chemokines, notably monocyte chemoattractant protein-1 and MIP-2, have been shown to exert protective and regenerative roles in this model of liver toxicity (9, 18), the role of IP-10 has not been previously examined. IP-10 levels were elevated in the serum, while its receptor, CXCR3, was elevated in the liver after APAP challenge. Importantly, CXCR3 levels were increased at a time when liver injury was maximal (8 h), as assessed by serum ALTs. While the cellular source of the systemically elevated IP-10 was not examined, it is likely that it is being released from the damaged hepatocytes, similar to that seen in autoimmune liver diseases (19) and hepatitis (20). Furthermore, primary murine hepatocytes can release IP-10 after isolation (20). The presence of increased CXCR3 expression in the liver prompted us to investigate the potential effects that this chemokine had on liver injury. When exogenous IP-10 was given 10 h after APAP challenge (corresponding to the time of maximal CXCR3 expression and liver injury), there was a marked improvement in liver function, and the histological appearance of the liver was dramatically improved. This therapeutic effect was striking in light of the fact that the standard of care in APAP poisoning, N-acetylcysteine, is no longer effective (6). Further exploration of this therapeutic application is needed to assess the potential clinical relevance of this chemokine.

The therapeutic effect of IP-10 appeared to be dependent on the up-regulation of CXCR2 on hepatocytes, the receptor for MIP-2, while not having a significant effect on the levels of MIP-2 (data not shown). MIP-2 and its receptor CXCR2 have previously been shown to play a central role in recovery from APAP-induced toxicity (9). These unanticipated results are striking because of the dichotomous roles that ELR vs non-ELR-CXC chemokines play in other biological events, particularly angiogenesis (21, 22). While members of the ELR-CXC chemokine family have been shown to induce endothelial cell chemotaxis, proliferation, and angiogenesis in vivo (23), the non-ELR-CXC chemokines are angiostatic (24). Furthermore, studies on hepatocyte proliferation have shown that the ELR-CXC chemokines promote hepatocyte proliferation, but the non-ELR-CXC chemokines are able to block this proliferative effect (8). Although these results seem to conflict, similar results have been seen in angiogenesis. For example, while the ligands for CXCR3, including IP-10, IFN-inducible T cell α chemoattractant, and monokine induced by IFN-γ, induced chemotaxis of human microvascular dermal endothelial cells at high concentrations, they are still capable of inhibiting the chemotactic response of human microvascular dermal endothelial cells to the ELR-CXC chemokines (25). While it is known that ELR-CXC chemokines, including IL-8 and MIP-2, can exist as homodimers (26, 27), it is possible that cotreatment with IP-10 and MIP-2 results in the formation of heterodimers (28) similar to the heterodimer formed by MIP-1α and MIP-1β (29), which may prevent the binding of the chemokine to its receptor.

Thus, in this study, we have found that IP-10 protein and mRNA levels were increased following APAP challenge, along with up-regulation of its receptor CXCR3 in the liver. These levels began to decline after 8 h, and maintenance of these levels with a 10-h post-treatment of rIP-10 resulted in a dramatic recovery of liver injury due to hepatocyte regeneration. This effect was dependent on CXCR2, which was up-regulated in the liver following APAP challenge. In conclusion, the therapeutic effect of IP-10 post-treatment at a time when conventional treatment is no longer effective was striking and suggests that IP-10 plays a role as a regulator of hepatocyte proliferation.