Mitochondrial metabolism and signaling direct dendritic cell function in antitumor immunity

Zhiyuan You; Jiyeon Kim; Chuansheng Guo; Haoran Hu; Nicole M Chapman; Xiaoxi Meng; Hao Shi; Yan Wang; Cliff Guy; Anil KC; Jia Li; Jordy Saravia; Gustavo Palacios; Sherri Rankin; Camenzind G Robinson; Hongbo Chi

doi:10.1126/science.adv6582

. Author manuscript; available in PMC: 2026 Apr 25.

Published in final edited form as: Science. 2026 Apr 2;392(6793):eadv6582. doi: 10.1126/science.adv6582

Mitochondrial metabolism and signaling direct dendritic cell function in antitumor immunity

Zhiyuan You ^1,^†, Jiyeon Kim ^1,^†, Chuansheng Guo ¹, Haoran Hu ¹, Nicole M Chapman ¹, Xiaoxi Meng ¹, Hao Shi ¹, Yan Wang ¹, Cliff Guy ¹, Anil KC ¹, Jia Li ¹, Jordy Saravia ¹, Gustavo Palacios ¹, Sherri Rankin ¹, Camenzind G Robinson ², Hongbo Chi ^1,^*

PMCID: PMC13109015 NIHMSID: NIHMS2161853 PMID: 41926583

Abstract

Antitumor immunity requires conventional type 1 dendritic cells (cDC1s). How cDC1s maintain functional fitness in the tumor microenvironment remains unclear. Here, we establish that intratumoral cDC1s exhibited discrete mitochondrial states, and OPA1-mediated mitochondrial energy and redox metabolism dictated cDC1 antitumor responses. Mechanistically, OPA1 orchestrated antigen presentation and CD8⁺ T cell priming function of cDC1s by promoting nuclear respiratory factor 1 (NRF1) expression and electron transport chain integrity, thereby supporting bioenergetics and NAD⁺/NADH balance. During tumor progression, mitochondrial membrane potential and volume, as well as OPA1–NRF1 signaling, declined in intratumoral cDC1s. Furthermore, intratumoral administration of cDC1s with polarized mitochondria showed strong immunotherapeutic benefits, particularly in combination with immune checkpoint blockade. Collectively, our findings reveal mitochondrial metabolism and signaling as putative targets to reinvigorate cDC1 function for cancer immunotherapy.

Introduction

The tumor microenvironment (TME) contains multiple cell types and factors that dynamically modulate immune cell function and antitumor responses (1). Among them, conventional type 1 dendritic cells (cDC1s) determine the quantity and quality of antitumor cytotoxic T lymphocyte (CTL) responses (2–4) and the outcome of immunotherapies (5–7). Specifically, to orchestrate antitumor immunity, cDC1s migrate from the TME and cross-present tumor-associated antigens to prime naïve CD8⁺ T cells in the tumor-draining lymph node (tdLN) (6, 8). Intratumoral cDC1s also mediate CTL recruitment to the TME (9), where they present local antigens and provide additional signals for CTL restimulation and antitumor effector responses (10, 11). Therefore, mechanisms that unleash the antitumor effects of cDC1s likely represent actionable targets to enhance adaptive immunity and immunotherapy against tumors.

Cellular metabolism is a key regulator of immune cell development, activation and homeostasis (12, 13), with nutrient and metabolic targeting emerging as promising strategies to boost antitumor immunity in the contexts of adoptive cell therapy and immune checkpoint blockade (ICB) (1, 14). Metabolic programming shapes DC development and functional fitness (15, 16). However, little is known about the role of mitochondrial metabolism in orchestrating the functional adaptation of cDCs, especially in the TME. Currently, it is appreciated that activation of pattern recognition receptors promotes metabolic rewiring of DCs to support their maturation (17), associated with an upregulation of glycolysis and fatty acid synthesis over oxidative phosphorylation (OXPHOS) to facilitate DC maturation and proinflammatory function (18, 19). Conversely, upregulation of OXPHOS is a hallmark for tolerogenic human monocyte-derived DCs compared to inflammatory counterparts (20, 21). However, these metabolic alterations occur in vitro or under nutrient-replete conditions, which are in direct contrast to the in vivo context of the nutrient-restricted TME (22). Consequently, how mitochondrial metabolism in DCs is rewired in the TME, and the functional outcomes and mechanistic basis, are largely unexplored. Therefore, it is critical to uncover mechanisms that are permissive for the metabolic and functional reprogramming of cDCs in the TME to improve cancer immunity and immunotherapy.

Here, we establish mitochondrial fusion protein OPA1 as a rheostat for dictating mitochondrial energy metabolism and redox balance in cDC1s for antitumor immunity. We reveal a OPA1–NRF1 signaling axis that links oxidative metabolism to suppression of autophagic degradation of MHC-I and antigen, thereby orchestrating cDC1 antigen presentation (immunological “signal 1”) and activation of CTL antitumor function. In addition, OPA1-mediated balance of NAD⁺/NADH ratio contributed to cDC1 function. Importantly, mitochondrial membrane potential and volume in intratumoral cDC1s declined during tumor progression, associated with downregulation of OPA1–NRF1 signaling, while administration of cDC1s with polarized mitochondria benefited cancer immunotherapy. Our results suggest that improving mitochondrial energy and redox metabolism represents a potential approach for reinvigorating cDC1 function and improving cancer therapy.

Results

Discrete mitochondrial states distinguish functional subpopulations of cDC1s

To investigate metabolic pathways important for cDC1-mediated antitumor responses, we challenged wild-type (WT) mice with OVA-expressing B16F10 (B16-OVA) melanoma cells and performed tandem mass tag-based multiplex proteomics profiling (23) of cDC1s from tumor and spleen. Mitochondrial respiration module was enriched in cDC1s isolated from B16-OVA tumors versus spleen (fig. S1A, and table S1). Accordingly, intratumoral cDC1s had higher oxygen consumption rate (OCR) and ATP production compared to splenic counterparts (fig. S1B). Enhanced OXPHOS signature was also evident in murine cDC1s from adenocarcinoma-bearing lungs (24) (fig. S1C) or cDC1s from a human pan-cancer dataset (25) (fig. S1D), indicating conserved upregulation of OXPHOS in intratumoral cDC1s.

To test if these alterations are specific to the TME or a general feature of cDC1 activation, we cultured splenic cDC1s with tumor cell-conditioned mediums (TCMs), followed by flow cytometry measurements of mitochondrial membrane potential [using Tetramethylrhodamine methyl ester perchlorate (TMRM)] and mitochondrial mass [using MitoTracker^™ Green (MG)]. Compared to control medium, TCMs from B16-OVA, B16F10, B16-Flt3L, MC38, Lewis lung carcinoma (LLC) or E.G7-OVA cells enhanced cDC1 mitochondrial membrane potential and TMRM to MG ratio, with more limited effects on mitochondrial mass (fig. S1E). In contrast, cDC1s treated with LPS (TLR4 agonist) or poly I:C (TLR3 agonist) had no effects on mitochondrial membrane potential, mass or TMRM/MG gMFI ratio (fig. S1F), suggesting context-dependent mitochondrial programming.

Co-staining with TMRM and MG uncovered subpopulations of cDC1s with discrete mitochondrial states in B16-OVA tumors (Fig. 1A, and fig. S1G). One subpopulation showed a high ratio of TMRM to MG staining ([TMRM/MG]^hi), indicative of cells with polarized mitochondria (26, 27), and the other had depolarized mitochondria ([TMRM/MG]^lo) (Fig. 1A). [TMRM/MG]^hi cDC1s from the tumor showed increased capacity to prime OT-I cells (Fig. 1B, and fig. S1H), and had increased expression of MHC-I and MHC-II molecules albeit not costimulatory ligands (fig. S1I). Of note, a recent study identifies IL-12p40⁺CCR7⁺ and CXCL9⁺CCR7⁻ (also expressing CXCL10) cDC1s with specialized roles in mediating CD8⁺ T cell differentiation states in the TME (28). We found that the frequencies of CXCL9⁺, CXCL10⁺ or CCR7⁺ cells were comparable between intratumoral [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s, while IL-12p40⁺ cells among [TMRM/MG]^hi cells were modestly increased (fig. S1, J and K). Beyond the B16-OVA model, two discrete mitochondrial states were also detected among intratumoral cDC1s from LLC tumors (fig. S1L), with [TMRM/MG]^hi cDC1s showing increased expression of MHC-I and MHC-II and enhanced capacity to activate OT-I cells (fig. S1, M and N). Similar results were also observed in the EO771 breast cancer model (Fig. 1C, and fig. S1, O and P) and an oncogene-driven hepatocellular carcinoma (HCC) model (29, 30) (Fig. 1D, and fig. S1, Q and R). Altogether, these results suggest that [TMRM/MG]^hi cDC1s are more immunogenic than [TMRM/MG]^lo cDC1s across multiple tumor types.

Besides tumors, [TMRM/MG]^hi and [TMRM/MG]^lo cDC1 subpopulations were also observed in the spleen and tdLN (including both resident and migratory cDC1s) of B16-OVA tumor-bearing mice (fig. S1, S and T). [TMRM/MG]^hi cDC1s from spleen or tdLN showed enhanced effects at inducing OT-I cell proliferation than [TMRM/MG]^lo counterparts (fig. S1, U and V), suggesting the conserved increase of immunogenicity of cDC1s with polarized mitochondria across tissues. Considering that cDC1s, especially the [TMRM/MG]^lo subpopulation, are rare populations in the tumor and tdLN, we used [TMRM/MG]^hi and [TMRM/MG]^lo subpopulations of splenic cDC1s for profiling. Transcriptome profiling followed by gene set enrichment analysis (GSEA) revealed that [TMRM/MG]^hi cDC1s showed increased gene signatures associated with antigen cross-presentation (fig. S1W) and OXPHOS (fig. S1X, and table S2). To further investigate the metabolic differences between [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s, we performed untargeted metabolomics analysis. Principal component analysis (PCA) revealed that these two subpopulations showed markedly distinct metabolic profiles (fig. S1Y). In particular, metabolite set enrichment analysis revealed that several mitochondrial metabolism-related pathways, including the malate–aspartate shuttle, glutamate metabolism, pyruvate metabolism and citric acid cycle, were upregulated in [TMRM/MG]^hi cDC1s compared to [TMRM/MG]^lo cells (fig. S1Z, and table S3), suggesting enhanced activity of several mitochondrial pathways, including those associated with energy production.

To directly assess the metabolic differences in intratumoral cDC1 subpopulations, we sorted [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s from B16-OVA tumors and performed Seahorse assay. We found [TMRM/MG]^hi cDC1s had elevated OCR and ATP production than [TMRM/MG]^lo counterparts (fig. S2A), suggesting increased OXPHOS. This observation was associated with increased mitochondrial volume [based on TOM20 staining via confocal imaging (31)] (fig. S2B). Moreover, electron microscopy (EM) analysis showed that mitochondria were more elongated in [TMRM/MG]^hi cDC1s than [TMRM/MG]^lo cells, with [TMRM/MG]^hi cDC1s displaying increased mitochondrial cristae number and length (fig. S2C). These data indicate that intratumoral [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s possess discrete mitochondrial morphologies and energy-generating capacities.

Finally, to determine whether two mitochondrial states of cDC1s are also present in other genetic mouse models of cancer or human cancer patients, we generated a cDC1 [TMRM/MG]^hi UP signature (see Methods, and table S4) and assessed its regulation in public single-cell RNA sequencing datasets. Plotting the activity score of this signature on uniform manifold approximation and projection (UMAP) plots of intratumoral cDC1s revealed [TMRM/MG]^hi and [TMRM/MG]^lo cDC1 subpopulations in mouse lung adenocarcinoma (24) (fig. S2D) and human tumors (32) (Fig. 1E). Thus, intratumoral cDC1s consist of [TMRM/MG]^hi and [TMRM/MG]^lo subpopulations in both mice and humans.

OPA1 orchestrates intratumoral cDC1 antitumor responses

Given the upregulation of mitochondrial respiration in intratumoral versus splenic cDC1s (fig. S1B), we next interrogated specific mitochondrial pathway involved. In our proteomics analysis, mitochondrial cristae formation pathway was the second highest ranked mitochondrial pathway enriched in intratumoral compared to splenic cDC1s (Fig. 1F, and table S5). Accordingly, EM imaging revealed that mitochondrial cristae number and length were increased in cDC1s from B16-OVA tumors (Fig. 1G). OPA1, a crucial driver of inner mitochondrial membrane fusion, cristae architecture and mitochondrial respiration (33), was upregulated in intratumoral compared to splenic cDC1s in the proteomics analysis (fig. S3A), which was validated by immunoblot analysis (fig. S3B). In contrast, the expression of DRP1 (promotes mitochondrial fission) or MFN1 and MFN2 (both induce outer mitochondrial membrane fusion) was largely comparable (fig. S3B), suggesting selective upregulation of OPA1 in intratumoral cDC1s.

Next, we generated mice with DC-specific deletion of OPA1 (Opa1^ΔDC) by crossing CD11c-Cre transgenic mice (34) with mice bearing floxed Opa1 alleles (35), which resulted in efficient deletion of OPA1 mRNA and protein selectively in cDCs (fig. S4, A and B). Multiple DC populations were unaltered in the lymphoid organs of Opa1^ΔDC compared to WT mice (fig. S4, C to H), suggesting that OPA1 deficiency did not alter their development or homeostasis. However, OPA1 deficiency decreased mitochondrial volume [based on TOM20 staining (31)] (fig. S5A) and caused mitochondrial fragmentation and cristae disorganization in intratumoral cDC1s (based on EM analysis) (Fig. 1H), which was associated with a decrease in mitochondrial membrane potential (fig. S5B). Finally, loss of OPA1 reduced and enhanced the proportions of [TMRM/MG]^hi and [TMRM/MG]^lo subpopulations among intratumoral cDC1s, respectively (Fig. 1I). Altogether, these data suggest the crucial role of OPA1 in mediating mitochondrial morphology and membrane potential in intratumoral cDC1s.

We next examined whether OPA1 orchestrates cDC1-mediated antitumor immunity using complementary tumor models. First, we challenged WT and Opa1^ΔDC mice with tumor cells, and found that deficiency of OPA1 in DCs markedly increased the growth and weight of B16-OVA (Fig. 1J, and fig. S5C), MC38 (fig. S5, D and E) and LLC (fig. S5, F and G) tumors. Additionally, HCC tumor burden (29, 30) was increased in Opa1^ΔDC compared to WT mice (Fig. 1, K and L), indicating a crucial role of OPA1 in DCs to restrict tumor growth. Of note, the percentage and number of intratumoral cDC1s or cDC2s was largely comparable between WT and Opa1^ΔDC B16-OVA tumor-bearing mice (fig. S5H), suggesting that OPA1 is dispensable for the overall cellularity of cDCs in the TME. In contrast, flow cytometry analysis of intratumoral T cells revealed reduced frequencies and numbers of CD8⁺ T cells in B16-OVA tumors from Opa1^ΔDC compared to WT mice (Fig. 1M). There was also markedly decreased number of effector-like (TCF1⁻TIM-3⁺), albeit not stem-like (TCF1⁺TIM-3⁻), cells (22, 36) among intratumoral CD8⁺ T cells from Opa1^ΔDC mice (fig. S5I). Moreover, the proportions and numbers of IFNγ⁺ and granzyme B (GZMB)⁺ cells among intratumoral CD8⁺ T cells were reduced in tumors from Opa1^ΔDC mice (Fig. 1N), suggesting impaired CD8⁺ T cell effector function.

To directly test the contribution of CD8⁺ T cells, we performed antibody-mediated CD8⁺ T cell depletion experiments (37). B16-OVA and MC38 tumor burdens were largely comparable in WT and Opa1^ΔDC mice upon CD8⁺ T cell depletion (fig. S5, J to M), suggesting that CD8⁺ T cells largely contribute to the improved antitumor responses in WT versus Opa1^ΔDC mice. Next, to test whether the defect in tumor control was due to loss of OPA1 specifically in cDC1s, we crossed XCR1-Cre mice, which express Cre recombinase selectively in cDC1s (38), with Opa1^fl/fl mice to generate mice with conditional deletion of OPA1 in cDC1s (Xcr1^cre/+Opa1^fl/fl). After challenge with B16-OVA, MC38 or LLC tumor cells, Xcr1^cre/+Opa1^fl/fl mice also exhibited increased tumor growth (Fig. 1, O to Q) and weight (fig. S5, N to P), suggesting that OPA1 supports cDC1 function in tumor control.

To test if OPA1 is important for DC function under conditions of chronic antigen stimulation, we infected WT and Opa1^DDC mice with the clone 13 strain of lymphocytic choriomeningitis virus (LCMV). In this chronic infection model, the accumulation, subset differentiation and cytokine production of virus-specific (GP33⁺) CD8⁺ T cells were largely comparable between WT and Opa1^ΔDC mice (fig. S5, Q to S). Collectively, these results highlight the crucial requirement of OPA1 in DCs for mediating antitumor, but not antiviral, CD8⁺ T cell responses under conditions of chronic antigen stimulation.

Given these results, we next investigated whether OPA1 contributes to the spatiotemporal regulation of intratumoral CD8⁺ T cell responses by cDC1s. Intratumoral cDC1s acquire tumor-associated antigens and then migrate to the tdLN, where they prime naïve tumor-reactive CD8⁺ T cell responses (6, 8). First, to examine whether OPA1 deletion alters the capacity of DCs to acquire tumor-associated antigens and undergo migration to the tdLN, we challenged WT and Opa1^ΔDC mice with B16F10-ZsGreen melanoma cells (8, 22). The proportions of ZsGreen⁺ cDCs in both tumor and tdLN were largely comparable between WT and Opa1^ΔDC mice (fig. S6, A and B), suggesting a dispensable role of OPA1 for DC antigen uptake in the TME or their migration to the tdLN. Second, to test if cDC1-dependent priming of naïve CD8⁺ T cells is altered in the tdLN, we used an established assay (6, 8, 22) (fig. S6C). We found that naïve OT-I cell priming in the tdLN was largely comparable in WT and Opa1^ΔDC mice bearing B16-OVA tumors (fig. S6D), suggesting a dispensable role of OPA1 in DCs to induce naïve CD8⁺ T cell priming in the tdLN. Third, as intratumoral cDC1s also contribute to the restimulation of effector CD8⁺ T cells in the TME (10), we assessed whether OPA1 in DCs orchestrates tumor-reactive effector CD8⁺ T cell responses in the TME using an established assay (10, 22) (fig. S6E). The accumulation of activated OT-I cells in the TME was reduced in Opa1^ΔDC mice (fig. S6F), associated with decreased proportions of effector-like (Ly108⁻TIM-3⁺ or TCF1⁻TIM-3⁺) OT-I cell subsets (fig. S6, G and H) and reduced expression of T-bet (fig. S6I). Accordingly, there were decreased frequencies of IFNg, IL-2 or TNFa-expressing OT-I cells in Opa1^ΔDC mice (fig. S6J), suggesting reduced effector function of intratumoral CD8⁺ T cells. These results together suggest that OPA1 is critical for DC-directed effector function of antigen-specific CD8⁺ T cells in the TME but is dispensable for tumor antigen uptake and migration to tdLN or priming naïve CD8⁺ T cells in the tdLN.

OPA1–NRF1 axis-mediated OXPHOS supports cDC1 immunogenic function

DCs regulate T cell-mediated immunity via processes related to antigen presentation, costimulatory ligand expression and cytokine production (i.e., signals 1, 2 and 3, respectively) (39). OPA1 deletion reduced the surface expression of SIINFEKL (i.e., OVA_257–264 peptide) bound to H-2K^b on intratumoral cDC1s but not cDC2s (Fig. 2A), whereas there was comparable expression of costimulatory ligands (CD40, CD80 and CD86) and proinflammatory cytokines (IL-12p40 and TNFa) (fig. S7, A and B). OPA1 deficiency-induced defects in antigen presentation capacity were mainly observed in intratumoral CCR7⁻ cDC1s (fig. S7C) that mediate restimulation of effector CD8⁺ T cells within the TME (11, 28), without alterations in intratumoral CCR7⁺ cDC1s (fig. S7C) or tdLN migratory cDC1s (fig. S7D), which are critical for naïve CD8⁺ T cell priming in the tdLN (6, 8). These results were consistent with the selective requirement of OPA1 in DCs to support CD8⁺ T cell responses in the TME but not in tdLN, as described above.

Fig. 2. — (A) B16-OVA tumor-bearing WT and *Opa1*^DDC mice were euthanized at day 14 after tumor inoculation (n = 6 each group). Quantification of the geometric mean fluorescence intensity (gMFI) of H-2K^b-SIINFEKL complex in intratumoral cDC1s and cDC2s. (B) CellTrace Violet^™ (CTV)-labeled naïve OT-I cells were adoptively transferred intravenously (i.v.) into WT or *Opa1*^ΔDC mice, followed by i.v. immunization with 20 μg OVA protein 24 hours later. After 3 days, the frequency of CTV^lo (proliferated) OT-I cells in spleens was quantified by flow cytometry analysis (n = 8 for WT; 10 for *Opa1*^ΔDC). (C) Sort-purified splenic WT or OPA1-deficient cDC1s were pulsed with heat-inactivated OVA-expressing *Listeria monocytogenes* (HKLM-OVA) in complete IMDM medium for 4 hours, followed by their irradiation and coculture with OT-I cells at a ratio of 1:10. Thymidine incorporation of OT-I cells was measured 72 hours later (n = 12 each group). (D) Sort-purified splenic WT or OPA1-deficient cDC1s were pulsed with B16F10 tumor cell lysate in complete IMDM medium for 4 hours, followed by their irradiation and coculture with pmel cells at a ratio of 1:10. Thymidine incorporation of pmel cells was measured 72 hours later (n = 4 for cDC1s; 6 for cDC2s). (E) ATAC-seq analysis of sort-purified splenic cDC1s from WT and *Opa1*^ΔDC mice (n = 4 per genotype). Transcription factor footprinting analysis was performed by comparing *Opa1*^ΔDC versus WT splenic cDC1s, and the transcription factors with more 1,000 binding sites were selected and then ranked by their Z-scores. Blue dots and red dots indicate transcription factors predicted to have decreased and increased activity, respectively, and NRF1 is identified as the most downregulated transcription factor. (F) Immunoblot analysis of NRF1 expression in cDC1s from WT and *Opa1*^ΔDC mice. ACTB was used as loading control (left). Quantification of the relative expression of NRF1 in cDC1s from WT and *Opa1*^ΔDC mice (right). (G) Venn diagram showing the significant overlap between downregulated genes in OPA1-deficient cDC1s versus WT cDC1s compared to putative NRF1 target genes (left). Functional enrichment analysis of the 188 overlapped genes, with the top downregulated Hallmark OXPHOS pathway labeled in blue color (right). (H) Quantification of the basal oxygen consumption rate (OCR) in splenic cDC1s expressing sgNTC or sgNrf1, which were sort-purified from “retrogenic mice” bearing B16-Flt3L tumors (for DC expansion) at day 9 after tumor inoculation, as measured by Seahorse metabolic flux assay; see also fig. S7G (n = 9 for sgNTC; 6 for sgNrf1). (I) Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed (left) or OVA_257–264 peptide-pulsed (right) splenic cDC1s expressing sgNTC or sgNrf1 (isolated from the B16-Flt3L tumor-bearing “retrogenic” mice for DC expansion, see also fig. S7G) for 72 hours (n = 8 for sgNTC for both OVA and OVA_257–264 peptide; 5 for sgNrf1 in response to OVA; 4 for sgNrf1 in response to OVA_257–264 peptide). (J) “Retrogenic” mice bearing WT or OPA1-deficient cDC1s that overexpress empty vector or NRF1 protein were generated and injected with B16-Flt3L tumors (for DC expansion), followed by sort purification of cDC1s 9 days later (see also fig. S7G). WT or OPA1-deficient cDC1s overexpressing empty vector or NRF1 were treated for 2 hours (during OVA protein pulse) with or without 500 nM rotenone (ETC complex I inhibitor), 1 mM 3-NPA (ETC complex II inhibitor), 0.5 μM antimycin A (ETC complex III inhibitor) or 1 μM oligomycin (ETC complex V inhibitor). OT-I cell proliferation was assessed after 72 hours (n = 3 for OPA1-deficient cDC1 with antimycin A treatment; 4 for remaining groups). (K) Heatmap showing the expression of putative NRF1 targets within in the Hallmark OXPHOS pathway that were downregulated in OPA1-deficient cDC1s compared to WT cells [as shown in (G); ranked by row z-score] (n = 3 for WT, 4 for *Opa1*^ΔDC). (L) Immunoblot analysis of the expression of NDUFB8 (ETC complex I; CI), SDHB (ETC complex II; CII), UQCRC2 (ETC complex III; CIII), MTCO1 (ETC complex IV; CIV) and ATP5A (ETC complex V; CV) in splenic cDC1s from WT and *Opa1*^ΔDC mice. ACTB was used as loading control. Densitometric quantification of NDUFB8 was performed and normalized to ACTB expression. The numbers show the relative NDUFB8 values compared to WT cDC1s. (M) Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed (left) or OVA_257–264 peptide-pulsed (right) splenic cDC1s expressing sgNTC or sgNdufaf1 (ETC complex I assembly factor; CI) (isolated from B16-Flt3L tumor-bearing “retrogenic” mice for DC expansion, see also fig. S7G) for 72 hours (n = 8 each group). (N) Thymidine incorporation of OT-I cells cultured with OVA protein-pulsed (left) or OVA_257–264 peptide-pulsed (right) splenic cDC1s expressing sgNTC, sgSdhb (component of ETC complex II; CII), sgUqcrq (component of ETC complex III; CIII) or sgAtpaf2 (ETC complex V assembly factor; CV) (isolated from B16-Flt3L tumor-bearing “retrogenic mice” for DC expansion, see also fig. S7G) for 72 hours (n = 5 for cDC1 with sgUqcrq or sgAtpaf2 cultured with OVA protein; 6 for all other groups). (O) Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed splenic cDC1s from WT and *Opa1*^ΔDC mice that were pretreated for 2 hours (during antigen pulse) with or without 500 nM rotenone, 1 mM 3-NPA, 0.5 μM antimycin A or 1 μM oligomycin. OT-I cell proliferation was assessed after 72 hours (n = 4 per group). (P) Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed splenic WT or OPA1-deficient cDC1s expressing sgNTC or sgNdufaf1, which were sort-purified from “retrogenic mice” bearing B16-Flt3L tumors (for DC expansion) at day 9 after tumor inoculation (as described in fig. S7G) (n = 4 for WT cDC1s expressing sgNdufaf1; 5 for WT cDC1s expressing sgNTC or OPA1-deficient cDC1s expressing sgNdufaf1; 6 for OPA1-deficient cDC1s expressing sgNTC). (Q) Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed splenic cDC1s expressing sgNTC or sgNrf1 (isolated from B16-Flt3L tumor-bearing “retrogenic mice” for DC expansion, see also fig. S7G) that were pretreated for 2 hours (during antigen pulse) with or without 500 nM rotenone, 1 mM 3-NPA, 0.5 μM antimycin A or 1 μM oligomycin. OT-I cell proliferation was assessed after 72 hours (n = 5 for sgNTC, 4 for sgNrf1). Data are shown as mean ± s.e.m. in (A to D, F, H to J, and M to Q). Two-way ANOVA in (A, C, D, J, O to Q), two-tailed unpaired Student’s t-test in (B, F, H, I and M) or one-way ANOVA in (N). Data are representative of two (C, D, H, J, and L, N to Q) or at least three (A, B, F, I and M) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant. Numbers in plots represent percentages of cell in gates (B).

To further examine the role of OPA1 in DC-mediated CD8⁺ T cell response, we adoptively transferred naïve OT-I cells into WT and Opa1^ΔDC mice followed by intravenous (i.v.) immunization with OVA protein (22, 31). Compared to WT mice, Opa1^ΔDC mice displayed impaired OT-I cell proliferation in the spleen (Fig. 2B). In contrast, OT-I cell proliferation in the LNs was unaltered upon i.v. OVA challenge (fig. S7E), in line with the observations from the tdLN (fig. S6D). Furthermore, OPA1-deficient cDC1s exhibited defects in mediating cell-associated antigen presentation (22, 31), as evidenced by impaired heat-killed OVA-expressing Listeria monocytogenes (HKLM-OVA)-induced OT-I cell proliferation (Fig. 2C). Likewise, OPA1-deficient cDC1s had reduced ability to induce pmel cell proliferation in response to B16F10 tumor-derived antigens (40) (Fig. 2D), indicating defective tumor antigen-elicited CD8⁺ T cell activation. Collectively, these data reveal an important role for OPA1 in mediating cDC1-dependent antigen presentation and activation of CD8⁺ T cells.

To uncover the mechanisms underlying OPA1-mediated effects, we performed assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq) analysis of WT and Opa1^ΔDC splenic cDC1s and found distinct chromatin accessibility profiles (fig. S7F). Transcription factor footprinting analysis (41) revealed reduced transcription factor nuclear respiratory factor 1 (NRF1) activity in OPA1-deficient cells (Fig. 2E, and table S6). Accordingly, NRF1 expression was substantially reduced in OPA1-deficient cDC1s (Fig. 2F). Furthermore, in transcriptome profiling, genes downregulated upon OPA1 deletion significantly overlapped with putative NRF1 target genes (42) (Fig. 2G, left). Functional enrichment analysis of these overlapped genes (188 genes in total, table S7) identified Hallmark OXPHOS as the top affected pathway (Fig. 2G, right, and table S8), suggesting the possible interplay between OPA1 and NRF1 in orchestrating OXPHOS in cDC1s.

To establish the functional relationship between OPA1 and NRF1 in promoting cDC1 OXPHOS and immunogenic function, we generated “retrogenic” mice to obtain cDC1s with NRF1 deletion by transducing Lineage⁻Sca-1⁺c-Kit⁺ (LSK) cells from Cas9-expressing mice with sgNrf1 RNA, followed by immune reconstitution (43) (fig. S7G). We found that NRF1-deficient cDC1s showed comparable expression of Opa1 as WT cells (fig. S7H). Together with the aforementioned observations that OPA1 deficiency impaired NRF1 activity and expression (Fig. 2, E and F), these results suggest that OPA1 acts upstream of NRF1. Similar to OPA1 deletion described above, NRF1-deficient cDC1s showed decreased mitochondrial OXPHOS (Fig. 2H) and impaired function in priming OT-I cell proliferation (Fig. 2I). More importantly, NRF1 overexpression partially rectified the defective OCR (fig. S7I) and immunogenic function (Fig. 2J; left, vehicle group) of OPA1-deficient cDC1s. Therefore, reduced NRF1 activity contributes to impairments in mitochondrial OXPHOS and CD8⁺ T cell priming function in OPA1-deficient cDC1s.

To explore the mechanism by which NRF1 contributes to OPA1-mediated OXPHOS, we analyzed the expression of genes co-regulated by OPA1 and NRF1 [putative NRF1 targets (42) downregulated in OPA1-deficient cDC1s], focusing on those in Hallmark OXPHOS pathway (Fig. 2G). This analysis revealed co-regulated genes involved in the electron transport chain (ETC) (e.g., Atp5k), citric acid cycle (e.g., Ogdh) and mitochondrial structure (e.g., Immt) (Fig. 2K). In addition, OPA1 deficiency reduced the expression of NDUFB8 (ETC complex I subunit) (Fig. 2L), which reflects decreased integrity of complex I (44, 45). OXPHOS is mediated by mitochondrial ETC complexes (46). Therefore, to functionally dissect the role of OXPHOS in cDC1-mediated T cell priming, we pulsed cDC1s with OVA or OVA_257–264 in the presence of different compounds targeting ETC complexes I, II, III and V [rotenone, 3-nitropropionic acid (3-NPA), antimycin A and oligomycin, respectively]. These inhibitors substantially reduced the ability of cDC1s to induce OT-I cell priming without affecting cell survival, with 3-NPA showing more modest effects (fig. S8, A and B). Next, we generated “retrogenic” mice to genetically target NDUFAF1 (essential for complex I assembly), SDHB (component of complex II), UQCRQ (component of complex III) or ATPAF2 (essential for complex V assembly) (47). sgNdufaf1-expressing cDC1s had defective capacity to mediate OT-I cell proliferation (Fig. 2M), in line with the defects observed with OPA1- or NRF1-deficient cDC1s described above. Deletion of UQCRQ or ATPAF2, and to a lesser extent SDHB, in cDC1s also resulted in priming defects (Fig. 2N). Importantly, ETC inhibitors or NDUFAF1 deletion mitigated the inhibitory effects of OPA1 deletion on cDC1-mediated OT-I cell priming (Fig. 2, O and P). Furthermore, ETC blockade eliminated the inhibitory effect of NRF1 deficiency on cDC1 priming function (Fig. 2Q). Moreover, NRF1 overexpression did not alter the immunogenicity of OPA1-deficient cDC1s treated with ETC inhibitors (Fig. 2J). Altogether, these data reveal a OPA1–NRF1 axis that supports downstream OXPHOS to mediate cDC1 function in priming CD8⁺ T cells.

OPA1 and NRF1 support cDC1 bioenergetics and suppress autophagy/lysosome-mediated MHC-I and antigen degradation

As OXPHOS is important for generating ATP, we analyzed cellular ATP (together with ADP and AMP) abundance. OPA1 deficiency profoundly diminished ATP levels and enhanced the ratios of ADP/ATP and AMP/ATP in cDC1s (Fig. 3A). Reduced ATP generation can activate AMPK signaling and autophagy (48). Indeed, OPA1-deficient cDC1s had upregulated AMPK (based on phosphorylation of AMPK Thr172 and ULK1 Ser555) and autophagy (based on the ratio of LC3-II to LC3-I and decreased P62 expression) activities (Fig. 3B). Similarly, AMPK activation and autophagy were upregulated in NRF1-deficient cDC1s (Fig. 3C) and cDC1s treated with ETC inhibitors to block OXPHOS (fig. S8C), suggesting the inhibitory effects of OPA1, NRF1 and OXPHOS on AMPK and autophagy pathways in cDC1s.

Fig. 3. — (A) ATP abundance (upper), and the ratios of ADP to ATP (middle) and AMP to ATP (lower) were measured in splenic cDC1s from WT and *Opa1*^ΔDC mice (n = 3 per group). (B and C) Immunoblot analysis of the indicated proteins in splenic cDC1s that were isolated from WT and *Opa1*^ΔDC mice (B) or in splenic cDC1s expressing sgNTC or sgNrf1 isolated from “retrogenic” mice bearing B16-Flt3L tumors (for DC expansion) at day 9 after tumor inoculation; see also fig. S7G (C). ACTB was used as a loading control. Densiometric quantification of p-AMPK, p-ULK1, P62, or LC3-II was performed and normalized to total protein (for p-AMPK and p-ULK1), ACTB (for P62) or LC3-I (for LC3-II) expression. Numbers show the relative values of each of protein compared to WT cDC1s (B) or cDC1s expressing sgNTC (C). (D) Splenic cDC1s from WT and *Opa1*^ΔDC mice were treated with or without chloroquine (CQ) overnight *in vitro* (n = 4 per group). Quantification of the geometric mean fluorescence intensity (gMFI) of surface MHC-I on WT and OPA1-deficient cDC1s without CQ treatment (left). Quantification of the ratio of MHC-I gMFI in CQ-treated (+CQ) compared to untreated (−CQ) splenic cDC1s from WT and *Opa1*^ΔDC mice (right). (E) Sort-purified splenic cDC1s from WT and *Opa1*^ΔDC mice were pulsed with DQ-OVA for the indicated times with or without CQ, followed by flow cytometry analysis. Quantification of the percentage of FITC⁺ cells (indicative of DQ-OVA degradation) among cDC1s is shown (n = 3 per group). (F) “Retrogenic” mice were generated by transducing LSK cells from Cas9-expressing WT and *Opa1*^ΔDC mice with sgNTC or sgAtg5, followed by their inoculation with B16-Flt3L tumors for DC expansion (see fig. S7G). cDC1s were isolated from these mice 9 days later. Quantification of surface MHC-I gMFI on splenic WT or OPA1-deficient cDC1s expressing sgNTC or sgAtg5 isolated from the above described “retrogenic” mice (n = 4 for WT and OPA1-deficient cDC1 with sgNTC groups; 3 for WT+sgAtg5 and *Opa1*^ΔDC+sgAtg5 groups). (G) Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed WT or OPA1-deficient cDC1s expressing sgNTC or sgAtg5 (generated as described in F) for 72 hours (n = 3 for *Opa1*^ΔDC+sgAtg5 group; 6 for all other groups). (H) NAD⁺/NADH ratio was measured in splenic cDC1s from WT and *Opa1*^ΔDC mice (n = 6 per group). (I) “Retrogenic” mice bearing WT or OPA1-deficient cDC1s that overexpress empty vector or mito-LbNOX protein were generated and injected with B16-Flt3L tumors for DC expansion, followed by sort purification of splenic cDC1s 9 days later (see also fig. S7G). Thymidine incorporation of OT-I cells cocultured with OVA protein-pulsed WT or OPA1-deficient cDC1s overexpressing empty vector or mito-LbNOX for 72 hours (n = 3 for WT+mito-LbNOX and *Opa1*^ΔDC+mito-LbNOX; 6 for WT+Vector and *Opa1*^ΔDC+Vector). (J) Seahorse metabolic flux analysis of oxygen consumption rate (OCR) of splenic cDC1s from WT, *Opa1*^ΔDC, *Dnm1l*^ΔDC or *Opa1*/*Dnm1l*^ΔDC mice. Cells were treated with the indicated mitochondrial inhibitors (Oligo, oligomycin; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; and Rot, rotenone; upper). Basal OCR (before Oligo treatment; middle) and maximal OCR (after FCCP treatment; lower) of splenic cDC1s isolated from WT, *Opa1*^ΔDC, *Dnm1l*^ΔDC or *Opa1*/*Dnm1l*^ΔDC mice (n = 6 for WT and *Opa1*^ΔDC groups; 9 for *Dnm1l*^ΔDC and *Opa1*/*Dnm1l*^ΔDC groups). (K) Immunoblot analysis of NRF1 expression in cDC1s from WT, *Opa1*^ΔDC, *Dnm1l*^ΔDC or *Opa1*/*Dnm1l*^ΔDC mice. ACTB was used as loading control (upper). Quantification of the relative expression of NRF1 in cDC1s from WT, *Opa1*^ΔDC, *Dnm1l*^ΔDC or *Opa1*/*Dnm1l*^ΔDC mice (lower). (L) B16-OVA tumor growth curves in WT (n = 6), *Opa1*^ΔDC (n = 4), *Dnm1l*^ΔDC (n = 6) and *Opa1*/*Dnm1l*^ΔDC (n = 5) mice. (M) MC38 tumor growth curves in WT (n = 11), *Opa1*^ΔDC (n = 8), *Dnm1l*^ΔDC (n = 8) and *Opa1*/*Dnm1l*^ΔDC (n = 5) mice. Data are shown as mean ± s.e.m. in (A, D to M). Two-tailed unpaired Student’s t-test in (A, D and H), two-way ANOVA in (E, L and M) or one-way ANOVA in (F, G, I, J and K). Data are representative of two (A to D, F, G, L and M) or at least three (E, I, J and K), or pooled from two (H) independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant.

As autophagy facilitates MHC-I degradation (49, 50), we examined the roles of OPA1 and mitochondrial pathways in this process. OPA1-deficient cDC1s showed decreased surface expression of MHC-I (Fig. 3D, left), which was associated with decreased MHC-I expression on [TMRM/MG]^hi, albeit not [TMRM/MG]^lo, cDC1s (fig. S8D). To determine whether increased autophagy/lysosome activity contributes to OPA1 deficiency-associated MHC-I downregulation, we treated WT and OPA1-deficient cDC1s with chloroquine (CQ) to block autophagy/lysosome-mediated degradation, followed by analysis of MHC-I expression. CQ treatment resulted in a greater accumulation of surface MHC-I on OPA1-deficient cDC1s compared to WT counterparts [based on the gMFI ratio of CQ-treated (CQ+) versus CQ-untreated (CQ–) cells] (Fig. 3D, right), suggesting that autophagy/lysosome-mediated effects contribute to MHC-I downregulation in OPA1-deficient cells. NRF1-deficient or ETC inhibitor-treated cDC1s also exhibited decreased surface MHC-I expression (fig. S8, E and F), thereby phenocopying OPA1 deficiency-associated defects.

Autophagy targets cargo to lysosomes for turnover (51), with enhanced lysosomal delivery or degradation being associated with increased antigen degradation (22, 52–55). OPA1 deficiency did not affect lysosomal content (based on gMFI of Lysotracker staining) (22) (fig. S8G). To assess antigen degradation, we used DQ-OVA, a self-quenched dye that emits green fluorescence after OVA is degraded (22, 52, 56) (i.e., cells will become FITC⁺). We found that the frequency of FITC⁺ cDC1s was increased in OPA1-deficient cDC1s, which was abolished by CQ treatment (Fig. 3E). In contrast, antigen uptake was not affected by OPA1 deficiency, as evidenced by the comparable signal of OVA-AlexaFluor 488 [a dye insensitive to OVA degradation (57)] (fig. S8H). These data suggest enhanced lysosomal degradation of OVA in OPA1-deficient cDC1s, likely through increased targeting of OVA to lysosome via autophagy (51). Of note, the transcript levels for multiple components in MHC-I-restricted antigen processing machinery (58), the immunoproteasome (59), and MHC-I molecules were largely comparable between WT and OPA1-deficient cDC1s (fig. S8I), suggesting the dispensable role of OPA1 for mRNA expression of antigen processing/presentation machinery.

Finally, to functionally define whether dysregulated autophagy/lysosome activity in OPA1-deficient cDC1s contributes to their priming capacity, we generated “retrogenic” mice to co-delete autophagy-associated gene Atg5 in OPA1-deficient cDC1s, by transducing LSK cells from Cas9-expressing Opa1^DDC mice with sgAtg5, followed by immune reconstitution (fig. S7G). Compared to OPA1-deficient cDC1s, OPA1 and ATG5 co-deficient cDC1s displayed higher surface expression of MHC-I (Fig. 3F), associated with their improved capacity to induce OT-I cell proliferation (Fig. 3G). Likewise, CQ treatment substantially enhanced the capacity of OPA1- or NRF1-defieicnt cDC1s to prime OT-I cell proliferation (fig. S8, J and K). CQ treatment also partially reversed the impaired priming function of cDC1s resulting from ETC inhibitor treatment or NDUFAF1 deletion (fig. S8, L and M). Therefore, dysregulated autophagy/lysosome activity contributes to the priming defects of cDC1s induced by OPA1 or NRF1 deficiency or ETC blockade.

OPA1-mediated NAD⁺/NADH balance partially contributes to cDC1 functional fitness

Beyond ATP synthesis, mitochondrial ETC flow also maintains the NAD⁺/NADH ratio (46). Coincident to the downregulation of ETC complex I (NADH dehydrogenase) (Fig. 2L), OPA1-deficient cDC1s also showed decreased NAD⁺/NADH ratio (Fig. 3H). To functionally test whether disrupted NAD⁺/NADH ratio contributes to OPA1 deficiency-induced cDC1 defects, we expressed mito-LbNOX [a mitochondria-localized NADH oxidase from Lactobacillus brevis (60)] in WT and OPA1-deficient cDC1s. This enzyme catalyzes the conversion of NADH to NAD⁺, but it neither pumps protons from the mitochondrial matrix to the intermembrane space nor facilitates electron flux. Thus, mito-LbNOX overexpression is permissive to dissect the NAD⁺ regenerative function of complex I from its role in OXPHOS and energy production (60). Overexpression of mito-LbNOX in OPA1-deficient cDC1s partially restored their ability to prime OT-I cell proliferation (Fig. 3I), suggesting that imbalanced NAD⁺/NADH ratio also contributes to cDC1 dysfunction by OPA1 deficiency. Altogether, OPA1 deficiency not only compromises energy production (via OXPHOS) but also disrupts the NAD⁺/NADH ratio, and these two effects may coordinately contribute to cDC1 dysfunction in the absence of OPA1.

Balanced mitochondrial fusion and fission contribute to cDC1 antitumor immunity

Balanced mitochondrial fusion and fission are important for maintaining proper mitochondrial morphology and bioenergetics (33). Given that intratumoral cDC1s lacking OPA1 accumulated fragmented mitochondria (Fig. 1H) and had reduced mitochondrial volume (fig. S5A), we next tested whether unopposed mitochondrial fission contributes to the defects of OPA1-deficient cDC1s, by generating mice with DC-specific deletion of mitochondrial fission protein DRP1 [by breeding CD11c-Cre mice with Dnm1l^fl/fl mice (61)] (Dnm1l^DDC mice) or DC-specific co-deletion of OPA1 and DRP1 (Opa1/Dnm1l^DDC mice). Co-deletion of DRP1 substantially increased mitochondrial volume in OPA1-deficient cDC1s (fig. S9A), suggesting that DRP1-mediated mitochondrial fission contributes to OPA1 deficiency-associated effects on mitochondrial morphology. Whereas DRP1-deficient cDC1s had modest defects in OCR, co-deletion of DRP1 considerably rectified the defect in OPA1-deficient cDC1s (Fig. 3J). In line with these observations, co-deletion of DRP1 largely restored the reduced NRF1 (Fig. 3K) and NDUFB8 (fig. S9B) levels in OPA1-deficient cDC1s. Furthermore, the increased AMPK and autophagy activities (fig. S9C), decreased surface MHC-I expression (fig. S9D) and imbalanced NAD⁺/NADH ratio (fig. S9E) in OPA1-deficient cDC1s were also largely restored by DRP1 co-deletion. Interestingly, whereas OPA1- or DRP1-deficient cDC1s were defective in priming OT-I cell proliferation both in vitro and in vivo, cDC1s lacking both OPA1 and DRP1 had largely normal priming ability (fig. S9, F and G). More importantly, Opa1/Dnm1l^ΔDC mice showed enhanced tumor control in both B16-OVA and MC38 tumor models compared to Opa1^ΔDC mice (Fig. 3, L and M), associated with increased accumulation of intratumoral CD8⁺ T cells (fig. S9H) and generation of effector-like (TCF1⁻TIM-3⁺) (fig. S9I) and IFNg- or GZMB-expressing (fig. S9J) cells. Similarly, B16-OVA and MC38 tumor growth was reduced in Xcr1^cre/+Opa1^fl/flDnm1l^fl/fl mice (with deletion of both OPA1 and DRP1 selectively in cDC1s) compared to Xcr1^cre/+Opa1^fl/fl mice (fig. S9, K and L). Collectively, these results indicate that OPA1- and DRP1-mediated balance of mitochondrial fusion and fission orchestrates cDC1 metabolic fitness, priming function and antitumor effects.

Intratumoral cDC1s decrease mitochondrial membrane potential and volume during tumor progression, associated with downregulation of OPA1 and NRF1 levels

It is appreciated that DC function becomes impaired in the TME (62, 63). We found that cDC1s isolated from B16-OVA tumors at a late timepoint (day 21) had reduced CD8⁺ T cell priming function compared to those isolated at an early timepoint (day 7) (Fig. 4A), whereas cDC1s from the spleen or tdLN had similar priming capacities at both timepoints (fig. S10A). Interestingly, this observed dysfunction of intratumoral cDC1s during tumor progression coincided with a progressive reduction in the proportion and number of [TMRM/MG]^hi cells [which represent cells with high OXPHOS capacity, mitochondrial volume and cristae number and length (fig. S2, A to C)] (Fig. 4B). Additionally, the TMRM/MG gMFI ratio in bulk cDC1s or in the [TMRM/MG]^hi cDC1 subpopulation was considerably decreased at day 14 or 21 compared to day 7 after tumor inoculation (fig. S10, B and C). Thus, both the quantity (i.e., proportion and number) and mitochondrial quality (as revealed by mitochondrial membrane potential) of [TMRM/MG]^hi cDC1s were decreased during tumor progression. Furthermore, MHC-I, CD40 and CD80 levels were reduced in intratumoral [TMRM/MG]^hi cDC1s at day 14 or 21 compared to day 7 (Fig. 4C). These changes coincided with declined effector function of tumor antigen-specific CD8⁺ T cells (fig. S10D). These results together suggest that intratumoral cDC1s from B16-OVA tumor model undergo progressive decline of mitochondrial membrane potential during tumorigenesis, associated with overall decreased functional fitness.

Fig. 4. — (A) Sort-purified intratumoral cDC1s from B16-OVA tumors at days 7 (n = 5) and 21 (n = 6) after tumor inoculation were pulsed with OVA protein in complete IMDM medium for 2 hours, followed by their irradiation and coculture with OT-I cells at a ratio of 1:10 for 72 hours. OT-I cell proliferation was measured by thymidine incorporation (n = 5 per group). (B) Quantification of the percentage (left) and number (normalized to tumor weight) (right) of the indicated populations among intratumoral cDC1s from B16-OVA tumors at the indicated timepoints, based upon TMRM and MG co-staining (n = 10 per group). (C) Quantification of the geometric mean fluorescence intensities (gMFIs) of MHC-I, MHC-II, CD40, CD80 or CD86 on [TMRM/MG]^hi cDC1s from B16-OVA at the indicated timepoints after tumor inoculation (n = 9 at day 14; 10 at day 7 or 21). (D) Violin plots show activity score of cDC1 [TMRM/MG]^hi UP signature (generated in-house from genes upregulated in [TMRM/MG]^hi cDC1s compared to [TMRM/MG]^lo cDC1s; see Methods) in cDC1s from human kidney cancer (left) (GSE154763) or human non-small lung cancer (NSCLC; right) (GSE139555) at different clinical [i.e., TNM (tumor, node, metastasis)] stages. Within each box, horizontal lines denote median values; boxes extend from the 25^th to the 75^th percentile of each group’s distribution of values. (E and F) ATAC-seq analysis of sort-purified cDC1s from B16-OVA tumors isolated on days 7 and 21 after tumor inoculation (n = 4 per group). Transcription factor footprinting (E) and motif enrichment (F) analyses were performed by comparing day 21 versus day 7, and the transcription factors were ranked by their Z-scores (E) or odds ratios (F). The transcription factors with > 1,000 binding sites were selected for ranking (E). Blue dots and red dots indicate transcription factors predicted to have decreased and increased activity at day 21, respectively (E and F), and NRF1 was identified as the most downregulated transcription factor in the footprinting analysis (E). (G) GSEA enrichment plots showing decreased cDC1 OPA1-activated signature (generated in-house from the downregulated genes in OPA1-deficient versus WT cDC1s; see Methods) (left) and increased OPA1-suppressed signature (generated in-house from upregulated genes in OPA1-deficient versus WT cDC1s; see Methods) (right) in intratumoral cDC1s from B16-OVA tumors at day 21 versus day 7 after tumor inoculation. FDR, false discovery rate; NES, normalized enrichment score. (H) Splenic and intratumoral cDC1s were isolated from B16-OVA tumor-bearing mice at days 7 and 21 after tumor inoculation, followed by immunoblot analysis of OPA1 and NRF1. ACTB was used as loading control. Densitometric quantification of OPA1 or NRF1 was performed and normalized to ACTB expression. Numbers show the relative expression of OPA1 or NRF1 in indicated cDC1 population relative to splenic cDC1s from B16-OVA-tumor-bearing mice at day 7. (I) Confocal imaging analysis of TOM20 in intratumoral cDC1s derived from B16-OVA tumors at days 7 and 21 after tumor inoculation. Scale bars: 10 μm (left). Quantification of volume of TOM20 per mitochondrion (right) (n ≥ 1,000 mitochondria per group). (J to L) Schematic of cDC1 therapy model (J). WT mice were inoculated with 1 × 10⁶ B16-OVA (K) or MC38 (L) tumor cells. Splenic total cDC1s, [TMRM/MG]^hi cDC1s or [TMRM/MG]^lo cDC1s were sorted from B16-Flt3L tumor-bearing mice (for DC expansion) and then were cultured with 100 μg/ml OVA protein and 20 μg/ml poly I:C (K) or MC38 cell lysate and 20 μg/ml poly I:C (L) in complete IMDM medium for 2 hours, followed by washing and subcutaneous transfer to tumor-bearing mice at day 5 after tumor inoculation. PBS was injected into non-transfer control mice (n = 5 per group). Tumor growth was monitored. (M and N) Splenic [TMRM/MG]^hi cDC1s or [TMRM/MG]^lo cDC1s were sorted and pulsed with OVA protein and poly I:C as in (K). B16-OVA tumor-bearing mice received adoptive transfer of the indicated cDC1 subpopulations (or PBS) on day 6 after tumor inoculation. Anti-PD-L1 (aPD-L1; 200 mg) (or isotype control antibody) was intraperitoneally injected at days 7, 10 and 13 after tumor inoculation (n = 5 per group). Tumor growth was monitored (M). Sixty days later, non-immunized (naïve) WT mice (n = 4) and tumor-free mice (n = 5) from the [TMRM/MG]^hi cDC1s plus aPD-L1 treatment group were rechallenged with B16-OVA. Tumor cell growth was monitored (N). (O and P) Splenic [TMRM/MG]^hi cDC1s or [TMRM/MG]^lo cDC1s were sorted and pulsed with OVA protein and poly I:C as in (K) or MC38 cell lysate and poly I:C as in (L). WT mice with established B16-OVA (O) or MC38 (P) tumors received adoptive transfer of the indicated cDC1 subpopulations (or PBS) at day 6 after tumor inoculation. aCTLA-4 (200 mg) (O) or aPD-1 (200 mg) (P) (or isotype control antibody) was intraperitoneally injected at days 7, 10 and 13 after tumor inoculation (n = 5 for each group). Tumor growth was monitored. Data are shown as mean ± s.e.m. in (A to C, I, and K to P). Two-tailed unpaired Student’s t-test in (A), two-way ANOVA in (B and K to P), one-way ANOVA in (C), two-tailed Wilcoxon rank sum test in (D), or Mann-Whitney test in (I). Data are representative of one (I), two (A to C, H, and M to P) or at least three (K and L). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS, not significant.

We next assessed the temporal regulation of mitochondrial features in the LLC, EO771 and oncogene-driven HCC tumor models. In LLC tumors, the proportion and number of [TMRM/MG]^hi cDC1s were progressively reduced during tumor progression (fig. S10E). In addition, the overall mitochondrial membrane potential of bulk cDC1s (fig. S10F) or [TMRM/MG]^hi cDC1s (but not [TMRM/MG]^lo cDC1s) (fig. S10G) was substantially decreased during tumor progression, as evidenced by downregulation of the TMRM/MG gMFI ratio. These changes were associated with decreased expression of MHC-I and CD40 in [TMRM/MG]^hi cDC1s from LLC tumors during tumor progression (fig. S10H). Second, in EO771 tumors, there were decreased abundance of [TMRM/MG]^hi cDC1s (fig. S10I) and downregulation of mitochondrial membrane potential in bulk cDC1s (fig. S10J) or [TMRM/MG]^hi cDC1s (fig. S10K) during tumor progression. Additionally, decreased levels of MHC-I and MHC-II were observed in [TMRM/MG]^hi cDC1s during EO771 tumor progression (fig. S10L). Third, in the oncogene-driven HCC model, whereas the percentages and numbers of [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s were not obviously altered during tumor progression (fig. S10M), the mitochondrial membrane potential of bulk cDC1s (fig. S10N) or [TMRM/MG]^hi cDC1s (fig. S10O) was decreased. Furthermore, [TMRM/MG]^hi cDC1s from the liver of HCC tumor-bearing mice had reduced expression of MHC-I, MHC-II and CD40 during tumor progression (fig. S10P), suggesting decreased immunogenicity. Finally, to explore whether human cDC1s undergo similar mitochondrial alterations during tumor progression, we examined the activity of the cDC1 [TMRM/MG]^hi UP signature in a publicly available dataset of cDC1s from human kidney cancer at different clinical stages (stages I, II, and III) (25). This signature was significantly reduced at stage III compared to stage I (Fig. 4D). Similar results were observed in human non-small cell lung carcinoma (NSCLC) at stage II versus stage I (64) (Fig. 4D). Together, these data reveal dynamic rewiring of intratumoral cDC1s towards a state characterized by reduced mitochondrial membrane potential, associated with decreased immunogenicity, during tumor progression.

To test whether the TME induces alterations in cDC1 mitochondrial features, we sorted bulk splenic cDC1s (from non-tumor-bearing mice) and performed intratumoral injection into B16-OVA tumor-bearing mice at day 7 or 21 after tumor inoculation, followed by flow cytometry analysis 24 hours later (fig. S11A). Compared to the cDC1s transferred into day 7 tumors, donor-derived cDC1s showed a decreased proportion and number of [TMRM/MG]^hi cDC1s when injected into day 21 tumors (fig. S11B). Moreover, in donor-derived cDC1s, mitochondrial membrane potential and mass, as well as the TMRM/MG gMFI ratio, were substantially decreased in cDC1s directly transferred into day 21 compared to day 7 tumors (fig. S11C). The reductions in TMRM, MG and TMRM/MG gMFI ratio were also observed when comparing donor-derived [TMRM/MG]^hi cDC1s (but not [TMRM/MG]^lo cDC1s except for reduced MG staining) at day 21 versus day 7 (fig. S11D). Intercellular transfer of tumor cell-derived mitochondria to intratumoral CD8⁺ T cells promotes their metabolic dysfunction (65). Therefore, we generated B16-OVA-mito-DsRed tumor cells (65) (fig. S11E) to trace mitochondrial transfer from tumor cells to immune cells. At day 21 after tumor inoculation, intratumoral CD8⁺ T cells acquired mitochondria from B16-OVA-mito-DsRed tumor cells in vivo (fig. S11F), in line with previous findings (65). Interestingly, a fraction (~3%) of intratumoral cDC1s was DsRed⁺ (fig. S11G), suggesting that cDC1s also acquire mitochondria from B16-OVA-mito-DsRed tumor cells. Nonetheless, mitochondrial membrane potential [based on MitoTracker Deep Red (MDR) staining] and the MDR/MG gMFI ratio were largely comparable between DsRed⁺ and DsRed⁻ cDC1s from B16-OVA-mito-DsRed tumors in vivo (fig. S11H), suggesting that mitochondrial membrane potential is not altered in intratumoral cDC1s upon acquiring mitochondria from B16-OVA tumors in vivo. Thus, the TME orchestrates the progressive decline of [TMRM/MG]^hi cDC1 number and mitochondrial membrane potential during tumor progression, independently of intercellular mitochondrial transfer from tumor cells to cDC1s.

Moreover, we examined whether mitochondrial features of cDC1s are dynamically changed upon LCMV clone 13 infection. As expected (66), the expression of activation markers such as MHC-II and CD40 was increased on splenic cDC1s at an early time point (day 1 versus day 0) (fig. S11I). In contrast, the mitochondrial membrane potential of cDC1s was reduced at day 1 post-infection (versus day 0), followed by an increase at day 7 or day 21 compared to day 1, as evidenced by the flow cytometry analysis of the frequencies of [TMRM/MG]^hi cDC1s (fig. S11J), the gMFIs of TMRM and MG (fig. S11K), and the TMRM/MG gMFI ratios (fig. S11K) at different time points, suggesting a discordant regulation of cDC1 activation and mitochondrial membrane potential in the context of chronic LCMV infection. These observations are distinct from those showing simultaneous decreases in DC activation markers (fig. S11L), functional fitness (Fig. 4A) and mitochondrial membrane potential (fig. S11M) in intratumoral cDC1s during tumor progression, pointing to selective mitochondrial metabolic adaptations in different contexts of chronic antigen stimulation.

Finally, to explore the regulation of the OPA1–NRF1 axis during tumor progression, we isolated cDC1s from B16-OVA tumors at days 7 and 21 after tumor inoculation and then profiled the chromatin states and transcriptome by performing ATAC-seq and microarray analyses, respectively. Transcription factor footprinting (Fig. 4E, and table S9) and motif enrichment (Fig. 4F, and table S10) analyses revealed that NRF1 transcriptional activity was downregulated in cDC1s isolated from late-stage tumors. Moreover, GSEA of transcriptome profiling revealed that OPA1-activated signature (see Methods, and table S11) was decreased in cDC1s from day 21 versus day 7 tumors, and the reciprocal pattern was observed for OPA1-suppressed signature (see Methods, and table S11) (Fig. 4G), suggesting that OPA1 activity is decreased during tumor progression. Accordingly, OPA1 and NRF1 protein levels were decreased in cDC1s from B16-OVA tumors during tumor progression (Fig. 4H). Furthermore, confocal imaging analysis revealed that intratumoral cDC1s showed decreased mitochondrial volume during tumor progression (Fig. 4I). Collectively, intratumoral cDC1s exhibit a progressive decrease of OPA1 and NRF1 expression, associated with reduced mitochondrial volume.

[TMRM/MG]^hi cDC1s synergize with ICB to elicit durable antitumor immunity

Given these mitochondrial alterations of cDC1s during tumor progression, we hypothesized that adoptive transfer of metabolically fit cDC1s into the TME may improve antitumor immunity. To this end, we purified splenic [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s, which exhibited marked differences in OXPHOS capacity (fig. S12A) and metabolomic profiles (fig. S1, Y and Z). Following antigen and poly I:C stimulation (22, 67), we performed intratumoral administration of cDC1s into B16-OVA or MC38 tumor-bearing mice (Fig. 4J). [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s showed the highest and lowest therapeutic effects, respectively, with intermediate effects observed for total cDC1s (Fig. 4, K and L). Thus, cDC1s with high mitochondrial membrane potential show markedly improved efficacy in cancer therapy models. To test whether the improved therapeutic efficacy of [TMRM/MG]^hi cDC1s is related to improved cell survival, we first analyzed cell survival in vitro and found that splenic [TMRM/MG]^hi and [TMRM/MG]^lo cDC1s showed a similarly progressive decrease in cell viability (fig. S12B). We next evaluated this possibility in vivo. To this end, we sorted splenic [TMRM/MG]^hi and [TMRM/MG]^lo cDC1 subpopulations from CD45.1⁺ donor mice, and then mixed CellTrace Violet^™ (CTV)-labeled [TMRM/MG]^hi cDC1s (CTV⁺) and CTV-unlabeled [TMRM/MG]^lo cDC1s (CTV⁻) at a 1:1 ratio, followed by intratumoral injection into CD45.2⁺ B16-OVA tumor-bearing recipient mice at day 6 after tumor inoculation (fig. S12C). Due to the short lifespan of adoptively transferred DCs (68), we analyzed DC survival and abundance 24 hours after adoptive transfer. The proportions and relative ratios of [TMRM/MG]^hi and [TMRM/MG^lo cDC1s were unchanged in the TME or tdLN compared to pre-transfer cells (fig. S12D). Furthermore, in the TME, the frequency of active caspase-3⁺ cells and expression of anti-apoptotic Bcl2 were also largely similar between these two subpopulations (fig. S12E), suggesting that the increased therapeutic effect of [TMRM/MG]^hi cDC1s is unlikely to be due to increased cell survival in vivo.

Despite the clinical breakthroughs with ICB therapies, the majority of patients exhibit poor responses or develop therapeutic resistance (69). To test whether the therapeutic efficacy of ICB can be improved via combination therapy with cDC1s of high mitochondrial membrane potential, B16-OVA tumor-bearing mice were treated with [TMRM/MG]^hi or [TMRM/MG]^lo cDC1s (on day 6 after tumor inoculation), followed by anti-PD-L1 treatment [on days 7, 10 and 13 after tumor inoculation, as described (67)]. Combination treatment with [TMRM/MG]^hi cDC1s plus anti-PD-L1 led to a greatly enhanced therapeutic effect compared to [TMRM/MG]^lo cDC1s plus anti-PD-L1 treatment or single treatments (Fig. 4M). Remarkably, all mice that received [TMRM/MG]^hi cDC1s plus anti-PD-L1 treatment completely rejected B16-OVA tumors (Fig. 4M), resulting in improved mouse survival (fig. S12F). Next, at 60 days after tumor inoculation, we rechallenged tumor-free mice that were previously treated with [TMRM/MG]^hi cDC1s plus anti-PD-L1 combination therapy (or control mice without prior tumor challenge). We found that these tumor-free mice rejected the B16-OVA tumor upon rechallenge (Fig. 4N), suggesting induction of durable immunological memory.

To further establish the therapeutic benefit of combinatorial treatment with [TMRM/MG]^hi cDC1s and ICB, we used two additional ICB therapies. In the B16-OVA tumor model, co-treatment of [TMRM/MG]^hi cDC1s with anti-CTLA-4 improved the tumor suppressive effects compared to single treatment or other control groups (e.g., co-treatment with [TMRM/MG]^lo cDC1s and ICB) (Fig. 4O), and substantially extended the survival of tumor-bearing mice (fig. S12G). In the MC38 adenocarcinoma model, we observed similar therapeutic effects of co-treatment with [TMRM/MG]^hi cDC1s plus anti-PD-1 on reducing tumor growth (Fig. 4P) and extending mouse survival (fig. S12H). Accordingly, tumor-free mice from such co-treatment were resistant to rechallenge with MC38 tumor cells (fig. S12I). Collectively, these data indicate that combinatorial treatment of cDC1s with polarized mitochondria and ICB greatly improves the therapeutic effects compared to either treatment alone.

Discussion

Despite the growing interest in understanding the mechanisms dictating DC antitumor responses (22, 70, 71), little is known about mitochondrial metabolic adaptation of cDC1s in the TME, or how this process shapes cDC1 function in antitumor immunity. Furthermore, whether mitochondrial metabolism-associated processes can imbue engineered cDC1s with improved therapeutic efficacy against tumors remains unknown. We identified a novel OPA1–NRF1 axis that links OXPHOS to the functional fitness of cDC1s by suppressing autophagic degradation of MHC-I and antigen, thereby providing new insights into the interplay between energy metabolism and immune signaling. Moreover, OPA1-mediated regulation of the NAD⁺/NADH ratio also contributes to functional fitness of cDC1s, suggesting a role for redox balance in this process. Importantly, intratumoral cDC1s undergo progressive mitochondrial dysfunction (as revealed by decrease in [TMRM/MG]^hi cellularity and reductions in cDC1 mitochondrial membrane potential and volume) during tumor progression, associated with the decrease of OPA1–NRF1 signaling, which may contribute to cDC1 functional impairment in antitumor immunity. We also found that intratumoral administration of cDC1s with polarized mitochondria markedly inhibits tumor growth and synergistically enhances responsiveness to ICB, leading to durable immune memory. These findings are in line with the observations that mitochondrial OXPHOS in DCs positively correlates with therapeutic outcomes in melanoma patients treated with DC vaccines (72). Thus, modulating mitochondrial metabolism (especially OXPHOS and NAD⁺/NADH ratio) and related signaling pathways (namely, the OPA1–NRF1 axis) hold promise for improving the efficacy of cancer immunotherapy.

Intratumoral cDC1s orchestrate antitumor CD8⁺ T cell responses (39), with new approaches emerging to rewire DC fate and function to enhance the efficacy of immunotherapy (73, 74). However, the low DC cellularity and dependence on tumor neoantigens, as well as limited understanding of the mechanisms orchestrating intratumoral DC functionality, impede the successful application of DC-based therapies (39). We showed that discrete mitochondrial states distinguish functional subpopulations of cDC1s and that intratumoral injection of splenic cDC1s with polarized mitochondria improves tumor therapy, although these cells may not fully recapitulate the characteristics of intratumoral cDC1s. Accordingly, metabolic engineering of intratumoral cDC1s, through the manipulation of mitochondria-related signaling pathways in situ such as by enhancing OPA1 or NRF1 activity, may fully unleash the potential of cDC1s in priming antitumor CD8⁺ T cell responses and potentiating cancer immunotherapies.

Despite the established roles of OPA1 in shaping mitochondrial morphology and assembly of ETC supercomplexes (33), whether OPA1 orchestrates mitochondrial metabolism via additional mechanisms remains poorly defined. We identified a novel OPA1–NRF1 axis that supports mitochondrial energy metabolism and cDC1 immunogenicity. Although PAMP stimulation dampens mitochondrial OXPHOS of DCs (18) and OXPHOS inhibition has negligible effects on DC activation under such settings (19), our data revealed an indispensable role of mitochondrial OXPHOS for cDC1 antigen presentation and antitumor immunity, highlighting metabolic adaptation of DCs in different contexts. As OXPHOS is a more efficient pathway for ATP generation than glycolysis (12), OPA1–NRF1-directed mitochondrial OXPHOS likely serves as a crucial mechanism to enhance intratumoral cDC1 bioenergetics in the nutrient-limited TME (22), with additional contributions from ETC flow-mediated NAD⁺/NADH balance to support cDC1 functional fitness. These findings extend the role of ETC in orchestrating antitumor immunity beyond influencing tumor-intrinsic fitness and immunogenicity (59, 75–78) and highlight that therapeutic strategies capable of manipulating mitochondrial ETC flow in cDC1s have the strong potential to improve antitumor responses and overcome immunotherapeutic resistance in human cancer treatment.

Materials and Methods:

Mice

C57BL/6, CD45.1⁺, OT-I, Cas9-transgenic, CD11c-Cre and XCR1-Cre mice were purchased from The Jackson Laboratory. Opa1^fl/fl and Dnm1l^fl/fl mice were kindly provided by Dr. Hiromi Sesaki (35, 61). 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. The “retrogenic” mice with indicated gene deletion were generated using a CRISPR–Cas9 delivery system as described before (43). Briefly, we isolated Lineage⁻Sca-1⁺c-Kit⁺ (LSK) cells from Cas9-expressing mice using a lineage cell depletion kit (130–110-470, Miltenyi Biotec) according to the manufacturer’s instructions and followed by cell sorting. LSK cells were cultured in SFEM medium (09600, Stemcell Technologies) and then transduced with retrovirus expressing a fluorescent reporter and control sgRNA (sgNTC: ATGACACTTACGGTACTCGT), or sgRNA targeting Ndufaf1 (TTATGACCTCTTGTAATGGG), Sdhb (CAGGTGCGGACCTATGGTGT), Uqcrq (CTCAAAGGGCGACAAGCTGT), Atpaf2 (ATGGGATCCCGTCATAGAGT), Nrf1 (GATGAGTACACGACGCGAGT) or Atg5 (GAGATATGGTTTGAATATGA), or that overexpressing NRF1 or mito-LbNOX, followed by adoptive transfer into lethally irradiated (1,100 rad) recipients to generate bone marrow chimeric mice. 3 weeks after bone marrow reconstitution, the “retrogenic” mice were then subcutaneously challenged with B16-Flt3L tumor cells for 9 days, follow by analysis of splenic cDC1s. All mice were maintained in specific pathogen-free conditions in the Animal Resource Center at St. Jude Children’s Research Hospital. Experiments and procedures were approved by and performed in accordance with the Institutional Animal Care and Use Committee of St. Jude Children’s Research Hospital (approval #470–100586).

Measurement of genome editing efficiency

Targeted amplicons were generated using gene-specific primers with partial Illumina adaptors, followed by two rounds of PCR: first for amplification and then for indexing as previously described (79). Briefly, 1×10⁵ LSK cells were lysed and used to generate gene-specific amplicons by first round PCR, followed by a second round of PCR to index the samples. Indexed amplicons were pooled with other loci to increase sequence diversity, and 10% PhiX control was added before sequencing on the Illumina MiSeq platform (2×250 bp). Samples were demultiplexed, and insertion and deletion (indel) frequencies were analyzed using CRIS.py (80). High editing efficiencies (69.8%–92.3%) were achieved for all target genes (table S12).

Cell purification and culture

Mouse spleens or tumors were digested with 1 mg/ml collagenase IV (LS004188, Worthington) plus 200 U/ml 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⁺MHC-II⁺XCR1⁺ CD11b⁻TCRb⁻CD49b⁻B220⁻CD64⁻Ly6C⁻) and cDC2s (CD11c⁺MHC-II⁺XCR1⁻CD11b⁺TCRb⁻CD49b⁻B220⁻CD64⁻Ly6C⁻) on a MoFlow (Beckman-Coulter), Reflection (i-Cyt) or Bigfoot (Thermo Fisher Scientific) cell sorter. Lymphocytes from spleen and peripheral lymph nodes were sorted for naïve OT-I cells (CD8a⁺CD62L^hiCD44^loCD25⁻), or polyclonal CD4⁺ T cells (CD45⁺CD4⁺B220⁻TCRb⁺), polyclonal CD8⁺ T cells (CD45⁺CD8a⁺B220⁻TCRb⁺) and B cells (CD45⁺CD19⁺B220⁺TCRb⁻) (for real-time PCR assays described below). Sorted cDCs were cultured in IMDM medium (12440053, Gibco^™) supplemented with 10% (v/v) dialyzed fetal bovine serum (FBS; A3382001, Thermo Fisher Scientific) plus 1% (v/v) penicillin-streptomycin (15140122, Thermo Fisher Scientific) and 55 μM b-mercaptoethanol (M6250, Aldrich) (IMDM complete medium). The cells were then subjected to the indicated experiment, as described in the figure legends. For preparation of tumor cell line-derived culture supernatant, B16-OVA, B16F10, B16-Flt3L, MC38, LLC or E.G7-OVA cells were cultured in IMDM medium supplemented with 10% (v/v) FBS plus 1% (v/v) penicillin-streptomycin, tumor cell culture supernatant was collected 48 hours later.

LCMV clone 13 infection

2 × 10⁶ PFU of LCMV clone 13 strain virus was injected intravenously. Mice with or without LCMV clone 13 infection were euthanized for analysis of cDC1s or CD8⁺ T cells by flow cytometry as indicated in the figure legends.

Metabolic assays via Seahorse

Oxygen consumption rate (OCR) was measured following the manufacturer’s instructions of Seahorse FluxPaks (103775–100, Agilent). In brief, equal numbers (1–2 × 10⁵) of sorted live cDCs were suspended in XF RPMI medium and then plated in a poly-L-lysine-coated XF96 pro plate. The OCR measurement under basal conditions (before drug treatments) and in response to 1 μM oligomycin (Oligo), 1.5 μM fluoro-carbonyl cyanide phenylhydrazone (FCCP) and 500 nM rotenone (Rot) were analyzed using Seahorse XF Pro analyzer (Seahorse Bioscience). Maximum OCR was the OCR value following FCCP treatment. ATP production was calculated by subtracting the OCR measurement (after oligomycin treatment) from basal OCR value (before oligomycin treatment).

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 in 4°C for appropriate 30 min. For T cell intracellular cytokine detection, cells were stimulated for 4 hours with phorbol 12-myristate 13-acetate (PMA) plus ionomycin in the presence of GolgiSTOP; for DC intracellular cytokine detection, DCs were treated with GolgiSTOP for 4 hours, before staining with 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. Mitochondrial mass and mitochondrial membrane potential were analyzed by staining cells with MitoTracker^™ Green FM (M7514, 100 nM, Invitrogen^™) and Tetramethylrhodamine, methyl ester (T668, 20 nM, Invitrogen^™), at 37°C in IMDM complete medium for 30 min, respectively. 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-labeled antibodies were used: PerCP-Cyanine5.5–anti-CD8a (53–6.7, 65–0081, 1:400), PE-Cy7–anti-CD11c (N418, 60–0114, 1:200), APC–anti-CD127 (A7R34, 20–1271, 1:100) (from Tonbo Biosciences); APC-eFluor 780–anti-MHC-II (M5/114.15.2, 47–5321-82, AB_1548783, 1:400), APC–anti-MHC-I (AF6–88.5.5.3, 17–59580-80, AB_1311280, 1:400), PE-Cyanine7–anti-TNFa (MP6-XT22, 25–7321-82, AB_11042728, 1:200), FITC–anti-CD40 (HM40–3, 11–0402-86, AB_465031, 1:400), FITC–anti-CD86 (GL1, 11–0862-82, AB_465148, 1:200), PE-cyanine 7–anti-T-bet (4B10, 25–5825-82, AB_11042699, 1:100), PerCP-Cyanine5.5–anti-IFNg (XMG1.2, 45–7311-82, AB_1107020, 1:200), PerCP-Cyanine5.5–anti-CD11c (N418, 45–0114-82, AB_925727, 1:400), PE–anti-IL-12/IL-23 p40 (C17.8, 12–7123-82, AB_466185, 1:200), PE-anti-CD11c (N418, 12–0114-82, AB_465552, 1:400), eFluor-450-anti-PDCA-1 (eBio927, 48–3172-82, AB_2043879, 1:400) (all from eBioscience); FITC–anti-CD45.1 (A20, 110706, AB_313495, 1:400), PE–anti-XCR1 (ZET, 148204, AB_2563843, 1:400), Brilliant Violet 421–anti-XCR1 (ZET, 148216, AB_2565230, 1:400), Brilliant Violet 510–anti-CD80 (16–10A1, 104741, AB_2810337, 1:200), Brilliant Violet 510–anti-CD4 (RM4–5, 100559, AB_2562608, 1:200), AF700–anti-CD8a (53–6.7, 100730, AB_493703, 1:200), Brilliant Violet 785–anti-TCRβ (H57–597, 109249, AB_2810347, 1:200), PE–anti-CD45.2 (104, 109808, AB_313445, 1:400), PE–anti-Bcl2 (BCL/10C4, 633508, AB_2290367, 1:100), PE/Dazzle 594–anti-PD-1 (29F.1A12, 135228, AB_2566006, 1:400), PE–anti-H-2K^b-SIINFEKL (25-D1.16, 141604, AB_10895905, 1:100), Alexa Fluor 647–anti-GZMB (GB11, 515405, AB_2294995, 1:100), PE-Cyanine7–anti-IFNg (XMG1.2, 505826, AB_2295770, 1:200), PE-Cyanine7–anti-CD62L (MEL-14, 104418, AB_313103, 1:400), PE-Cyanine7–anti-MHC-II (M5/114.15.2, 107630, AB_2069376, 1:400), PE-Cyanine7–anti-CXCL10 (IP-100, 519507, AB_2904432, 1:100), Brilliant Violet 421–anti-TNFa (MP6-XT22, 506328, AB_2562902, 1:200), Brilliant Violet 421–anti-CD11b (M1/70, 101236, AB_11203704, 1:400), Brilliant Violet 711–anti-TIM-3 (RMT3–23, 119727, AB_2716208, 1:400), Brilliant Violet 711–anti-Ly-6C (HK1.4, 128037, AB_2562630, 1:400), Brilliant Violet 650–anti-CD11b (M1/70, 101259, AB_2566568, 1:400), Brilliant Violet 650–anti-CD44 (IM7, 103049, AB_2562600, 1:400), Brilliant Violet 650–anti-CD86 (GL-1, 105036, AB_2686973, 1:400), Brilliant Violet 605–anti-CD64 (X54–5/7.1, 139323, AB_2629778, 1:200), Brilliant Violet 605–anti-CD8a (53–6.7, 100743, AB_2561352, 1:400), APC–anti-XCR1 (ZET, 148206, AB_2563932, 1:400), APC–anti-CD44 (IM7, 103012, AB_312963, 1:400), Brilliant Violet 605–anti-KLRG1 (2F1/KLRG1, 138419, AB_2563357, 1:400), APC–anti-Ly108 (330-AJ, 134610, AB_2728155, 1:400), APC-anti-CXCL9 (MIG, 519504, AB_2561409, 1:100), Brilliant Violet 650–anti-CX3CR1 (SA011F11, 149033, AB_2565999, 1:400), Brilliant Violet 785–anti-CD45.2 (104, 109839, AB_2562604, 1:400), Brilliant Violet 421–anti-CCR7 (4B12, 120120, AB_2561446, 1:100) (all from BioLegend); PE–anti-IL-2 (JES6–5H4, 554428, AB_395386, 1:200), Brilliant Violet 605–anti-Ly108 (13G3, 745250, AB_2742834, 1:200), FITC–anti-CD103 (M290, 557494, AB_396731, 1:400) (from BD Biosciences); Alexa Fluor 647–anti-TCF1 (C63D9, 6709, 1:100), Pacific Blue–anti-phospho-S6 (D57.2.2E, 8520, 1:100), (all from Cell Signaling Technology). LysoTracker Red DND-99 is from Thermo Fisher Scientific (L7528; 1:2,000). The Db/gp33 tetramer is from NIH Tetramer Core Facility.

Tumor models and treatments

B16F10, LLC, EO771 and E.G7-OVA cell lines were purchased from ATCC. B16-OVA and MC38 cell lines were provided by D. Vignali (University of Pittsburgh). B16-Flt3L cell line was provided by D. Green (St. Jude Children’s Research Hospital). B16F10 cell line expressing ZsGreen (B16-ZsGreen) was generated by lentiviral transduction of pHIV-ZsGreen construct (18121, Addgene) and were sorted for expression of ZsGreen (22). These cell lines are not on the list of commonly misidentified cell lines (International Cell Line Authentication Committee). All cell lines were maintained at 37°C with 5% CO₂ in DMEM (11965092, Gibco) supplemented with 10% (v/v) FBS and 1% (v/v) penicillin-streptomycin. Mice were injected subcutaneously with 5 × 10⁵ MC38, B16-OVA or LLC cells in the right flank. To establish the orthotopic mammary tumor model, 2.5 × 10⁵ EO771 cells were inoculated in the left #5 mammary fat pad. To establish the AKT/NRas^G12V-driven hepatocellular carcinoma (HCC) mouse model, 6-week-old male mice were injected with 5 μg pT3-EF1a-myrAKT1-HA (31789, Addgene), 5 μg pT-Caggs-NRas^G12V (20205, Addgene) and 2.5 μg pCMV (CAT)T7-SB100 (34879, Addgene) as previously described (30). Tumors were measured regularly with digital calipers and tumor volumes were calculated using the formula: length × width × [(length × width)^0.5] × π/6. To isolate intratumoral lymphocytes, tumors were harvested at day 14 after inoculation, excised, minced, and digested with 1 mg/ml collagenase IV (Worthington) and 200 U/ml DNase I (Sigma) for 1 h at 37°C.

cDC1 adoptive transfer and ICB treatment

For DC transfer experiments, freshly isolated splenic cDC1s were used following an established strategy (67) with modification. In brief, B16-Flt3L cells (2.5 × 10⁶) were injected subcutaneously into both flanks of WT mice to expand cDC1. Spleens were harvested 9 days after tumor inoculation, splenic DCs were enriched with CD11c Microbeads (130–125-835, Miltenyi Biotec) and then total cDC1s or [TMRM/MG]^hi and [TMRM/MG]^lo cDC1 subpopulations were sorted based on the ratio of TMRM to MG. MC38 cell lysate was prepared as described previously (67) with modifications. Tumor cells were cultured at 1× 10⁷ cells per ml in complete IMDM medium, and the cells were irradiated under an ultraviolet (254 nm) lamp for 30 mins followed by incubation under normal cell culture conditions overnight. The cells were subjected to 5 freeze (− 80 °C)/thaw (37 °C) cycles, followed by gentle homogenization using a 2 ml homogenizer with 20 stokes. Then, the lysate was further sonicated and passed through a 40 μm cell strainer, and centrifuged at 500g for 5 min before coculture with DCs. Purified cDC1s were pulsed with 100 mg/ml OVA protein or MC38 cell lysate at a 1:2 ratio (DCs:tumor cells) together with 20 mg/ml poly I:C (InvivoGen) in complete IMDM medium for 2 hours. Then, the cDC1s were washed with HBSS twice and transferred intratumorally (0.5–1 × 10⁶ cells per mouse) via subcutaneous injection adjacent to the tumors at day 5 after 1 × 10⁶ B16-OVA or MC38 inoculation. For combination of cDC1 transfer and ICB treatment, mice with B16-OVA or MC38 inoculation received adoptive transfer of cDC1s on day 6 after tumor inoculation. Then, anti-PD-L1 antibody (10F.9G2, AB_10949073, Bio X Cell), anti-CTLA-4 (9H10, AB_10950184, Bio X cell) (both for B16-OVA tumor) or anti-PD-1 antibody (J43, AB_1107747, Bio X Cell) (for MC38 tumor) [or rat IgG2b isotype control (LTF-2, AB_1107780, Bio X Cell)] was injected intraperitoneally three times at a dose of 200 mg in 100 ml PBS on days 7, 10 and 13 after inoculation of tumor cells.

To assess survival or mitochondrial alterations of cDC1s in vivo, total cDC1s or their [TMRM/MG]^hi or [TMRM/MG]^lo subpopulations were sorted from non-immunized (naïve) WT mice. After washing with HBSS twice, the cells were adoptively transferred intratumorally (5 × 10⁵ cells per mouse) via subcutaneous injection adjacent to the B16-OVA tumors at indicated day as described in the legend. The tumors and tdLNs were then isolated for flow cytometry analysis.

Adoptive T cell transfer

Naïve OT-I cells were isolated using a naïve CD8⁺ T cell isolation kit (130–096-543; Miltenyi Biotec) according to the manufacturer’s instructions. For activated OT-I cell transfer, purified naïve OT-I cells were activated using 10 μg/ml anti-CD3 (2C11; Bio X Cell, BE0001–1) and 5 μg/ml 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% dialyzed FBS in the presence of human recombinant IL-2 (20 IU/ml; National Cancer Institute), mouse IL-7 (12.5 ng/ml; PeproTech) and IL-15 (25 ng/ml; PeproTech) for 2–3 days before adoptive transfer. Where indicated, 2 × 10⁶ naïve OT-I cells were adoptively transferred into recipient mice bearing B16-OVA tumors on day 12 after tumor inoculation, and intratumoral OT-I cells were analyzed at day 19 after tumor inoculation (day 7 after adoptive transfer) as indicated.

Antigen presentation assays and ELISA

For in vitro assays, cDC1s were sorted from spleen or tumor, pulsed with 200 μg/ml OVA protein (Low Endo, LS003059, Worthington) or 250 pg/ml OVA_257–264 peptide (vac-sin, InvivoGen) for 2 hours, washed twice with IMDM complete medium and irradiated (3,000 rad), and then cocultured with naïve (CD44^loCD62L^hi) OT-I cells for three days. For HKLM-OVA antigen cross-presentation, sorted splenic cDC1s and cDC2s were cocultured with 1 × 10⁷ HKLM-OVA and OT-I cells for 3 days. B16F10 endogenous antigen cross-presentation was performed as described previously (40) with modification. Briefly, sorted splenic cDC1s and cDC2s were cocultured with B16F10 cell lysate at a 1:5 ratio (DCs: tumor cells) for 4 hours, washed twice and irradiated (3,000 rad), and then cultured with naïve pmel cells for three days. ³H-thymidine (PerkinElmer) was added to the culture 8 hours before cells were harvested to measure thymidine incorporation. Where indicated, cDC1s were incubated with OVA protein in IMDM medium with or without 35 mM chloroquine (124893, Sigma-Aldrich), 1 mM oligomycin (S1478, Selleckchem), 500 nM rotenone (R8875, Sigma-Aldrich), 1 mM 3-NPA (164603, Sigma-Aldrich), 10 mM Antimycin A (A8674, Sigma-Aldrich). For chloroquine treatment conditions, cDC1s were pretreated with or without chloroquine for 2 hours before OVA protein pulsing. Cells were washed extensively to remove the inhibitors, irradiated, and then cocultured with OT-I cells for proliferation analysis as described above.

For in vivo antigen presentation assays, 1 × 10⁶ CTV-labeled naïve CD45.1⁺ OT-I cells were transferred into mice intravenously, followed by intravenous injection with 20 mg OVA 24 hours later. After OVA immunization, spleens were harvested three days later, and the proliferation of splenic OT-I cells was examined by CTV dilution.

For ELISA, culture supernatants from in vitro DC and T cell coculture assays were collected, and then the abundances of IL-2 and IFNg were measured using IL-2 (88–7024-22, Thermo Fisher Scientific) and IFNg (88–7314-22, Thermo Fisher Scientific) ELISA kits according to manufacturer’s instructions.

Immunoblot analyses

Intratumoral or splenic cDC1s (1–2 ´ 10⁵) were sorted and lysed in RIPA buffer (9806, Cell Signaling Technology). The lysates were 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 in TBS for 1 hour 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 Analyzer (LI-COR) and AMERSHAM Imager 600 (GE Healthcare). The following antibodies were used: anti-ACTB (3700), anti-OPA1 (80471), anti-DRP1 (5391), anti-MFN1 (14739), anti-MFN2 (9482), anti-ULK1 (8054), anti-ULK1-Ser555 (5869), anti-AMPK (2532), anti-AMPK-Thr172 (50081), anti-LC3B (43566) (from Cell Signaling Technology); anti-P62 (18420–1-AP, AB_10694431, Proteintech); anti-NRF1 (GTX103179; GeneTex); total OXPHOS rodent WB antibody cocktail was purchased from Abcam (ab110413). All primary antibodies were used at 1:1,000 dilution for immunoblot analysis.

ATP, ADP, AMP, NAD⁺ and NADH measurements

ATP levels were determined using ATP Determination kit (A22066, Thermo Scientific). ADP and AMP levels were determined using methods as reported previously with modification (81). Splenic cDC1s (1–2 ´ 10⁶) were harvested and lysed in ice-cold native lysis buffer (25 mM Tris-HCL, 10 mM KCl, 3 mM MgCl₂, 0.5 mM DTT), followed by 2 min sonication. Samples were centrifuged at 13,000 g for 5 min. First, AMP and ADP were converted to ATP in a 100 μl reaction containing 1 unit of myokinase (M3003, Sigma), 0.25 unit of pyruvate kinase (P9136, Sigma), 0.2 mM dCTP (18253013, Invitrogen^™), 0.3 mM phosphoenolpyruvate (B20358.06, Thermo Scientific), and 10 μl of samples. The reaction was incubated at room temperature for 30 min. Total ATP concentration was measured and recorded as ‘reading A’. Second, only ADP was converted to ATP in a 100 μl reaction containing 0.25 unit of pyruvate kinase, 0.3 mM P-enolpyruvate, and 10 μl of samples. The reaction was incubated at room temperature for 30 min. The ATP concentration was measured, and the reading was recorded as ‘reading B’. Third, ATP concentrations in 10 μl of samples and ATP standard were determined, and the readings were ‘readings C’. The AMP concentration equals ‘reading A’ minus ‘reading B’, whereas the ADP concentration equals ‘reading B’ minus ‘readings C’, and ‘readings C’ are the actual ATP concentration in the samples. NAD⁺ and NADH levels were analyzed in extracts of splenic cDC1s (1 ´ 10⁵) using the NAD⁺/NADH Assay Kit (Fluorometric) (CBI-MET-5030; Cell Technology Inc.) according to the manufacturer’s instructions.

DQ-OVA degradation assay and antigen uptake in vitro

Sorted cDC1s from WT and Opa1^ΔDC mice were incubated with 10 μg/ml DQ-OVA (D12053, 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 analyzed for DQ-OVA release as assessed positive FITC (FITC⁺) staining. To assess antigen uptake, WT and OPA1-deficient cDC1s were incubated with 10 μg/ml OVA-AlexaFluor 488 (O34781, Thermo Fisher Scientific) for 120 min, followed by flow cytometry analysis for AlexaFluor 488 positive (AlexaFluor 488⁺) cells.

Immunofluorescence

Sort-purified splenic or intratumoral cDC1s were allowed to adhere to poly-L-lysine coated coverslips prior to fixation with 4% PFA for 10 min. Cells were then permeabilized with PBS containing 0.1% Triton X-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-TOM20 antibody (1 μg/ml, Abcam, 186735) prior to detection with fluorescently labeled secondary antibody and fluorescently labeled phalloidin to detect F-Actin (Thermo Fisher Scientific; catalog A12380, 1 U/ml). Coverslips were mounted in Vectashield Vibrance mounting media with DAPI (Vector Laboratories, H-1800) and were imaged using a Marianas spinning disk confocal (Intelligent Imaging Innovations) equipped with a Sora (Yokagawa), Prime 95B sCMOS camera (Photometrics) and a 1.45 NA 100X oil objective. Images were acquired and analyzed using Slidebook (v6.0.24).

Electron microscopy

Cells for electron microscopy were grown on LabTek thermanox chambered slides and fixed in 0.1 M phosphate buffer containing 2.5% glutaraldehyde and 2% paraformaldehyde. Samples were post fixed in reduced 2% osmium tetroxide and contrasted in 2% uranyl acetate followed by dehydration in an ascending series of ethanol washes from 50% to 100%. Samples were then infiltrated with EmBed-812/ethanol mixtures, embedded in 100% EmBed-812, and polymerized at 60°C for 48 hours. Following resin polymerization, resin was removed from the permanox slide, cells located and marked under a microscope, the resin blocks cut down to isolate the cells and mounted for ultrathin sectioning. 50 nm sections were cut on a Leica (Wetzlar, Germany) ARTOS ultramicrotome and collected on formvar/carbon coated copper slot grids for imaging. Samples were imaged on a Zeiss (Oberkochen, Germany) Gemini 460 SEM using at 19 kV and 202 pA utilizing a segmented STEM detector operating in darkfield mode (segments 2 through 5 active) with the imaging LUT inverted. Unless otherwise stated, all reagents from Electron Microscopy Sciences (Hatfield, PA). The number and length of cristae in per mitochondrion or per mitochondrial area were quantified using Image J.

RNA isolation, gene expression profiling and microarray analysis

RNA was isolated and purified from various cell types using the RNeasy Micro Kit (74004, Qiagen) following 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 primers were listed below: Opa1-F: TGGAAAATGGTTCGAGAGTCAG; Opa1-R: CATTCCGTCTCTAGGTTAAAGCG, Tapbp-F: GGCCTGTCTAAGAAACCTGCC, Tapbp-R: CCACCTTGAAGTATAGCTTTGGG, Tap1-F: GGACTTGCCTTGTTCCGAGAG, Tap1-R: GCTGCCACATAACTGATAGCGA, Tap2-F: CTGGCGGACATGGCTTTACTT, Tap2-R: CTCCCACTTTTAGCAGTCCCC, Sec61g-F: CAGGTAATGCAGTTTGTGGAGC, Sec61g-R: TGGATCAGTTTCACGAAGAAGC, Sec61b-F: TCCCAGTGCTGGTGATGAGT, Sec61b-R: GCGTGTACTTGCCCCAAAT, Sec61a1-F: GGAAGTCATCAAGCCATTCTGT, Sec61a1-R: GCATCCAGTAGAACGGGTCAG, Erap1-F: TAATGGAGACTCATTCCCTTGGA, Erap1-R: AAAGTCAGAGTGCTGAGGTTTG, Pdia3-F: CGCCTCCGATGTGTTGGAA, Pdia3-R: CAGTGCAATCCACCTTTGCTAA, Canx-F: ATGGAAGGGAAGTGGTTACTGT, Canx-R: GCTTTGTAGGTGACCTTTGGAG, Cst3-F: AGGAGGCAGATGCCAATGAG, Cst3-R: GGGCTGGTCATGGAAAGGA, Cstb-F: AGGTGAAGTCCCAGCTTGAAT, Cstb-R: GTCTGATAGGAAGACAGGGTCA, Calr-F: CCTGCCATCTATTTCAAAGAGCA, Calr-R: GCATCTTGGCTTGTCTGCAA, Cast-F: GGAAGGACAAACCAGAGAAGC, Cast-R: AGGGGCAGCTATCCAAATCTT, Psmb8-F: ATGGCGTTACTGGATCTGTGC, Psmb8-R: CGCGGAGAAACTGTAGTGTCC, Psmb9-F: CATGAACCGAGATGGCTCTAGT, Psmb9-R: TCATCGTAGAATTTTGGCAGCTC, Psmb10-F: GAGGAATGCGTCCTTGGAACA, Psmb10-R: CACAACCGAATCGTTAGTGGC, H2k1-F: TTGAATGGGGAGGAGCTGAT, H2k1-R: GCCATGTTGGAGACAGTGGA, H2d1-F: AGTGGTGCTGCAGAGCATTACAA, H2d1-R: GGTGACTTCACCTTTAGATCTGGG, B2m-F: TTCTGGTGCTTGTCTCACTGA, B2m-R: CAGTATGTTCGGCTTCCCATTC, Actb-F: GGCACCACACCTTCTACAAT, Actb-R: CTTTGATGTCACGCACGATTTC.

For microarray analyses, the following populations were sorted: 1. cDC1 subpopulations from spleen of WT mice based on TMRM and MG co-staining profiles (namely, [TMRM/MG]^hi and [TMRM/MG]^lo) (n = 4 each group); 2. Splenic cDC1s from WT or Opa1^ΔDC mice (n = 3 for WT mice; 4 for Opa1^ΔDC mice); 3. cDC1s isolated from B16-OVA tumors at days 7 and 21 after tumor inoculation (n = 4 per group). 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 (v1.4.1). Differential gene expression analysis was performed by R package limma (v3.46.0). False discovery rate (FDR) was estimated by Benjamini–Hochberg method.

Proteomics

0.8–3 × 10⁶ cells intratumoral and splenic cDC1s (n = 3 each group) were sorted from B16-OVA tumor-bearing mice and homogenized in lysis buffer (50 mM HEPES, pH 8.5, 8 M urea, 0.5% sodium deoxycholate, 1´ PhosSTOP phosphatase inhibitor, and 1 mM DTT). Then, the protein extracts were proteolyzed with Lys-C (Wako, 1:100 w/w) for 2 hours at room temperature, diluted to 2 M urea with 50 mM HEPES (pH 8.5), and further digested overnight with trypsin (Promega, 1:50 w/w) at room temperature. Peptide disulfide bond was reduced with 1 mM DTT at room temperature for another 2 hours and subsequently alkylated with 10 mM iodoacetamide (IAA) at room temperature for 30 min in the dark. The digestion was terminated by the addition of trifluoroacetic acid at a final concentration of 0.5%. After centrifugation at 21,000 g for 10 min, the supernatant was desalted with the UltraMicroSpin column (The Nest Group), and then dried by Speedvac. The peptides were resuspended in 50 mM HEPES (pH 8.5) and labeled with the TMTpro 18-plex kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. All channels were mixed equally and desalted before further analysis. The TMT-labeled peptides were analyzed by an optimized two-dimensional liquid chromatography-tandem mass spectrometry (LC/LC-MS/MS) platform. All the mass spectrometry data were processed with an in-house JUMP software suite (82), a tag-based hybrid search engine to improve sensitivity. The protein database was generated by combining downloaded Swiss-Prot, TrEMBL, and UCSC databases and removing redundancy, followed by concatenation with a decoy database. Major parameters included 15 ppm mass tolerance for precursor ions and 20 ppm for product ions, full trypticity, static modification of the TMTpro tags (+304.20714 Da) on Lys residues and peptide N termini and carbamidomethyl modification on cysteine (+57.02146 Da), dynamic modification for Met oxidation (+15.99492 Da), maximal miscleavage sites (n = 2), and maximal modification sites (n = 3). The resulting peptide spectral maps (PSMs) were filtered to reduce protein FDR below 1%. Peptides generated from multiple homologous proteins were assigned to the canonical protein form in the manually curated Swiss-Prot database (83) based on the rule of parsimony. If no canonical form was defined, the peptide was assigned to the protein with the highest PSM number. The protein quantification was based on TMT reporter ion intensities with y₁-ion based correction of TMT data to reduce the effect of ratio compression.

The upregulated proteins in intratumoral versus splenic cDC1s were defined based on the cutoffs of FDR < 0.01 and log₂FC > 0.5. Enriched pathways for upregulated proteins in cDC1s isolated from B16-OVA tumors were analyzed by funcEnrich.Fisher function in NetBID (v2.0.3). Fisher’s exact test was used to test pathway enrichment against the MitoCarta3.0 database (84). DAVID (85) Bioinformatics Resource (v2021, knowledgebase v2023q3) was used for functional annotation clustering analysis for the upregulated proteins in intratumoral cDC1s compared to splenic cDC1s with default settings and FDR displayed.

Metabolomics

Splenic cDC1 subpopulations were isolated, counted and all samples were adjusted to 3 × 10⁶ cells per tube. Then, cells were centrifuged at 800 × g for 5 min at room temperature and washed once with ice-cold saline. The supernatant was discarded, and the cell pellets were flash-frozen in liquid nitrogen and stored at −80 °C for extraction of metabolites. To extract the molecules with different polarities, an adapted three-phase solvent system was utilized to obtain total hydrophilic metabolites and lipids (86). Briefly, cell pellets were resuspended with 150 μl of saline, followed by adding 1.2 ml of chloroform/methanol/water (3:4:1, v/v/v) and then homogenized using a Bead Ruptor Elite (OMNI international, Kennesaw GA, USA) for 30 sec at 8 m/s. The homogenate was allowed to rest at 4 ºC for 30 sec and then centrifuged for 10 min at 21,000 g at 4 ºC. After centrifugation, the upper aqueous phase was transferred into a new tube, frozen on dry-ice and then lyophilized. The dried extracts containing hydrophilic metabolites were dissolved using 30 μl of water/acetonitrile (95:5, v/v) supplemented with 10 mM ammonium acetate, transferred into autosampler vials and then 2 μl per injection were analyzed by LC-MS. A Vanquish Horizon UHPLC (Thermo Fisher Scientific, Waltham, MA, USA) was used for the LC separations, using stepped gradient conditions as follows: 0–16.5 min 1 to 50% B; 16.5–18 min 50 to 99% B; 18–36 min 99% B; 36–39 min 99 to 1% B; 39–45 min 1% B. Mobile phase A was water supplemented with 10 mM ammonium acetate. Mobile phase B was acetonitrile. The column used was an xBride BEH Amide Column (2.1 × 150 mm, 2.5 μm) (Waters Corp., Milford, MA, USA), operated at 40 °C. The flow rate was 100 μl/min, and the injection volume was 2 μl. A Thermo Scientific Q Exactive hybrid quadrupole-Orbitrap mass spectrometer (QE-MS) (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a HESI-II Probe was employed as detector. For each sample two chromatographic runs were carried out subsequently, acquiring data for negative and positive ions separately. The QE-MS was operated using a data-dependent LC-MS/MS method (Top10 dd-MS²) for both positive and negative ion modes. The mass spectrometer was operated at a resolution of 140,000 (FWHM, at m/z 200), AGC targeted of 3 × 10⁶, Max injection time 100 msec. The instrument’s operating conditions were: Scan range 60–900 m/z, sheath gas flow 20, auxiliary gas flow 5, sweep gas 1, spray voltage 3.6 kV for positive mode or 2.5 kV for negative mode, capillary temperature 320 ºC, S-lenses RF level 55, auxiliary gas heater 320 ºC. For the top 10 dd-MS² conditions a resolution of 35,000 was used, AGC targeted of 1 × 10⁵, max injection time 100 msec, MS² isolation width 1.0 m/z, and NCE 35.

The software Compound Discoverer 3.3 (CD3.3) (Thermo Fisher Scientific, Waltham, MA, USA), and MetaboAnalyst 6.0 were used to process untargeted metabolomics analyses. For CD3.3 a pre-defined workflow from the software was employed: Untargeted Metabolomics with statistics Detect Unknowns with ID using online databases and mzLogic. The spectra alignment node used the first sample as the reference file. The compound detection option was set for a signal-to-noise ratio ≥2, min peak intensity 1 × 10⁶, parent ion mass tolerance of ±5 ppm, RT tolerance of 0.25 min and preferred fragments data selection M+H, M−H. The data were normalized by the constant mean algorithm. The metabolite identifications and pathway analysis nodes used the default features in this metabolomics workflow. The entire result table obtained separately for the negative and positive mode data were then exported as excel files and finally combined to generate a single master table with all the metabolites. The LC-MS metabolomics data file was also analyzed using MetaboAnalyts 6.0 to identify all the significantly differential features (87). The differential features that were not identified by CD3.3 were manually screened to find biologically significant identities using their m/z exact mass in the HMDB LC-MS search tool with a mass tolerance for the acceptable identities below 35 ppm of error. Data represent the means ± standard error of the mean. Treated and untreated groups differences were determined using Compound Discoverer 3.3 and MetaboAnalyst 6.0, considered significant for P ≤ 0.05; NS denotes non-significant differences.

ATAC-seq and data analysis

Library preparation

The ATAC-seq library was prepared as previously described (22). In brief, splenic cDC1s from WT and Opa1^ΔDC mice (n = 4 per genotype) or cDC1s isolated from B16-OVA tumors at days 7 and 21 after tumor inoculation (n = 4 per group) 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 previously (22). In brief, the paired-end fastq files obtained from NovaSeq were trimmed for Nextera adapter by trimmomatic (v0.36, paired-end mode, with parameter LEADING:10 TRAILING:10 SLIDINGWINDOW:4:18 MINLEN:25 ILLUMINACAP=2:30:10). BWA (v0.7.16) was used to align reads to mouse genome mm10 with default parameters. PCR duplicated reads [marked by Picard (v2.9.4)], mitochondrial reads, and mm10 blacklist regions were filtered out from the resulted BAM files after alignment. 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 (v2.1.1.20160309, with ‘–extsize 200 –nomodel’ and default parameters if not indicated) 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 WT and OPA1-deficient cDC1s; or intratumoral cDC1s from B16-OVA at days 7 and 21 if they overlapped by 100-bp and then were counted from each of the 8 samples by bedtools (v2.25.0). Transcription factor footprinting activity were inferred and visualized using the RGT HINT software (v0.13.2) (41). Changes in binding activity for transcription factors were estimated with a motif in JASPAR. Transcription factors with less than 1,000 binding sites, which is likely caused by noise, were filtered from analysis. For motif analysis, we selected 1,000 unchanged regions (log₂FC < 0.05 and FDR-adjusted P > 0.5) as controls. Motif scanning was performed using FIMO from the MEME suite (v4.11.3; threshold 1e-4, motif-pseudo 0.0001) (88), with the TRANSFAC 2019 database (Vertebrata only, excluding 3D structure-based motifs). Two-tailed Fisher’s exact test was used to assess motif enrichment in differentially accessible versus control regions.

Gene signature curation, functional enrichment and gene set enrichment analysis

The below gene signatures were curated from public or in-house datasets:

The “cDC1 [TMRM/MG]^hi UP signature” was curated from upregulated genes (log₂FC > 0.5 and FDR < 0.05) in the microarray analysis of splenic [TMRM/MG]^hi cDC1s compared to splenic [TMRM/MG]^lo cDC1s (for the “cDC1 [TMRM/MG]^hi UP signature”). A total of 411 genes composed the “cDC1 [TMRM/MG]^hi UP signature” (table S4).
“Putative NRF1 target genes” signature was curated from the ChIP-seq narrow peak file of motor brain tissue, which was downloaded from GEO under accession number GSE161808 (42). Peak annotation was performed using annotatePeaks.pl function from HOMER (v4.9.1) with default parameters. Genes with peak located at the transcription start site (TSS)-promoter regions were considered as “Putative NRF1 target genes” (contains 2461 genes). Overlapped genes (188 genes in total) (table S7) between “Putative NRF1 target genes” and downregulated genes (log₂FC < −0.5 and FDR < 0.05) in splenic OPA1-deficient cDC1s compared to WT cDC1s (1,500 genes) were evaluated. Two-tailed Fisher’s exact test P values of the overlaps were reported. Enriched pathways for the abovementioned overlapped genes were analyzed by funcEnrich.Fisher function NetBID (v2.0.3).
The “cDC1 OPA1-activated signature” or “cDC1 OPA1-suppressed signature” (table S11) was curated from the top 200 downregulated (log₂FC < −0.5 and FDR < 0.05) or top 200 upregulated (log₂FC > 0.5 and FDR < 0.05) genes in the microarray analysis of splenic OPA1-deficient cDC1s compared to splenic WT cDC1s, respectively.

For gene set enrichment analysis (GSEA) of the microarray datasets, the normalized expression matrix and gene sets were input into GSEA (v4.3.2), and gene sets were ranked based on their enrichment scores calculated using the two-tailed Kolmogorov–Smirnov test with permutation for ‘gene_set’. Hallmark signatures from Molecular Signatures Database (MSigDB) or in-house curated gene signatures, including “cDC1 [TMRM/MG]^hi UP signature”, “Putative NRF1 target genes”, “cDC1 OPA1-activated signature” or “cDC1 OPA1-suppressed signature” (described above) were used. Signatures were considered enriched if FDR < 0.05 and |normalized enrichment score (NES)| > 1.4.

Public scRNA-seq dataset analysis

The raw gene expression matrix and/or metadata were derived from GEO accession number GSE131957 (24) (mouse lung adenocarcinoma dataset), GSE154763 (25) (contains samples from human lymphoma, thyroid carcinoma, pancreatic adenocarcinoma, myeloma, esophageal carcinoma, kidney cancer, ovarian or fallopian tube carcinoma, and uterine corpus endometrial carcinoma), GSE139555 (human non-small cell lung carcinoma) (64) or under EMBL-EBI accession number E-MTAB-8107/6149/6653 (32) (contains samples from human colorectum, ovary, and lung tumors). These scRNA-seq data were analyzed with Seurat (v4.2.0), and cells with abnormally low features or unique molecular identifier (UMI) counts were removed. Cells with abnormally high UMI counts (i.e., potentially multi-cell droplets) or high mitochondrial read percentages (i.e., potentially dead or damaged cells) were also removed if the downloaded gene read count matrix were not preprocessed. Different cancer types within individual dataset were integrated by using Harmony (v1.2.0). Normalization (scale.factor = 1 × 10⁶), feature selection and scaling were performed on the prefiltered RNA matrix with the Seurat R package (v4.2.0), followed by principal component analysis (PCA; Seurat RunPCA function, default settings). A UMAP was generated (Seurat RunUMAP, dims = 1:30), and clustering was performed (Seurat FindNeighbors (dims = 1:30), followed by Seurat FindClusters. If no metadata were provided, mouse cDC1s were defined as Clec9a⁺Xcr1⁺ cells. Human cDC1s were defined as CLEC9A⁺XCR1⁺ cells.

Differential gene expression analysis was performed by comparing “Tumor tissue” to “Normal tissue” using a two-tailed Wilcoxon rank-sum test and Bonferroni correction with the FindMarkers function in Seurat R package (v4.2.0). For GSEA of scRNA-seq, genes were first ranked in an order of descending log₂FC values (derived from the differential expression analysis). Then, pre-ranked GSEA (GseaPreranked) was performed using the Broad GSEA command line software (v4.3.2) against MsigDB signatures or in-house curated signatures described above. Signatures were considered enriched if met FDR ≤ 0.05 and |NES| > 1.4. Activity scores of gene signatures were calculated using the AddModuleScore function implemented in the Seurat package. Two-tailed Wilcoxon rank-sum test was used to calculate the P values for activity scores. The activities of gene signatures were visualized using the violin plots or UMAP generated by VlnPlot and FeaturePlot functions in the Seurat or by ggplot2 R package (v3.3.5).

Statistical analysis for biological experiments

For biological experiment (non-omics) analyses, data were analyzed using Prism 10 software (GraphPad) by two-tailed unpaired Student’s t-test, one-way ANOVA, or two-way ANOVA as indicated in figure legend. Mann–Whitney test or one-way ANOVA (nonparametric) was used for comparing mitochondrial volumes. The log-rank (Mantel–Cox) test was used for comparing mouse survival curves. Two-tailed unpaired Student’s t-test was used for transcription factor footprinting analysis of ATAC-seq data. Two-tailed Wilcoxon rank sum test was applied for differential expression or activity score analysis of scRNA-seq data. P < 0.05 was considered significant. Data are presented as mean ± s.e.m. Age- and sex-matched mice with predetermined genotypes were randomly assigned to control and experimental groups before treatments. No other randomization was performed. Data collection and analysis were not performed blind to the conditions of the experiments. No data were excluded from the analyses.

Supplementary Material

Table S2

NIHMS2161853-supplement-Table_S2.xlsx^{(13.4KB, xlsx)}

Table S3

NIHMS2161853-supplement-Table_S3.xlsx^{(15.3KB, xlsx)}

Table S1

NIHMS2161853-supplement-Table_S1.xlsx^{(73.6KB, xlsx)}

Table S4

NIHMS2161853-supplement-Table_S4.xlsx^{(20.7KB, xlsx)}

Table S5

NIHMS2161853-supplement-Table_S5.xlsx^{(36.7KB, xlsx)}

Table S6

NIHMS2161853-supplement-Table_S6.xlsx^{(65.7KB, xlsx)}

Table S7

NIHMS2161853-supplement-Table_S7.xlsx^{(12.2KB, xlsx)}

Table S8

NIHMS2161853-supplement-Table_S8.xlsx^{(18.4KB, xlsx)}

Table S9

NIHMS2161853-supplement-Table_S9.xlsx^{(60.1KB, xlsx)}

Table S10

NIHMS2161853-supplement-Table_S10.xlsx^{(49KB, xlsx)}

Table S12

NIHMS2161853-supplement-Table_S12.xlsx^{(10KB, xlsx)}

Table S11

NIHMS2161853-supplement-Table_S11.xlsx^{(14.6KB, xlsx)}

Supplementary Materials

NIHMS2161853-supplement-Supplementary_Materials.pdf^{(5.6MB, pdf)}

Reproducibility Checklist

NIHMS2161853-supplement-Reproducibility_Checklist.docx^{(73.5KB, docx)}

Figs. S1 to S12

Tables S1 to S12

MDAR Reproducibility Checklist

Acknowledgments

We acknowledge B. Youngblood, and J. Raynor, H. Song and E. Norton in the Chi lab for critical reading of the manuscript and insigthful feedback. H. Sesaki for Opa1 and Dnm1l floxed mice. M. Hendren and R. Walton for animal colony management and technical support; K. Yu, H. Tan and St. Jude Proteomics and Metabolomics Center for protein profiling; St. Jude Immunology flow cytometry core facility for cell sorting; and the Hartwell Center for microarray profiling.

Funding:

ALSAC (HC)

St. Jude Children’s Research Hospital (HC, CGR)

National Institutes of Health grant R01CA253188 (HC)

National Institutes of Health grant R01AI105887 (HC)

National Institutes of Health grant P30CA021765 (CGR).

Footnotes

Competing interests: H.C. consults for Kumquat Biosciences, Inc. and TCura Bioscience. H.C. and C.G. are listed as inventors on patent application 63/482,616 submitted by St. Jude Children’s Research Hospital that covers the use of glutamine for cancer immunotherapy. The remaining authors declare no competing interests.

Data and Materials availability:

The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information. All microarray and ATAC-seq described in the manuscript have been deposited in the NCBI Gene Expression Omnibus (GEO) database and are accessible through the GEO SuperSeries access number GSE285346. Only standard algorithms were used to perform bioinformatic analyses, and codes are available from the authors upon request.

References

1.Chapman NM, Chi H, Metabolic rewiring and communication in cancer immunity. Cell Chem Biol 31, 862–883 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Anderson DA 3rd, Dutertre CA, Ginhoux F, Murphy KM, Genetic models of human and mouse dendritic cell development and function. Nat Rev Immunol 21, 101–115 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Cabeza-Cabrerizo M, Cardoso A, Minutti CM, Pereira da Costa M, Reis e Sousa C, Dendritic Cells Revisited. Annu Rev Immunol 39, 131–166 (2021). [DOI] [PubMed] [Google Scholar]
4.Bottcher JP et al. , NK Cells Stimulate Recruitment of cDC1 into the Tumor Microenvironment Promoting Cancer Immune Control. Cell 172, 1022–1037 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Sanchez-Paulete AR et al. , Cancer Immunotherapy with Immunomodulatory Anti-CD137 and Anti-PD-1 Monoclonal Antibodies Requires BATF3-Dependent Dendritic Cells. Cancer Discov 6, 71–79 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Salmon H et al. , Expansion and Activation of CD103(+) Dendritic Cell Progenitors at the Tumor Site Enhances Tumor Responses to Therapeutic PD-L1 and BRAF Inhibition. Immunity 44, 924–938 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Mahadevan KK et al. , Type I conventional dendritic cells facilitate immunotherapy in pancreatic cancer. Science 384, eadh4567 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Roberts EW et al. , Critical Role for CD103(+)/CD141(+) Dendritic Cells Bearing CCR7 for Tumor Antigen Trafficking and Priming of T Cell Immunity in Melanoma. Cancer Cell 30, 324–336 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Spranger S, Dai D, Horton B, Gajewski TF, Tumor-Residing Batf3 Dendritic Cells Are Required for Effector T Cell Trafficking and Adoptive T Cell Therapy. Cancer Cell 31, 711–723 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Broz ML et al. , Dissecting the tumor myeloid compartment reveals rare activating antigen-presenting cells critical for T cell immunity. Cancer Cell 26, 638–652 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Meiser P et al. , A distinct stimulatory cDC1 subpopulation amplifies CD8(+) T cell responses in tumors for protective anti-cancer immunity. Cancer Cell 41, 1498–1515 (2023). [DOI] [PubMed] [Google Scholar]
12.Chapman NM, Chi H, Metabolic adaptation of lymphocytes in immunity and disease. Immunity 55, 14–30 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Palsson-McDermott EM, O’Neill LAJ, Targeting immunometabolism as an anti-inflammatory strategy. Cell Res 30, 300–314 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Collins N, Belkaid Y, Control of immunity via nutritional interventions. Immunity 55, 210–223 (2022). [DOI] [PubMed] [Google Scholar]
15.Giovanelli P, Sandoval TA, Cubillos-Ruiz JR, Dendritic Cell Metabolism and Function in Tumors. Trends Immunol 40, 699–718 (2019). [DOI] [PubMed] [Google Scholar]
16.Moller SH, Wang L, Ho PC, Metabolic programming in dendritic cells tailors immune responses and homeostasis. Cell Mol Immunol 19, 370–383 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.You Z, Chi H, Lipid metabolism in dendritic cell biology. Immunol Rev 317, 137–151 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Krawczyk CM et al. , Toll-like receptor-induced changes in glycolytic metabolism regulate dendritic cell activation. Blood 115, 4742–4749 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Everts B et al. , TLR-driven early glycolytic reprogramming via the kinases TBK1-IKKvarepsilon supports the anabolic demands of dendritic cell activation. Nat Immunol 15, 323–332 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Malinarich F et al. , High mitochondrial respiration and glycolytic capacity represent a metabolic phenotype of human tolerogenic dendritic cells. J Immunol 194, 5174–5186 (2015). [DOI] [PubMed] [Google Scholar]
21.Adamik J et al. , Distinct metabolic states guide maturation of inflammatory and tolerogenic dendritic cells. Nat Commun 13, 5184 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Guo C et al. , SLC38A2 and glutamine signalling in cDC1s dictate anti-tumour immunity. Nature 620, 200–208 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Tan H et al. , Integrative Proteomics and Phosphoproteomics Profiling Reveals Dynamic Signaling Networks and Bioenergetics Pathways Underlying T Cell Activation. Immunity 46, 488–503 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Maier B et al. , A conserved dendritic-cell regulatory program limits antitumour immunity. Nature 580, 257–262 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Cheng S et al. , A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell 184, 792–809 (2021). [DOI] [PubMed] [Google Scholar]
26.Gabriel SS et al. , Transforming growth factor-beta-regulated mTOR activity preserves cellular metabolism to maintain long-term T cell responses in chronic infection. Immunity 54, 1698–1714 (2021). [DOI] [PubMed] [Google Scholar]
27.Yu YR et al. , Disturbed mitochondrial dynamics in CD8(+) TILs reinforce T cell exhaustion. Nat Immunol 21, 1540–1551 (2020). [DOI] [PubMed] [Google Scholar]
28.Piot C et al. , Spatial Organisation of Tumour cDC1 States Correlates with Effector and Stem-Like CD8(+) T Cells Location. Eur J Immunol 55, e70011 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Ivics Z, Hackett PB, Plasterk RH, Izsvak Z, Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91, 501–510 (1997). [DOI] [PubMed] [Google Scholar]
30.Ho C et al. , AKT (v-akt murine thymoma viral oncogene homolog 1) and N-Ras (neuroblastoma ras viral oncogene homolog) coactivation in the mouse liver promotes rapid carcinogenesis by way of mTOR (mammalian target of rapamycin complex 1), FOXM1 (forkhead box M1)/SKP2, and c-Myc pathways. Hepatology 55, 833–845 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Du X et al. , Hippo/Mst signalling couples metabolic state and immune function of CD8alpha(+) dendritic cells. Nature 558, 141–145 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Qian J et al. , A pan-cancer blueprint of the heterogeneous tumor microenvironment revealed by single-cell profiling. Cell Res 30, 745–762 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Tabara LC, Segawa M, Prudent J, Molecular mechanisms of mitochondrial dynamics. Nat Rev Mol Cell Biol 26, 123–146 (2025). [DOI] [PubMed] [Google Scholar]
34.Caton ML, Smith-Raska MR, Reizis B, Notch-RBP-J signaling controls the homeostasis of CD8- dendritic cells in the spleen. J Exp Med 204, 1653–1664 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Zhang Z et al. , The dynamin-related GTPase Opa1 is required for glucose-stimulated ATP production in pancreatic beta cells. Mol Biol Cell 22, 2235–2245 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Di Pilato M et al. , CXCR6 positions cytotoxic T cells to receive critical survival signals in the tumor microenvironment. Cell 184, 4512–4530 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Lim SA et al. , Lipid signalling enforces functional specialization of T(reg) cells in tumours. Nature 591, 306–311 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Ferris ST et al. , cDC1 prime and are licensed by CD4(+) T cells to induce anti-tumour immunity. Nature 584, 624–629 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Wculek SK et al. , Dendritic cells in cancer immunology and immunotherapy. Nat Rev Immunol 20, 7–24 (2020). [DOI] [PubMed] [Google Scholar]
40.Sharma P et al. , Hyperglycosylation of prosaposin in tumor dendritic cells drives immune escape. Science 383, 190–200 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Li Z et al. , Identification of transcription factor binding sites using ATAC-seq. Genome Biol 20, 45 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Liu S et al. , NRF1 association with AUTS2-Polycomb mediates specific gene activation in the brain. Mol Cell 81, 4663–4676 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.LaFleur MW et al. , A CRISPR-Cas9 delivery system for in vivo screening of genes in the immune system. Nat Commun 10, 1668 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Jang S, Javadov S, OPA1 regulates respiratory supercomplexes assembly: The role of mitochondrial swelling. Mitochondrion 51, 30–39 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Benegiamo G et al. , The genetic background shapes the susceptibility to mitochondrial dysfunction and NASH progression. J Exp Med 220, e20221738 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Bennett CF, Latorre-Muro P, Puigserver P, Mechanisms of mitochondrial respiratory adaptation. Nat Rev Mol Cell Biol 23, 817–835 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Vercellino I, Sazanov LA, The assembly, regulation and function of the mitochondrial respiratory chain. Nat Rev Mol Cell Biol 23, 141–161 (2022). [DOI] [PubMed] [Google Scholar]
48.Herzig S, Shaw RJ, AMPK: guardian of metabolism and mitochondrial homeostasis. Nat Rev Mol Cell Biol 19, 121–135 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Loi M et al. , Macroautophagy Proteins Control MHC Class I Levels on Dendritic Cells and Shape Anti-viral CD8(+) T Cell Responses. Cell Rep 15, 1076–1087 (2016). [DOI] [PubMed] [Google Scholar]
50.Yamamoto K et al. , Autophagy promotes immune evasion of pancreatic cancer by degrading MHC-I. Nature 581, 100–105 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Xia H, Green DR, Zou W, Autophagy in tumour immunity and therapy. Nat Rev Cancer 21, 281–297 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Zietara N et al. , Immunoglobulins drive terminal maturation of splenic dendritic cells. Proc Natl Acad Sci U S A 110, 2282–2287 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Samie M, Cresswell P, The transcription factor TFEB acts as a molecular switch that regulates exogenous antigen-presentation pathways. Nat Immunol 16, 729–736 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Cebrian I et al. , Sec22b regulates phagosomal maturation and antigen crosspresentation by dendritic cells. Cell 147, 1355–1368 (2011). [DOI] [PubMed] [Google Scholar]
55.Han D et al. , Anti-tumour immunity controlled through mRNA m(6)A methylation and YTHDF1 in dendritic cells. Nature 566, 270–274 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Theisen DJ et al. , WDFY4 is required for cross-presentation in response to viral and tumor antigens. Science 362, 694–699 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Burgdorf S, Kautz A, Bohnert V, Knolle PA, Kurts C, Distinct pathways of antigen uptake and intracellular routing in CD4 and CD8 T cell activation. Science 316, 612–616 (2007). [DOI] [PubMed] [Google Scholar]
58.Dudziak D et al. , Differential antigen processing by dendritic cell subsets in vivo. Science 315, 107–111 (2007). [DOI] [PubMed] [Google Scholar]
59.Liang J et al. , Selective deficiency of mitochondrial respiratory complex I subunits Ndufs4/6 causes tumor immunogenicity. Nat Cancer 6, 323–337 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Titov DV et al. , Complementation of mitochondrial electron transport chain by manipulation of the NAD+/NADH ratio. Science 352, 231–235 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Wakabayashi J et al. , The dynamin-related GTPase Drp1 is required for embryonic and brain development in mice. J Cell Biol 186, 805–816 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Luri-Rey C et al. , Cross-priming in cancer immunology and immunotherapy. Nat Rev Cancer 25, 249–273 (2025). [DOI] [PubMed] [Google Scholar]
63.Bottcher JP, Reis e Sousa C, The Role of Type 1 Conventional Dendritic Cells in Cancer Immunity. Trends Cancer 4, 784–792 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Wu TD et al. , Peripheral T cell expansion predicts tumour infiltration and clinical response. Nature 579, 274–278 (2020). [DOI] [PubMed] [Google Scholar]
65.Ikeda H et al. , Immune evasion through mitochondrial transfer in the tumour microenvironment. Nature 638, 225–236 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Norris BA et al. , Chronic but not acute virus infection induces sustained expansion of myeloid suppressor cell numbers that inhibit viral-specific T cell immunity. Immunity 38, 309–321 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Wculek SK et al. , Effective cancer immunotherapy by natural mouse conventional type-1 dendritic cells bearing dead tumor antigen. J Immunother Cancer 7, 100 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Zhou Y et al. , Vaccine efficacy against primary and metastatic cancer with in vitro-generated CD103(+) conventional dendritic cells. J Immunother Cancer 8, e000474 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Kalbasi A, Ribas A, Tumour-intrinsic resistance to immune checkpoint blockade. Nat Rev Immunol 20, 25–39 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Cubillos-Ruiz JR et al. , ER Stress Sensor XBP1 Controls Anti-tumor Immunity by Disrupting Dendritic Cell Homeostasis. Cell 161, 1527–1538 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Dixon KO et al. , TIM-3 restrains anti-tumour immunity by regulating inflammasome activation. Nature 595, 101–106 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Adamik J et al. , Immuno-metabolic dendritic cell vaccine signatures associate with overall survival in vaccinated melanoma patients. Nat Commun 14, 7211 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Zhivaki D et al. , Correction of age-associated defects in dendritic cells enables CD4(+) T cells to eradicate tumors. Cell 187, 3888–3903 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Ascic E et al. , In vivo dendritic cell reprogramming for cancer immunotherapy. Science 386, eadn9083 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Mangalhara KC et al. , Manipulating mitochondrial electron flow enhances tumor immunogenicity. Science 381, 1316–1323 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Mamedov MR et al. , CRISPR screens decode cancer cell pathways that trigger gammadelta T cell detection. Nature 621, 188–195 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Bezwada D et al. , Mitochondrial complex I promotes kidney cancer metastasis. Nature 633, 923–931 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Martinez-Reyes I et al. , Mitochondrial ubiquinol oxidation is necessary for tumour growth. Nature 585, 288–292 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Sentmanat MF, Peters ST, Florian CP, Connelly JP, Pruett-Miller SM, A Survey of Validation Strategies for CRISPR-Cas9 Editing. Sci Rep 8, 888 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Connelly JP, Pruett-Miller SM, CRIS.py: A Versatile and High-throughput Analysis Program for CRISPR-based Genome Editing. Sci Rep 9, 4194 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Cao J et al. , Low concentrations of metformin suppress glucose production in hepatocytes through AMP-activated protein kinase (AMPK). J Biol Chem 289, 20435–20446 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Wang X et al. , JUMP: a tag-based database search tool for peptide identification with high sensitivity and accuracy. Mol Cell Proteomics 13, 3663–3673 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Bairoch A, Apweiler R, The SWISS-PROT protein sequence database and its supplement TrEMBL in 2000. Nucleic Acids Res 28, 45–48 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Rath S et al. , MitoCarta3.0: an updated mitochondrial proteome now with sub-organelle localization and pathway annotations. Nucleic Acids Res 49, D1541–D1547 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Sherman BT et al. , DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res 50, W216–W221 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Milne SB, Mathews TP, Myers DS, Ivanova PT, Brown HA, Sum of the Parts: Mass Spectrometry-Based Metabolomics. Biochemistry 52, 3829–3840 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
87.Pang Z et al. , MetaboAnalyst 6.0: towards a unified platform for metabolomics data processing, analysis and interpretation. Nucleic Acids Research 52, W398–W406 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
88.Bailey TL et al. , MEME SUITE: tools for motif discovery and searching. Nucleic Acids Res 37, W202–208 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S2

NIHMS2161853-supplement-Table_S2.xlsx^{(13.4KB, xlsx)}

Table S3

NIHMS2161853-supplement-Table_S3.xlsx^{(15.3KB, xlsx)}

Table S1

NIHMS2161853-supplement-Table_S1.xlsx^{(73.6KB, xlsx)}

Table S4

NIHMS2161853-supplement-Table_S4.xlsx^{(20.7KB, xlsx)}

Table S5

NIHMS2161853-supplement-Table_S5.xlsx^{(36.7KB, xlsx)}

Table S6

NIHMS2161853-supplement-Table_S6.xlsx^{(65.7KB, xlsx)}

Table S7

NIHMS2161853-supplement-Table_S7.xlsx^{(12.2KB, xlsx)}

Table S8

NIHMS2161853-supplement-Table_S8.xlsx^{(18.4KB, xlsx)}

Table S9

NIHMS2161853-supplement-Table_S9.xlsx^{(60.1KB, xlsx)}

Table S10

NIHMS2161853-supplement-Table_S10.xlsx^{(49KB, xlsx)}

Table S12

NIHMS2161853-supplement-Table_S12.xlsx^{(10KB, xlsx)}

Table S11

NIHMS2161853-supplement-Table_S11.xlsx^{(14.6KB, xlsx)}

Supplementary Materials

NIHMS2161853-supplement-Supplementary_Materials.pdf^{(5.6MB, pdf)}

Reproducibility Checklist

NIHMS2161853-supplement-Reproducibility_Checklist.docx^{(73.5KB, docx)}

Data Availability Statement

[R1] 1.Chapman NM, Chi H, Metabolic rewiring and communication in cancer immunity. Cell Chem Biol 31, 862–883 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] 2.Anderson DA 3rd, Dutertre CA, Ginhoux F, Murphy KM, Genetic models of human and mouse dendritic cell development and function. Nat Rev Immunol 21, 101–115 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R3] 3.Cabeza-Cabrerizo M, Cardoso A, Minutti CM, Pereira da Costa M, Reis e Sousa C, Dendritic Cells Revisited. Annu Rev Immunol 39, 131–166 (2021). [DOI] [PubMed] [Google Scholar]

[R4] 4.Bottcher JP et al. , NK Cells Stimulate Recruitment of cDC1 into the Tumor Microenvironment Promoting Cancer Immune Control. Cell 172, 1022–1037 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] 5.Sanchez-Paulete AR et al. , Cancer Immunotherapy with Immunomodulatory Anti-CD137 and Anti-PD-1 Monoclonal Antibodies Requires BATF3-Dependent Dendritic Cells. Cancer Discov 6, 71–79 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Salmon H et al. , Expansion and Activation of CD103(+) Dendritic Cell Progenitors at the Tumor Site Enhances Tumor Responses to Therapeutic PD-L1 and BRAF Inhibition. Immunity 44, 924–938 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Mahadevan KK et al. , Type I conventional dendritic cells facilitate immunotherapy in pancreatic cancer. Science 384, eadh4567 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Roberts EW et al. , Critical Role for CD103(+)/CD141(+) Dendritic Cells Bearing CCR7 for Tumor Antigen Trafficking and Priming of T Cell Immunity in Melanoma. Cancer Cell 30, 324–336 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Spranger S, Dai D, Horton B, Gajewski TF, Tumor-Residing Batf3 Dendritic Cells Are Required for Effector T Cell Trafficking and Adoptive T Cell Therapy. Cancer Cell 31, 711–723 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Broz ML et al. , Dissecting the tumor myeloid compartment reveals rare activating antigen-presenting cells critical for T cell immunity. Cancer Cell 26, 638–652 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Meiser P et al. , A distinct stimulatory cDC1 subpopulation amplifies CD8(+) T cell responses in tumors for protective anti-cancer immunity. Cancer Cell 41, 1498–1515 (2023). [DOI] [PubMed] [Google Scholar]

[R12] 12.Chapman NM, Chi H, Metabolic adaptation of lymphocytes in immunity and disease. Immunity 55, 14–30 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Palsson-McDermott EM, O’Neill LAJ, Targeting immunometabolism as an anti-inflammatory strategy. Cell Res 30, 300–314 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.Collins N, Belkaid Y, Control of immunity via nutritional interventions. Immunity 55, 210–223 (2022). [DOI] [PubMed] [Google Scholar]

[R15] 15.Giovanelli P, Sandoval TA, Cubillos-Ruiz JR, Dendritic Cell Metabolism and Function in Tumors. Trends Immunol 40, 699–718 (2019). [DOI] [PubMed] [Google Scholar]

[R16] 16.Moller SH, Wang L, Ho PC, Metabolic programming in dendritic cells tailors immune responses and homeostasis. Cell Mol Immunol 19, 370–383 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.You Z, Chi H, Lipid metabolism in dendritic cell biology. Immunol Rev 317, 137–151 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Krawczyk CM et al. , Toll-like receptor-induced changes in glycolytic metabolism regulate dendritic cell activation. Blood 115, 4742–4749 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Everts B et al. , TLR-driven early glycolytic reprogramming via the kinases TBK1-IKKvarepsilon supports the anabolic demands of dendritic cell activation. Nat Immunol 15, 323–332 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Malinarich F et al. , High mitochondrial respiration and glycolytic capacity represent a metabolic phenotype of human tolerogenic dendritic cells. J Immunol 194, 5174–5186 (2015). [DOI] [PubMed] [Google Scholar]

[R21] 21.Adamik J et al. , Distinct metabolic states guide maturation of inflammatory and tolerogenic dendritic cells. Nat Commun 13, 5184 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Guo C et al. , SLC38A2 and glutamine signalling in cDC1s dictate anti-tumour immunity. Nature 620, 200–208 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] 23.Tan H et al. , Integrative Proteomics and Phosphoproteomics Profiling Reveals Dynamic Signaling Networks and Bioenergetics Pathways Underlying T Cell Activation. Immunity 46, 488–503 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Maier B et al. , A conserved dendritic-cell regulatory program limits antitumour immunity. Nature 580, 257–262 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] 25.Cheng S et al. , A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell 184, 792–809 (2021). [DOI] [PubMed] [Google Scholar]

[R26] 26.Gabriel SS et al. , Transforming growth factor-beta-regulated mTOR activity preserves cellular metabolism to maintain long-term T cell responses in chronic infection. Immunity 54, 1698–1714 (2021). [DOI] [PubMed] [Google Scholar]

[R27] 27.Yu YR et al. , Disturbed mitochondrial dynamics in CD8(+) TILs reinforce T cell exhaustion. Nat Immunol 21, 1540–1551 (2020). [DOI] [PubMed] [Google Scholar]

[R28] 28.Piot C et al. , Spatial Organisation of Tumour cDC1 States Correlates with Effector and Stem-Like CD8(+) T Cells Location. Eur J Immunol 55, e70011 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Ivics Z, Hackett PB, Plasterk RH, Izsvak Z, Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91, 501–510 (1997). [DOI] [PubMed] [Google Scholar]

[R30] 30.Ho C et al. , AKT (v-akt murine thymoma viral oncogene homolog 1) and N-Ras (neuroblastoma ras viral oncogene homolog) coactivation in the mouse liver promotes rapid carcinogenesis by way of mTOR (mammalian target of rapamycin complex 1), FOXM1 (forkhead box M1)/SKP2, and c-Myc pathways. Hepatology 55, 833–845 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Du X et al. , Hippo/Mst signalling couples metabolic state and immune function of CD8alpha(+) dendritic cells. Nature 558, 141–145 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Qian J et al. , A pan-cancer blueprint of the heterogeneous tumor microenvironment revealed by single-cell profiling. Cell Res 30, 745–762 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R33] 33.Tabara LC, Segawa M, Prudent J, Molecular mechanisms of mitochondrial dynamics. Nat Rev Mol Cell Biol 26, 123–146 (2025). [DOI] [PubMed] [Google Scholar]

[R34] 34.Caton ML, Smith-Raska MR, Reizis B, Notch-RBP-J signaling controls the homeostasis of CD8- dendritic cells in the spleen. J Exp Med 204, 1653–1664 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R35] 35.Zhang Z et al. , The dynamin-related GTPase Opa1 is required for glucose-stimulated ATP production in pancreatic beta cells. Mol Biol Cell 22, 2235–2245 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Di Pilato M et al. , CXCR6 positions cytotoxic T cells to receive critical survival signals in the tumor microenvironment. Cell 184, 4512–4530 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Lim SA et al. , Lipid signalling enforces functional specialization of T(reg) cells in tumours. Nature 591, 306–311 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Ferris ST et al. , cDC1 prime and are licensed by CD4(+) T cells to induce anti-tumour immunity. Nature 584, 624–629 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] 39.Wculek SK et al. , Dendritic cells in cancer immunology and immunotherapy. Nat Rev Immunol 20, 7–24 (2020). [DOI] [PubMed] [Google Scholar]

[R40] 40.Sharma P et al. , Hyperglycosylation of prosaposin in tumor dendritic cells drives immune escape. Science 383, 190–200 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R41] 41.Li Z et al. , Identification of transcription factor binding sites using ATAC-seq. Genome Biol 20, 45 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] 42.Liu S et al. , NRF1 association with AUTS2-Polycomb mediates specific gene activation in the brain. Mol Cell 81, 4663–4676 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.LaFleur MW et al. , A CRISPR-Cas9 delivery system for in vivo screening of genes in the immune system. Nat Commun 10, 1668 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] 44.Jang S, Javadov S, OPA1 regulates respiratory supercomplexes assembly: The role of mitochondrial swelling. Mitochondrion 51, 30–39 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] 45.Benegiamo G et al. , The genetic background shapes the susceptibility to mitochondrial dysfunction and NASH progression. J Exp Med 220, e20221738 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] 46.Bennett CF, Latorre-Muro P, Puigserver P, Mechanisms of mitochondrial respiratory adaptation. Nat Rev Mol Cell Biol 23, 817–835 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Vercellino I, Sazanov LA, The assembly, regulation and function of the mitochondrial respiratory chain. Nat Rev Mol Cell Biol 23, 141–161 (2022). [DOI] [PubMed] [Google Scholar]

[R48] 48.Herzig S, Shaw RJ, AMPK: guardian of metabolism and mitochondrial homeostasis. Nat Rev Mol Cell Biol 19, 121–135 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Loi M et al. , Macroautophagy Proteins Control MHC Class I Levels on Dendritic Cells and Shape Anti-viral CD8(+) T Cell Responses. Cell Rep 15, 1076–1087 (2016). [DOI] [PubMed] [Google Scholar]

[R50] 50.Yamamoto K et al. , Autophagy promotes immune evasion of pancreatic cancer by degrading MHC-I. Nature 581, 100–105 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.Xia H, Green DR, Zou W, Autophagy in tumour immunity and therapy. Nat Rev Cancer 21, 281–297 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Zietara N et al. , Immunoglobulins drive terminal maturation of splenic dendritic cells. Proc Natl Acad Sci U S A 110, 2282–2287 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R53] 53.Samie M, Cresswell P, The transcription factor TFEB acts as a molecular switch that regulates exogenous antigen-presentation pathways. Nat Immunol 16, 729–736 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Cebrian I et al. , Sec22b regulates phagosomal maturation and antigen crosspresentation by dendritic cells. Cell 147, 1355–1368 (2011). [DOI] [PubMed] [Google Scholar]

[R55] 55.Han D et al. , Anti-tumour immunity controlled through mRNA m(6)A methylation and YTHDF1 in dendritic cells. Nature 566, 270–274 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Theisen DJ et al. , WDFY4 is required for cross-presentation in response to viral and tumor antigens. Science 362, 694–699 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] 57.Burgdorf S, Kautz A, Bohnert V, Knolle PA, Kurts C, Distinct pathways of antigen uptake and intracellular routing in CD4 and CD8 T cell activation. Science 316, 612–616 (2007). [DOI] [PubMed] [Google Scholar]

[R58] 58.Dudziak D et al. , Differential antigen processing by dendritic cell subsets in vivo. Science 315, 107–111 (2007). [DOI] [PubMed] [Google Scholar]

[R59] 59.Liang J et al. , Selective deficiency of mitochondrial respiratory complex I subunits Ndufs4/6 causes tumor immunogenicity. Nat Cancer 6, 323–337 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R60] 60.Titov DV et al. , Complementation of mitochondrial electron transport chain by manipulation of the NAD+/NADH ratio. Science 352, 231–235 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] 61.Wakabayashi J et al. , The dynamin-related GTPase Drp1 is required for embryonic and brain development in mice. J Cell Biol 186, 805–816 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] 62.Luri-Rey C et al. , Cross-priming in cancer immunology and immunotherapy. Nat Rev Cancer 25, 249–273 (2025). [DOI] [PubMed] [Google Scholar]

[R63] 63.Bottcher JP, Reis e Sousa C, The Role of Type 1 Conventional Dendritic Cells in Cancer Immunity. Trends Cancer 4, 784–792 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] 64.Wu TD et al. , Peripheral T cell expansion predicts tumour infiltration and clinical response. Nature 579, 274–278 (2020). [DOI] [PubMed] [Google Scholar]

[R65] 65.Ikeda H et al. , Immune evasion through mitochondrial transfer in the tumour microenvironment. Nature 638, 225–236 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R66] 66.Norris BA et al. , Chronic but not acute virus infection induces sustained expansion of myeloid suppressor cell numbers that inhibit viral-specific T cell immunity. Immunity 38, 309–321 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R67] 67.Wculek SK et al. , Effective cancer immunotherapy by natural mouse conventional type-1 dendritic cells bearing dead tumor antigen. J Immunother Cancer 7, 100 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R68] 68.Zhou Y et al. , Vaccine efficacy against primary and metastatic cancer with in vitro-generated CD103(+) conventional dendritic cells. J Immunother Cancer 8, e000474 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R69] 69.Kalbasi A, Ribas A, Tumour-intrinsic resistance to immune checkpoint blockade. Nat Rev Immunol 20, 25–39 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R70] 70.Cubillos-Ruiz JR et al. , ER Stress Sensor XBP1 Controls Anti-tumor Immunity by Disrupting Dendritic Cell Homeostasis. Cell 161, 1527–1538 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R71] 71.Dixon KO et al. , TIM-3 restrains anti-tumour immunity by regulating inflammasome activation. Nature 595, 101–106 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R72] 72.Adamik J et al. , Immuno-metabolic dendritic cell vaccine signatures associate with overall survival in vaccinated melanoma patients. Nat Commun 14, 7211 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R73] 73.Zhivaki D et al. , Correction of age-associated defects in dendritic cells enables CD4(+) T cells to eradicate tumors. Cell 187, 3888–3903 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R74] 74.Ascic E et al. , In vivo dendritic cell reprogramming for cancer immunotherapy. Science 386, eadn9083 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R75] 75.Mangalhara KC et al. , Manipulating mitochondrial electron flow enhances tumor immunogenicity. Science 381, 1316–1323 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R76] 76.Mamedov MR et al. , CRISPR screens decode cancer cell pathways that trigger gammadelta T cell detection. Nature 621, 188–195 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R77] 77.Bezwada D et al. , Mitochondrial complex I promotes kidney cancer metastasis. Nature 633, 923–931 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R78] 78.Martinez-Reyes I et al. , Mitochondrial ubiquinol oxidation is necessary for tumour growth. Nature 585, 288–292 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R79] 79.Sentmanat MF, Peters ST, Florian CP, Connelly JP, Pruett-Miller SM, A Survey of Validation Strategies for CRISPR-Cas9 Editing. Sci Rep 8, 888 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R80] 80.Connelly JP, Pruett-Miller SM, CRIS.py: A Versatile and High-throughput Analysis Program for CRISPR-based Genome Editing. Sci Rep 9, 4194 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R81] 81.Cao J et al. , Low concentrations of metformin suppress glucose production in hepatocytes through AMP-activated protein kinase (AMPK). J Biol Chem 289, 20435–20446 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R82] 82.Wang X et al. , JUMP: a tag-based database search tool for peptide identification with high sensitivity and accuracy. Mol Cell Proteomics 13, 3663–3673 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R83] 83.Bairoch A, Apweiler R, The SWISS-PROT protein sequence database and its supplement TrEMBL in 2000. Nucleic Acids Res 28, 45–48 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R84] 84.Rath S et al. , MitoCarta3.0: an updated mitochondrial proteome now with sub-organelle localization and pathway annotations. Nucleic Acids Res 49, D1541–D1547 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R85] 85.Sherman BT et al. , DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res 50, W216–W221 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R86] 86.Milne SB, Mathews TP, Myers DS, Ivanova PT, Brown HA, Sum of the Parts: Mass Spectrometry-Based Metabolomics. Biochemistry 52, 3829–3840 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R87] 87.Pang Z et al. , MetaboAnalyst 6.0: towards a unified platform for metabolomics data processing, analysis and interpretation. Nucleic Acids Research 52, W398–W406 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R88] 88.Bailey TL et al. , MEME SUITE: tools for motif discovery and searching. Nucleic Acids Res 37, W202–208 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Mitochondrial metabolism and signaling direct dendritic cell function in antitumor immunity

Zhiyuan You

Jiyeon Kim

Chuansheng Guo

Haoran Hu

Nicole M Chapman

Xiaoxi Meng

Hao Shi

Yan Wang

Cliff Guy

Anil KC

Jia Li

Jordy Saravia

Gustavo Palacios

Sherri Rankin

Camenzind G Robinson

Hongbo Chi

Abstract

Introduction

Results

Discrete mitochondrial states distinguish functional subpopulations of cDC1s

Fig. 1. OPA1 orchestrates discrete mitochondrial states of cDC1s and their function in activation of antitumor CD8+ T cell responses.

OPA1 orchestrates intratumoral cDC1 antitumor responses

OPA1–NRF1 axis-mediated OXPHOS supports cDC1 immunogenic function

Fig. 2. OPA1–NRF1 axis-mediated OXPHOS promotes cDC1 antigen presentation.

OPA1 and NRF1 support cDC1 bioenergetics and suppress autophagy/lysosome-mediated MHC-I and antigen degradation

Fig. 3. Mitochondrial OXPHOS-driven autophagy inhibition and NAD+ regeneration contribute to OPA1-mediated cDC1 functional fitness.

OPA1-mediated NAD+/NADH balance partially contributes to cDC1 functional fitness

Balanced mitochondrial fusion and fission contribute to cDC1 antitumor immunity

Intratumoral cDC1s decrease mitochondrial membrane potential and volume during tumor progression, associated with downregulation of OPA1 and NRF1 levels

Fig. 4. Intratumoral cDC1s experience a progressive decline of mitochondrial membrane potential and volume during tumor progression, associated with downregulation of OPA1 and NRF1 levels.

[TMRM/MG]hi cDC1s synergize with ICB to elicit durable antitumor immunity

Discussion

Materials and Methods:

Mice

Measurement of genome editing efficiency

Cell purification and culture

LCMV clone 13 infection

Metabolic assays via Seahorse

Flow cytometry

Tumor models and treatments

cDC1 adoptive transfer and ICB treatment

Adoptive T cell transfer

Antigen presentation assays and ELISA

Immunoblot analyses

ATP, ADP, AMP, NAD+ and NADH measurements

DQ-OVA degradation assay and antigen uptake in vitro

Immunofluorescence

Electron microscopy

RNA isolation, gene expression profiling and microarray analysis

Proteomics

Metabolomics

ATAC-seq and data analysis

Library preparation

Data analysis

Gene signature curation, functional enrichment and gene set enrichment analysis

Public scRNA-seq dataset analysis

Statistical analysis for biological experiments

Supplementary Material

Acknowledgments

Funding:

Footnotes

Data and Materials availability:

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Fig. 1. OPA1 orchestrates discrete mitochondrial states of cDC1s and their function in activation of antitumor CD8⁺ T cell responses.

Fig. 3. Mitochondrial OXPHOS-driven autophagy inhibition and NAD⁺ regeneration contribute to OPA1-mediated cDC1 functional fitness.

OPA1-mediated NAD⁺/NADH balance partially contributes to cDC1 functional fitness

[TMRM/MG]^hi cDC1s synergize with ICB to elicit durable antitumor immunity

ATP, ADP, AMP, NAD⁺ and NADH measurements