SLC38A2 and glutamine signalling in cDC1s dictate anti-tumour immunity

Abstract

Cancer cells evade T cell-mediated killing through tumour-immune interactions whose mechanisms are not well understood^1,2. Dendritic cells (DCs), especially type-1 conventional DCs (cDC1s), mediate T cell priming and therapeutic efficacy against tumours³. DC functions are orchestrated by pattern recognition receptors^3-5, although other signals involved remain incompletely defined. Nutrients are emerging mediators of adaptive immunity^6-8, but whether nutrients affect DC function or communication between innate and adaptive immune cells is largely unresolved. Here we establish glutamine as an intercellular metabolic checkpoint that dictates tumour-cDC1 crosstalk and licenses cDC1 function in activating cytotoxic T cells. Intratumoral glutamine supplementation inhibits tumour growth by augmenting cDC1-mediated CD8⁺ T cell immunity, and overcomes therapeutic resistance to checkpoint blockade and T cell-mediated immunotherapies. Mechanistically, tumour cells and cDC1s compete for glutamine uptake via the transporter SLC38A2 to tune anti-tumour immunity. Nutrient screening and integrative analyses show that glutamine is the dominant amino acid in promoting cDC1 function. Further, glutamine signalling via FLCN impinges on TFEB function. Loss of FLCN in DCs selectively impairs cDC1 function in vivo in a TFEB-dependent manner and phenocopies SLC38A2 deficiency by eliminating the anti-tumour therapeutic effect of glutamine supplementation. Our findings establish glutamine-mediated intercellular metabolic crosstalk between tumour cells and cDC1s that underpins tumour immune evasion, and reveal glutamine acquisition and signalling in cDC1s as limiting events for DC activation and putative targets for cancer treatment.

Fig. 1. Intratumoral glutamine supplementation promotes cDC1-mediated anti-tumour immunity.

a, Levels of glutamine and glucose in plasma and TIF of mice bearing MC38 tumours at day 15 (n = 4 per group). b,c, Growth and endpoint weight of MC38 (b; n = 8 per group) and B16-OVA (c; n = 10 per group) tumours (day 24 and 18, respectively) after intratumoral PBS or glutamine supplementation. d, MC38 tumour growth in Rag1^−/− mice after PBS or glutamine treatment (n = 7 per group). e, MC38 tumour growth and mouse survival after indicated treatments (n = 12 for Gln + anti-PD-1, n = 13 for all other groups). f, Growth of B16-OVA tumours in mice receiving intratumoral PBS or glutamine with activated OT-I cells (indicated by arrow) (n = 10 per group). g, MC38 tumour growth in tumour-free (having received prior glutamine + anti-PD-1 treatment; n = 8) or naive mice (n = 5) upon challenge with MC38 cells. h, Indicated T cell populations at day 15 in MC38 tumours treated with PBS (n = 7) or glutamine (n = 5). i, DCs, CD45⁺ non-macrophage immune cells, macrophages and CD45⁻ cells were sorted from PBS- and glutamine-treated MC38 tumours and mixed for scRNA-seq analysis. Violin plots show activity scores of early activation and effector/cytokine signalling signatures in intratumoral CD8⁺ T cells from MC38 tumours treated with PBS (n = 1,113 cells) or glutamine (n = 2,031 cells). Box plots show the median (centre line) with interquartile range of 25% to 75%. j,k, IFNγ⁺, TNF⁺ and granzyme B⁺ (GZMB⁺) (j) or effector-like (TIM-3⁺TCF1⁻) and stem-like (TIM-3⁻TCF1⁺) (k) CD8⁺ T cells at day 15 from MC38 tumours treated with PBS (n = 7) or glutamine (n = 5). l, MC38 tumour growth in indicated mice treated with PBS (n = 10 for wild-type, n = 8 for Batf3^−/−) or glutamine (n = 9 for wild-type, n = 8 for Batf3^−/−). WT, wild-type. m, Growth rate of B16-OVA tumours after transfer of OVA-pulsed cDC1s activated in the presence or absence of glutamine (n = 9 for DCs treated with glutamine, n = 8 for DCs treated without glutamine). Non-transfer control mice (n = 10) received PBS. Data are mean ± s.e.m., except in i. a, Two-tailed paired Student’s t-test. b,c,h,j,k, Two-tailed unpaired Student’s t-test (b,c, tumour weight). b–g,l,m, Two-way ANOVA for tumour size. e, Mantel–Cox test for survival. i, Two-tailed Wilcoxon rank sum test. Data are representative of two (a,d–h,j,l,m) or at least three (b,c,k) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Fig. 2. Glutamine interplay between tumour cells and cDC1s modulates anti-tumour immunity.

a–d, [³H]Thymidine (TdR) incorporation by OT-I (a,c,d) or OT-II (b) cells after coculture with cDC1s (n = 4 per group) pulsed with OVA in amino acid (AA)-replete medium (+AA) or medium lacking an individual amino acid (a,b), in culture supernatants derived from MC38 cells cultured in glutamine-free medium supplemented with indicated glutamine concentrations (c), or in MC38 culture supernatant supplemented with an individual amino acid (d). Arrows in a,b indicate fold change for +AA versus −Gln. e, CD86 and MHCII expression on BMDCs after 24 h of Transwell coculture with MC38 cells in medium containing 2 or 0.6 mM glutamine (n = 3 per group). MFI, mean fluorescence intensity. f, Expression of glutamine transporters in indicated mouse cell types (from GSE121861). NK cells, natural killer cells. g, Immunoblot analysis of SLC38A2 and β-actin in control and SLC38A2-deficient MC38 cells. h, Growth of control and SLC38A2-deficient MC38 tumours in wild-type mice (n = 10 per group). i,j, Indicated T cell populations (i) or IFNγ⁺, TNF⁺ or granzyme B⁺ (GZMB⁺) CD8⁺ T cells (j) from control and SLC38A2-deficient MC38 tumours at day 15 (n = 7 (i) or 6 (j) per group). k, Growth of control and SLC38A2-deficient MC38 tumours in Batf3^−/− mice (n = 10 per group). l,m, [³H]TdR incorporation by OT-I (l,m) or OT-II (l) cells after coculture with OVA-pulsed (l) or HKLM-OVA-pulsed (m) wild-type and SLC38A2-deficient cDC1s (n = 8 per genotype). n, CFSE^low OT-I cells in indicated OVA-immunized chimeric mice at day 3 after challenge (n = 5 per group). o,p, Growth of MC38 tumours in wild-type (n = 10) and Slc38a2^ΔDC (n = 9) mice (o) or in mice given intratumoral PBS (n = 9 for wild-type mice, n = 8 for Slc38a2^ΔDC mice) or glutamine (n = 7 per genotype) (p). q,r, OVA-specific CD8⁺ T cells (q) or IFNγ⁺ and TNF⁺ CD8⁺ T cells (r) in MC38-OVA tumours from wild-type (n = 8) and Slc38a2^ΔDC (n = 10) mice at day 19. s, MC38 tumour growth in wild-type (n = 7) and Xcr1^cre/+Slc38a2^fl/fl (n = 5) mice. Data are mean ± s.e.m. c,e,i,j,l,m,q,r, Two-tailed unpaired Student’s t-test. n, One-way ANOVA. h,k,o,p,s, Two-way ANOVA. Data are representative of one (g), two (e,i–k,m–s) or at least three (a–d,h,l) independent experiments. *P < 0.05, **P < 0.01. ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Fig. 3. Glutamine promotes the priming effect and anti-tumour immunity of cDC1s via FLCN.

a,b, The interaction of haemagglutinin (HA)–FLCN and Flag–FNIP2 in HEK293T cells cultured in glutamine-free medium (a) or after glutamine starvation and replacement (b) for the indicated time. Numbers indicate the relative intensity of the Flag band. Exp., exposure. c–e, [³H]Thymidine incorporation by OT-I (c, e) or OT-II (d) cells after coculture with OVA-pulsed (c,d) or HKLM-OVA-pulsed (e) splenic wild-type or FLCN-deficient cDC1s (c,e) or cDC2s (d,e) (n = 9 (c), n = 8 (d) or n = 6 (e) per genotype). f, CFSE^low OT-I cells in indicated OVA-immunized chimeric mice at day 3 after challenge (n = 6 per group). g,h, Growth of MC38 tumours in wild-type (n = 6) and Flcn^ΔDC (n = 5) mice (g) or wild-type (n = 10) and Xcr1^cre/+Flcn^fl/fl (n = 8) mice (h). i, MC38 tumour growth in the indicated mice given intratumoral PBS or glutamine (n = 8 mice per group). j,k, [³H]Thymidine incorporation by OT-I (j) or OT-II (k) cells cocultured with wild-type and FLCN-deficient cDC1s (j) or cDC2s (k) pulsed with OVA in fresh medium, MC38 supernatant or MC38 supernatant plus glutamine (n = 4 per group). l, T cell populations in MC38 tumours from the indicated mice at day 15 (n = 6 per genotype). m, IFNγ⁺, TNF⁺ and granzyme B⁺ (GZMB⁺) CD8⁺ T cells in MC38 tumours from wild-type (n = 6) and Flcn^ΔDC (n = 8) mice at day 15. n,o, OVA-specific CD8⁺ T cells (n) and TNF⁺IFNγ⁺ CD8⁺ T cells (o) in MC38-OVA tumours from wild-type (n = 6) and Flcn^ΔDC (n = 5) mice at day 19. Data are mean ± s.e.m. c–e,l–o, Two-tailed unpaired Student’s t-test. f, One-way ANOVA. g–k, Two-way ANOVA. Data are representative of two (a,b,e,f,j,k,n,o) or at least three (c,d,g–i,l,m) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Fig. 4. Co-deletion of TFEB restores the priming effect of FLCN-deficient or glutamine-deprived cDC1s.

a, Transcription factor footprinting analysis of ATAC-seq peaks in wild-type and FLCN-deficient splenic cDC1s, ranked by activity score. Selected transcription factors are indicated. Red and blue dots depict upregulated and downregulated activities of transcription factors, respectively, with select transcription factors indicated. b,c, GSEA enrichment plots showing upregulation of KEGG lysosome pathway (b) and putative TFEB target genes in FLCN-deficient cDC1s (c). FDR, false discovery rate; NES, normalized enrichment score. d, Wild-type and FLCN-deficient splenic cDC1s (n = 3 per genotype) were incubated with DQ-OVA for the indicated times in DQ-OVA degradation assays. Numbers indicate mean fluorescence intensity (MFI) of DQ-OVA. e, Immunoblot analysis of cathepsin D (pro and mature forms), FLCN, TFEB and GAPDH expression in cDC1s from the indicated mice. f, [³H]Thymidine incorporation by OT-I cells after coculture with splenic cDC1s from wild-type (n = 21), Flcn^ΔDC (n = 12), Tfeb^ΔDC (n = 15) and Flcn/Tfeb^ΔDC (n = 12) mice. g, CFSE^low OT-I cells in OVA-immunized wild-type (n = 8), Flcn^ΔDC (n = 9), Tfeb^ΔDC (n = 6) or Flcn/Tfeb^ΔDC (n = 10) mice on day 3 after challenge. h, MC38 tumour growth in wild-type (n = 9), Flcn^ΔDC (n = 7), Tfeb^ΔDC (n = 9) and Flcn/Tfeb^ΔDC (n = 7) mice. i,j, Effector-like (TIM-3⁺TCF1⁻) or stem-like (TIM-3⁻TCF1⁺) subsets (i) and MFI of T-bet (j) in CD8⁺ T cells in MC38 tumours from wild-type (n = 8), Flcn^ΔDC (n = 7), Tfeb^ΔDC (n = 6) and Flcn/Tfeb^ΔDC (n = 7) mice. Graph in j shows histogram of T-bet, with quantified MFI indicated. k, TFEB protein levels in cytosolic and nuclear fractions in wild-type and FLCN-deficient splenic cDC1s. Numbers indicate TFEB abundance in cytosolic or nuclear TFEB (normalized to GAPDH or lamin B, respectively) versus those in wild-type cDC1s treated with glutamine. l, [³H]Thymidine incorporation by OT-I cells after coculture with wild-type and TFEB-deficient splenic cDC1s pulsed with OVA in glutamine-replete or glutamine-deficient medium (n = 6 per genotype). Data are mean ± s.e.m. a,d, Two-tailed unpaired Student’s t-test. f,g,i,j,l, One-way ANOVA. h, Two-way ANOVA. Data are representative of two (d,e,j,l) or at least three (h,i,k) independent experiments or pooled from two (g) or three (f) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 1. Glutamine supplementation enhances anti-tumour immunity.

a, Glutamine (Gln) and glucose levels in plasma and tumour interstitial fluid (TIF) of B16-OVA tumour-bearing mice at day 15 (n = 4 each group). b,c, Amino acid levels in plasma and TIF of MC38 (b) or B16-OVA (c) tumour-bearing mice at day 15 (n = 4 each group). Cys, cysteine or cystine. d, Relative glutamine levels across time in B16-OVA TIF after intratumoral glutamine supplementation (n = 5 each group). e, B16-OVA tumour growth in mice receiving indicated treatments (n = 9 for Gln treatment; 10 for other groups). f, Indicated cell populations in PBS- or Gln-treated MC38 tumours at day 15 (n = 6 each group). g, Indicated T cell populations in PBS- or Gln-treated B16-OVA tumours at day 15 (n = 10 each group). h,i, DCs, CD45⁺ non-macrophage immune cells, macrophages, and CD45⁻ cells were sorted from PBS- and Gln-treated MC38 tumours and mixed at 5:4:1:1 ratio for scRNA-seq analysis. UMAP plot of cell subclusters (h). Violin plots showing Gzmb and Prf1 expression in intratumoral CD8⁺ T cells (i). j, In vivo killing of OVA_257–264-pulsed splenocytes (CFSE⁺) in MC38-OVA tumour-bearing mice treated with PBS or Gln (n = 5 each group). k, UMAP plots showing distribution of intratumoral CD8⁺ T cells subclusters in PBS- and Gln-treated MC38 tumours. Right, TIM-3⁻TCF1⁺ and TIM-3⁺TCF1⁻ CD8⁺ T cell subcluster quantification. l, Violin plots showing activity scores of early activation and effector/cytokine signalling gene signatures in TIM-3⁺TCF1⁻ and TIM-3⁻TCF1⁺ CD8⁺ T cells. m, B16-OVA tumour growth after receiving activated OT-I cells expanded in medium with or without Gln (n = 6 each group; adoptive transfer indicated by an arrow). Non-transfer control mice (n = 4) received PBS. n, The enrichment of the MHCI antigen presentation gene set in indicated cell populations (from scRNA-seq in h) as assessed by GSEA. NES, normalized enrichment score. o, Violin plot showing activity score of indicated gene signature in intratumoral cDC1 from PBS- or Gln-treated MC38 tumours. p, CD40, CD80, CD86, and MHCII expression on intratumoral cDC1s and cDC2s from PBS- or Gln-treated MC38 tumours at day 15 (n = 5 each group). MFI, mean fluorescence intensity. q, B16-OVA tumour growth in Batf3^−/− mice treated with intratumoral PBS or Gln (n = 8 per group). r, B16-OVA tumour-bearing mouse survival after transfer of OVA-pulsed cDC1 activated in presence or absence of Gln (n = 9 for DC treated with Gln; 8 for DC treated without Gln). Non-transfer control mice (n = 10) received PBS. Data are means ± s.e.m. except in i, l, o, where box shows median (centre line) with interquartile range of 25% to 75%. a–c, Two-tailed paired Student’s t-test. f,g,j,p, Two-tailed unpaired Student’s t-test. d, One-way ANOVA. e,m,q, Two-way ANOVA. i,l,o, Two-tailed Wilcoxon rank sum test. n, Two-tailed Kolmogorov–Smirnov test. r, Mantel–Cox test. Data are representative of one (d,j) or two (a–c, e–g, m, p–r) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, **** P < 0.0001. NS, not significant. Source data

Extended Data Fig. 2. Glutamine supplementation promotes intratumoral CD8⁺ T cell activation and effector function.

a,b, ZsGreen⁺ migratory cDC1s and cDC2s in tumour-draining lymph nodes (dLNs; a) or tumour (b) of PBS- or Gln-treated B16-ZsGreen tumour-bearing mice at day 16 (n = 10 each group). Gln, glutamine. c,d, Schematic depicting adoptive transfer of CFSE-labelled naïve OT-I cells into MC38-OVA tumour-bearing mice treated with PBS (n = 7) or Gln (n = 8) (c). CFSE^low OT-I cells in tumour dLNs and non-dLNs (ndLNs) (d). e–k Schematic depicting adoptive transfer of CD45.1⁺ activated OT-I cells into B16-OVA tumour-bearing mice treated with PBS or Gln (n = 10 each group) (e). Intratumoral OT-I cell number (f). Effector-like (TIM-3⁺Ly108⁻ or TIM-3⁺TCF1⁻) and stem-like (TIM-3⁻Ly108⁺ or TIM-3⁻TCF1⁺) (g,h) subsets of OT-I cells. T-bet mean fluorescence intensity (MFI) in OT-I cells (i). Granzyme B⁺ (GZMB⁺) (j) and TNF⁺IFNγ⁺ (k) OT-I cells. Data are means ± s.e.m. a,b,d,f,i–k, Two-tailed unpaired Student’s t-test. g,h, Two-way ANOVA. Data are from one experiment (a,b,d,f–k). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 3. Glutamine promotes priming capacity of DCs.

a, Schematic of functional amino acid screening assay used in Fig. 2a, b and Extended Data Fig. 3b–e. b–e, IL-2 (b,c) and IFNγ (b) production or [³H] Thymidine (TdR) incorporation by OT-I (b,d) or OT-II (c,e) cells after coculture with splenic cDC1s (b,c) or cDC2s (d,e) pulsed with OVA in amino acid-replete medium (+AA) or medium lacking an individual amino acid (n = 4 each group). Arrow, fold change (+AA versus −Gln) (n = 4 each group). Gln, glutamine. f, [³H] TdR incorporation by OT-I or OT-II cells after coculture with cDC1 (CD24^high)- and cDC2 (CD24^low)-like FLT3L-BMDCs pulsed with OVA in medium with or without Gln (n = 5 each group). BMDCs, bone marrow-derived DCs. g,h, [³H] TdR incorporation by OT-I cells after coculture with splenic cDC1s pulsed with OVA or OVA_257–264 (g) or HKLM-OVA (h) in Gln-replete or -deficient medium (g, n = 8 each group; h, n = 9 each group). i, Live (7-AAD⁻Annexin V⁻) splenic cDC1s and cDC2s after culture with or without Gln for indicated times (n = 4 per group). j,k, [³H] TdR incorporation by OT-I or OT-II cells after coculture with cDC1s (j) or cDC2s (k) pulsed with OVA in Gln-free medium (–AA) supplemented with an individual amino acid (n = 4 each group). Arrow, fold change (+Gln versus −AA). l, Relative changes in amino acid levels in MC38 cell culture supernatant versus fresh medium (FM; containing 0.6 mM Gln) (n = 3 per group). Red bar indicates Gln. m,n, [³H] TdR incorporation by OT-II cells after coculture with cDC1s pulsed with OVA in FM or the medium supernatant (MC38 sup) derived from MC38 cells cultured in glutamine-free medium supplemented with indicated glutamine concentrations (m) or supernatant supplemented with an individual amino acid (n) (n = 4 each group). o, Intracellular Gln levels in splenic cDC1s cultured under indicated conditions (n = 5 each group). p, Relative changes in amino acid levels in B16F10 cell culture supernatant versus fresh medium (containing 0.6 mM Gln) (n = 3 each group). Red bar indicates Gln. q, [³H] TdR incorporation by OT-I cells after coculture with splenic cDC1s pulsed with OVA in FM or B16F10 cell culture supernatant (B16 sup) supplemented with or without Gln (n = 4 each group). r, CD86 and MHCII expression on cDC1s incubated in the indicated media (n = 6 each group). MFI, mean fluorescence intensity. s, IL-12p40⁺ cDC1s and cDC2s after lipopolysaccharide (LPS) or poly I:C stimulation in Gln-replete or -deficient medium (n = 4 each group). Data are means ± s.e.m. f–h,m,s, Two-tailed unpaired Student’s t-test. o,q,r, One-way ANOVA. Data are representative of one (l,p), two (f–i,q) or at least three (b–e,j,k,m–o,s), or pooled from three (r) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 4. SLC38A2 deficiency in tumour cells promotes anti-tumour immunity.

a, Expression of glutamine transporters in indicated human cell populations (from GSE72056). b, Microarray analysis (from ImmGen database) showing the relative expression of the indicated glutamine transporters across different immune cell types. c, Real-time PCR analysis of mRNA levels of glutamine transporters in MC38 cells, cDC1s and cDC2s (n = 3 each group). d, Immunoblot analysis of SLC38A2 and β-Actin in intratumoral cDC1s, CD8⁺ T and tumour cells from B16-FLT3L tumours. kDa, kilodaltons. e, Intracellular ¹³C-glutamine (Gln) levels in control and SLC38A2-deficient MC38 cells after 10 min incubation (n = 4 each group). f, Intracellular Gln levels in control and SLC38A2-deficient MC38 cells (n = 4 each group). g, Immunoblot analysis of SLC38A2 and β-Actin in control and SLC38A2-deficient B16-OVA cells. h, Intracellular levels of indicated ¹³C-labelled amino acids control and SLC38A2-deficient B16-OVA cells after 10 min incubation (n = 4 for each genotype). Ala, alanine; Ser, serine; Thr, threonine; Cys, cysteine; Asn, asparagine. i, Growth of control (n = 7) or SLC38A2-deficient (n = 9) B16-OVA tumours in wild-type mice. j, Levels of indicated SLC38A2 substrates in tumour interstitial fluid (TIF) from control and SLC38A2-deficient B16-OVA tumours at day 15 (n = 5 each group). Cys, cysteine or cystine. k, Indicated T cell populations in control and SLC38A2-deficient MC38 tumours at day 15 (n = 7 each group). l, Control and SLC38A2-defcient tumour growth in Rag1^−/− mice (n = 14 each group). Data are means ± s.e.m. e,f,h,k, Two-tailed unpaired Student’s t-test. i,j,l, Two-way ANOVA. Data are representative of one (c,d,g,j) or two (e,f,h,i,k, l) independent experiments. *P < 0.05, **P < 0.01, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 5. DC and T cell phenotypes in Slc38a2^ΔDC mice.

a, Real-time PCR analysis of Slc38a2 mRNA levels in different splenic immune cell populations from indicated mice (n = 4 per genotype). WT, wild-type. b, Intracellular ¹³C-glutamine (Gln) levels in wild-type or SLC38A2-deficient cDC1s (n = 5 each genotype) and cDC2s (n = 6 for wild-type; 4 for Slc38a2^ΔDC) after 10 min of incubation. c, Intracellular levels of ¹³C-labelled amino acids in wild-type and SLC38A2-deficient splenic cDC1s and cDC2s after 10 min of incubation (n = 4 per genotype). Ala, alanine; Ser, serine; Thr, threonine; Cys, cysteine; Asn, asparagine. cDCs were sort-purified from wild-type and Slc38a2^ΔDC mice. d,e, Splenic cDCs and pDCs (d) or cDC1s and cDC2s (e) in wild-type (n = 6) and Slc38a2^ΔDC (n = 4) mice. f,g, Total (f) or CD44^highCD62L^low (g) CD4⁺ and CD8⁺ T cells from wild-type (n = 6) and Slc38a2^ΔDC (n = 4) mice. PLNs, peripheral lymph nodes; MLNs, mesenteric lymph nodes. h, IFNγ-, IL-2-, IL-4- or IL-17A-producing CD4⁺, and IFNγ- or IL-2-producing CD8⁺ T cells from wild-type (n = 6) and Slc38a2^ΔDC (n = 4) mice. i, cDC1 and cDC2 chimerism ratios (normalized to internal B cells) in indicated mixed BM chimeras (n = 5 each group). j,k, Active caspase 3⁺ (j) or Ki67⁺ (k) cDC1s and cDC2s from Slc38a2^ΔDC:CD45.1⁺ mixed BM chimeras (n = 5 each group). l, [³H] Thymidine (TdR) incorporation by OT-I and OT-II cells after coculture with wild-type or SLC38A2-deficient OVA-pulsed cDC2s (n = 8 per genotype). m,n, IL-2 or IFNγ production by OT-I or OT-II cells after coculture with OVA-pulsed wild-type or SLC38A2-deficient cDC1s (m) and cDC2s (n) as indicated (n = 6 per genotype for cDC1s, n = 8 per genotype for cDC2s). o,p, [³H] TdR incorporation by OT-I or OT-II cells after coculture with OVA_257–264- or OVA_323–339-pulsed WT or SLC38A2-deficient cDC1s (o) and cDC2s (p) as indicated (n = 8 per genotype) q, Expression of indicated molecules on splenic cDC1s and cDC2s from wild-type (n = 8) and Slc38a2^ΔDC (n = 7) mice. r, IL-12p40⁺ wild-type or SLC38A2-deficient cDC1s and cDC2s after lipopolysaccharide (LPS) or poly I:C stimulation (n = 4 per genotype). Data are means ± s.e.m. a–r, Two-tailed unpaired Student’s t-test. Data are representative of one (c), two (a,b,i–k,r) or three (l–p), or pooled from two (d–h, q) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 6. SLC38A2 deficiency in DCs impairs anti-tumour adaptive immunity.

a, B16-OVA tumour growth in wild-type (n = 9) and Slc38a2^∆DC (n = 11) mice. WT, wild-type. b,c, Indicated T cell populations (b) or IFNγ⁺ or Granzyme B⁺ (GZMB⁺) CD8⁺ T cells (c) in MC38 tumours from wild-type (n = 8) and Slc38a2^∆DC (n = 9) mice at day 15. d,e, ZsGreen⁺ migratory cDC1s and cDC2s in tumour-draining lymph nodes (dLNs) (d) or tumours (e) from indicated mice bearing B16-ZsGreen tumours (n = 10 per genotype). f, CFSE^low OT-I cells in tumour dLNs or non-dLNs (ndLNs) of MC38-OVA tumour-bearing wild-type (n = 6) and Slc38a2^∆DC (n = 8) mice at day 2 after naïve OT-I cell adoptive transfer. g–k, Activated OT-I cells were transferred into B16-OVA tumour-bearing wild-type (n = 7) and Slc38a2^∆DC (n = 12) mice. OT-I cells were analysed at day 7 after adoptive transfer. OT-I cell number (g). Effector-like (TIM-3⁺Ly108⁻ or TIM-3⁺TCF1⁻) and stem-like (TIM-3⁻Ly108⁺ or TIM-3⁻TCF1⁺) (h,i) OT-I cells. T-bet mean fluorescence intensity (MFI) in OT-I cells (j). GZMB⁺ OT-I cells (k). l, B16-OVA tumour growth in mice receiving sgNTC- or sgSlc38a2-transduced OT-I cells at day 12 (n = 4 for sgNTC; 3 for sgSlc38a2; transfer indicated by an arrow). Non-transfer control mice (n = 3) received PBS. m, MC38 tumour growth in wild-type and Cd4^creSlc38a2^fl/fl mice (n = 7 per genotype). Data are means ± s.e.m. b–g,j,k, Two-tailed unpaired Student’s t-test. a,h,i,l,m, Two-way ANOVA. Data are representative of one (d–k) or two (a–c,l,m) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant. Source data

Extended Data Fig. 7. GATOR1−GATOR2 complex assembly or interaction is not influenced by glutamine, and FLCN deficiency in DCs impairs anti-tumour immunity.

a,b, Immunoblot analysis of GATOR1 or GATOR2 complex proteins in glutamine (Gln)-starved HEK293T cells after Gln refeeding for 15 min (Gln add-back). No starvation, cells maintained in Gln-sufficient medium. Immunoprecipitation (IP) was performed using anti-DEPDC5 (a) or anti-WDR24 (b) antibody. kDa, kilodaltons. c, Real-time PCR analysis of Flcn mRNA levels in indicated cell populations from wild-type and Flcn^∆DC mice (n = 4 per genotype for cDC1s and cDC2s; 2 per genotype for other cell types). WT, wild-type. d, Number of splenic cDC1s and cDC2s from indicated mice (n = 6 per genotype). e, cDC1 and cDC2 chimerism ratios (normalized to internal B cells) in indicated mixed BM chimeras (n = 5 each group). f,g, Active caspase 3⁺ (f) or Ki67⁺ (g) cDC1s and cDC2s from Flcn^ΔDC:CD45.1⁺ mixed BM chimeras (n = 5 each group). h,i, IL-2 (h,i) and IFNγ (h) production by OT-I (h) and OT-II (i) cells after coculture with OVA-pulsed splenic wild-type or FLCN-deficient cDC1s (h; n = 6 per genotype) or cDC2s (i; n = 8 per genotype). j, IL-12p40⁺ wild-type and FLCN-deficient cDC1s or cDC2s after lipopolysaccharide (LPS) or poly I:C stimulation (n = 4 per genotype). k, MC38 tumour weight (at day 20) in wild-type (n = 6) and Flcn^∆DC (n = 5) mice. l, Tumour growth curves in wild-type (n = 7) and Flcn^∆DC (n = 4) mice inoculated with B16-OVA cells. m, MC38 tumour growth in indicated chimeric mice (n = 8 for Batf3^−/−:WT chimeras; 6 for other chimeras). n, [³H] Thymidine (TdR) incorporation by OT-I and OT-II cells after coculture with cDC1s or cDC2s pulsed with OVA in Gln-sufficient (+Gln) or Gln-free (−Gln) medium (n = 6 each group). o,p, Intratumoral CD45⁺ cells and cDCs were sorted from MC38 tumour-bearing wild-type and Flcn^∆DC mice at day 15 and mixed at ratio of 2:1 for scRNA-seq analysis. UMAP plots of cells indicated by cell clusters (o). Frequencies of indicated cell populations (n = 2 per genotype) (p). q, Analysis of indicated intratumoral immune cell populations in MC38 tumour-bearing wild-type and Flcn^∆DC mice (n = 10 per genotype). r, CD8⁺ T to T_reg cell ratio in MC38 tumours from indicated mice at day 15 (n = 6 per genotype). s, Bar graph showing the number of differentially expressed genes (DEGs; Flcn^∆DC versus wild-type) in each scRNA-seq cell cluster (o). t,u, Violin plots showing the activity scores of indicated gene signatures in intratumoral cDC1s (profiled by scRNA-seq) from PBS (n = 279 cells)- or Gln (n = 90 cells)-treated MC38 tumours. v, GSEA enrichment plot showing downregulation of indicated gene signature in intratumoral FLCN-deficient versus wild-type cDC1s (profiled by scRNA-seq). FDR, false discovery rate; NES, normalized enrichment score. Data are means ± s.e.m. c–k,q,r, Two-tailed unpaired Student’s t-test. l–n, Two-way ANOVA. t,u, Two-tailed Wilcoxon rank sum test. Data are representative of one (m), two (a–g,j,q) or at least three (h,i,k,l,n,r) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 8. FLCN deficiency in DCs impairs effector function of intratumoral CD8⁺ T cells.

a, Violin plots showing activity scores of indicated gene signatures in intratumoral CD8⁺ T cells (profiled by scRNA-seq). b,c, CD44 (b) and CD69 (c) expression on CD8⁺ T cells in MC38 tumours from wild-type (n = 6) and Flcn^∆DC (n = 7) mice at day 15. MFI, mean fluorescence intensity; WT, wild-type. d, PD-1⁺CD8⁺ T cells in MC38 tumours from wild-type (n = 6) and Flcn^∆DC (n = 8) mice at day 15. e, UMAP plots showing distributions of effector-like (TIM-3⁺TCF1⁻) and stem-like (TIM-3⁻TCF1⁺) cells among intratumoral CD8⁺ T cells from indicated MC38 tumour-bearing mice. Right, indicated intratumoral CD8⁺ T cell subclusters quantification. f,g, TIM-3⁺TCF1⁻ and TIM-3⁻TCF1⁺ CD8⁺ T cells (f) and T-bet MFI in intratumoral CD8⁺ T cells (g) in MC38 tumours from wild-type (n = 6) and Flcn^∆DC (n = 8) mice at day 15. h, Ki67⁺ CD8⁺ T cells in MC38 tumours from indicated mice (n = 8 per genotype). i, IFNγ⁺, TNF⁺ or Granzyme B⁺ (GZMB⁺) CD8⁺ T cells from mice as in d,f. j,k, ZsGreen⁺ migratory cDC1s and cDC2s in tumour-draining lymph nodes (dLNs; j) or tumours (k) from indicated B16-ZsGreen tumour-bearing mice at day 16 (n = 9 per genotype). l, CFSE^low OT-I cells in tumour dLNs and non-dLNs (ndLNs) of indicated MC38-OVA tumour-bearing mice (n = 6 per genotype) at day 2 after naïve OT-I cell transfer. m–r, Activated OT-I cells were transferred into B16-OVA tumour-bearing wild-type (n = 14) and Flcn^∆DC (n = 12) mice. OT-I cells were analysed at day 7 after adoptive transfer. OT-I cell number (m). TIM-3⁺Ly108⁻ and TIM-3⁻Ly108⁺ (n) or TIM-3⁺TCF1⁻ and TIM-3⁻TCF1⁺ (o) OT-I cells. T-bet in OT-I cells (p). GZMB⁺ (q) or TNF⁺IFNγ⁺ (r) OT-I cells. Data are means ± s.e.m. a, Two-tailed Wilcoxon rank sum test. b–d,g–m,p–r, Two-tailed unpaired Student’s t-test. f,n,o, Two-way ANOVA. Data are representative of one (j–r), two (c,g,h) or at least three (b,d,f,i) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 9. Enhanced lysosomal activation in FLCN-deficient cDC1s is rescued by TFEB co-deletion.

a, Principal component analysis (PCA) plot of ATAC-seq data for wild-type and FLCN-deficient cDC1s. Percentage of variance is shown. WT, wild-type. b,c, GSEA of transcriptome analysis of FLCN-deficient versus wild-type cDC1s using KEGG pathway (b) and Hallmark gene sets combined with putative TFEB target genes (c). False discovery rate (FDR) and normalized enrichment score (NES) of top 15 pathways (ranked by NES) were visualized. d, TFEB protein levels in cytosolic and nuclear fractions in wild-type and FLCN-deficient splenic cDC1s. Numbers indicate TFEB abundance in cytosolic or nuclear TFEB (normalized to GAPDH or Lamin B1, respectively) versus wild-type. kDa, kilodaltons. e, Flow cytometry analysis (left) and quantification (right) of Lysotracker staining in splenic cDC1s from wild-type (n = 7) and Flcn^ΔDC (n = 7) mice. MFI, mean fluorescence intensity. f, Imaging analysis of volume of EEA1⁺ vesicles per cell (n = 28 for wild-type; 29 for Flcn^ΔDC) or an individual EEA1⁺ vesicle (n = 496 for wild-type; 512 for Flcn^ΔDC) in wild-type or FLCN-deficient cDC1s. Scale bar, 10 μm. g, MFI of LAMP1 in wild-type (n = 4) or FLCN-deficient (n = 3) splenic cDC1s. h, Heatmap showing expression of top 20 differentially expressed genes (ranked by Log₂FC) from KEGG lysosome pathway in FLCN-deficient (n = 3) versus wild-type (n = 2) splenic cDC1s. i, Cathepsin D (pro and mature forms) and β-Tubulin expression in wild-type and FLCN-deficient cDC1s. j, Accessibility of the Ctsd gene and its upstream regions in wild-type and FLCN-deficient cDC1s assessed by ATAC-seq. k, pHrodo Green Dextran MFI in splenic wild-type and FLCN-deficient cDC1s (n = 4 per genotype). l, MFI of Lysotracker in splenic cDC1s from wild-type (n = 7), Flcn^ΔDC (n = 4), Tfeb^ΔDC(n = 5) and Flcn/Tfeb^ΔDC (n = 4) mice. m, IL-2 and IFNγ production by OT-I cells after coculture with OVA-pulsed splenic cDC1s from the indicated mice (n = 6 per genotype). Data are means ± s.e.m. e–g,k, Two-tailed unpaired Student’s t-test. l,m, One-way ANOVA. Data are representative of two (d,f,g,i,k), or three (m), or pooled from three (e,l) independent experiments. *P < 0.05, **P < 0.01, ****P < 0.0001. NS, not significant. Source data

Extended Data Fig. 10. Roles of FLCN–TFEB signalling axis in mediating anti-tumour immunity and glutamine availability.

a, MC38 tumour weight (at day 22) in wild-type (n = 9), Flcn^ΔDC (n = 7), Tfeb^ΔDC (n = 9) and Flcn/Tfeb^ΔDC(n = 7) mice. WT, wild-type. b, B16-OVA tumour growth in wild-type (n = 11), Flcn^ΔDC (n = 11), Tfeb^ΔDC (n = 8) and Flcn/Tfeb^ΔDC (n = 7) mice. c, d, Total CD4⁺ and CD8⁺ T cells (c) or IFNγ⁺, TNF⁺ and Granzyme B⁺ (GZMB⁺) CD8⁺ T cells (d) in MC38 tumours from wild-type (n = 8), Flcn^ΔDC (n = 7), Tfeb^ΔDC (n = 6) and Flcn/Tfeb^ΔDC (n = 7) mice at day 15. e, Overlap of putative TFEB target genes and significantly upregulated genes in FLCN-deficient (versus wild-type) cDC1s. f, GSEA plot depicting the enrichment of gene set containing the 26 overlapped genes identified in e, in cDC1s treated with glutamine (Gln)-free medium versus complete medium. g, Heatmap showing expression of the leading-edge genes from Extended Data Fig. 10f in cDC1s treated with Gln-free medium and complete medium (n = 4 per group). h, Schematic of glutamine intercellular crosstalk between cDC1s and tumour cells, and nutrient signalling in cDC1s in modulating anti-tumour immunity. cDC1s and tumour cells both express SLC38A2 to mediate glutamine uptake to tune anti-tumour immunity, with tumour cells expressing higher levels of SLC38A2 than cDC1s. In cDC1s, glutamine induces FLCN–FNIP2 complex assembly and inhibits TFEB activity to promote the antigen presentation capacity of cDC1s. Consequently, glutamine-dependent signalling in cDC1s enhances cytokine production and generation of cytotoxic effector-like CD8⁺ T cells in the TME. The figure was generated using BioRender. Data are means ± s.e.m. d, Two-tailed unpaired Student’s t-test. a,c, One-way ANOVA. b, Two-way ANOVA. Data are representative of two (a–d) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant. Source data