C57/BL6 mice (female, 6- to 7-week-old) were purchased from Shanghai SLAC Laboratory Animal Co (Shanghai, China). All animal procedures were performed in accordance with the Guidelines for Care and Use of Laboratory Animals of Shanghai Institute of Materia Medica (SIMM), Chinese Academy of Sciences and approved by the Institutional Animal Care and Use Committee of SIMM.

Mice were randomly divided into three groups: wild-type (WT) group (n=4, Vehicle-WT), vehicle group (Vehicle-AIH, n=6) and JHU083-treated group (JUH083-AIH, n=6). Induction of AIH was based on the treatment of mice with ConA (8mg/kg body weight (BW), C2010, Sigma-Aldrich, USA). Mice in Vehicle-WT group were treated with 5% methyl cellulose (MC) dissolved in saline by oral gavage 24h and 1h before administration of saline via intravenous injection. The Vehicle-AIH group received administration of 5% MC by oral 24h and 1h before ConA injection, while mice in JUH083-AIH group were treated with JHU083 (0.3mg/kg BW, HY-122218, MedChemExpress, NJ, USA) dissolved in 5% MC via gavage 24h and 1h before ConA treatment. Blood and liver tissues were collected from mice that were sacrificed 6h after ConA treatment. Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH) and alkaline phosphatase (ALP) were analyzed using an automatic biochemical analyzer (AU480, Beckman Coulter, Pasadena, CA, USA) to assess liver injury. For survival analysis, concentration of ConA was adjusted to a lethal dose (20mg/kg) and mice were under observation until 48h after ConA injection.

The tissues were fixed in 4% paraformaldehyde and embedded in paraffin. Tissue sections were prepared and stained with hematoxylin-eosin (H&E). Terminal Deoxynucleotidyl Transferase (TdT)-mediated dUTP Nick-End Labeling (TUNEL) staining was performed in paraffin-embedded liver sections using the TUNEL Kit (C1090, Beyotime, Haimen, China). Finally, the slides were scanned using a fluorescence microscope (1013688, PE Vectra3, Perkin Elmer, Waltham, MA, USA).

Spleens were dissected from C57BL/6 mice and cell suspensions were prepared as following steps. Spleens were milled and centrifuged at 400g for 5 minutes. Red blood cells were lysed and splenocytes were washed with 10ml phosphate buffer solution (PBS). Cells were resuspended with complete medium (RPMI-1640 medium supplemented with 10% inactivated fetal bovine serum, 1% penicillin-streptomycin solution, 10mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 1mM glutamine, 50μM β-mercaptoethanol and 1mM sodium pyruvate) for further culture or for cytometric assays.

Isolated splenocytes were activated with ConA (1.5μg/ml) with or without 2.5μM DON (HY-108357, MedChemExpress, NJ, USA). After 24h, the expression of CD69 and CD25 were analyzed by flow cytometry to evaluate the activation of CD4(+) and CD8(+) T cells. For T cell differentiation analysis, splenocytes were stimulated using ConA (1.5μg/ml) with or without DON (2.5μM). GlogiPlug (555028, BD Biosciences, USA) was added 18h after for the next 24 h of incubation. The cells were first stained with monoclonal antibodies CD4 and CD8 for 20 min at room temperature, then fixed using a Fixation/Permeabilization Solution Kit (555028, BD Biosciences, USA), and finally stained with IFN-γ, IL4, IL17 and Granzyme B antibodies. Flow cytometry analysis was conducted by flowJo_V10 software (BD Biosciences, USA). For in vivo T cells activation and differentiation tests, splenocytes were separated from mice in WT-Vehicle, Vehicle-AIH and JHU083-AIH group and stained with above antibodies for flow cytometry analysis. The antibodies used in the flow cytometry were as follows: CD4 (FITC, E-AB-F1097C), CD8a (PE, E-AB-F1104D), CD25 (PE-Cy5, E-AB-F1102G), CD69 (PE-Cy7, E-AB-F1187H), IL4 (PE, E-AB-F1204UD), IL17A (APC, E-AB-F1272E) were purchased from Elabscience Biotechnology (Wuhan, China), antibodies IFNγ (APC-Cy7, cat# 561479) was purchased from BD Biosciences (UK), and Granzyme B (PE-Cy7, cat# 372213) was purchased from Biolegend (San Diego, CA, USA). SLC7A5 antagonist LAT1-IN1 (HY-108540) was purchased from MedChemExpress (NJ, USA).

The IFN-γ Mouse Uncoated ELISA Kit (EK280/3-48, Multi Science, China) and IL17 Mouse Uncoated ELISA Kit (EK217/2-48, Multi Science, China) were used to measure the levels of IFN-γ and IL17 in the serums of mice in Vehicle-WT, Vehicle-AIH and JHU083-AIH group and the supernatants of unstimulated splenocytes or splenocytes that were stimulated with ConA in the presence of vehicle or DON. Briefly, the serums and supernatants were collected and added to the well of corresponding microplates for 2h. After the samples were washed with wash solution, 100μL of mouse IFN-γ or IL17 conjugate was added to each well for 2h. The washing process was repeated, and 200μL of Substrate Solution was added and incubated for 30 min. Then, 50μL of Stop Solution was added to each well, and the optical density was determined within 30 min using a microplate reader (SpectraMax M5, Molecular Devices, Santa Clara, CA, USA) set to 450nm.

Protein samples were homogenized with SDS loading buffer and boiled to denature the proteins. Samples with different sizes were separated by 10% SDS-polyacrylamide gel electrophoresis, and transferred to NC membranes. After blocking with 5% non-fat milk solution for 1h, membranes were incubated with β-Actin, P-mTOR, mTOR, P-P70S6K P70S6K at 4°C overnight. Then, membranes were incubated with HRP-linked anti-rabbit IgG (7074, Cell Signaling Technology, Trask Lane Danvers, MA, USA) at room temperature for 1h. Finally, the membranes were stained with ECL detection buffer (1705061, Bio-Rad, Hercules, CA, USA) and visualized using ChemiDoc system (Bio-Rad, Hercules, CA, USA). Primary antibodies including β-Actin (12620), P-mTOR (2971S), mTOR (2972S), P-P70S6K (97596), P70S6K (34475) were purchased from Cell Signaling Technology (Trask Lane Danvers, MA, USA).

Total RNA was extracted using TRIzol reagent (9109, TaKaRa, Dalian, China) followed by cDNA synthesis using RT Master Mix (RR036B, TaKaRa, Dalian, China). Subsequently, cDNA was used to measure the mRNA levels of IFN-γ, IL4, IL17, SLC7A5 and SLC1A5 using a Real-Time PCR System (Roche, Basel, Switzerland). β-Actin was used as the normalization control. All reactions were performed in triplicates. The relative quantifications were measured by the comparative CT method. The primer sequences used were as follows: IFN-γ: forward: CAGCAACAGCAAGGCGAAA, reverse: CTGGACCTGTGGGTTGTTGAC; IL-4: forward: CGCCATGCACGGAGATG, reverse: CGAGCTCACTCTCTGTGGTGTT; IL17: forward: ATCTGTGTCTCTGATGCTGTTGCTG, reverse: TGGAACGGTTGAGGTAGTCTGAGG; SLC7A5: forward: CTGGATCGAGCTGCTCATC, reverse: GTTCACAGCTGTGAGGAGC; SLC1A5: forward: ATCGCACAACTAAACGGGGT, reverse: ACTGCTTCCAGGATGATGGC; β-Actin: forward: AACAGTCCGCCTAGAAGCAC, reverse: CGTTGACATCCGTAAAGACC.

Experimental data were expressed as the mean ± standard error of mean (SEM), and comparisons between two groups were performed using one-way ANOVA analysis. Data were analyzed using GraphPad Prism 8.0.2. Statistical significance was determined at P < 0.05 (*); P < 0.01 (**); and P < 0.001 (***).

In order to verify the hepatoprotective effect of glutaminase antagonist JHU083 in a ConA-induced AIH model, the survival times of vehicle and JHU083 (0.3mg/kg BW) (28) treated mice in this mice model were investigated. Vehicle-AIH and JHU083-AIH mice were intravenously injected with a high dose of ConA (20mg/kg BW). Vehicle-AIH mice began to die 4h after injection and all mice died within 38h, while survival rate of JHU083-AIH mice was 70% at 24h after ConA injection and did not change after extension to 48h. On the basis of the survival curve, it was proved that JHU083 treatment greatly reduced the death rate of ConA-induced AIH mice (Figure 1A). Although AIH mice were ameliorated after treatment with JHU083 under the induction of 20mg/kg ConA (Supplementary Figure 1), the pathological damages under such ConA concentration would be too severe. Therefore, we chose a non-lethal dosage instead to compare the alteration of transaminase, conditions of cells and pathological damages. Then, ConA was adjusted to a lower dose (8mg/kg), and blood and liver specimens were collected 6h post ConA injection. In JHU083-AIH group, liver congestion and necrosis were obviously mitigated in a macroscopic level (Figure 1B) and serum ALT, AST, LDH and ALP (Figure 1C) were significantly lower than mice in Vehicle-AIH group. Liver specimens were stained with H&E. At 6h after ConA administration, marked liver hyperemia and necrosis of liver cells were observed in vehicle mice rather than in JHU083 treated mice (Figure 1D). No pathological injury was found in Vehicle-WT mice. Apart from these, the TUNEL assay indicated that TUNEL-positive cells decreased significantly after JHU083 treatment (Figure 1E), indicating decreased apoptosis of liver cells. These results proved that blocking glutamine metabolism ameliorated the progression of ConA-induced AIH in mice.

According to Ruoning Wang (29), we transplanted CFSE-labeled splenocytes into mice before ConA injection. After 6h ConA stimulation, there was no proliferation in the transplanted splenocytes (Supplementary Figure 2). Thus, we focused on the activation and differentiation of T cells. Given that activated effector T cells need glutamine metabolism to provide energy (11), we first measured the activation of T cells upon glutamine metabolism blocking in vivo. CD25 and CD69 are two markers expressed on the surface of activated T cells (30). To investigate the effect of JHU083 on the activation of T cells in AIH mice, we evaluated the activation of CD4(+) and CD8(+) T cells in spleens extracted from mice in WT-Vehicle, Vehicle-AIH and JHU083-AIH group via the proportions of CD25 or CD69 positive cells. Consistent with our findings on pathology, JHU083 significantly reduced percentages of activated CD4(+) and CD8(+) T cells (Figures 2A, B).

To further explore the effect of JHU083 on the differentiation of CD4(+) and CD8(+) T cells in AIH mice, the differentiation marker of CD4(+) (IFNγ for Th1 cells, IL4 for Th2 cells and IL17A for Th17 cells) and CD8(+) T cells (IFNγ and Granzyme B for CTLs) were evaluated using flow cytometry. We observed that the frequency of IFNγ- and IL17A-expressing CD4(+) T cells instead of IL4-expressing CD4(+) T cells were significantly lower in the spleens of JHU083-AIH mice compared to that of Vehicle-AIH mice (Figure 3A). Lower IFNγ and IL17 levels were presented in the serums of JHU083-AIH mice than in Vehicle-AIH mice (Figure 3B). Besides, the frequency of CD8(+) T cells expressing IFNγ (Figure 3C) or Granzyme B (Figure 3D) was significantly reduced after JHU083 treatment. The above outcomes indicated that blocking glutamine metabolism had negative effects on activation and differentiation of T cells in vivo.

Consistent with the results in vivo, ConA did not induce proliferation within 24h in vitro (Supplementary Figure 3). To evaluate the effects of blocking glutamine metabolism on T cells in vitro, we used DON to substitute JHU083 to investigate whether DON might suppress the activation of T cells. Splenocytes of mice were separated and cultured in RPIM 1640 medium and activated by 1.5μg/ml ConA in the presence or absence of DON (2.5μM) (Supplementary Figure 4). 24h after activation, splenocytes were collected and stained with CD4, CD8 and activation markers CD25 and CD69, and the proportions of CD25(+) or CD69(+) cells and median fluorescence intensity (MFI) of CD25(+) or CD69(+) cells were analyzed. Although CD25 can be also expressed on regulatory T cells, we assessed the proportion of CD4+Foxp3+ Treg cells and found that DON did not affect the proportion of Treg cells (Supplementary Figure 5). As shown in Figures 4A, 4B, splenocytes treated with DON had lower expressions of CD25 and CD69 in CD4(+) and CD8(+) T cells.

To further characterize the effect of DON on the differentiation of CD4(+) and CD8(+) T cells, the above-mentioned differentiation markers of CD4(+) and CD8(+) T cells were tested. DON significantly inhibited the differentiation of Th1 and Th17 cells rather than Th2 cells (Figure 5A). Consistently, DON reduced mRNA expression (Figure 5B) and production (tested by ELISA, Figure 5C) of IFNγ and IL17. In addition, we proved that DON suppressed CD8(+) T cells to turn into CTL, which was evidenced by lower productions of IFNγ- (Figure 5D) and Granzyme B-expressing T cells (Figure 5E). Thus, GLS blocking played a critical role in inhibiting T cells activation and differentiation in vitro.

Mammalian target of rapamycin (mTOR) signaling is a master regulator of cell metabolism and promotes anabolic processes such as protein synthesis (31). The activation of mTOR signaling is dependent on growth factors and nutrients in cancer cells (32). For instance, leucine can promote mTOR signaling activation via binding to sestrin2 and arginine binds to CASTOR1 (33). Glutamine activates mTOR signaling in a RAG-dependent manner via glutaminolysis (34). In addition, many studies have noted that the activation and differentiation of T cell are strongly correlated with mTOR signaling (24–27). We hypothesized that blocking glutamine metabolism may affect mTOR signaling in T cells in an amino acid dependent way.

Thus, we examined the phosphorylated protein levels of mTOR signaling using proteins extracted from splenocytes that were activated by 1.5μg/ml ConA with or without treating 2.5μM DON for 24h. After activating by ConA, the protein levels of phosphorylated mTOR and 70 kDa ribosomal protein S6 kinase (P70S6K) increased sharply. But in the presence of DON, the levels of phosphorylated mTOR and P70S6K instead of their unphosphorylated proteins significantly decreased (Figures 6A, B).

The levels of amino acids are closely correlated with corresponding amino acid transporters. To explore how DON makes an impact on mTOR signaling, we focused on glutamine and leucine transporters because it was reported that the function of leucine and glutamine exchanger solute carrier family 7 member 5 (SLC7A5) and glutamine transporter solute carrier family 7 member 5 (SLC1A5) affected both intracellular amino levels and activation of mTOR signaling (35–39). We discovered that DON significantly reduced the mRNA expression of SLC7A5 but had no effect on SLC1A5 (Figure 7A). These results suggested that SLC7A5 may play a critical part in the process of GLS inhibition to the suppression of activation of mTOR signaling. To further confirm the function of SLC7A5, we inhibited it with LAT1-IN1 (30mM) (40), the competitive antagonist of SLC7A5. After LAT1-IN1 treatment, the activation of CD4(+) and CD8(+) T cells, evidenced by the expressions of CD25 and CD69, was significantly weakened (Figures 7B, C). In addition, the expression levels of phosphorylated mTOR and P70S6K were inhibited by LAT1-IN1 (Figures 7D, E). Therefore, our findings demonstrated that GLS blocking not only suppresses T cell activation and differentiation, but also decreases mRNA expression of glutamine and leucine exchanger SLC7A5, reducing intracellular leucine level and further inhibiting activation of mTOR signaling.