SLC38A2 and glutamine signalling in cDC1s dictate anti-tumour immunity – Nature

Mice

The research conducted in this study complies with all of the relevant ethical regulations. The animal protocols were approved by and performed in accordance with the Institutional Animal Care and Use Committee (IACUC) of St. Jude Children’s Research Hospital. Mice were housed and bred in specific pathogen-free conditions in the Animal Resource Center at St. Jude Children’s Research Hospital. Mice were on 12-hour light–dark cycles that coincide with daylight in Memphis, TN, USA. The St. Jude Children’s Research Hospital Animal Resource Center housing facility was maintained at 20–25 °C and 30–70% humidity. C57BL/6, CD45.1⁺, OT-I, OT-II, Rag1^−/−, Rosa–Cas9 knock-in⁵¹, Batf3^−/− (ref. ²⁹), Cd4^cre (ref. ⁵²), Cd11c^cre (ref. ⁵³) and Xcr1^cre (ref. ³⁸) mice were purchased from The Jackson Laboratory or as described previously⁵⁴. Slc38a2^fl/fl mice were purchased from INFRAFRONTIER/EMMA. Flcn^fl/fl mice were kindly provided by L. Schmidt⁵⁵. Tfeb^fl/fl mice were kindly provided by A. Ballabio⁵⁶. The mice were backcrossed to the C57BL/6 background; sex- and age-matched mice were used throughout the study at 7–12 weeks old, and both male and female mice were used. The genetically modified mice were viable and developed normally. To generate mixed bone marrow chimeras, bone marrow cells from wild-type, Slc38a2^ΔDC or Flcn^ΔDC mice were mixed with cells from Batf3^−/− mice at a 1:1 ratio and transferred into lethally irradiated (11 Gy) CD45.1⁺ mice, followed by reconstitution for 6–8 weeks¹⁶. In certain experiments, bone marrow cells from wild-type or Flcn^ΔDC mice were transferred into lethally irradiated (11 Gy) CD45.1⁺ mice. For chimeras used in Extended Data Figs. 5i–k and 7e–g, bone marrow cells from wild-type, Slc38a2^ΔDC or Flcn^ΔDC mice were mixed with cells from CD45.1⁺ mice at a 1:1 ratio and transferred into lethally irradiated (11 Gy) C57BL/6 mice and analysed 8 weeks later. Age- and sex-matched mice with pre-determined genotypes (not blinded to investigators) were randomly assigned to control and experimental groups.

Cell purification and culture

Mouse spleens were digested with 1 mg ml⁻¹ collagenase IV (LS004188, Worthington) plus 200 U ml⁻¹ DNase I (DN25, Sigma) for 45 min at 37 °C, and CD11c⁺ DCs were enriched using CD11c MicroBeads (130-125-835, Miltenyi Biotec) according to the manufacturer’s instructions. Enriched cells were stained and sorted for cDC1s (CD11c⁺CD8α⁺CD24⁺TCRβ⁻CD49b⁻B220⁻) and cDC2s (CD11c⁺MHCII⁺CD8α⁻CD11b⁺TCRβ⁻CD49b⁻B220⁻F4/80⁻Ly6C⁻) on a MoFlow (Beckman Coulter) or Reflection (i-Cyt) cell sorter. Lymphocytes from spleen and peripheral lymph nodes were sorted for naive OT-I cells (CD8⁺CD62L^highCD44^lowCD25⁻) and naive OT-II cells (CD4⁺CD62L^highCD44^lowCD25⁻) from OT-I and OT-II mice, respectively. Sorted DCs were cultured with specific medium as indicated in figure legends. Medium with or without individual amino acids was generated with RPMI 1640 powder (R8999-04A, US Biological) by supplementation of individual amino acids. The medium was supplemented with 10% (v/v) dialysed fetal bovine serum (FBS; A3382001, Thermo Fisher Scientific). For preparation of tumour cell line-derived culture supernatant, MC38 or B16F10 cells were cultured in glutamine-free RPMI 1640 medium (15-040-CV, Corning) supplemented with 10% (v/v) dialysed FBS plus 1% (v/v) penicillin-streptomycin (15140122, Thermo Fisher Scientific), and different concentrations of glutamine (0.3, 0.6 or 2 mM; 25030081, Thermo Fisher Scientific) as indicated in the figures. Tumour cell culture supernatant was collected 48 h later.

In vitro BMDC culture

Bone marrow cells were flushed from mouse tibias and femurs, and red blood cells were lysed using ACK lysis buffer. Cells were then plated in RPMI 1640 medium supplemented with 10% (v/v) FBS, 1% (v/v) penicillin-streptomycin and 55 μM β-mercaptoethanol (RPMI 1640 complete medium). FLT3L BMDCs were cultured as previously described¹⁶. In brief, bone marrow cells were cultured in RPMI 1640 complete medium with 200 ng ml⁻¹ FLT3L-Ig (BE0098, Bio X Cell) for 7–9 days. FLT3L BMDCs were sorted as cDC1s (B220⁻CD11c⁺CD24⁺CD172α⁻) and cDC2s (B220⁻CD11c⁺CD24⁻CD172α⁺) for further experiments. iCD103⁺ BMDCs were generated as previously described for Transwell assays⁵⁷. In brief, bone marrow cells were plated in RPMI 1640 complete medium supplemented with 200 ng ml⁻¹ FLT3L-Ig and 2 ng ml⁻¹ mGM-CSF (315-03, Peprotech). Half of the fresh medium was supplemented to the cultures at day 5, and non-adherent cells were collected and replated in fresh medium at day 9. Loosely adherent cells were collected at days 15–17 for Transwell assays.

Flow cytometry

For analysis of surface markers, cells were first incubated with Fc block (2.4G2, Bio X Cell) for 10 min in phosphate-buffered saline (PBS) containing 2% (w/v) FBS, and then stained with the appropriate antibodies on ice for appropriate 30 min. For intracellular cytokine detection, cells were stimulated for 4 h with phorbol 12-myristate 13-acetate (PMA) plus ionomycin or OVA_257–264 in the presence of monensin before staining with a fixation/permeabilization kit (554774, BD Biosciences) according to the manufacturer’s instructions. For intracellular IL-12p40 detection, enriched splenic DCs were stimulated for 4 h with LPS or poly I:C in the presence of GolgiStop before staining using a fixation/permeabilization kit (554774, BD Biosciences) according to the manufacturer’s instructions. Transcription factor staining was performed with FOXP3/transcription factor staining buffer set (00-5523-00, eBioscience) according to the manufacturer’s instructions. Lysotracker staining was performed with LysoTracker Red DND-99 dye (L7528, Invitrogen) according to the manufacturer’s instructions. pHrodo Green Dextran (P35368, Invitrogen) staining was performed according to the manufacturer’s instructions. Flow cytometry data were acquired on LSRII, LSR Fortessa or Symphony A3 (BD Biosciences) using BD FACSDiva software (v8) and analysed using FlowJo software (Tree Star; v10). 7-Aminoactinomycin D (7AAD; A9400, 1:200, Sigma) or fixable viability dye (65-0865-14, 1:1,000, eBioscience) was used for dead-cell exclusion. The following fluorescent conjugate-labelled antibodies were used: PE-Cy7–anti-CD11c (N418, 60-0114, 1:200, Tonbo Biosciences); FITC–anti-FOXP3 (FJK-16s, 11-5773-82, 1:200), PE-Cyanine7–anti-T-bet (4B10, 25-5825-82, 1:100), APC-eFluor 780–anti-MHCII (M5/114.15.2, 47-5321-82, 1:400), PE-Cyanine7–anti-CD24 (M1/69, 25-0242-82, 1:400), FITC–anti-CD86 (GL1, 11-0862-82, 1:200), PE–anti-IL-12/IL-23 p40 (C17.8, 12-7123-82, 1:200), PE–anti-LAMP1 (eBioH4A3, 12-1079-42, 1:400) (all from eBioscience); Brilliant Violet 510–anti-CD4 (RM4-5, 100559, 1:200), AF700–anti-CD8α (53-6.7, 100730, 1:200), Brilliant Violet 785–anti-TCRβ (H57-597, 109249, 1:200), PE–anti-CD45.2 (104, 109808, 1:400), PE/Dazzle 594–anti-PD-1 (29F.1A12, 135228, 1:400), Alexa Fluor 647–anti-granzyme B (GB11, 515405, 1:100), PE-Cyanine7–anti-IFNγ (XMG1.2, 505826, 1:200), Brilliant Violet 421–anti-TNF (MP6-XT22, 506328, 1:200), APC–anti-IL-4 (11B11, 504106, 1:200), Pacific Blue–anti-IL-17A (TC11-18H10.1, 506918, 1:200), Brilliant Violet 711–anti-TIM-3 (RMT3-23, 119727, 1:400), Brilliant Violet 650–anti-CD44 (1M7, 103049, 1:400), PE-Cyanine7–anti-CD62L (MEL-14, 104417, 1:400), APC–anti-CD69 (H1.2F3, 104514, 1:200), Brilliant Violet 650–anti-CD11b (M1/70, 101259, 1:200), APC–anti-XCR1 (ZET, 148206, 1:400), APC–anti-CD103 (2E7, 121414, 1:400), Pacific Blue–anti-Ki67 (16A8, 652422, 1:400) (all from BioLegend); PE–anti-IL-2 (JES6-5H4, 554428, 1:200), Brilliant Violet 605–anti-Ly108 (13G3, 745250, 1:200) (from BD Biosciences); Alexa Fluor 647–anti-TCF1 (C63D9, 6709, 1:100, Cell Signaling Technology). Migratory cDC1s and cDC2s in tumour dLN were respectively gated as F4/80⁻Ly6C⁻CD11c^loMHCII^hiCD103⁺CD11b^– and F4/80⁻Ly6C⁻CD11c^loMHCII^hiCD103^–CD11b⁺, as described⁵⁸. Intratumoral cDC1s and cDC2s were respectively gated as CD45⁺F4/80⁻Ly6C⁻CD11c⁺MHCII⁺CD103⁺CD11b^– and CD45⁺F4/80⁻Ly6C⁻CD11c⁺MHCII⁺CD103^–CD11b⁺, as described¹⁹.

Antigen presentation assays

For in vitro assays, cDC1s and cCD2s were sorted from spleen, pulsed with 200 μg ml⁻¹ OVA protein (Low Endo, LS003059, Worthington), 250 pg ml⁻¹ OVA_257–264 (vac-sin, InvivoGen) or 3 μg ml⁻¹ OVA_323–339 (vac-isq, InvivoGen) for 2 h, then washed twice and cultured with naive CD44^lowCD62L^high OT-I or OT-II cells for three days. For HKLM-OVA antigen cross-presentation assay, sorted cells were cocultured with 1 × 10⁷ HKLM-OVA and OT-I cells for 3 days as previously described⁵⁹. [³H]Thymidine (PerkinElmer) was added to the culture 8 h before cells were collected to measure proliferation. Where indicated, cDC1s or cDC2s were incubated with OVA in RPMI 1640 medium lacking an individual amino acid or amino acid-free medium supplemented with an individual amino acid for 2 h, irradiated and then cocultured with T cells. For treatment with tumour cell culture supernatant, cDC1s or cDC2s were pulsed with OVA in the presence of MC38 or B16F10 culture supernatant or tumour cell culture supernatant supplemented with an individual amino acid for 2 h, followed by irradiation and coculture with OT-I or OT-II cells. For in vivo priming assay, 1 × 10⁶ CFSE-labelled naive CD45.1⁺ OT-I cells were transferred into mice intravenously, followed by intravenous injection with 20 μg OVA 24 h later. Three days after OVA immunization, spleens were collected, and the proliferation of OT-I cells was examined by CFSE dilution with flow cytometry.

ELISA

Culture supernatant from in vitro antigen presentation assays was collected, and the levels of IL-2 and IFNγ were determined using IL-2 (88-7024-22, Thermo Fisher Scientific) and IFNγ (88-7314-22, Thermo Fisher Scientific) ELISA kits according to manufacturer’s instructions.

DQ-ovalbumin degradation assay

Splenic cDC1s were sorted from wild-type and Flcn^ΔDC mice as described above. Sorted cells were incubated with 50 μg ml⁻¹ DQ-Ovalbumin (DQ-OVA; D-12053, Thermo Fisher Scientific) for 0, 30, 60 or 120 min. DQ-OVA is a self-quenched OVA conjugate that emits green fluorescence upon hydrolysis by proteases. Cells were washed with PBS at the indicated times and analysed for DQ-OVA release as assessed positive FITC (FITC⁺) staining as described previously⁵⁹.

Tumour model and treatments

B16F10 cell line was purchased from ATCC. MC38, MC38-OVA and B16-OVA cell lines were provided by D. Vignali. B16-FLT3L cell line was provided by D. Green. B16F10 cell line expressing ZsGreen (B16-ZsGreen) was generated by lentiviral transduction of pHIV-ZsGreen construct (18121, Addgene) into B16F10 tumour cells, which were sorted based on expression of ZsGreen. These cell lines are not on the list of commonly misidentified cell lines (International Cell Line Authentication Committee). Cell lines used in this study were not independently authenticated or tested for mycoplasma contamination. All cell lines were maintained at 37 °C with 5% CO₂ in DMEM supplemented with 10% (v/v) FBS and 1% (v/v) penicillin-streptomycin. Mice were injected subcutaneously with 5 × 10⁵ MC38, B16-OVA or B16-ZsGreen cells in the right flank. After tumour inoculation, mice were randomized and assigned to different groups for treatments. Glutamine was injected into tumours at a dose of 200 mg kg⁻¹ per mouse daily starting from day 5 after tumour inoculation and for 10 consecutive days thereafter. Anti-PD-1 antibody (J43, Bio X Cell) or rat IgG2b isotype control (LTF-2, Bio X Cell) was injected intraperitoneally three times at a dose of 200 μg in 100 μl PBS at days 7, 10 and 13 after inoculation of MC38 cells. Anti-PD-L1 antibody (10F.9G2, Bio X Cell) or rat IgG2b isotype control (LTF-2, Bio X Cell) was injected intraperitoneally three times at a dose of 200 μg in 100 μl PBS at days 9, 12 and 15 after inoculation of B16-OVA cells. Mice with complete tumour rejections from intratumoral glutamine injection and anti-PD-1 combination therapy were rechallenged with 1 × 10⁶ MC38 cells after 60 days. Tumours were measured every two days with digital callipers and tumour volumes were calculated by the formula: length × width × (length × width)^0.5 × π/6. Tumour size limits were approved to reach a maximum of 3,000 mm³ or ≤20% of body weight (whichever was lower) by the IACUC at St. Jude Children’s Research Hospital.

To analyse tumour antigen-specific immune responses, 1 × 10⁶ MC38-OVA cells were injected subcutaneously into mice. Tumour antigen-specific CD8⁺ T cells were analysed by H-2K^b-OVA tetrameter staining for 30 min at room temperature. To prepare intratumoral lymphocytes, tumours were collected at day 15 after inoculation, excised, minced and digested with 1 mg ml⁻¹ collagenase IV (Worthington) and 200 U ml⁻¹ DNase I (Sigma) for 1 h at 37 °C. To analyse DC migration, tumour dLNs (including inguinal and axillary LN) were collected and digested with 1 mg ml⁻¹ collagenase IV (Worthington) and 200 U ml⁻¹ DNase I (Sigma) for 30 min at 37 °C.

Generation of CRISPR–Cas9 knockout tumour cell lines

MC38 or B16-OVA cells were transduced with lentivirus of pLenti-Cas9-GFP (86145, Addgene). Cas9-expressing (GFP⁺) cells were sorted, and expression of Cas9 protein was confirmed by immunoblot analysis (data not shown). Cas9-expressing MC38 or B16-OVA cells were then transduced with lentivirus expressing Ametrine and control sgRNA (sgNTC: ATGACACTTACGGTACTCGT) or sgRNA targeting Slc38a2 (sgSlc38a2: ATTAAATACTGACATTCCAA) as previously described⁶⁰. After sorting of Ametrine⁺ cells, cells were expanded, and deletion of SLC38A2 was verified by immunoblot analysis. For tumour growth, 1 × 10⁶ sgNTC- or sgSlc38a2-transduced, Cas9-expressing MC38 or B16-OVA cells were injected subcutaneously into mice.

DC transfer and adoptive T cell transfer for tumour therapy

For DC transfer experiments, freshly isolated splenic cDC1s were used following an established strategy³⁰. In brief, B16-FLT3L cells (2.5 × 10⁶) were injected subcutaneously into both flanks of wild-type mice to expand cDC1s. Spleens were collected 10 days after tumour inoculation, and cDC1s were enriched using the CD8⁺ DC isolation kit (130-091-169, Miltenyi Biotec). Purified cDC1s were pulsed with 100 μg ml⁻¹ OVA (low Endo, Worthington) together with 20 μg ml⁻¹ poly I:C (InvivoGen) for 2 h in RPMI 1640 medium containing 10% dialysed FBS with or without glutamine. cDC1s were washed and transferred (1 × 10⁶ cells per mouse) subcutaneously adjacent to the tumours at day 5 after B16-OVA inoculation.

For OT-I cell transfer experiments, naive OT-I cells were isolated using a naive CD8α⁺ T cell isolation kit (130-096-543; Miltenyi Biotec) according to the manufacturer’s instructions. Purified naive OT-I cells were activated using 10 μg ml⁻¹ anti-CD3 (2C11; Bio X Cell, BE0001-1) and 5 μg ml⁻¹ anti-CD28 (37.51; Bio X Cell, BE0015-1) antibodies. Activated OT-I cells were then expanded in Click’s medium (Irvine Scientific) containing 10% dialysed FBS supplemented with or without glutamine in the presence of human recombinant IL-2 (20 IU ml⁻¹; PeproTech), mouse IL-7 (12.5 ng ml⁻¹; PeproTech) and IL-15 (25 ng ml⁻¹; PeproTech) for 2–3 days before adoptive transfer. CFSE-labelled naive OT-I cells were transferred into PBS- or glutamine-supplemented MC38-OVA-bearing wild-type mice or MC38-OVA-bearing wild-type and Slc38a2^ΔDC or wild-type and Flcn^ΔDC mice on day 7 after tumour inoculation, followed by analysis of their proliferation (based on CFSE dilution) in the tumour dLNs on day 2 after adoptive transfer. Alternatively, activated OT-I cells were transferred into PBS- or glutamine-supplemented B16-OVA-bearing wild-type mice or B16-OVA-bearing wild-type and Slc38a2^ΔDC or wild-type and Flcn^ΔDC mice on day 12 after tumour inoculation, followed by their analysis in the tumour on day 7 after adoptive transfer as indicated in the figures and their legends. Where indicated, naive Cas9-expressing OT-I cells from Cas9 mice were activated and transduced with sgRNA targeting Slc38a2 or control sgRNA as described above. Ametrine⁺ transduced cells were sorted before adoptive transfer into recipients.

In vivo killing assay

In vivo killing assay was performed as previously described²⁴. In brief, splenocytes were pulsed with OVA_257–264 or PBS at 37 °C for 1 h. These antigen- or PBS-pulsed splenocytes were then labelled with CFSE or CellTrace Violet, respectively, at 37 °C for 15 min, then mixed at 1:1 ratio and a total of 1 × 10⁷ cells were transferred into MC38-OVA-bearing wild-type mice at day 12 that were treated with PBS or glutamine intratumorally daily starting at day 5 after tumour inoculation, followed by analysis of in vivo cytotoxicity against these splenocytes at 24 h after injection.

Immunoprecipitation and immunoblot analysis

For FLCN immunoprecipitation in Fig. 3a, HEK293T cells were starved with glutamine-free medium for 0.5, 1, 2, 3 h or not starved; and in Fig. 3b, HEK293T cells were starved with glutamine-free medium for 3 h, followed by the addition of 2 mM glutamine for 10 or 15 min. The cells were then lysed in CHAPS buffer (0.3% CHAPS, 10 mM β-glycerol phosphate, 10 mM pyrophosphate, 40 mM HEPES pH 7.4, 2.5 mM MgCl₂) supplemented with protease inhibitor cocktail (04693124001, Roche) for 30 min. The cell lysates were cleared by centrifugation and mixed with anti-HA magnetic beads (88837, Thermo Fisher Scientific) at 4 °C for 4 h. For immunoprecipitation of GATOR1 or GATOR2 complex, the cleared cell lysates were incubated with anti-DEPDC5 (for GATOR1) and anti-WDR24 (for GATOR2) antibodies and control IgG (3000-0-AP, ProteinTech) at 4 °C overnight, followed by a further incubation with protein A/G agarose beads (sc-2003, Santa Cruz) for 2 h. Immunoprecipitated complexes were washed three times with CHAPS buffer and subjected to immunoblot analysis. For immunoblot analysis, tumour cell lines (0.5 × 10⁶) or splenic cDC1s (0.1–0.15 × 10⁶) were collected and lysed in RIPA buffer (9806, Cell Signaling Technology), resolved in 4–12% Criterion XT Bis-Tris Protein Gel (Bio-Rad) and transferred to PVDF membrane (1620177, Bio-Rad). Membranes were blocked using 5% BSA for 1 h and then incubated with primary antibodies overnight (see below). After washing three times with TBST, the membranes were incubated with 1:5,000-diluted HRP-conjugated anti-mouse IgG (W4021, Promega) for 1 h. Following another three washes, the membranes were imaged by ODYSSEY Fc Imager (LI-COR). For immunoblot analysis of SLC38A2, cell lysates were treated with PNGase F (P0704S, New England Biolabs) to remove N-linked oligosaccharides according to the manufacturer’s instructions. The following antibodies were used: anti-β-Actin (3700), anti-GAPDH (D16H11), anti-Lamin B1 (D4Q4Z), anti-HA (3724), anti-MIOS (13557), anti-WDR59 (53385) (all were used at 1:1,000 dilution and from Cell Signaling Technology); anti-Cathepsin D (AF1029, R&D); anti-FLCN (ab124885), anti-DEPDC5 (ab213181), anti-SEH1L (ab218531) (all were used at 1:1,000 and from Abcam); anti-SEC13 (sc-514308); anti-NPRL2 (sc-376986) (both were used at 1:1,000 and from Santa Cruz); anti-Flag (F1804, 1:1,000, Sigma); anti-TFEB (A303-673A, 1:1,000, Bethyl Laboratories); anti-SLC38A2 (BMP081, 1:1,000, MBL); anti-WDR24 (20778-1-AP, 1:1,000, ProteinTech); and anti-NPRL3 (NBP-97766, 1:1,000, Novus Biologicals).

Cytosolic and nuclear cell fractionation

For TFEB cytosolic and nuclear cell fractionation analysis in Fig. 4k and Extended Data Fig. 9d, freshly isolated splenic cDC1s (3 × 10⁶) from B16-FLT3L tumour-bearing mice were incubated in glutamine-sufficient medium or starved with glutamine-free medium for 3 h. The cells were washed twice with ice-cold PBS and collected into cytosol extraction buffer (150 mM NaCl; 50 mM HEPES, pH 7.4; and 0.025% (w/v) digitonin) supplemented with protease and phosphatase inhibitor cocktail (Roche). The samples were incubated on ice for 10 min, followed by centrifugation at 980 g for 5 min at 4 °C to pellet the nuclei, and the supernatant (cytoplasmic fraction) was further cleared by centrifugation at 13,000 g for 5 min. The nuclear pellet for each sample was washed three times with cytosol extraction buffer and lysed in RIPA buffer supplemented with protease and phosphatase inhibitor cocktail for 40 min on ice. After centrifugation at 14,000 g for 20 min, the resulting supernatant was used as the nuclear fraction.

Immunofluorescence

Sort-purified cDC1s were allowed to adhere to poly-l-lysine-coated coverslips prior to fixation with 4% paraformaldehyde (PFA) for 10 min. Cells were then permeabilized with PBS containing 0.1% Triton-100 for 3 min prior to blocking with PBS containing 2% bovine serum albumin, 5% normal goat serum and 0.05% Tween-20. Cells were incubated overnight at 4 °C in blocking buffer containing anti-EEA1 antibody (3288, C45B10, 1:250; Cell Signaling Technology) followed by Alexa Fluor 488-conjugated anti-rabbit secondary antibody (A11008, 1 μg ml⁻¹; Thermo Fisher Scientific) and Alexa Fluor 568-conjugated phalloidin to detect F-Actin (A12380, 1 U ml⁻¹; Thermo Fisher Scientific). Coverslips were mounted in Vectashield Vibrance (Vector Labs), and images were acquired using a Marianas spinning disk confocal microscope (Intelligent Imaging Innovations) equipped with SoRa (Yokagawa), Prime 95B CMOS camera (Photometrics) and a 1.45 NA 100× oil objective. Images were acquired using Slidebook software (version 6.0.24; 3i) and analysed using Imaris software (version x64 9.5.1; Bitplane).

Metabolomics and mass spectrometry for detection of glucose and amino acids

Plasma and TIF were collected as previously described⁷. In brief, subcutaneous tumour tissues were cut into pieces and then centrifuged through a 0.22-μm nylon filter (CLS8169, Corning). The flow-through was collected as TIF. The matched blood was collected from the orbital venous plexus, and plasma supernatant was collected by centrifugation. TIF and plasma were flash-frozen with liquid nitrogen and stored at –80 °C before analysis. Tumour cell culture supernatant was collected from medium cultured with MC38 or B16F10 cells in RPMI 1640 medium supplemented with 0.6 mM glutamine. 1 × 10⁶ sorted splenic cDC1s or cDC2s from wild-type or Slc38a2^ΔDC mice were collected. sgNTC- or sgSlc38a2-transduced, Cas9-expressing MC38 cells were cultured in DMEM supplemented with 10% (v/v) FBS and 1% (v/v) penicillin-streptomycin. Then, the cells were collected and washed once with ice-cold PBS, and the metabolites were extracted using 750 μl of methanol:acetonitrile:water (5:3:2, v/v/v) and the supernatant was dried by lyophilization. Aliquots of 20−50 μl from plasma and TIF were extracted with at least 15-fold excess volume of the methanol:acetonitrile:water solution, and the supernatant was then collected and dried by lyophilization. Dried extracts containing the hydrophilic metabolites were dissolved in 40 μl of water:acetonitrile (8:2, v/v) and 10 μl were used in the procedure to derivatize amino acids as described previously⁶¹ with some modifications. In brief, the samples were placed into glass autosampler vials and then 35 μl of sodium borate buffer (100 mM, pH 9.0) was added and mixed by pipetting. Next, 10 μl of the 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC, 10 mM in acetonitrile)-derivatizing reagent (Cayman Chemical) was added. The vial was sealed, mixed by vortexing, and then incubated at 55 °C for 15 min. The vial was cooled to room temperature and then 1 μl was analysed by liquid chromatography with tandem mass spectrometry (LC–MS/MS). An ACQUITY Premier UPLC System (Waters Corp) was used for the LC separations, using non-linear gradient conditions as follows: 0−0.4 min 3% B; 0.4−8 min 3 to 96% B (using curve no. 8 of the inlet condition in MassLynx); 8−12 min 96% B; 12−12.5 min 96 to 3% B; 12.5−14 min 3% B. Mobile phase A was water supplemented with 0.15% acetic acid, and mobile phase B was acetonitrile with 0.15% acetic acid. The column used was an Accucore C30 (50 × 2.1 mm, 2.6 μm) (Thermo Fisher Scientific), operated at 50 °C. The flow rate was 300 μl min⁻¹ and the injection volume used was 1 μl. All LC–MS/MS solvents and reagents were the highest purity available (water, acetonitrile, acetic acid, boric acid, sodium hydroxide) and were purchased from Thermo Fisher Scientific. A Xevo TQ-XS Triple Quadrupole Mass Spectrometry (TQ-XS) (Waters Corp) equipped with a multi-mode ESI/APCI/ESCi ion source was employed as detector. The TQ-XS was operated in the positive ion mode using the multiple reaction monitoring mass spectrometry method (MRM). The MRM conditions were set to a minimum of 15 points per peak, with automatic dwell time. The operating conditions of the source were: Capillary Voltage 3.8 kV, Cone Voltage 40 V, Desolvation Temp 550 °C, Desolvation Gas Flow 1,000 l h⁻¹, Cone Gas Flow 150 l h⁻¹, Nebuliser 7.0 Bar, Source Temp 150 °C. Authentic amino acids standards were purchased from Sigma-Aldrich and employed to establish the MRM conditions and calibration curves. The acquired MRM data were processed using the software application Skyline (version 21.2; MacCoss Lab Software).

Quantification of amino acid uptake

Sorted splenic cDC1s and cDC2s or sgNTC- and sgSlc38a2-transduced Cas9-expressing MC38 cells were washed once with PBS and were then plated into 6-well plates at 1 × 10⁶ cells per well in RPMI 1640 medium containing 10% dialysed FBS and 2 mM [¹³C₅]glutamine for 10 min. In certain experiments, cells were incubated with medium containing [¹³C]glutamine, [¹³C]alanine, [¹³C]serine, [¹³C]threonine, [¹³C]cysteine or [¹³C]asparagine (Cambridge Isotope Laboratories) for 10 min. The cells were subsequently washed once with ice-cold PBS, and the polar metabolites were extracted using 1 ml of methanol:acetonitrile:water (5:3:2, v/v/v) and the supernatant was dried by lyophilization. The dried extracts containing the hydrophilic metabolites were dissolved in 30 μl of water:acetonitrile (8:2, v/v) and 10 μl were used for the glutamine-derivatization procedure as described previously⁶¹ with minor modifications. In brief, the samples were placed into glass autosampler vials and then 35 μl of sodium borate buffer (100 mM, pH 9.0) was added, followed by mixing with pipetting. Next, 10 μl of the 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC, 10 mM in acetonitrile)-derivatizing reagent (Cayman Chemical) was added. The vial was sealed, mixed by vortex, and incubated at 55 °C for 15 min. The vial was cooled to room temperature, and then 15 μl of the sample was analysed by LC–MS/MS. A Vanquish Horizon UHPLC (Thermo Fisher Scientific) was used for the LC separations, using non-linear gradient conditions as follows: 0−1 min 3% B; 1−22 min 3 to 96% B (using curve no. 8, Thermo Scientific SII for Xcalibur); 22−25 min 96% B; 25−26 min 96 to 3% B; 26−30 min 3% B. Mobile phase A was water supplemented with 0.15% acetic acid, and mobile phase B was acetonitrile with 0.15% acetic acid. The column used was an Accucore C30 (250 × 2.1 mm, 2.6 μm) (Thermo Fisher Scientific), operated at 50 °C. The flow rate was 300 μl min⁻¹ and the injection volume used was 15 μl. All LC–MS/MS solvents and reagents were the highest purity available (water, acetonitrile, acetic acid, boric acid, sodium hydroxide) purchased from Thermo Fisher Scientific. A Q Exactive hybrid quadrupole-Orbitrap mass spectrometer (QE-MS) (Thermo Fisher Scientific) equipped with a HESI-II probe was employed as detector. The QE-MS was operated in the positive ion mode using targeted selected ions monitoring followed by a data-dependent MS/MS method (tSIM/dd-MS²). The QE-MS was operated at a resolution of 140,000 (FWHM, at 200 m/z), AGC targeted of 1 × 10⁶, max injection time 100 ms. For the dd-MS² conditions a resolution of 35,000 was used, AGC targeted of 1 × 10⁵, maximum injection time 50 ms, loop count 8, MS² isolation width 0.4 m/z and NCE 35. The operating conditions of the source were: Sheath gas flow 45; aux gas flow 8; sweep gas 1; spray voltage 3.8 kV in positive ion mode; capillary temperature 325 °C; S-lenses RF level 55; aux gas heater at 325 °C. Authentic unlabelled and [¹³C₅]glutamine standards were purchased from Sigma-Aldrich. The relative contents of intracellular glutamine [M+0] and [¹³C₅]glutamine [M+5] were determined from the tSIM/dd-MS² data as the corresponding parent/daughter ions 317.1230/171.0554 for [M+0] and 322.1663/171.0554 for [M+5]. The data was processed using the Xcalibu software (Thermo Fisher Scientific).

RNA isolation and gene expression profiling

RNA was isolated and purified from various cell types using the RNeasy Micro Kit (74004, Qiagen) according to the manufacturer’s instructions. cDNA synthesis was performed using the High Capacity cDNA Reverse Transcription Kit (4368813, Thermo Fisher Scientific) according to the manufacturer’s instructions. Real-time PCR was performed on the QuantStudio 7 Flex System (Applied Biosystems) using the PowerSYBR Green PCR Master Mix (4367659, Thermo Fisher Scientific). The sequences for mouse Flcn primers were previously described⁵⁵. The primers for detection of glutamine transporters are listed below: Slc1a5-F: CATCAACGACTCTGTTGTAGACC, Slc1a5-R: CGCTGGATACAGGATTGCGG; Slc6a14-F: GACAGCTTCATCCGAGAACTTC, Slc6a14-R: ATTGCCCAATCCCACTGCAT; Slc6a19-F: CAGGTGCTCAGGTCTTCTACT, Slc6a19-R: CGATCACAGAATCCATCTCACAA; Slc7a5-F: ATATCACGCTGCTCAACGGTG, Slc7a5-R: CTCCAGCATGTAGGCGTAGTC; Slc7a6-F: GCCTGCGTATGTCTGCTGA, Slc7a6-R: GCCCATGATAATGATGGCAATGA; Slc7a7-F: CACCACCAAGTATGAAGTGGC, Slc7a7-R: CCCTTAGGGGAGACAAAGATGC; Slc7a8-F: TGTGACTGAGGAACTTGTGGA, Slc7a8-R: GTGGACAGGGCAACAGAAATG; Slc7a9-F: GAGGAGACGGAGAGAGGATGA, Slc7a9-R: CCCCACGGATTCTGTGTTG; Slc38a1-F: AGCAACGACTCTAATGACTTCAC, Slc38a1-R: CCTCCTACTCTCCCGATCTGA; Slc38a2-F: TAATCTGAGCAATGCGATTGTGG, Slc38a2-R: AGATGGACGGAGTATAGCGAAAA; Slc38a3-F: GGAGGGGCTTCTACCAGTG, Slc38a3-R: GGAAAAGGATGATGCCCGTATTG; Slc38a4-F: GCGGGGACAGTATTCAGGAC, Slc38a4-R: GGAACTTCTGACTTTCGGCAT; Slc38a5-F: CTACAGGCAGGAACGCGAAG, Slc38a5-R: GGTTGAACACTGACATTCCGA; Actb-F: GGCACCACACCTTCTACAAT, Actb-R: CTTTGATGTCACGCACGATTTC. For microarray analysis, splenic cDC1s were sort-purified from wild-type (n = 2) and Flcn^ΔDC mice (n = 3) as described above. RNA was extracted and purified, and 125 ng RNA was used to profile with Affymetrix Mouse Clariom S Assay. For microarray analysis, the gene expression probe signals were quantile-normalized and summarized by the RMA algorithm by Affymetrix Expression Console (version 1.4.1), then the differential gene expression analysis was performed by R package limma (version 3.46.0). False discovery rate (FDR) was estimated by Benjamini–Hochberg method. Heat maps were generated using ComplexHeatmap (version 2.6.2) to show the average expression of genes from biological replicates of the same genotype. Microarray data have been deposited into the GEO series database under accession GSE210155.

ATAC-seq and data analysis

Library preparation

The ATAC-seq library was prepared as previously described⁶⁰. In brief, splenic cDC1s from wild-type and Flcn^ΔDC mice (n = 4 per genotype) were isolated as described above. A total of 5 × 10⁴ cells for each sample were used for the ATAC-seq library construction. After lysing in 50 μl ATAC-seq lysis buffer (10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl₂, 0.1% IGEPAL CA-630) on ice for 10 min, the resulting nuclei pellet was resuspended in 50 μl transposase reaction mix (25 μl 2 × TD buffer, 22.5 μl nuclease-free water, and 2.5 μl transposase) and incubated for 30 min at 37 °C. The tagged DNA was cleaned up using the Qiagen MinElute kit (Qiagen). A first round PCR with 5 cycles was performed to amplify and barcode the tagged DNA. The optimal cycle of further amplification was determined by real-time PCR (KAPA SYBRFast system; Kapa Biosystems). The final PCR products were purified using AMPure XP beads (Beckman Coulter). The fragment distribution of each library was checked by a TapeStation System (Agilent Technologies) and then sequenced on an Illumina NovaSeq with ~300 million reads per sample.

Data analysis

ATAC-seq analysis was performed as described previously⁶⁰. In brief, the paired-end fastq files obtained from NovaSeq were trimmed for Nextera adapter by trimmomatic (version 0.36, paired-end mode, with parameter LEADING:10 TRAILING:10 SLIDINGWINDOW:4:18 MINLEN:25). BWA (version 0.7.16) was used to align reads to mouse genome mm10 with default parameters. Resulting BAM files were filtered to remove duplicated reads (marked by Picard (version 2.9.4)) and to remove mitochondrial reads. After adjustment of Tn5 shift (reads were offset by +4 bp for the sense strand and −5 bp for the antisense strand), the reads were separated into nucleosome-free, mononucleosome, dinucleosome and trinucleosome by fragment size. All samples in this study had approximately 1 × 10⁷ nucleosome-free reads, indicative of good data quality. Next, these nucleosome-free reads were used for peak calling by MACS2 (version 2.1.1.20160309, with default parameters with ‘–extsize 200–nomodel’) with a higher cut-off (MACS2 −q 0.05). The consensus peaks for each group were further generated by keeping peaks that were presented in at least 50% of the replicates. The reproducible peaks were merged between wild-type and FLCN-deficient cDC1s if they overlapped by 100-bp and then were counted from each of the 8 samples by bedtools (version 2.25.0). Transcription factor footprinting activity were inferred and visualized using the RGT HINT software (version 0.13.2)⁶². Raw and processed ATAC-seq data have been deposited into the GEO series database under accession GSE210155.

scRNA-seq and data analysis

Library preparation

For scRNA-seq analysis in Fig. 1 and Extended Data Fig. 1, wild-type mice were challenged with MC38 colon adenocarcinoma cells, and treated with PBS or glutamine daily starting from day 5. DCs (CD45⁺CD64⁻Ly6C⁻CD11c⁺MHCII⁺), CD45⁺ non-macrophage immune cells (CD45⁺CD64⁻), macrophages (CD45⁺CD64⁺), and CD45⁻ tumour and other non-immune cells in the tumour tissues were sorted at 15 d after tumour challenge and mixed at a 5:4:1:1 ratio (to ensure that sufficient numbers of the less abundant DCs and non-macrophage immune cells were profiled; n = 2 biological replicates per group). For scRNA-seq analysis in Extended Data Figs. 7 and 8, wild-type and Flcn^ΔDC mice (n = 2 per genotype) were challenged with MC38 cells. CD45⁺ cells and DCs (CD45⁺CD64⁻Ly6C⁻CD11c⁺MHCII⁺) in the tumour tissues were sorted at 15 days after tumour challenge and mixed at a 2:1 ratio. The cell mixture was centrifuged at 2,000 rpm for 5 min and then resuspended in 1× PBS (Thermo Fisher Scientific) plus 0.04% BSA (Amresco) with a final concentration of 1 × 10⁶ cells per ml. The single-cell suspensions were loaded onto a Chromium Controller and encapsulated into droplets. Chromium Next GEM Single Cell 3′ (version 3.1; for scRNA-seq analysis in Fig. 1 and Extended Data Fig. 1) or Next GEM Single Cell 5′ (version 2; for scRNA-seq analysis in Extended Data Figs. 7 and 8) and Gel Bead Kit (10x Genomics) were used for the library preparation following manufacture’s instruction. The final libraries were quality-checked by 2100 Bioanalyzer (Agilent Technologies) and quantified by Qubit Fluorometer (Invitrogen). The resulting libraries were sequenced on NovaSeq (Illumina) with paired-end reads of 26 (for Chromium Next GEM Single Cell 5′ kit) or 28 (for Chromium Next GEM Single Cell 3′ kit) cycles (for read 1, 90 cycles for read 2 and 10 cycles for index 1 and 2 separately). An average of 500 million reads per sample were obtained.

Data preprocessing and quality control

After raw sequencing data were de-multiplexed by bcl2fastq (version 2.20.0.422), the Cell Ranger Single-Cell software suite (version 6.0; 10x Genomics) was used to process with the scRNA-seq FASTQ files. In brief, the FASTQ files were aligned to the mm10 mouse reference genome (ENSEMBL GRCm38). Gene expression was quantified by reads confidently mapped to the genome and assigned to cells by the cell barcodes. The output from Cell Ranger was imported into R (version 4.0.5) and analysed with Seurat (version 4.0.2). Cells with fewer than 200 genes detected, or with low unique molecular identifiers (UMI) counts (potentially dead cells) or unusually high UMI counts (potentially two or more cells in a single droplet) were removed. Cells with high percent (>5%) of reads mapping to mitochondrial genes (potentially dead cells) were also removed. Genes detected in fewer than three cells were discarded.

Clustering and cluster annotation

For unsupervised clustering and visualization, we normalized the expression level of each gene using NormalizeData with scale.factor as 1 × 10⁶ built in Seurat pipeline as described previously⁶⁰. In brief, principal component analysis (PCA) was performed using the top 2,000 highly variable genes. The top 30 principal components were used to build a shared nearest neighbour (SNN) graph, and cells were clustered using the Louvain algorithm as implemented in a FindClusters function from the Seurat package with resolution as 0.5. The cluster-specific differentially expressed genes were calculated by FindAllMarkers function from Seurat. For the CD8⁺ T cell subset analysis, we subsetted the CD8⁺ T cells by gating on the high expression of the CD3 subunit genes (Cd3e or Cd3d) and Cd8b gene and performed unsupervised clustering using the same graph-based clustering method. The CD8⁺ T cell subsets were further characterized by the high expression of Tcf7 (encodes TCF1) or Havcr2 (encodes TIM-3). For DCs, we first subsetted cDCs with high expression of Ptprc and Flt3. A second-round of dimensionality reduction and unsupervised clustering were then performed. The cDC1 cell cluster was characterized by expression of Clec9a and Xcr1. cDC2 cell cluster was characterized by expression of Cd209a. cDC1 and cDC2 gene signatures were then generated by identifying the top and bottom 200 genes (ranking log₂ fold change) of differential expression between cDC1s versus cDC2s. DCs that expressed Ccr7^63,64 were further subclustered into cDC1-derived and cDC2-derived cells based on cDC1 and cDC2 gene signatures. The cDC1-derived Ccr7⁺ DC cluster was then merged with the cDC1 cluster, while cDC2-derived Ccr7⁺ DC cluster was then merged with the cDC2 cluster for subsequent differential expression analysis, which was performed by FindMarkers function from Seurat package. The activity scores of gene signatures were calculated by AddModuleScore function. The difference in gene signature activity was examined by non-parametric, two-tailed Wilcoxon rank sum test and visualized using violin plots. Raw and processed scRNA-seq data have been deposited into the GEO series database under accession GSE210155.

Gene set enrichment analysis and signature curation

Genes were ranked by the fold change generated by the differential expression analysis. The pre-ranked gene set enrichment analysis (GSEA) was performed as previously described⁶⁵ against gene sets from KEGG, BIOCARTA, PID, REACTOME, C7 immunological, GO and HALLMARK collections from the Molecular Signatures Database (mSigDB) (https://www.broadinstitute.org/gsea/msigdb/, version 7.4) and signatures curated from published papers, as follows. For CD8⁺ T cells, gene signatures of ‘early activation’ and ‘effector/cytokine’ were curated by a previous publication²⁴. For cDC1s, the ‘MHCI antigen presentation’ signature and ‘DC activation’ signature were described in previous publications^24,64. The set of ‘putative TFEB target genes’ signature was derived from a public dataset, which identified TFEB targets by integrating TFEB ChIP-seq analysis and TFEB overexpression⁶⁶.

Public bioinformatics dataset analysis

To examine the expression of glutamine transporters in tumour cells, DCs, macrophages, B cells, natural killer cells, CD8⁺ and CD4⁺ T cells and other immune cells from the TME, a human melanoma dataset³⁵ (GSE72056) and a mouse tumour scRNA-seq dataset³⁴ (GSE121861, profiling B16F10 melanoma, EMT6 breast mammary carcinoma, LL2 Lewis lung carcinoma, CT26 and MC38 colon carcinoma and Sa1N fibrosarcoma) were analysed with Seurat (version 4.0.2). Tumour cells and CD45⁺ immune cells from different mouse tumour models were pooled for analysis in GSE121861. Expression of glutamine transporters in the indicated cell types was visualized by DotPlot function. To compare the expression of Slc1a5, Slc6a14, Slc6a19, Slc7a5, Slc7a6, Slc7a7, Slc7a8, Slc7a9, Slc38a1, Slc38a2, Slc38a3, Slc38a5, Slc38a7 and Slc38a8 in different immune cell types, the Immgen Microarray Gene Skyline data⁶⁷ was downloaded and visualized by the heatmap function in ComplexHeatmap R package.

Public microarray datasets profiling glutamine transporters in different immune cell subsets are available from the Immgen database (https://www.immgen.org/). KEGG, BIOCARTA, PID, REACTOME, C7 immunological, GO and HALLMARK collections were from the Molecular Signatures Database (mSigDB) (https://www.broadinstitute.org/gsea/msigdb/).

Statistical analysis for biological experiments

Analyses of biological experiments (non-omics) were performed using Prism software (version 8; GraphPad) by two-tailed paired Student’s t-test, two-tailed unpaired Student’s t-test, or one-way ANOVA with Newman–Keuls’s test. Two-way ANOVA was performed for comparing tumour growth curves. The Mantel–Cox test was used for comparing mouse survival curves. Two-tailed Wilcoxon rank sum test was applied for differential expression or activity score analysis of scRNA-seq data. Two-tailed Kolmogorov–Smirnov test by GSEA was used for pathway activity score analysis of scRNA-seq data. Two-tailed unpaired Student’s t-test was used for transcription factor footprinting analysis of ATAC-seq peaks. P < 0.05 was considered significant, and the exact P values are provided in the source data that accompanies this manuscript. Data are presented as mean ± s.e.m. No statistical method was used to pre-determine the sample sizes, but our sample sizes are similar to those reported in other publications^16,17. Age- and sex-matched mice with pre-determined genotypes were randomly assigned to control and experimental groups. No other randomization was performed. Data collection and analysis were not performed blind to the conditions of the experiments.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.