Prf^−/− and IFN-γR^−/− mice were purchased from the Jackson Laboratory. C57BL/6J mice were purchased from Guangdong Medical Laboratory Animal Center (Guangzhou, China). All mice were males and received either a normal control diet (SFD) or HFD (60 kcal % fat; Research Diets) beginning at an age of 6–8 weeks old. All mice were maintained under specified pathogen-free conditions at Jinan University (Guangzhou, China). Animal procedures were approved by and performed in accordance with the Jinan University's Institutional Laboratory Animal Care and Use Committee guidelines.

The protocol used for isolating murine liver mononuclear cells (MNCs) was as described previously (22). Liver tissue was obtained from mice, and the tissue was dissociated to procure MNCs. To obtain liver MNCs, murine livers were pressed through a 200-gauge stainless steel mesh and suspended in either RPMI-1640 medium or PBS. The cells were then centrifuged at 50 g for 1 min. The cell suspension was collected and centrifuged again at 974 g for 10 min. The cell pellet containing MNCs was then resuspended in 40% Percoll (GE Healthcare, Uppsala, Sweden), after which the cell suspension was overlaid on 70% Percoll and centrifuged at 1,260 g for 30 min. The resulting cell pellets were collected from the interphase following two additional washings in PBS or RPMI-1640 medium.

Mice were fasted overnight. Then, whole blood was collected, and serum alanine aminotransferase (ALT) and cholesterol levels were determined using an automatic biochemistry analyzer (7600-020, Hitachi, Japan).

Mice were fasted overnight, and 0.1 g of liver tissue was harvested from the mice in 1 ml of PBS. Liver tissue was then homogenized by hand and centrifuged at 3,000 rpm for 10 min, after which the supernatant was carefully collected. All steps were performed at 4°C. IL-6, IFN-γ, and TNF-α levels in liver supernatants were determined using a commercially available mouse enzyme-linked immunosorbent assay (ELISA) kit (eBioscience, San Diego, CA, USA) according to the manufacturer's instructions.

Non-parenchymal cells were transferred to a new well and treated with 1:1000 GolgiPlug, 1 ng/ml ionomycin, and 50 ng/ml PMA for 4–6 h. Intracellular and cell surface staining was performed as described in the fixation/permeabilization kit (554714; BD) protocol. Cells were stained with the surface markers PEcy7-anti-mouse CD3, PE-anti-mouse NK1.1, FITC-anti-mouse CD4, and PerCPCY5.5-anti-mouse CD8 for 15 min at 4°C. Cells were stained for cytokines with BV421 anti–mouse IFN-γ and APC-IL-17A for 30 min at 4°C, washed with PBS, and analyzed using FACS verse flow cytometry (BD). Data were analyzed using FlowJo (TreeStar).

To deplete CD4⁺ T cells, 200-μg doses of anti-CD4 monoclonal antibody (clone: GK1.5; Sungene Biotech) per mouse were intraperitoneally injected weekly during HFD challenge. Sterile-filtered PBS was used as a control.

Liver tissue was harvested and fixed in 4% (w/v) paraformaldehyde, and 4 mm-thick sections that had been affinized and rehydrated were stained with hematoxylin and eosin (H&E). Hepatic lipid content was determined using frozen sections embedded in Tissue-Tek O.C.T. compound and stained with Oil Red O (Sigma-Aldrich, St. Louis, MO, USA). Images were acquired on a Leica DM3000 microscope.

Liver tissue was harvested, fixed in 4% (w/v) paraformaldehyde, and cut in 4 mm-thick sections. Liver sections were then perfused with 30 ml of 4% paraformaldehyde for fixation. Sections were then incubated with the following dilutions of mouse-specific primary antibodies: 1:200 anti-F4/80 (ab16911, Abcam) and 1:200 iNOS antibody (GTX74171, Gentex). For visualization, 1:200 fluorescent Alexa Fluor 594 and FITC 488 secondary antibodies (Invitrogen Vector) were used for both individual staining and co-staining at room temperature for 2 h. After washing, tissue sections were fixed with Vectashield containing DAPI for visualization. A laser cofocal microscopy (TCS SP8, Leica) was used to capture images and conduct further analysis. For the microscopy images displaying M1 (iNOS+ F4/80+) or total macrophages (F4/80+), 4 slides per mouse liver tissue were prepared and 4 fields were captured from each slide. The quantification of M1 or total macrophages was conducted in these 16 fields and designated as one biological independent sample, and the percentage of M1 in total macrophages was calculated and shown.

The protocol for quantifying hepatic triglyceride (TG) levels was carried out as described previously (23). Briefly, 20–30 mg of liver tissue was homogenized in 500 μl of PBS and mixed with chloroform/methanol 2:1 (vol/vol). The organic phase was transferred, air-dried overnight, and resuspended in 1% Triton X-100 in absolute ethanol. The concentration of TGs was then quantified using a serum triglyceride determination kit (Sigma, Triglyceride Reagent T2449 and Free Glycerol Reagent F6428).

Total liver RNA was isolated using TRIzol Reagent (DP424, Tiangen, China). cDNA synthesis was performed using a Prime Script RT Reagent Kit (Takara, Shiga, Japan). Levels of mRNA expression were quantified by real-time PCR (RT-PCR). RT-PCR was performed using TB Green (Takara). Primer sequences are shown in the Table 1.

Data are presented as the mean ± SEM. Statistical significance between two groups was evaluated using a two-tailed unpaired Student's t-test. Values of P < 0.05 were considered to be statistically significant. The data shown in each panel of these figures were collected from a single experiment; each experiment was repeated for at least three times and showed consistent results. Moreover, the statistical analysis was conducted on each single experiment.

NAFLD is common in the elderly, in whom it carries a more substantial burden of hepatic (non-alcoholic steatohepatitis, cirrhosis, and hepatocellular carcinoma) and extra-hepatic manifestations and complications (cardiovascular disease, extrahepatic neoplasms) than in younger age groups (24). Aged Prf^−/− mice have been reported to have accumulation of senescent cells and development of chronic systemic and local inflammation in the liver (25, 26). We hypothesized that aged Prf^−/− mice might also have more severe hepatic morbidities since inflammation correlates with liver dysfunction. To test this hypothesis, we first determined liver weights and liver injury (ALT) levels in aged WT and Prf^−/− mice at 14 months of age. As expected, the aged Prf^−/− mice showed significantly increased liver weight (Figure 1A), elevated liver damage, and increased lipid accumulation as shown by levels of ALT that trended as increased and significantly increased liver TG levels (Figure 1B). Furthermore, liver histological analysis revealed more severe hepatic steatosis and significantly increased accumulation of lipid in aged Prf^−/− mice compared with WT mice (Figure 1C). These results indicated that the perforin deficiency aggravates liver injury and steatosis in aged mice.

The role of perforin in NAFLD was then investigated using an HFD-induced NAFLD model. Prf^−/− and WT mice at 6–8 weeks of age were fed on SFD or HFD for 10 weeks to induce NAFLD. As expected, HFD challenge was associated with elevated body weight and ALT activation in WT mice (Figures 2A,B). The increase in body weight in response to HFD challenge was comparable between WT and Prf^−/− mice; however, Prf^−/− mice had significantly enlarged livers (Figure 2A). Additionally, the Prf^−/− mice had more severe liver damage as indicated by higher ALT levels. The livers of HFD Prf^−/− mice exhibited significantly increased lipid accumulation (Figure 2C). Histological analysis of livers indicated that HFD Prf^−/− mice had more severe hepatic steatosis and lipid accumulation when compared with WT controls (Figure 2D). Moreover, RT-PCR analysis of liver samples from HFD mice showed that the expression levels of genes involved in lipid production such as fatty acid binding protein 4 (Fabp4) and peroxisome proliferator-activated receptor gamma (PPARγ) were significantly increased, whereas expression levels of lipid catabolism-related genes such as carnitine palmitoyl transferase 1 (CPT-1α) and aldehyde oxidase (AOX-1) were significantly decreased in HFD Prf^−/− mice (Figure 2E). These results indicated that perforin deficiency with HFD challenge aggravated liver injury and steatohepatitis.

Pro-inflammatory T cells promote M1 macrophage activation and intensively contribute to HFD-induced NAFLD (27). To explore the mechanism that drives more severe NAFLD in Prf^−/− mice, we analyzed the composition of the immune cell infiltrate in the liver by flow cytometry. Perforin deficiency did not alter the infiltration of CD4, CD8, NK, NK1.1+ T cells, or total macrophages in the liver (Figures 3A–C). However, the cell number of CD11c+ macrophages was significantly increased (Figure 3D). We next evaluated inflammatory cytokine production by these immune cell subsets. Interestingly, we observed that IFN-γ production from CD4 T, but not CD8 T cells, NK cells, or NK1.1+ T cells, was significant increased, while IL-17 was barely detectable and largely unaffected in all subsets (Figures 3E–H). We characterized the cell number in each category and found no significant difference of CD4 T cells, CD8 T cells, and NK and NK1.1+T cells. The cell numbers of CD11c+ macrophages (M1) and IFN-γ+CD4 T cells were significantly increased in perforin KO liver, which equivalent as the percentage analysis (Figure 3I).

We also determined the levels of pro-inflammatory cytokines secreted by the liver in HFD-challenged Prf^−/− and WT mice. As expected, Prf^−/− livers produced more IL-6, TNF-α, and IFN-γ compared with livers from WT controls (Figure 4A). Immunofluorescence analysis showed that perforin deficiency robustly promoted the enrichment of M1 macrophage in the liver, which was consistent with the previous percentage and cell number analysis (Figure 4B). These findings suggest that upon HFD challenge, the IFN-γ level and M1 macrophage-mediated inflammation are enhanced in Prf^−/− mice.

To determine whether increased IFN-γ production from CD4 T cells in Prf^−/− mice was associated with the exacerbated liver phenotypes that develop after HFD challenge, we depleted CD4 T cells in Prf^−/− mice and then fed the mice with HFD. As expected, CD4 T cell depletion predisposed Prf^−/− mice to decreased liver weights, lipid accumulation, and diminished liver damage (Figures 5A–C). Notably, levels of the pro-inflammatory cytokine TNF-α, as well as macrophage accumulation, were also significantly decreased following CD4 T cell depletion in Prf^−/− mice; so was the IFN-γ level, though CD4 T cells are not the only cells producing IFN-γ (Figures 5D,E). Furthermore, the mRNA expression levels of genes involved in lipogenesis such as Fabp4, CEBPα, PPARγ, SREBP1c, Chrebpβ, and Scd1 were decreased following CD4 T cell depletion in Prf^−/− mice, whereas lipolysis-related genes such as AOX1, CPT1α, and LPL were unchanged following CD4 T cell depletion in Prf^−/− mice (Figure 5F). These findings indicate that CD4 T cells play a critical role in perforin-mediated regulation of NAFLD progression.

Since the level of IFN-γ was significantly increased in the livers of Prf^−/− mice, we next explored whether CD4 T cells contribute to exacerbated NAFLD in these mice via IFN-γ activity. Therefore, we crossed Prf^−/− mice with IFN-γ receptor-deficient mice to get double knockout mice (IFN-γR^−/− and Prf^−/−). Following HFD challenge, IFN-γR^−/− and Prf^−/− mice gained similar amounts of body weight but had significantly decreased liver weights when compared to Prf^−/− mice (Figures 6A,B). Notably, IFN-γR^−/− and Prf^−/− mice showed significantly rescued NAFLD symptoms, including diminished hepatic steatosis, cellular ballooning, and lipid accumulation (Figure 6C). We also found that IFN-γR^−/−& Prf^−/− mice had reduced serum ALT, cholesterol, and liver TG levels, as well as diminished pro-inflammatory cytokine production, while the level of IFN-γ was no significantly changed after IFN-γ receptor deficiency (Figure 6D). Moreover, the cell number of pro-inflammatory (F4/80⁺iNOS⁺) macrophage in the livers of IFN-γR^−/−Prf^−/− mice was also dramatically decreased when compared to Prf^−/− mice (Figure 6E). Taken together, these findings strongly support an important role for elevated IFN-γ in promoting NAFLD progression in the context of perforin deficiency, given that ablation of IFN-γ signaling had a protective effect on the liver in an NAFLD mouse model.

To define the functional properties of CD4 T cells from Prf^−/− mice, total spleen lymphocytes from WT and Prf^−/− mice were stimulated in vitro with anti CD3/anti-CD28 in the presence of Golgi-Stop. CD4 but not CD8 T cells from Prf^−/− mice showed increased levels of IFN-γ production upon CD3/CD28 stimulation (Figures 7A,B). To further study whether elevated IFN-γ production by CD4 T cells in Prf^−/− mice was an intrinsic property of these mice, naïve CD4 T cells were sorted from WT and Prf^−/− spleens and directly differentiated into Th1 cells. Interestingly, naïve CD4 T cells from Prf^−/− mice showed an elevated ability to differentiate into Th1 cells (Figure 7C). These findings support the conclusion that CD4 T cells undergo an intrinsic functional change in Prf^−/− mice.