BDCA1+ cDC2s, BDCA2+ pDCs and BDCA3+ cDC1s reveal distinct pathophysiologic features and impact on clinical outcomes in melanoma patients

Abstract

Objectives: Dendritic cells play a pivotal but still enigmatic role in the control of tumor development. Composed of specialised subsets (cDC1s, cDC2s, pDCs), DCs are critical in triggering and shaping antitumor immune responses. Yet, tumors exploit plasticity of DCs to subvert their functions and escape from immune control. This challenging controversy prompted us to explore the pathophysiological role of cDCs and pDCs in melanoma, where their precise and coordinated involvement remains to be deciphered.

Methods: We investigated in melanoma patients the phenotypic and functional features of circulating and tumor-infiltrating BDCA1⁺ cDC2s, BDCA2⁺ pDCs and BDCA3⁺ cDC1s and assessed their clinical impact.

Results: Principal component analyses (PCA) based on phenotypic or functional parameters of DC subsets revealed intra-group clustering, highlighting specific features of DCs in blood and tumor infiltrate of patients compared to healthy donors. DC subsets exhibited perturbed frequencies in the circulation and actively infiltrated the tumor site, while harbouring a higher activation status. Whereas cDC2s and pDCs displayed an altered functionality in response to TLR triggering, circulating and tumor-infiltrating cDC1s preserved potent competences associated with improved prognosis. Notably, the proportion of circulating cDC1s predicted the clinical outcome of melanoma patients.

Conclusion: Such understanding uncovers critical and distinct impact of each DC subset on clinical outcomes and unveils fine-tuning of interconnections between DCs in melanoma. Elucidating the mechanisms of DC subversion by tumors could help designing new therapeutic strategies exploiting the potentialities of these powerful immune players and their cross-talks, while counteracting their skewing by tumors, to achieve immune control and clinical success.

Keywords: cDC1; cDC2; immune subversion; melanoma; pDC; prognosis factor.

Figure 1

Decreased frequencies of circulating DC subsets in melanoma patients and infiltration level of the tumor site determine the clinical outcome of patients. PBMC and tumor‐infiltrating cells from melanoma patients together with PBMC from HD and control tissues were labelled with specific antibodies allowing depicting the three DC subsets and submitted to flow cytometry analysis. (a) Comparative frequencies of BDCA1⁺ cDC2s, BDCA2⁺ pDCs and BDCA3⁺ cDC1s within alive CD45⁺ cells on the blood of healthy donors (HD, open circles, n = 56 to 67) and patients (Pt, filled circles, n = 17), non‐tumor tissue (tonsils, open triangles, n = 9) and tumor infiltrate of melanoma patients (filled triangles, n = 23). Results are expressed as percentages of positive cells. Bars indicate mean. P‐values were calculated using Mann–Whitney (dashed lines) and Kruskal–Wallis (straight lines) nonparametric tests. * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001. (b) Relative proportions of each DC subsets within all DCs in patients’ blood (n = 17) and tumor infiltrates (n = 23). Bars indicate mean. P‐values were calculated using the Mann–Whitney test. (c) Correlation matrix between the three DC subsets frequencies in HD blood (left panel), patient blood (middle panel) and tumor infiltrate (right panel). Spearman correlations r factors with their significant P‐values (< 0.05) after Bonferroni–Holm’s correction are noted within the squares. (d) Comparative OS (from diagnostic time – left panel) and PFS (from sampling time – right panel) of patients with low or high circulating pDCs or cDC1s, respectively. Groups were separated according to the median percentage of circulating pDCs (0.196%, n = 8 or 9 patients/group) or cDC1s (0.016%, n = 7–10 patients/group). (e) Comparative PFS (from sampling time) of patients with low or high tumor‐infiltrating cDC2s. Groups were separated according to the median percentage of infiltrating cDC2s (0.244%, n = 12 patients/group). (d, e) Comparison using log‐rank test.

Figure 2

Peripheral and tumor‐infiltrating DC subsets from melanoma patients displayed an overall activated basal status. The expression of the co‐activation molecules CD80, CD40 and CD86 on DC subsets was analysed by flow cytometry on PBMCs and tumor‐infiltrating cells of melanoma patients, HD or non‐tumor tissue controls. (a) Expression levels of the co‐stimulatory molecules CD80, CD40 and CD86 on the three DC subsets from the blood of healthy donors (HD, open circles, n = 22) and melanoma patients (Pt, filled circles, n = 17), tumor infiltrates of melanoma patients (filled triangles, n = 23) and non‐tumor tissues (tonsils, open triangles, n = 9). Results are expressed as percentages of positive cells within the corresponding DC subset. Bars indicate mean. P‐values were calculated using Mann–Whitney (dashed lines) and Kruskal–Wallis with post hoc Dunn’s multiple comparison (stars) nonparametric tests. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. (b) Comparative PFS (from diagnosis time) of patients with low or high tumor‐infiltrating CD80⁺ cDC2s (left panel), CD86⁺ pDCs (middle panel), and comparative OS (from diagnosis time) of patients with low or high tumor‐infiltrating CD40⁺ cDC1s (right panel). Groups were separated using the median percentage of tumor‐infiltrating CD80⁺ cDC2s (34.96%), CD86⁺ pDCs (80.15%) and CD40⁺ cDC1s (93.42%) (n = 10–12 patients/group) among each DC subset, respectively. Comparison using log‐rank test.

Figure 3

The frequency and activation status of circulating and tumor‐infiltrating DC subsets in melanoma patients allowed their distinct clustering from healthy donors. (a) Heat map based on the expression of co‐stimulatory molecule (CD80, CD40, CD86) by the three DC subsets in each sample type (HD blood, patient blood and tumor infiltrate). (b) Principal component analysis (PCA) based on DC subsets co‐stimulatory molecules expression for the three DC subsets (left panel) or only cDC2s (right panel – including graph of variables). (c) Correlation matrix between the three DC subsets expressing co‐stimulatory molecules in HD blood (left panel), patient blood (middle panel) and tumor infiltrate (right panel). Spearman correlations with significant P‐values (< 0.05) after Bonferroni–Holm’s correction are circled in black.

Figure 4

Circulating and tumor‐infiltrating BDCA1⁺ cDC2s and BDCA2⁺ pDCs from melanoma patients displayed defective maturation after TLR stimulation. Cell suspensions from blood (HD, n = 17; Pt, n = 15) or tumor infiltrates (Pt, n = 14) were stimulated or not for 21h with or without TLR ligands (polyI:C, R848 or CpG_A) alone or mixed together (mix), and the expression of the co‐stimulatory molecules CD80, CD40 and CD86 was measured on BDCA1⁺ cDC2s and BDCA2⁺ pDCs using flow cytometry. Results are expressed as percentages of positive cells within the corresponding subset. Bars indicate mean. Stars indicate significant differences compared to the control condition without stimulation (−) from each group. P‐values were calculated using Mann–Whitney tests (dashed lines) and Kruskal–Wallis with post hoc Dunn’s multiple comparison (stars) nonparametric tests. * P ≤ 0.05, *** P ≤ 0.001.

Figure 5

Upon TLR triggering, TNFα and IFNα productions by BDCA1⁺ cDC2s and BDCA2⁺ pDCs, respectively, from blood and tumor were impaired, whereas IFNλ1 and TNFα productions by circulating and tumor‐infiltrating BDCA3⁺ cDC1s remained fully functional in the context of melanoma. Cell suspensions from blood (HD, n = 15, open circles; Pt, n = 17, filled circles) or tumor infiltrates (Pt, n = 16, filled triangles) were stimulated for 5h with or without TLR‐L (polyI:C, R848 or CpG_A) alone or mixed together, and the production of cytokines was evaluated by intracellular cytokine staining using flow cytometry. Results are expressed as percentages of cytokine‐expressing cells within the corresponding DC subset. Bars indicate mean. Stars indicate significant difference with control without stimulation from each group. P‐values were calculated using Mann–Whitney (dashed lines) and Kruskal–Wallis with post hoc Dunn’s multiple comparison (stars) nonparametric tests. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Figure 6

High productions of IL12p40/p70, TNFα and IFNλ1 by circulating and tumor‐infiltrating cDC2s, pDCs and cDC1s after TLR stimulation positively impacted melanoma patients’ clinical evolution. Cell suspensions from blood (HD, n = 15; Pt, n = 17) or tumor infiltrates (Pt, n = 16) were stimulated for 5h with or without TLR‐L (polyI:C, R848 or CpG_A) alone or mixed together, and the production of cytokines was evaluated by intracellular cytokine staining using flow cytometry. The proportions of cytokine‐expressing cells were correlated with the clinical parameters of the corresponding patients. (a) Comparative PFS (from diagnosis time – left panel) and OS (from diagnosis time – middle and right panels) of patients with low or high circulating IL12p40/p70⁺ cDC2s and IFNλ1 + cDC1s in absence of ex vivo stimulation, and TNFα⁺ pDCs after R848 stimulation. Groups were separated using the median percentage of the corresponding parameters (IL12p40/p70⁺ cDC2s: 10.58%, TNFα+ pDCs: 29.51%, and IFNλ1⁺ cDC1s :1.22% (n = 7 or 9 patients/group). (b) Comparative PFS (from diagnosis or sampling time) of patients with low or high TNFα⁺ cDC2s or pDCs and IFNλ1⁺ or TNFα⁺ cDC1s after stimulation of tumor‐infiltrating cells with mix TLR‐L or polyI:C, respectively. Groups were separated using the median percentage of tumor‐infiltrating TNFα⁺ cDC2s (22.80%) or pDCs (45.43%) after TLR‐L mix stimulation, and IFNλ1⁺ (36.23%) or TNFα⁺ (18.62%) cDC1s after polyI:C stimulation (n = 3 or 7 patients/group). (a, b) Comparison using log‐rank test. (c) Heat map based on intracellular cytokine expressions (IL‐12p40/p70, IFNα, IFNλ1, TNFα) by the three DC subsets following stimulation or not with the mixture of TLR‐L (polyI:C, R848 and CpG_A) in each sample type (HD blood, patient blood and tumor infiltrate). (d) PCA based on of intracellular cytokine expressions (IL‐12p40/p70, IFNα, IFNλ1, TNFα) by the three DC subsets after 5h of culture with or without TLR triggering (polyI:C, R848 or CpG_A alone or mixed together) in HD blood, patient blood and tumor infiltrates.

Figure 7

Enhanced secretions of IL12p70, type I and III IFNs both with and without ex vivo TLR triggering arised from circulating and tumor‐infiltrating cells of melanoma patients. Cell suspensions from blood (HD, n = 18, open circles; Pt, n = 15, filled circles) or tumor infiltrates (Pt, n = 15, filled triangles) were stimulated for 20 h with or without TLR ligands (polyI:C, R848 or CpG_A) alone or mixed together, and the culture supernatants were examined for the presence of IL‐12p70, IFNα, IFNβ, IFNλ1 and IFNλ2 by Luminex technology. Results are expressed in pg mL^–1. Bars indicate mean. Stars indicate significant differences of the stimulated conditions compared to unstimulated ones within each group. P‐values were calculated using Mann–Whitney (dashed lines) and Kruskal–Wallis with post hoc Dunn’s multiple comparison (stars) nonparametric tests. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Figure 8

Graphical summary of the main features of circulating and tumor‐infiltrating DC subsets in melanoma patients and their impact on clinical outcome. The left part of the figure displays phenotypic and functional perturbations observed in both blood (upper part) and tumor (bottom part) of melanoma patients when compared to healthy donors. The right part of the figure highlights DC‐based prognosis factors of clinical evolution. This figure has been created with BioRender science illustration tool.