Male C57BL/6 WT mice (8-week-old, 20–25 g) were obtained from Clea Japan (Tokyo, Japan). Male Cd1d-deficient (Cd1d^-/-) mice (8-week-old) were generated as described previously (23, 24). Mice were maintained at constant humidity (50% ± 5%) and temperature (25°C ± 1°C) on a 12 h light/dark cycle, and were provided food and water ad libitum. All experimental procedures were approved by the Animal Experimentation and Ethics Committee of the Kitasato University School of Medicine (2020–111) and were performed in accordance with institutional guidelines for animal experimentation, which are based on the Guidelines for Proper Conduct of Animal Experiments published by the Science Council of Japan.

The animals underwent either sham surgery or I/R. Partial hepatic ischemia was induced as previously described (3). Briefly, mice were anesthetized with a combination cocktail containing 0.3 mg/kg of medetomidine hydrochloride (Nippon Zenyaku Kogyo Co., Ltd., Fukushima, Japan), 4.0 mg/kg of midazolam (Astellas Pharma Inc., Tokyo, Japan), and 5.0 mg/kg of butorphanol (Meiji Seika Pharma Co., Ltd., Tokyo, Japan). Laparotomy was performed and the blood supply to the median and left hepatic lobes was occluded for 60 min using an atraumatic vascular clamp (ROBOZ Surgical Instrument, Washington, DC, USA). Reperfusion was initiated by removing the clamps. After reperfusion was initiated, 4 mL/kg of physiological saline was injected and the abdomen was closed with 4-0 nylon sutures. After the procedure, the effect of medetomidine was reversed by ip injection of 0.75 mg/kg atipamezole (Nippon Zenyaku Kogyo Co., Ltd.), and the animals were allowed to recover. Sham control mice underwent the same protocol without vascular occlusion.

Mouse livers were subjected to ischemia followed by reperfusion. Blood was drawn and livers were excised at 6, 24, 48, 72, and 96 h after reperfusion. Serum was used to determine alanine aminotransferase (ALT) activity using a Dri-Chem 7000 Chemistry Analyzer System (Fujifilm, Tokyo, Japan). Immediately after blood collection, the livers were excised and rinsed with saline. A small section of each liver was placed in 10% formaldehyde, and the remaining liver was frozen in liquid nitrogen and stored at -80°C.

Animals received α-GalCer (0.1 mg/kg body weight, ip; Funakoshi, Tokyo, Japan) dissolved in 0.1% DMSO and diluted in phosphate-buffered saline (PBS) at the induction of ischemia. The dose of α-GalCer was chosen based on our previous results (25).

In separate experiments, mice were injected ip with 500 μg of a neutralizing monoclonal antibody specific for mouse IL-4 (BioLegend, San Diego, CA, USA) (26) or 500 μg of a neutralizing monoclonal antibody specific for mouse IFN-γ (BioLegend) (27) 1 h before the induction of ischemia. As a control group, mice were administered isotype-matched control IgG (BioLegnd).

Excised liver tissues were fixed immediately with buffered 10% formaldehyde for histological analysis. Sections (4 µm thick) were prepared from paraffin-embedded tissues and subjected to either hematoxylin and eosin (H&E) staining or immunostaining. Images of H&E-stained sections were captured using a microscope (Biozero BZ-700 Series; Keyence, Osaka, Japan). The level of necrosis (as a percentage of the total area) was estimated by measuring the necrotic area relative to the entire histological section using the ImageJ software (US National Institutes of Health, Bethesda, MD, USA). The results were expressed as percentages.

Sections were also stained with an antibody against proliferating cell nuclear antigen (PCNA; Thermo Fisher Scientific, Waltham, MA, USA). Immune complexes were detected using Histofine Simple Stain MAX PO (MULTI) (Nichirei, Tokyo, Japan), following the manufacturer’s protocol. The number of PCNA⁺ hepatocytes was counted in six fields (400×) per animal. The percentage of PCNA⁺ hepatocytes was then calculated, and the results were expressed as percentages.

Tissue samples were fixed with periodate-lysine-paraformaldehyde fixative at room temperature for 3 h. Following cryoprotection with 30% sucrose prepared in 0.1 M phosphate buffer (pH 7.2), fixed liver samples were embedded in Tissue-Tek O.C.T. compound (Sakura Finetek USA, Torrance, CA, USA) and frozen at –80°C. Samples were cut into 8 μm sections using a cryostat and incubated for 1 h at room temperature with Dako Protein Block Serum-Free solution (Glostrup, Denmark) to block any nonspecific binding. The sections were then incubated overnight at 4°C with a rat anti-CD1d monoclonal antibody (BD Bioscience, San Jose, CA, USA) and a rabbit anti-F4/80 polyclonal antibody (Cell Signaling Technology, Beverly, MA, USA). After washing three times in PBS, the sections were incubated for 1 h at room temperature with the following secondary antibodies: Alexa Fluor 488-conjugated donkey anti-rabbit IgG and Alexa Fluor 594-conjugated donkey anti-rat IgG (Invitrogen, Carlsbad, CA, USA). Images of the stained sections were captured using a fluorescence microscope (Biozero BZ-700; Keyence).

Total RNA was extracted from mouse tissues and homogenized using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). Single-stranded cDNA was generated from 1 μg of total RNA by reverse transcription using the ReverTra Ace qPCR RT kit (Toyobo, Osaka, Japan), according to the manufacturer’s instructions. Quantitative PCR was performed using TB Green Premix Ex Taq II (Tli RNaseH Plus; Takara Bio, Shiga, Japan). The gene-specific primers used for real-time RT-PCR were designed using the Primer 3 software (http://primer3.sourceforge.net/) based on data from GenBank. The primer sequences are listed in Table 1 . Data were normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in each sample.

Under anesthesia induced by ip injection of a mixture of 0.75 mg/kg medetomidine, 4.0 mg/kg midazolam, and 5.0 mg/kg butorphanol, the liver was perfused with 10 mL of 1× Hank’s balanced salt solution via the portal vein. The excised livers were immediately placed in RPMI at room temperature, minced using scissors, and incubated at 37°C for 20 min in RPMI containing 0.05% collagenase (Type IV; Sigma Chemical Co., St. Louis, MO, USA). The tissues were then pressed through a 70 μm cell strainer. Cells were centrifuged at 1290 × g for 10 min at 4°C. Pelleted cells were resuspended in PBS. Non-parenchymal cells were isolated from liver homogenates by density-gradient centrifugation on 33% Percoll (GE Healthcare Life Sciences, Piscataway, NJ, USA), as previously reported (28). Viable nucleated cells were counted by trypan blue exclusion and resuspended at a uniform cell density.

The cells were incubated with the 2.4G2 mAb (anti-cγRIII/II) to block non-specific binding of the primary mAb. Then, the cells were stained with a combination of the following fluorochrome-conjugated antibodies: FITC-conjugated anti-TCRβ (H57-597, BioLegend), anti-F4/80 (BM8, BioLegend), APC-conjugated anti-CD1d antibody (1B1, BioLegend), APC/CY7-conjugated anti-Ly6G (1A8, BioLegend), PE/Cy7-conjugated anti-CD11b (M1/70, BioLegend), PE-conjugated anti-CD45 (30-F11, BioLegend), Brilliant Violet 510-conjugated anti-Ly6C (HK1.4, BioLegend), Brilliant Violet 421-conjugated anti-CD45 (30-F11, BioLegend), and α-GalCer (PBS-57)-loaded CD1d tetramer and control CD1d-tetramer were kindly provided by The NIH Tetramer Core Facility (NTCF) at Emory University (Atlanta, GE, USA). Tubes containing the mixture were placed in the dark on ice for 30 min. Pellets were washed twice with PBS. For flow cytometric analysis, cells were initially gated on forward scatter (FSC) and side scatter (SSC), and then gated on CD45⁺ cells. Cells positive for 7-aminoactinomycin D (BioLegend), which stains dead cells, were gated electronically and removed from the analysis. Samples were analyzed on a FACSVerse cytometer (BD Biosciences, Franklin Lakes, NJ, USA). Data were analyzed using Kaluza software v2.1 (Beckman Coulter, Brea, CA, USA). We quantified and the cells as a percentage of total live CD45⁺ cells (CD45⁺7ADD^- cells).

For intracellular cytokine staining, cells suspensions from liver tissue were stained for surface marker expression before fixation and permeabilization using the FOXP3 Fix/Perm Buffer Set (BioLegend). Treated cells were stained with APC-conjugated anti-IL-4 mAb (11B11, BioLegend), anti-IFN-γ mAb (XMG1.2, BioLegend), and PE/Cy7 conjugated anti-IL-13 mAb (eBio 13A, eBioscience, San Diego, CA, USA). Stained cells were acquired with FACS Verse flow cytometers (BD Biosciences) and analyzed using the FlowJo software (FlowJo, LLC, Ashland, OR, USA).

BM cells were isolated from the femurs and tibias of 8-week-old mice. The bones were flushed with PBS and the erythrocytes were lysed by treatment with RBC lysis buffer (BioLegend). BM cells were cultured in RPMI 1640 medium containing 10% fetal bovine serum and plated in 6-well plates at a density of 1 × 10⁶/mL. Cultures were incubated at 37°C in a humidified atmosphere containing 5% CO₂. To generate BM-derived macrophages, cells were cultured for 7 days in the presence of 20 ng/mL murine recombinant macrophage colony-stimulating factor (BioLegend). iNKT cells (TCRβ⁺α-GalCer-loaded CD1d tetramer⁺ cells) were purified by cell sorting using a BD FACSAria™ II (BD Biosciences). For the co-culture experiments, sorted hepatic iNKT cells (1 × 10⁵/well) and macrophages on day 7 of culture (3 × 10⁵/well) were seeded into 12-well plates in RPMI 1640 medium supplemented with 5% FCS. The co-culture was maintained in the presence of α-GalCer (100 ng/mL) for 72 h (29). Antibody-mediated neutralization experiments were performed in the presence of an additional 5 µg/mL of anti-mouse IL-4 monoclonal antibody (BioLegend) (30), 5 µg/mL of anti-mouse IFNγ monoclonal antibody (BioLegend) (31), or 5 µg/mL of isotype control IgG (BioLegend).

Both non-adherent (floating cells) and adherent cells were then harvested. The cells were homogenized in TRIzol (Life Technologies) and their mRNA levels were measured using real-time RT-PCR. Supernatants from co-cultured iNKT cells and macrophages were collected to measure cytokine levels using the cytometric bead array (BD Bioscience) and data were analyzed using the BD Biosciences FCAP software (V3.0).

All results are presented as mean ± standard deviation (SD). All statistical analyses were performed using the GraphPad Prism software, version 8 (GraphPad Software, La Jolla, CA, USA). Data were compared between two groups using unpaired two-tailed Student’s t-tests and between multiple groups using one-way analyses of variance followed by Tukey’s post-hoc tests. A p-value < 0.05 was considered statistically significant. Significance is depicted with asterisks on graphs as follows: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

To examine whether activation of iNKT cells facilitated liver repair after hepatic I/R injury, mice were treated with α-GalCer at the induction of ischemia of the liver. In mice treated with α-GalCer or the vehicle, ALT levels peaked at 6 h post-reperfusion and decreased gradually ( Figure 1A ). There was no statistically significant difference in ALT levels at 6 h and 24 h between the two groups. At 48 h post-reperfusion, ALT levels in α-GalCer-treated mice were lower than those in vehicle-treated mice. Consistently, changes in hepatic necrosis area were similar to those in ALT levels at 24 h ( Figure 1B ), while the necrotic area in α-GalCer-treated mice at 48 h post-reperfusion was lower than that in vehicle-treated mice ( Figure 1B ). We also employed immunocytochemistry for PCNA to examine hepatocyte proliferation during hepatic I/R injury. In vehicle-treated mice, the percentage of PCNA⁺-hepatocytes increased 48 h post-reperfusion, which is consistent with the results in previous reports showing that the liver enters a proliferative phase 48 h post-reperfusion (3). In contrast, the number of PCNA⁺-hepatocytes in α-GalCer-treated mice began to increase at 24 h post-perfusion, and the percentage of PCNA⁺ hepatocytes in α-GalCer-treated mice was higher than that in vehicle-treated mice at 24 h ( Figure 1C ). These results indicated that α-GalCer administration appeared to accelerate the resolution of live inflammation and liver repair after hepatic I/R injury, as evidenced by decreases in ALT levels and necrotic area, and increases in the number of PCNA⁺-hepatocytes.

In sham-operated mice treated with α-GalCer, ALT levels were moderately elevated at 24 h and 48 h after sham-operation and decreased at 72 h ( Figure S1 ). The ALT levels in sham-operated mice were remarkably lower than those in α-GalCer-treated mice subjected to hepatic I/R injury. Although H&E staining of the livers obtained from sham-operated mice treated with α-GalCer exhibited the formation of small inflammatory foci, hepatic necrosis similar to that observed in mice subjected to hepatic I/R injury ( Figure 1B ) was not observed in sham-operated mice. In addition, PCNA expression in the hepatocytes was not up-regulated.

When iNKT cells were identified as TCRβ⁺ and CD1d-tetramer⁺, using flow cytometry analysis ( Figure 1D , left panel), the percentage of iNKT cells in total live CD45⁺ cells was decreased at 6 h post-reperfusion and markedly decreased at 24 h and 48 h post-reperfusion in both α-GalCer- and vehicle-treated mice ( Figure 1D , right panel). Although the population of iNKT cells in vehicle-treated mice remained low at 72 h post-reperfusion, it was slightly elevated at 96 h post-reperfusion. In contrast, the percentage of iNKT cells in α-GalCer-treated mice at 72 and 96 h was larger than that in vehicle-treated mice.

During the injury and repair phases of hepatic I/R injury, genes related to pro-inflammatory macrophages and reparative macrophages in liver tissues are expressed (3, 28). Accordingly, we compared the expression of mRNA encoding pro-inflammatory macrophage-related genes (tumor necrosis factor, Tnf; interleukin-1β, Il1b; interleukin-6, Il6; nitric oxide synthetase-2, Nos2; and interferon-γ, Ifng) and mRNA encoding reparative macrophage-related genes (mannose receptor, Mrc1; resistin-like molecule-α, Retnla; interleukin-4, Il4; and interleukin-13, Il13) between the two groups ( Figure 2A ). The mRNA levels of pro-inflammatory macrophages, including Tnf, Il1b, Il6, and Ifng, in the livers of α-GalCer-treated mice at 6 h post-reperfusion were higher than those in the livers of vehicle-treated mice. In addition, the mRNA levels of Nos2 and Ifng in α-GalCer-treated mice increased during the repair phase of hepatic I/R injury. With respect to the genes related to reparative macrophages, the levels of Mrc1, Retnla, and Il4 in α-GalCer-treated mice in the repair phase of hepatic I/R injury were higher than those in vehicle-treated mice. Further, the expression levels of mRNAs encoding IL-4 and IL-13 in the livers of α-GalCer-treated mice at 6 h post-reperfusion were higher than those in the livers of vehicle-treated mice.

To further identify macrophage subsets during hepatic I/R injury, Ly6G^high and CD11b⁺ populations were gated out from live CD45⁺ cells to exclude neutrophils. Next, we examined the subpopulations of liver macrophages isolated from mice treated with α-GalCer or the vehicle ( Figure 2B ) and found two different phenotypes of macrophages: pro-inflammatory macrophages, namely the Ly6G-/Ly6C^high/CD11b^high/F4/80^high cells (Ly6C^high macrophages), and reparative macrophages, namely the Ly6G-/Ly6C^low/CD11b^high/F4/80^high cells (Ly6C^low macrophages). Flow cytometry analysis revealed that the percentage of Ly6C^high macrophages in vehicle-treated livers was elevated at 6 h, and these levels were sustained until they decreased at 96 h ( Figure 2B ). The percentage of Ly6C^high macrophages at 6 h and 24 h post-reperfusion in α-GalCer-treated livers was further increased than that in vehicle-treated livers ( Figure 2B ). The time course of changes in Ly6C^high macrophages in α-GalCer-treated livers appeared to be similar to that in vehicle-treated livers ( Figure 2B ).

Regarding the changes in the reparative macrophages, the percentage of Ly6C^low macrophages in vehicle-treated livers appeared to increase from 24 h, reaching maximal levels at 72 h ( Figure 2B ). In contrast, the percentage of Ly6C^low macrophages in the livers of α-GalCer-treated mice showed significant increase at 24 h, reached a peak at 48 h, and then decreased. Taken together, these results indicate that α-GalCer treatment increased both pro-inflammatory and reparative macrophage populations and induced early switching from a pro-inflammatory to a reparative macrophage phenotype during hepatic I/R injury.

Flow cytometry also revealed neutrophils (Ly6G⁺/CD11b⁺) to be the major infiltrating leukocyte population in livers exposed to hepatic I/R injury ( Figure S2B ). The accumulation of neutrophils was evident and similar at 6 h post-reperfusion in both α-GalCer- and vehicle-treated livers. The neutrophil level subsequently decreased, reaching a nadir at 96 h post-reperfusion. At 48 h and 72 h post-reperfusion, the percentage of neutrophils in α-GalCer-treated mice was lower than that in vehicle-treated mice. KC populations (Ly6G-/Ly6C-/CD11b^high/F4/80^high) in mice treated with either α-GalCer or the vehicle were markedly reduced after hepatic I/R injury. KCs in vehicle-treated mice began to re-appear and gradually increased in population at 72 h and 96 h post-reperfusion ( Figure S2C ). The populations of KCs in α-GalCer-treated mice remained low after hepatic I/R injury, and the levels at 96 h were lower than those in vehicle-treated mice.

The switch in macrophage phenotypes in the injured region is responsible for liver repair after acute liver injury (2, 32). Cytokines, including IL-4, IL-13, and IFN-γ, are involved in the transition from pro-inflammatory to reparative macrophages. To determine whether iNKT cells played a role in modulating the local reprogramming of macrophages after injury, we measured the levels of these cytokines produced by activated iNKT cells during hepatic I/R injury ( Figure 3 ). α-GalCer treatment enhanced iNKT cell-derived IL-4 and IFN-γ levels, but not IL-13 levels, at 6 h post-reperfusion, whereas iNKT cells produced lower levels of these cytokines without of α-GalCer stimulation.

Because IL-4 and/or IFN-γ appear to be involved in macrophage polarization toward a reparative or pro-inflammatory phenotype, respectively, we examined whether these cytokines contributed to stimulating the phenotypic switch of macrophages associated with liver repair ( Figure 4 ). First, we examined the effect of IL-4 blockade on macrophage shift and liver repair. Anti-IL-4 neutralizing antibodies or control IgG were administered to mice 1 h before the induction of ischemia concomitant with administration of α-GalCer. Treatment with anti-IL-4 antibodies increased the percentage of Ly6C^high macrophages at 24 h and 48 h post-reperfusion and, at the same time, decreased the percentage of Ly6C^low macrophages ( Figure 4A ). Additionally, the accumulation of neutrophils in the livers did not differ between the two treatment groups ( Figure S3 ). There was also no statistical difference in the KC population between the two treatment groups ( Figure S3B ). Treatment with anti-IL-4 antibodies increased ALT levels and necrotic area and decreased the number of PCNA⁺-hepatocytes at 48 h post-reperfusion ( Figure 4B ). Further, the levels of genes related to pro-inflammatory macrophages, including Tnf, Il1b, Il6, Nos2, and Ifng, in mice treated with anti-IL-4 antibodies were higher than those in mice treated with control IgG. There were significant differences in the expression of genes related to reparative macrophages, including Mrc1, Retnla, and Il13, between the two treatments. Collectively, these results suggested that IL-4-mediated stimulation of macrophage polarization from a pro-inflammatory to a reparative phenotype contributes to liver repair.

We also wanted to know whether IFN-γ mediates macrophage polarization toward a pro-inflammatory phenotype in α-GalCer-treated mice. To this end, α-GalCer-treated mice were pre-treated with anti-IFN-γ neutralizing antibodies prior to the induction of hepatic I/R injury. Flow cytometry analyses revealed that treatment with anti-IFN-γ antibodies decreased the percentage of Ly6C^high macrophages in α-GalCer-treated livers at 24 h and 48 h post-reperfusion compared with that in IgG-treated livers, but no changes in Ly6C^low macrophage percentages were observed ( Figure 4A ). These results suggest that IFN-γ might have mediated the switch of macrophages into a pro-inflammatory phenotype, but it did not affect the transition of macrophages into a reparative phenotype. Anti-IFN-γ antibody treatment reduced the percentage of neutrophils at 24 h post-reperfusion, compared with control IgG treatment, but did not change the percentage of KC populations ( Figure S3 ).

Treatment with anti-IFN-γ antibodies decreased the ALT levels and necrotic area and increased the number of PCNA⁺-hepatocytes at 48 h post-reperfusion ( Figure 4B ). The expression levels of genes related to pro-inflammatory macrophages, including Tnf and Nos2, in IFN-γ antibody-treated mice were lower than those in IgG-treated mice. In addition, the expression levels of genes related to reparative macrophages, including Mrc1, Retnla, and Il13, were higher in IFN-γ antibody-treated mice than in IgG-treated mice.

Taken together, these results suggested that IFN-γ contribute to an increase in the percentage of pro-inflammatory macrophages, which was associated with the enhancement of liver inflammation and suppression of hepatocyte proliferation.

We examined whether the α-GalCer-activated iNKT cells interacted with macrophages expressing CD1d in situ. Immunofluorescence analysis revealed that CD1d was expressed in F4/80⁺ cells (macrophages) at 6 h post-reperfusion ( Figure S4A ). Further, flow cytometry analysis showed that the Ly6C^high macrophage populations were strongly positive for CD1d ( Figure S4B ). Because there were few Ly6C^low macrophages at 6 h, CD1d expression levels were not detected in these cells. At 24 and 48 h post-reperfusion, both Ly6C^high and Ly6C^low macrophages expressed CD1d. These results suggested that macrophages were a source of CD1d expression, which interacted with iNKT cells in the early phase of hepatic I/R injury.

Next, we investigated whether activated iNKT cells affected macrophage differentiation using co-culture systems of BM-derived macrophages with iNKT cells ( Figure S5 ). First, we determined the characteristics of adherent and non-adherent cells in the culture wells. Adherent cells displayed an increase and decrease in Itgam and Cd3e levels, respectively, whereas non-adherent cells showed the opposite phenotype ( Figure S5A, B ). These results indicated that adherent cells were recognized as macrophages and that non-adherent cells were iNKT cells in this co-culture system.

We also examined whether iNKT cells stimulated a shift in macrophage phenotype in vitro ( Figure 5 ). Naïve macrophages treated with α-GalCer did not up-regulate the mRNA expressions related to the phenotypes of pro-inflammatory and reparative macrophages. In addition, naïve macrophages co-cultured with iNKT cells in the absence of α-GalCer did not up-regulate the expression of mRNA related to a pro-inflammatory macrophageo phenotype (Tnf, Il1b, Il6, and Ifng, but not Nos2) or reparative macrophage phenotype (Mrc1, Retnla, Il4,and Il13). In contrast, naïve macrophages co-cultured with iNKT cells in the presence of α-GalCer increased the expression levels of genes related to a pro-inflammatory macrophage phenotype (Il1b, Il6, and Ifng) and a reparative macrophage phenotype (Mrc1, Retnla, Il4,and Il13). These results indicated that activated iNKT cells interacted with macrophages to differentiate them into pro-inflammatory and reparative phenotypes. Further, we examined the expression of CD1d in macrophages. Cd1d1 was expressed in naïve macrophages, and α-GalCer treatment alone did not affect this expression. The expression of Cd1d1 in macrophages was up-regulated in macrophages co-cultured with activated iNKT cells, suggesting that the interaction of iNKT cells with macrophages through CD1d converted the macrophage phenotypes.

We assessed whether the iNKT cells produced IL-4 and IFN-γ, which polarize the macrophage phenotypes ( Figure S6 ). iNKT cells significantly increased the mRNA ( Figure S6A ) and protein levels ( Figure S6B ) of these cytokines in the co-culture system.

In co-cultured macrophages with activated iNKT cells, treatment with IL-4 antibodies increased the expression of genes related to pro-inflammatory macrophages, including Tnf and Il1b, and decreased the expression of genes related to reparative macrophages, including Mrc1 and Il13 ( Figure 6A ). However, treatment with IL-4 antibodies decreased the mRNA levels of Il6 and Nos2, but caused no change in Retnla and Il4 levels. Anti-IL-4 antibodies did not change Il4 mRNA levels, but decreased Ifng mRNA levels in iNKT cells ( Figure 6B ). In addition, treatment with anti-IFN-γ antibodies decreased the expression of genes related to a pro-inflammatory macrophage phenotype (Il1b, Il6, Nos2, and Ifng, but not Tnf) in macrophages, and increased the expression of genes related to a reparative macrophage phenotype (Retnla and Il4, not Mrc1 and Il13) in macrophages ( Figure 6A ). Anti-IFN-γ antibodies increased the mRNA levels of Il4 from iNKT cells, but decreased mRNA levels of Ifng ( Figure 6B ).

Finally, we examined whether iNKT-cell deficiency affects liver repair after hepatic I/R injury in Cd1d^-/- mice. Cd1d^-/- mice exhibited delayed liver repair as indicated by increases in ALT levels and hepatic necrotic area and decrease in the number of PCNA⁺-hepatocytes at 48 h and 72 h post-reperfusion ( Figure 7 ). Flow cytometry analyses demonstrated that the number Ly6C^high macrophages in Cd1d^-/- mice at 24 h was larger than that in WT mice, and that the number of Ly6C^low macrophages in Cd1d^-/- mice at 48 h and 72 h was lower than that in WT mice. Expression of genes encoding TNFα, IL-1β, IL-6, and iNOS in the livers of Cd1d^-/- mice was higher than that in the livers of WT mice at 48 h post-reperfusion. In contrast, the expression of mRNAs encoding MR, Fizz1, IL-4, and IL-13 in the livers of Cd1d^-/- mice was lower than that in the livers of WT mice during the repair phase of hepatic I/R injury. In addition, compared with WT mice, Cd1d^-/- mice displayed an increase in mRNA levels of Ifng and a decrease in those of Il4 at 6 h post-reperfusion. To further evaluate the role of iNKT cells in liver repair, Cd1d^-/- mice were treated with α-GalCer or a vehicle as a control. As shown in Figure S7A , there were no statistical differences in the levels of ALT, necrotic area, and PCNA⁺-hepatocytes between vehicle- and α-GalCer-treated Cd1d^-/- mice at 72 h post-reperfusion. In addition, the expression levels of genes related to pro-inflammatory and reparative macrophages did not differ between the two treatments ( Figure S7B ).