Deficiency of Interleukin-15 Enhances Susceptibility to Acetaminophen-Induced Liver Injury in Mice

Hepatocytes have a direct necrotic role in acetaminophen (APAP)-induced liver injury (AILI), prolonged secondary inflammatory response through innate immune cells and cytokines also significantly contributes to APAP hepatotoxicity. Interleukin 15 (IL-15), a multifunction cytokine, regulates the adaptive immune system and influences development and function of innate immune cells. To better understand the role of IL-15 in liver injury, we treated wild-type (WT) and IL-15-knockout (Il15−/−) mice with a hepatotoxic dose of APAP to induce AILI and evaluated animal survival, liver damage, APAP metabolism in livers and the inflammatory response. Production of pro-inflammatory cytokines/chemokines was greater in Il15−/− than WT mice. Subanalysis of hepatic infiltrated monocytes revealed greater neutrophil influx, along with greater hepatic induction of inducible nitric oxide synthase (iNOS), in Il15−/− than WT mice. In addition, the level of hepatic hemeoxygenase 1 (HO-1) was partially suppressed in Il15−/− mice, but not in WT mice. Interestingly, elimination of Kupffer cells and neutrophils did not alter the vulnerability to excess APAP in Il15−/− mice. However, injection of galactosamine, a hepatic transcription inhibitor, significantly reduced the increased APAP sensitivity in Il15−/− mice but had minor effect on WT mice. We demonstrated that deficiency of IL-15 increased mouse susceptibility to AILI. Moreover, Kupffer cell might affect APAP hepatotoxicity through IL-15.


Introduction
Acetaminophen (APAP) is an over-the-counter analgesic widely used worldwide. However, APAP-induced liver injury (AILI) represents the most common hepatogenous poisoning secondary to drug overdose. Excess APAP saturates the sulfation and glucuronidation of the metabolic pathway and results in generation of toxic N-acetyl-p-benzoquinone imine (NAPQI) by cytochrome P450 (CYP) [1], thereby depleting hepatic glutathione (GSH) [2]. Residual unconjugated NAPQI induces covalent binding of intracellular proteins and causes further formation of reactive oxygen species (ROS) [3], thus resulting in apoptosis and necrosis of hepatocytes [4]. Induction of intracellular inflammation regulatory proteins such as hemeoxygenase 1 (HO-1) attenuates APAP toxicity [5]. In addition, the downstream innate immune response, by immune cells and associated cytokines, modulates the progression of liver injury [6].
Recently, increased APAP sensitivity was attributed to enhanced inflammation in mice lacking DCs, but the detailed mechanism remained speculative [9]. Depletion or inactivation of KCs by chemicals in an AILI model had controversial results, with a protective effect in one study [11] but a negative result in another [10]. Furthermore, mice lacking of cytokines such as interleukin 10 (IL-10) [14], IL-6 [15] or IL-13 [16] were found susceptible to APAP hepatotoxicity, whereas induction of pro-inflammatory mediators such as tumor necrosis factor alpha (TNFa) [17], interferon gamma (IFNc) [18], IL-18 or IL-1b [19] and nitric oxide (NO) enhanced AILI in mice. Collectively, the roles of innate immune cells, especially antigen-presenting cells, and cytokines in AILI are complicated and still unclear.
IL-15, a multifunction cytokine mainly produced by antigenpresenting cells such as macrophages, DCs, B cells or endothelial cells, regulates the adaptive immune system and plays an important role in innate immunity [20,21]. IL-15 can direct the development of CD8+ memory T cells, NK and NKT cells [20] and modulate the function of macrophages and DCs [22]. In addition, IL-15 can inhibit apoptosis of neutrophils [23] and regulate the production of inflammatory cytokines such as TNFa, IL-6, IL-1b and IL-10 in macrophages in response to lipopolysaccharide stimulation [24]. Synthetic IL-15 could moderately diminish liver injury in concanavalin A or Fas ligand-induced hepatitis [25,26], whereas DC-derived IL-15 enhanced endotoxin shock injury through the liver [27]. Moreover, IL-15 mediates the crosstalk between conventional and plasmacytoid DCs for immune activation [28]. Interestingly, IL-15 promoted hepatocyte mitosis and liver proliferation in healthy mice and those with hepatectomy, respectively [29].
In this study, we aimed to study the role of IL-15 in a sterile APAP-induced fulminant hepatitis model in IL-15-knockout (Il15 2/2 ) and wild-type (WT) mice. By treatment of mice with a hepatotoxic dose of APAP and evaluation of animal survival, liver damage, APAP metabolism in livers and the inflammatory response, we were able to elucidate whether IL-15 was involved in APAP hepatitis.

Animals and materials
Male C57BL/6J mice were purchased from National Laboratory Animal Center, Taipei, Taiwan. C57BL/6NTac/Il15 2/2 mice, obtained from Taconic Farms (Terrytown, NY), were backcrossed to C57BL/6J background for 4 generations. This substrain, as C57BL/6J/Il15 2/2 (presented as Il15 2/2 in our study) mouse, was used in our later study. All mice (9-12 weeks old) were kept in a pathogen-free condition in compliance with institutional animal care and use committee guidelines (project I.D., IACUC NO. 10-043). All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO, USA) unless otherwise stated. Recombinant murine IL-15, with specific activity of $2610 5 units/mg, was purchased from Perprotech (Rocky Hill, NJ, USA). Rat anti-IL-15 neutralizing antibody was obtained from eBiscience (San Diego, CA, USA). Recombinant IL-15 was intraperitoneally or subcutaneously injected 20 minutes before APAP challenge. IL-15 neutralizing antibody was intraperitoneally injected 30 minutes prior to APAP administration. APAP was dissolved in warm saline and intraperitoneally injected into overnight-fasted mice to induce hepatitis, and 350 mg/kg of APAP (sub-lethal dose) were generally used in this study unless otherwise mentioned. Animal survival was monitored for 5 days. 1.5 mg/kg of vinblastine (vin) or 30 mg/kg of gadolinium chloride (GdCl 3 ) was intravenously given to animals for neutrophil or KC elimination, respectively.

Evaluation of liver injury
Serum alanine and aspartate aminotransferase (ALT/AST) levels were determined by the test kit DRY-CHEM SLIDE ALT/ AST (Fujifilm, Tokyo, Japan). The liver was sectioned and fixed overnight in 10% formalin solution, dehydrated, paraffin-embedded, cut at 4-mm thickness and stained with haematoxylin and eosin (H&E) for histological examination. The liver necrotic area was quantified by 2 individual pathologists with Image-Pro Plus (Media Cybernetics, Bethesda, MD) and the necrotic area (%) was expressed as necrosis area/total area of liver section.

Hepatocyte isolation and culture, and KC-enriched fractionation
Hepatocytes were isolated from mice by a 2-step collagenase perfusion protocol. The hepatic portal vein was ligated and perfused with Ca 2+ /Mg 2+ -free Hank's balanced salts solution (HBSS; Gibco, Invitrogen, Carlsbad, CA) with 5 mM EGTA, than 0.025% type IV collagenase containing HBSS. After perfusion, liver was excised and hepatocytes were suspended in serum free L-15 medium and pass through 100 mm strainer. The filtrate was centrifuged at 45 g for 2 min at 4uC. The supernatant as the non-parenchymal cell (NPC) fraction was centrifuged at 350 g for 5 min at 4uC and, used to analyze mononuclear cell population or KC enrichment.
After perfusion, hepatocyte viability (more than 85%) was determined by trypan blue exclusion. Hepatocytes were suspended in 10% fetal bovine serum containing William's Essential Medium (WEM) and cultured in 0.02% collagen-coated plate; after 4 hrs, the indicated concentration of APAP was added to analyze cytotoxicity to hepatocytes. The APAP toxicity was quantified by use of the Cell Counting Kit-8 (CCK-8).

Determination of cytokines by ELISA
Murine IL-15, IL-1b, IFNc and TNFa levels in serums or livers were measured by ELISA kits (R&D research). Serum samples were collected by heart puncture and liver proteins were collected from frozen liver samples homogenized in protein lysis buffer.

Immunohistochemical staining of liver samples
The levels of hepatic inducible NO synthase (iNOS) and nitrotyrosine were evaluated in deparaffinized and antigenretrieved liver sections by LSAB2 system-HRP kits (DakoCytomation, Produktionsvej, Denmark). Liver sections were incubated with primary antibodies against iNOS or nitrotyrosine (Upstate Biotechnology, Lake Placid, NY) as the manufacture's instructions. Immunohistochemical staining was analyzed by use of Image-Pro Plus (Media Cybernetics, Bethesda, MD), with positive area (%) = positive signal area/total area of liver section.

Real-time PCR
Hepatic RNA was isolated by Trizol reagent (Gibco, Invitrogen, Carlsbad, CA) from ,50 mg of liver sample. cDNA was reversely transcribed by the H minus reverse transcription kit (Fermentas, Thermo Fisher Scientific, Waltham, MA). Total 0.05 mg RNAtranscripted cDNA was processed by the SYBR green method with ABI-7500 (Applied Biosystems, Inc., Foster city, CA). The primer sequences for target genes are listed in Table S1. The expression of target genes was normalized to that of glyceralde-hyde 3-phosphate dehydrogenase (GAPDH) and calculated relative to the control.

Statistical analysis
Data are presented as mean 6 SEM. Statistical significance was determined by the Student's t test and the log-rank test by use of GraphPad Prism 5 (GraphPad Software, La Jolla, CA). P,0.05 was considered statistically significant. (See Additional Supporting Methods S1.)

Deficiency of IL-15 enhanced susceptibility to APAPinduced fulminant hepatitis in mice
Although IL-15 was previously shown to attenuate Jo-2 and concanavalin A-induced liver injuries [25,26], its role in druginduced liver injury remained elusive. In our model of hepatitis induced by 400 mg/kg (lethal dose) of APAP in mice, the 5-day mortality was higher for Il15 2/2 than WT mice [100% (8/8) vs. 60% (6/10)] ( Figure 1A). To confirm this result, we examined the serum activities of the liver enzymes ALT and AST in Il15 2/2 and WT mice. Four hours after APAP injection, the enzyme levels significantly increased and peaked between 8 and 12 hours post-APAP injection in Il15 2/2 mice, as compared to WT counterparts ( Figure 1B,C). Another non-lethal dose (200 mg/kg) of APAP was further given to demonstrate this increased APAP sensitivity in Il15 2/2 mice. Compared with WT controls, Il15 2/2 mice showed higher serum ALT levels at 8 hr post-APAP injection ( Figure 1D), which were further confirmed by histopathological examination of damaged livers, with more extensive centrilobular necrosis in Il15 2/2 mice ( Figure 1E). Thus, deficiency of IL-15 increased the susceptibility to AILI in mice.

IL-15 was induced in the KC-enriched fraction in APAPinduced hepatic injury
Because of enhanced APAP hepatotoxicity in Il15 2/2 mice, we further examined IL-15 serum levels and hepatic expression in mice. The serum IL-15 levels significantly increased at 4 hr and reached to peak at 8 hr post-APAP in WT mice (Figure 2A), which were in parallel to liver damages by ALT levels. In contrast, expression of hepatic IL-15 was reduced at 8 but not 2 hr post-APAP in WT mice ( Figure 2B), indicating IL-15 was not upregulated in hepatic parenchymal cells during AILI. Monocytes (e.g., KCs) were previously reported to be the dominant IL-15producing cells [30]. Hence, we examined the expression of IL-15 in the KC-enriched cell fraction. WT mice showed 7-to 8-fold increased mRNA expressions of IL-15 at 8 hr after APAP injection ( Figure 2C). Thus, IL-15 was predominantly up-regulated in KCs but not in hepatocytes during AILI.
Enhanced susceptibility to AILI in Il15 2/2 mice is not mediated by altered metabolism and ROS detoxification of APAP in hepatocytes To exclude that APAP metabolites participate in the heightened APAP hepatotoxicity in Il15 2/2 mice, we examined the expressions of CYP2E1 and CYP1A2, the main enzymes to oxidize excess APAP to toxic NAPQI [1] in liver. Hepatic CYP2E1 and CYP1A2 expressions were similar in WT and Il15 2/2 mice at 0 hr ( Figure 3A,B). Additionally, GSH depletion in liver indirectly reflects the elimination of APAP metabolites. Both groups of mice showed similar GSH levels at 0.5 and 2 hr, with the lowest GSH level at 2 hr post-APAP ( Figure 3C). Interestingly, delayed GSH recovery was observed at 8 and 12 hr post-APAP in Il15 2/2 mice but not in WT mice. We further analyzed in vitro APAP sensitivity of isolated hepatocytes from individual mice. The hepatocyte viability after 4, 6 and 8 hrs of APAP treatment did not differ between WT and Il15 2/2 mice ( Figure 3D).
To explore the causes of diminished GSH recovery in Il15 2/2 mice, we determined synthetic efficiency of GSH by analyzing GCLC expression [31]. However, GCLC levels did not differ in both groups of mice at 0 and 8 hr post-APAP ( Figure 4A,B). This finding indicated the delayed GSH recovery in Il15 2/2 mice was probably attributed to the less residual functional hepatocytes after APAP injection. Nuclear factor erythroid 2-related factor 2 (Nrf2)related and ROS detoxification genes also mediate the sensitivity of APAP hepatotoxicity [32,33]. Therefore, we further examined the activation of Nrf2-related genes, such as NAD(P)H:quinone oxidoreductase 1 (NQO-1), glutathione S-transferase pi 1 (GstPi-1), multidrug resistance protein 2 (MRP-2), MRP-3, etc. and ROS detoxification genes, such as superoxide dismutase-1 (SOD-1), superoxide dismutase-2 (SOD-2), catalase, etc. Except for the minimal suppression of SOD-1 in Il15 2/2 mice, both groups did not differ in mRNA expression levels of these genes ( Figure 4C,D). Thus, metabolism of APAP and induction of ROS detoxification genes in hepatocytes do not contribute to the heightened APAP sensitivity in Il15 2/2 mice.
APAP treatment induced a stronger inflammatory response in Il15 2/2 mice than WT mice Because inflammation caused by necrotic hepatocytes participated in AILI [6], we further examined the APAP-induced inflammatory response in mice. The mRNA levels of hepatic proinflammatory cytokines IL-1b, TNFa and IL-6 were higher at 8 hr post-APAP in Il15 2/2 mice than WT counterparts, with a higher IL-1b level at 2 hr ( Figure 5A). Moreover, the serum levels of IL-1b, TNFa and IFNc and liver levels of IL-1b and IFNc were also higher in Il15 2/2 mice than WT controls ( Figure S1). Il15 2/2 mice showed stronger induction of the adhesion molecules, intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion protein-1 (VCAM-1) as well as neutrophilic chemokines, such as macrophage inflammatory protein-1 alpha (MIP-1a), KC/ GRO and MIP-2a at 8 hr post-APAP ( Figure 5B,C).
Because of the enhanced production of hepatic neutrophilic chemokines in Il15 2/2 mice, we investigated the profile of infiltrated monocytes in mouse livers. NPCs with CD11b high F4/ 80 high expression were referred to KCs [34], and the population of CD11b high F4/80 low cells were confirmed to be neutrophils by anti-Gr-1 antibody (data not shown) [34]. In the resting state, .90% and ,4% of hepatic monocytes were KCs and neutrophils, respectively. APAP-induced neutrophil infiltration of liver was greater in Il15 2/2 than WT mice at 8 hr after challenge, while the infiltration of KCs was relatively lesser ( Figure S2). Therefore, APAP treatment increased inflammatory cytokine production and the extent of inflammatory cell infiltration in livers of Il15 2/2 mice.
APAP-injected Il15 2/2 mice showed discordant induction of hepatic iNOS and HO-1 It has been reported that, NO, a product of iNOS, contributes to the APAP-induced hepatotoxicity [35]. We therefore tested hepatic iNOS induction by immunostaining in APAP-treated mice. The Il15 2/2 mouse livers showed stronger iNOS inductions than those of WT livers at 4 and 8 hr post-APAP injection ( Figure 6A,C), with a similar trend of nitrotyrosine accumulation in both groups of mice ( Figure 6B,C). Collectively, APAP-induced  Effect of IL-15 on APAP Hepatitis PLOS ONE | www.plosone.org reactive nitrogen species (RNS) formation was more pronounced in Il15 2/2 than WT mice.
HO-1, also known as Hsp32, had been reported to protect against AILI [5]. After APAP treatment, hepatic HO-1 mRNA was induced within 2 hr and peaked at 4 hr, then returned to normal after 24 hr post-APAP in WT mice ( Figure 7A). Despite similar hepatic HO-1 mRNA levels at 2 hr post-APAP in WT and Il15 2/2 mice, hepatic HO-1 levels were suppressed at 8 hr post-APAP in Il15 2/2 mice ( Figure 7B). This finding was further confirmed by decreased HO-1 protein expression in livers of Il15 2/2 mice at the same time ( Figure 7C). Previously, HO-1 was predominantly induced in parenchymal cells after APAP injection in rat [5]. Compared with WT counterparts, Il15 2/2 mice showed suppressed induction of HO-1 in hepatocytes ( Figure 7D,E). Likewise, Hsp70 induction protected against APAP hepatitis in mice [36]. In our study, both mRNA and protein levels of hepatic Hsp70 were lesser in Il15 2/2 than WT mice after APAP injection ( Figure S3).

Increased susceptibility to AILI is abrogated by liverspecific transcriptional inhibitor in mice
Pro-inflammatory cytokines such as IL1-b, TNFa and IFNc induce iNOS expressions in hepatocytes [37]. To further clarify the effecter cells in heightened APAP hepatitis of Il15 2/2 mice, galactosamine (GalN) was given to mice for specific inhibition of hepatocyte transcription [38]. Unexpectedly, pre-treatment with GalN abrogated the increased hepatotoxicity in Il15 2/2 mice in terms of serum ALT level at 6 hr post-APAP, with a minor effect on WT counterparts ( Figure 8A, B & C). The reduced APAP hepatitis in Il15 2/2 mice, by GalN pre-treatment, was further confirmed by the serum ALT levels at 4 hr (APAP+GalN: 42506580 U/l, APAP+saline 1125062038 U/l, P,0.05) and 8 hr (APAP+GalN: 1701061493 U/l, APAP+saline: 2986064273 U/l, P,0.05), respectively. However, the GalN reduction effect on APAP hepatitis in WT mice was not as much as that in Il15 2/2 counterparts. With the histopathological examination and liver necrotic quantization, GalN pre-treatment reduced the APAP-related extensive cell-content release of the injured hepatocytes in Il15 2/2 mice ( Figure 8B,C). Similar GSH depletions at 0.5 hr post-APAP were observed in WT and Il15 2/2 mice, with or without GalN pre-treatment ( Figure 8D), which further indicated APAP metabolism, was not affected by GalN pre-treatment. Moreover, without APAP challenge, GalN pretreatment did not affect hepatic GSH levels in WT and Il15 2/2 mice. Thus, heightened susceptibility to AILI in Il15 2/2 mice can be ameliorated by GalN pre-treatment and this effect seems to be not through alteration of hepatic GSH contents at early stage of APAP hepatitis. The increased IL-15 levels in serum and KCs fraction during AILI in WT mice implicated its involvement in APAP hepatitis (Figure 2A,C). Previously, subcutaneous injection of recombinant IL-15 protected mice against concanavalin A-induced hepatitis [25]. Therefore, recombinant murine IL-15 was injected, subcutaneously or intraperitoneally, into Il15 2/2 mice to test the effect of the exogenous cytokine on APAP hepatotoxicity. Intriguingly, there was no significant difference among mice pretreated with recombinant IL-15, subcutaneously or intraperitoneally, or saline based on serum ALT level at 8 hr post-APAP ( Figure 9A). In addition, hepatic histopathological examination showed unremarkable result among these groups ( Figure 9B). Similarly, by effect assessments of serum ALT levels or hepatic histopathologies at 8 hr post-APAP ( Figure 9C,D), pretreatment of WT mice with IL-15 neutralizing antibody did not change the APAP sensitivity. These results suggested that reconstitution or neutralization of IL-15 would not influence the susceptibility of mice to APAP, implicating a probable intracellular, but not extracellular, role of IL-15 within KC during AILI, or otherwise, a secondary phenomenon of elevated serum IL-15 in APAP hepatitis. In view of the up-regulated IL-15 expression in KC fraction of WT mice after APAP challenge ( Figure 2C), we further examined the role of innate immunity in IL-15 protection against AILI.
KCs but not neutrophils might play a major role in the heightened sensitivity to APAP in Il15 2/2 mice Innate immune cells play an important role in APAP hepatitis [39]. To assay whether innate immune cells, such as neutrophils and KCs, are involved in exacerbated AILI of Il15 2/2 mice, we used vinblastine and GdCl 3 for further experiments. Vinblastine, a mitosis inhibitor, was used to induce neutropenia in a mouse ischemia/reperfusion model [40]. Neutrophil depletion was confirmed by flow cytometry in mice (data not shown). Although increased neutrophil infiltration of liver was observed after APAP treatment in Il15 2/2 mice, pre-treatment of vinblastine could not alter the vulnerability of Il15 2/2 mice to AILI, based on the serum  Figure 10A). Similarly, vinblastine could not change the sensitivity to APAP hepatitis in WT mice ( Figure 10A). The role of KCs in APAP hepatotoxicity still remained controversial [41], therefore we further evaluated the role of KCs in our model. A low dose of GdCl 3 (7 mg/kg) was reported to inactivate KCs and modestly reduced APAP hepatitis [11], whereas the high dose of GdCl 3 (30 mg/kg) has been reported to eliminate KCs in mice, thereby abrogating cadmium (Cd) induced hepatotoxicity [42]. Therefore, 30 mg/kg of GdCl 3 , a dose eliminating KCs by Indian ink uptake assay (data not shown) without liver damage, was used in our further studies. The KC-eliminated WT mice, by GdCl 3 , showed an increased susceptibility to APAP ( Figure 10B,D), indicating a protective role of KCs in AILI and the accompanied inflammation ( Figure 10E,F). This finding was compatible with previous protective role of KCs in AILI by Ju et al. in C57BL/6J WT mice [11]. Moreover, GdCl 3 pre-treatment significantly reduced serum IL-15 elevation of WT mice during AILI ( Figure 10C,D). Thus, KCs were probably the dominant IL-15 producer in APAP hepatitis. Since mice lack of KCs shared similar sensitivity to APAP hepatotoxicity with mice deficiency of IL-15 ( Figure 10B), it implicated a regulatory correlation between IL-15 and KCs in AILI. Furthermore, elimination of KCs in Il15 2/2 mice did not affect the AILI severity ( Figure 10B), hepatic TNFa and IL-1b ( Figure 10E), as well as liver nitrotyrosine formation (Figure 10F), inferring a potential role of IL-15 in mediating KC effect on AILI. Further study is needed to clarify this assumption.
The efficacy of APAP metabolism ( Figure 3) and the expressions of hepatic Nrf2-related and ROS detoxification genes ( Figure 4) were similar in WT and Il15 2/2 mice. Therefore, the increased sensitivity of Il15 2/2 mice might not depend on liver parenchymal cells at the earlier stage of APAP toxicity. Indeed, after APAP challenge, we found productions of pro-inflammatory cytokines IL-1b, TNFa and IFNc, as well as vascular adherence molecules and chemokines [6,39], were greater in Il15 2/2 than WT mice Induction of iNOS, by pro-inflammatory cytokines such as IL-1b, IFNc and TNFa in hepatocytes results in RNS formation [37], which further potentiates the APAP-induced hepatotoxicity [35]. It had been reported anti-oxidative proteins, Hsp70 and HO-1, protected mice against APAP hepatotoxicity [5,36]. In a rat heatshock model, Hoetzel et al. demonstrated that an NO donor could diminish HO-1 or Hsp70 induction by suppressing activator protein-1 (AP-1) DNA binding activity or modulating protein translation, respectively [44]. Our results in Il15 2/2 mice ( Figure 6) suggested that excess hepatic iNOS and RNS production might potentiate APAP toxicity by suppressing HO-1 and Hsp70 induction (Figure 7,S3).
Through transcriptional arrest of hepatocytes, GalN with endotoxin or TNFa had been reported to induce acute apoptotic hepatitis in mice [45], whereas GalN pre-treatment markedly diminished the increased APAP hepatitis in Il15 2/2 mice versus a minor advantage in WT controls ( Figure 8A-C). However, the apoptotic pathway was interrupted in our AILI model through mitochondria damage [46] and, moreover, GalN might block the secondary hepatic transcriptional activation from inflammatory mediators. In addition, the divergent responses between Il15 2/2 and WT mice to AILI might be partly attributed to the excess inflammation in the former. The detailed mechanism of GalN protection against APAP hepatitis needed further investigation. But, the protective effect of GalN was, at least, not through alteration of hepatic GSH contents in mice at early stage of AILI ( Figure 8D).
Exogenous IL-15 administration into Il15 2/2 mice or IL-15 neutralizing antibody injection into WT counterparts could not alter the APAP hepatotoxicity (Figure 9), implicating the extracellular IL-15 was insufficient to affect the present fulminant AILI model. Moreover, increased serum IL-15 levels during AILI might be considered as the compensatory effect of hepatic inflammation and the activation of KCs. Although exogenous IL-15 did not affect APAP hepatitis, endogenous IL-15 seemed to play an interesting and pivotal role in AILI of our mice.
Prolonged innate inflammatory response induced by APAPdamaged hepatocyte determines the prognosis of liver injury [6]. APAP-induced inflammation was modulated by innate immune cells such as neutrophils [8], KCs [11] and DCs [9]. Neutrophil elimination did not protect against APAP hepatotoxicity in WT and Il15 2/2 mice ( Figure 10A), which indicates a minor role of neutrophils in AILI [39]. Previously, KC produced hepatoprotective factors (e.g., IL-10 and cyclooxygenase derived mediators) and depletion of KCs enhanced APAP hepatitis [11]. In ischemia-reperfusion (IR) liver injury, GdCl 3 (30 mg/kg) protected mice via reduction of TNFa and IFNc levels in liver effluents of balb/c mice [47], indicating the inflammatory contribution of KCs. However, Jaeschke et. al. [39] suggested the protective role of KCs, but not inflammatory effecter, during AILI, and hepatic inflammation might be enhanced by other NPCs, such as endothelial cells [19,39]. During AILI, cyclooxygenase products from KCs not only suppressed inflammation but also up-regulated HSPs [39], whereas enhanced hepatic inflammation was found in APAP-injected Il15 2/2 mice, with decreased HSP induction ( Figure 5, 7, S3). Additionally, elimination of KCs by GdCl 3 did not alter the severity of AILI in Il15 2/2 mice, but increased the APAP sensitivity of WT mice, leading to a similar susceptibility in both groups of mice ( Figure 10B). Moreover, pretreatment of WT mice with GdCl 3 diminished serum IL-15 elevation during AILI ( Figure 10C) and showed a negative correlation between serum IL-15 level and liver damage extent ( Figure 10D). Furthermore, with GdCl 3 pretreatment, hepatic TNFa, IL-1b and nitrotyrosine levels were up-regulated in WT mice, indicating the inflammation prevention by KCs in our model ( Figure 10E,F). However, depletion of KCs did not affect the enhanced hepatic inflammation in Il15 2/2 mice after APAP challenge ( Figure 10E,F). These findings suggested the probable upstream and downstream relationship of KCs and IL-15 in the protection against AILI. IL-15 had been found to determine the activation level of macrophages in mouse rheumatoid arthritis model [48]. In addition, IL-15, at extremely low concentration, selectively suppresses pro-(e.g., TNFa, IL-6 and IL-1) but not antiinflammatory (e.g., IL-10) cytokine production in LPS-activated macrophage [24]. Thus, IL-15 might potentially influence KC function and thereby modulate the inflammatory response during AILI.
There were two isoforms of IL-15, short and long signal peptide isoforms, existing in human and murine cells [49,50,51,52]. After translation, the short (intracellular) IL-15 isoform was not secreted but rather distributed within the cytoplasm and the nucleus [52]. Although the biological function of intracellular IL-15 had not been fully elucidated in KCs, it could function as a histone deacetylase inhibitor regulating the expression of IL-12 receptor b1 in macrophage [53]. Moreover, intracellular IL-15 could indirectly regulate mMCP-2 induction by altering the expressions of transcription factors C/EBPb and YY1 in mouse mast cells [54]. Therefore, based on previous reports and our results, IL-15 might function uniquely within KCs in AILI.
In previous studies, IL-15 was found to enhance severity of endotoxin shock liver injury [27], colitis [55] and virus infection [56]. In contrast, IL-15 protected against nephrotoxic serum nephritis [57] and experimental autoimmune encephalomyelitis [58]. Increased serum IL-15 level was also noted in patients with liver transplantation [59] or hepatitis C virus infection [60]. Here, we demonstrated an enhanced APAP susceptibility in Il15 2/2 mice, which might result from an overactive inflammatory response. In view of these results, further studies might be justified to elucidate the biological function of intracellular IL-15 within KCs during AILI.