Dendritic cell-targeted therapy expands CD8 T cell responses to bona-fide neoantigens in lung tumors

Abstract

Cross-presentation by type 1 DCs (cDC1) is critical to induce and sustain antitumoral CD8 T cell responses to model antigens, in various tumor settings. However, the impact of cross-presenting cDC1 and the potential of DC-based therapies in tumors carrying varied levels of bona-fide neoantigens (neoAgs) remain unclear. Here we develop a hypermutated model of non-small cell lung cancer in female mice, encoding genuine MHC-I neoepitopes to study neoAgs-specific CD8 T cell responses in spontaneous settings and upon Flt3L + αCD40 (DC-therapy). We find that cDC1 are required to generate broad CD8 responses against a range of diverse neoAgs. DC-therapy promotes immunogenicity of weaker neoAgs and strongly inhibits the growth of high tumor-mutational burden (TMB) tumors. In contrast, low TMB tumors respond poorly to DC-therapy, generating mild CD8 T cell responses that are not sufficient to block progression. scRNA transcriptional analysis, immune profiling and functional assays unveil the changes induced by DC-therapy in lung tissues, which comprise accumulation of cDC1 with increased immunostimulatory properties and less exhausted effector CD8 T cells. We conclude that boosting cDC1 activity is critical to broaden the diversity of anti-tumoral CD8 T cell responses and to leverage neoAgs content for therapeutic advantage.

Fig. 1. Increasing neoAgs in KP drives cDC1-dependent anti-tumoral responses to neoAgs.

A KP cells were transiently transfected with CAS9 and sgRNA targeting the Mlh1 locus to generate Mlh1-deficient cells, or CAS9 alone (KP^ctrl) (Created with BioRender.com). B, C 5 × 10⁵ KP^ctrl or KP^neo tumor cells were implanted subcutaneously, and tumor outgrowth was measured at the indicated time points. B Mice were treated with anti-CD8 antibodies or isotype control (IgG) (n = 5, data are from one out of the two independent experiments). C Tumor cells were implanted into wild-type or Batf3^−/− (n = 6, data are from one out of two independent experiments). D The Venn diagram represents the number of shared and unique neoantigens expressed at mRNA level in KP^ctrl and KP^neo cells. neoAgs commonly expressed in KP^ctrl and KP^neo (shared, sh) and neoAgs expressed de-novo in KP^neo (unique, neo) are plotted according to expression levels (TPM) and affinity for MHC class-I (IC₅₀). E, F At day 12 after challenge, splenic CD8 T from KP^ctrl or KP^neo tumor-bearing mice were restimulated with selected unique (E) or shared (F) individual neoAgs and tested for IFN-g by ELISpot. Individual dots are technical replicates from one representative experiment (pooled CD8 T cells from 3 mice), out of three performed. G KP^ctrl or KP^neo tumor tissues were harvested at day 26 and labeled with sh1 MHC class I-specific dextramers to identify neoAg-specific CD8⁺ T cells. Data from 1 experiment with 6 animals for KP^ctrl and 7 animals for KP^neo group. H) Wild type or Batf3^−/− mice were challenged with KP^neo tumors. At day 12, CD8 T cells were isolated from the spleen to test reactivity against selected shared and unique peptides by IFN-γ ELISpot. Individual dots are technical replicates from one representative experiment (pooled CD8 T cells from 3 mice), out of three performed. Two-way ANOVA followed by Sidak’s post-test in (B), or Tukey’s post-test in C, two-tailed Mann-Whitney U test in (G). All data are plotted as mean ± SEM. Source data are provided as a Source Data File.

Fig. 2. Flt3L/αCD40 therapy induces regression of KP^neo tumors and has a mild impact on KP^ctrl tumors.

A Scheme of anti-PD-L1 (αPD-L1) or IgG treatment in KP^ctrl or KP^neo tumors (left). Tumor outgrowth (right). B Flt3L and anti-CD40 (FL/αCD40) therapy or IgG administration scheme in KP^ctrl and KP^neo tumors (left). Tumor outgrowth (right). C Growth of individual tumors showed in (B), indicating the fraction of rejected tumors in each group. D Frequencies of effector/effector memory (CD44⁺/CD62L^-) CD8 T cells in the tdLN at day 12 after tumor challenge. E Frequencies (left) and absolute numbers (right) of tumor-infiltrating CD8 T cells at day 12. F Percentage (left) and absolute numbers (right) of tumor-infiltrating CD8 T cells at the endpoint (day 21). G Tumor CD8/Treg ratio at day 21. A–G Data from one out of two independent experiments is presented (n = 6 mice per group, per experiment). H, I IFN-γ ELISpot showing the specificities of CD8 T cells under FL/αCD40 therapy, induced by KP^ctrl tumors and tested for shared peptides (H) or by KP^neo tumors and tested for unique neoAgs (I). n = 3, data represent spots from pooled CD8 T cells from 3 mice per group, one out of three independent experiments. Two-way ANOVA followed by Tukey’s post-test in A-G; two-tailed Multiple t-test with Holm-Sidak correction method in (H, I). All data are plotted as mean ± SEM. Source data are provided as a Source Data File.

Fig. 3. The impact of FL/αCD40 therapy requires Batf3 cDC1.

A, B KP^ctrl (A) and KP^neo (B) were implanted in wild-type or Batf3^−/− animals and treated with FL/αCD40 as depicted in 2 A. Tumor outgrowth (n = 4 mice per group). C, D Absolute numbers of tumor-infiltrating CD8 T cells at day 12 (n = 3 mice per group) and 21 (n = 4 mice per group), respectively. E Relative expression of the indicated genes (RT-qPCR) in KP^ctrl and KP^neo tumors tissues upon treatment with FL/αCD40, in WT and Batf3^−/− mice. Heatmap showing the z-score of the indicated genes (n = 4). F IFN-γ ELISpot showing the specificities of CD8 T cells to unique peptides, induced by KP^neo tumors under DC-therapy in WT and Batf3^−/− mice, respectively. n = 3, data represent spots from pooled CD8 T cells from 3 mice per group. Two-way ANOVA followed by Sidak’s post-test in (A–C); or Tukey’s post-test in D; two-tailed Multiple t-test with Holm-Sidak correction method in F. All data are plotted as mean ± SEM and represent one out of two independent experiments. Source data are provided as a Source Data File.

Fig. 4. cDC1 positively correlates with effector CD8 T cells in LUAD across a range of TMB values.

A TMB distribution across lung cancer patients in the TCGA-LUAD PanCancer Atlas dataset. The median value was used to stratify patients as TMB-low or TMB-high. B Correlation between cDC1 and CD8 effector T cell scores. Correlation analysis was performed by two-tailed Pearson correlation coefficient (r). A, B Total number of patients=506; where, TMB-high=253 and TMB-low=252. C Heatmap showing the expression of the cDC1 signature genes across the dataset (z- score average for the indicated genes). cDC1 bottom and top quartiles are highlighted (n = 127 each). D Violin plots showing CD8 effector T cell scores for bottom and top cDC1 quartiles within TMB-low and TMB-high groups. Data represent median (continuous line) and interquartile (dashed lines) (For TMB-low and cDC1 bottom group n = 54, TMB-low and cDC1 top n = 79, TMB-high and cDC1 bottom n = 73, and TMB-high and cDC1 top n = 48). Statistical analysis was performed by two-way ANOVA followed by Tukey’s posttest. E Prognostic values of the cDC1 signature for cancer patient overall survival (on months) comparing bottom and top cDC1 quartiles in the TMB-low and TMB-high groups (TMB-low, n = 127; TMB-high, n = 127). Survival curves were compared using the Log-rank test (Mantel-Cox). Source data are provided as a Source Data File.

Fig. 5. FL/αCD40 therapy triggers CD8 T cell responses and tumor regression in KPneo lung orthotopic tumors.

A KP^ctrl and KP^neo tumors were implanted orthotopically in the lung and treated as depicted in the scheme. B Quantification of tumor burden 9 days after tumor challenge. Tumor burden was calculated as the ratio between tumor nodules and total lung area (n = 4). C Absolute numbers of cDC1 in tumor-bearing lungs with KP^ctrl or KP^neo tumors upon FL/αCD40 or IgG quantified by FC (n = 4). D Representative tissue cryosections showing localization of cDC1 within KP^neo tumor nodules (dotted lines) in XCR1-Venus mice. Representative images from one out of four animals per group, from one independent experiment out of three performed. Scale bars represent 50 µm. E Frequencies (left) and absolute numbers (right) of CD8 T cells in lung tissues were quantified by FC (n = 4). F Representative IHC images of lung tissues labeled with anti-CD8 antibodies, from one out of four animals per group, from one independent experiment out of three performed. Insets depict CD8 T cells infiltrating tumor nodules. Scale bars represent 50 µm. G Quantification of the number of CD8⁺ cells infiltrating lung nodules showed in (F) (n = 4). H, I Relative frequencies of PD1⁺ CD8⁺ T cells (H) and granzyme B (GzmB)-producing CD8 T cells (I) in lungs of animals carrying KP^ctrl or KP^neo tumors, treated with therapy or IgG (n = 4). J Relative frequencies of IFN-γ-producing CD8 T cells in the lungs of the indicated groups were measured after ex-vivo restimulation and ICS (n = 4). Two-way ANOVA followed by Tukey’s post-test in (B, E, G, H, I, J); or Sidak’s post-test in (C). All data are plotted as mean ± SEM, and represent one out of two independent experiments. Source data are provided as a Source Data File.

Fig. 6. DC therapy relieves exhaustion of CD8⁺ T cells in lung orthotopic tumors.

Mice were implanted with orthotopic KP^neo tumors and treated with FL/αCD40 or IgG, CD45⁺ cells were isolated from the lungs at day 9 post-tumor challenge and subjected to scRNA-seq. A Dot plot showing scaled gene expression of selected genes defining CD8⁺ T clusters. B Proportion of CD8 T-cell clusters. C Heatmaps showing log₂FC of selected differentially expressed genes (DEG) in FL/αCD40 vs. IgG in C3, C5 and C6 clusters (average log₂FC FL/αCD40 vs. IgG FDR < 0.01). D Violin plots show combined mean expression values for the indicated genes (score) for the cytotoxic T cell markers (Gzma, Gzmb, Prf1, Tnf, Ifnγ, Cxcr6) or exhaustion markers (Pdcd1, Ctla4, Tigit, Lag3, Havcr2, Tox, Nrp1) in cells of the indicated clusters in FL/αCD40 vs. IgG. A–D scRNAseq data correspond to a pool of 4 animals per group. E Absolute numbers of Ki67⁺ CD8 T cells evaluated by ICS on KP^ctrl- and KP^neo-bearing lungs (n = 4). F, G Representative FC plots and quantification of the frequencies of TCF-1⁺PD-1⁺ (F, n = 4 mice) and Tim3⁺Lag3⁺ (G, n = 5 mice) CD8⁺ T cells. Two-way ANOVA followed by Tukey’s post-test in (E, F, G); two-tailed Mann-Whitney U test in (D). Data are plotted as mean ± SEM, E–G represent one out of two independent experiments.

Fig. 7. DC-therapy induces expansion and functional rescue of cross-presenting cDC1 in lung tissues.

KP^neo tumors were implanted orthotopically in the lung and treated with FL/αCD40 or IgG as depicted in 5 A. CD45⁺ cells were isolated at day 9 after tumor challenge and analyzed by scRNA-seq. A Dot plot showing scaled gene expression of selected genes defining DC clusters. B Heatmap showing the normalized expression of selected genes across different DCs clusters and the association of individual genes to the indicated processes. C Violin plots show combined mean expression values for the indicated genes (score) for the exhaustion markers (Axl, Ccl19, Ccl22, Aldh1) in cells of the indicated clusters in FL/αCD40 vs. IgG. A–C scRNAseq data correspond to a pool of 4 animals per group. D Representative FC plots and quantification of the frequencies and absolute numbers of Venus⁺ cDC1 cells. E Representative FC plots and quantification of the frequencies and absolute numbers of Ki67⁺ cDC1 cells. F Absolute numbers of IL-12⁺ cDC1 cells in KP^neo treated with FL/αCD40 or IgG. G, H Representative histograms and quantification of PD-L1 or Tim3 expression in cDC1 subsets, respectively. D–H For IgG group n = 4 mice, for FL/αCD40 n = 3 mice, one out of two independent experiments. I Scheme describing cross-presentation assay, where sorted cDC1 from KP^neo tumor-bearing lungs were pulsed with SIINFEKL ex vivo and co-cultured with CTV-labeled OT-I T cells (Created with BioRender.com). J Representative histograms and quantification of OT-I proliferation. K IFN-γ production by OT-I after 48 h of co-culture. J, K n = 3, data represent experimental replicates from pooled sorted cDC1 from 3 mice per group, per experiment, one out of two independent experiments. two-tailed Mann-Whitney U test in (C); two-tailed Multiple t tests in (D–K). Data are plotted as mean ± SEM, represent one out of two independent experiments.