The 27 samples of patients with OPSCC that were studied here were part of a larger observational study (P07-112) investigating cellular immunity in patients with head and neck cancer.3 The study was conducted in accordance with the Declaration of Helsinki and approved by the local medical ethical committee of the Leiden University Medical Center and in agreement with the Dutch law. HPV typing and p16^ink4a immunohistochemistry (IHC) staining was performed on formalin-fixed paraffin-embedded (FFPE) tumor sections as described previously.6 An overview of patient characteristics is given in online supplementary table 1.

Healthy donor buffy coats (Sanquin, Amsterdam, The Netherlands), heparinized OPSCC blood and fresh tumor samples were obtained and handled as described previously.3 8 9 In brief, venous blood samples were drawn prior to surgery, and processed within 6 hours. Peripheral blood mononuclear cells (PBMCs) were isolated using ficoll density-gradient centrifugation according to standard operating procedures. OPSCC tumors were cut into small pieces and used for single cell dissociation by enzymatic digestion and to obtain tumor-infiltrating lymphocytes (TILs). Obtained single cell tumor cell suspensions and TIL bulk cultures were cryopreserved and stored in the vapor phase of liquid nitrogen until further use.

Cultured TILs were analyzed for the presence of HPV16-specific T cells as described previously.3 In brief, reactivity of the cultured TILs were tested against autologous monocytes loaded with HPV16 E6/E7 synthetic long peptide (22-mers with 14 amino acids overlap) in a 5-day [3 hour]-thymidine-based proliferation and cytokine production assay. Proliferation was measured by [3 hour]-thymidine incorporation during the last 16 hours. Antigen-specific cytokine production was measured by using a cytometric bead array (CBA, Th1/Th2 kit, BD Bioscience, Breda, The Netherlands). The cut-off value for cytokine production was 20 pg/mL, except for IFNγ for which it was 100 pg/mL. For the proliferation assay, a response was considered positive when the stimulation index, calculated by dividing the average counts of the test wells by the average counts of the medium control wells, was at least 3. Positive cytokine production was defined as at least a twofold increase compared with that of unstimulated cells. Patients were considered immune response-positive (IR⁺) when a positive response was detected in the cultured TILs by antigen-specific proliferation and/or cytokine production.

Characterization of myeloid infiltration was first carried out with triple immunofluorescent staining in 20 OPSCC tumors as described previously.6 8 In brief, following deparaffinization and antigen retrieval (Tris-EDTA buffer, pH 9.0), FFPE tissue sections were stained using anti-CD14, anti-CD33 and anti-CD163 and isotype-specific secondary fluorescent antibodies goat-anti-mouse IgG2a Alexa 488, goat-anti-mouse IgG2b Alexa 546 and goat-anti-mouse IgG1 Alexa 647 (all from Molecular probes; 1:200; see also online supplementary table 2). Nuclear counterstain was obtained with 4′,6-diamidino-2-phenylindole (DAPI). Based on the morphology of cancer cell nests and autofluorescence of keratinocytes the immune cells per mm² were manually counted as intraepithelial or stromal using the LSM V.5 Image Examiner software (average of five images at a 250× magnification). By use of overlapping colors the following infiltrating myeloid cells were distinguished: CD14⁺CD33⁻CD163⁻, CD14⁺CD33⁺CD163⁻, CD14⁺CD33⁻CD163⁺, CD14⁺CD33⁺CD163⁺, CD14⁻CD33⁺CD163⁻ and CD14⁻CD33⁻CD163⁺.

Immunofluorescent four color staining was performed for 10 selected OPSCC tumors as described previously.16 17 Antibody specificity and optimal antigen retrieval were assessed by single-plex IHC using tonsils. Following incubation with superblock buffer (Thermo Fisher Scientific, Waltham, Massachusetts, USA), the primary antibodies and isotype/species-specific secondary fluorescent antibodies were applied (online supplementary table 2). Tissue slides were imaged at 20× magnification with the Vectra 3.0 Automated Quantitative Pathology Imaging System (PerkinElmer). Imaging analysis was performed with the InForm Cell Analysis software (PerkinElmer). Five random multispectral imaging fields of interest were selected for acquisition from each tumor and myeloid cells were manually counted within these fields. The following phenotypes were identified: HLA-DR⁺CD11c⁺CD14/CD68⁺CD163⁻ macrophages, HLA-DR⁺CD11c⁺CD14/CD68⁺CD163⁺ macrophages, HLA-DR⁺CD11c⁺CD14^‒CD68⁻CD163⁻ DC and HLA-DR⁺CD11c⁺CD14^‒CD68⁻CD163⁺ DC.

Cryopreserved single cell OPSCC tumor samples were thawed and used for myeloid cell subset determination by flow cytometry. Cryopreserved healthy donor PBMCs were thawed, depleted of CD3⁺ T and CD19⁺ cells by the use of CD3-microbeads (Miltenyi Biotec), subsequently CD19 pan B dynabeads (Thermo Fisher Scientific) and then used for myeloid cell sorting by flow cytometry. After thawing and/or CD3/CD19 depletion, samples were stained with the LIVE-DEAD Fixable yellow or near-IR dead cell stain kit (Thermo Fisher Scientific) for 20 min at room temperature. Following washing, Fc receptors were blocked by incubating for 10 min on ice with phosphate buffered salt (PBS)/0.5% bovine serum albumin (BSA)/10% fetal calf serum (FCS), after which the cells were washed, and stained for 30 min on ice with fluorochrome-conjugated antibodies. Details on antibodies used for myeloid cell subsets determination and flow activated cell sorter (FACS)-sorting of PBMCs are listed in online supplementary table 3. Acquisition of cells was performed on a BD LSR Fortessa and flow cytometry-based cell sorting was done using a BD FACS Aria III. Data were analyzed using DIVA software (V.8.02; BD Biosciences).

CD11c⁺HLA-DR⁺ and CD14⁺CD163^‒, CD14⁺CD163⁺, CD14^‒CD163^‒ and CD14^‒CD163⁺ myeloid cells were FACS-sorted from CD3^‒CD19^‒ PBMCs from healthy donors and cultured in Iscove's Modified Dulbecco's Media (IMDM; Lonza) supplemented with 2% human AB (hAB) serum (Capricorn Scientific), 100 IU/mL penicillin (pen; Thermo Fisher Scientific), 100 µg/mL streptomycin (strep; Thermo Fisher Scientific) and 2 mmol/L l-glutamin (L-glut; Lonza) at 20,000 cells per well in triplo in a 96-well round bottom plate in the presence of 100 ng/mL lipopolysaccharide (LPS) (Sigma-Aldrich), 25 µg/mL poly(I:C, Invivogen) and 2.5 µg/mL R848 (Enzo life sciences). After 20 hours, supernatants were harvested, pooled and frozen until analysis. Cytokines were determined using IL-12p70 and IL-18 ELISA (R&D systems) and a custom-made 13-plex cytokine assay (Granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-1α, IL-1β, IL-10, IL-23, IL-6, IL-8, MIP-1α, MIP-3α, TGF-α, TNF-α, M-CSF and VEGF-A, ProcartaPlex Immunoassay, Thermo Fisher Scientific).

To assess the T cell stimulatory capacity of the flow cytometry-sorted CD163⁺ and CD163^‒ myeloid subsets, these cells were cultured for 20 hours at 15.000–20.000 cells per well in triplo in IMDM + pen/strep/L-glut+2% hAB serum a 96-wells round bottom plate in the presence of 100 ng/mL LPS and 2.5 µg/mL R848, after which allogeneic Cell Trace Violet (CTV; Thermo Fisher Scientific)-labeled allogeneic PBMCs were added at 1:10, 1:20 and 1:40 myeloid-to-PBMC ratio. At day 4, 50 µL of the culture supernatants was harvested and frozen for cytokine analysis, and then 10 µg/mL brefeldin A (Sigma-Aldrich) was added during the last 16 hours of culture. The next day, T cell proliferation was assessed by measuring CTV dilution and IFN-γ production by intracellular cytokine staining (ICS) using flow cytometry (online supplementary table 3). Also, supernatants were assessed for the presence of IFN-γ by ELISA (Mabtech) and IL-2, IL-3, IL-4, IL-5, IL-9, IL-13, IL-17A and IL-22 by a custom-made 8-plex cytokine assay (ProcartaPlex Immunoassay, Thermo Fisher Scientific).

Healthy donor-derived PBMCs were cultured in triplicate wells for 24 hours in the presence or absence of 50 ng/mL IL-12 (Miltenyi Biotec) and/or IL-18 (R&D systems) in 96-wells round bottom plates and 5 µg/mL Brefeldin A was added during the last 4 hours of culture. Cytokine production by the stimulated T cells was determined by ICS as described previously (online supplementary table 3).3 9

To assess IL-12/IL-18 responsiveness of T cells in an antigen-specific manner, we made use of HPV16-specific Th clones, which were generated, maintained and tested as described previously.18 19 Responsiveness of the Th clones to IL-12 and IL-18 was assessed for 18 different Th clones that were obtained from three different HPV16⁺ SCC tumors (one OPSCC and two CxCa tumors). To this end, autologous B-LCL were loaded overnight with various concentrations of the cognate peptide (0, 0.005, 0.05, 0.5 µg/mL), and subsequently cultured for 24 hours with the HPV16-specific Th clones at a 1:5 B-LCL:Th ratio in the presence or absence of IL-12 and IL-18. All stimulations were done in duplicate wells of a 96-wells round bottom plate and 5 µg/mL Brefeldin A was added during the last 4 hours of culture.

To assess the antigen-specific stimulatory potential of the CD163⁺ and CD163^‒ macrophage and DC populations, these cells were also tested for their capacity to stimulate the abovementioned HPV16-specific Th clones. To this end, CD11c⁺HLA-DR⁺ and CD14⁺CD163^‒, CD14⁺CD163⁺, CD14^‒CD163^‒ and CD14^‒CD163⁺ myeloid cells were sorted from HLA-DPB1 01:01⁺ or HLA-DPB1 14:01⁺ buffy coats, loaded for 1.5 hours with increasing concentrations of the cognate peptide (0, 0.05, 0.5 and 5 µg/mL), and matured by culturing them for an additional 4 hours with 100 ng/mL LPS and 2.5 µg/mL R848 in IMDM + pen/strep/L-glut+2% hAB serum. Next, HLA-DPB1 01:01 or HLA-DPB1 14:01 restricted HPV16-specific Th clones were added at 1:10 myeloid-to-Th cell ratio, and antigen reactivity of the Th cells was determined after 16 hours by ICS using flow cytometry (online supplementary table 3). To test the role for IL-12 and IL-18 in this setting, 10 µg/mL of anti-IL-12 (clone C8.6; Thermo Fisher Scientific) and anti-IL-18 (clone 125–2 hour, MBL International Corporation) antibodies were added during the last 2 hours of the LPS/R848 stimulation.

The bulk RNA sequencing data GEO: GSE13673114 was used. To this end, paired-end reads obtained from bulk RNA sequencing of CD163⁺ cDC2 were mapped to human transcript sequences obtained from Gencode V.28 using Salmon (V.0.9.1).20 21 Transcript abundances quantified by Salmon were summarized to gene-level counts and normalized gene-level abundances in transcript per million (TPM) units using the tximport R/Bioconductor package (V.1.2.0). For the analysis of differentially expressed genes (DEGs), the gene counts were imported into DESeq222 and a negative binomial generalized linear model (GLM) was fitted to the counts data. Dispersion parameters in the GLM were estimated by a local smoothed dispersion fit by providing the argument fitType=‘local’ to the estimateDispersions function of DESeq2. A Wald test was performed on the model coefficients to identify DEGs between CD163⁺CD14⁻ and CD163⁺CD14⁺ cells. DEGs selected based on adjusted p value <0.05 and at least twofold change were analyzed for pathway enrichment using Ingenuity Pathway Analysis. For visualization, gene abundances in TPM were log2 transformed (ie, TPM + 1.0 to avoid log-zero) and averaged over replicates to obtain mean expression and listed as median log2 TPM values in an Excel file (version 1803) (online supplementary table 4). Genes associated with cDC phenotype, HLA, TNF/TNFR superfamily, cytokine and chemokine (receptor) activity, TLR binding and TNF and IFN signaling for display in heatmaps.

Non-parametric (Wilcoxon signed-rank or Mann-Whitney test for two samples and Friedman or Kruskal-Wallis with Dunn’s multiple comparison test for multiple samples) and parametric (paired or unpaired t test for two samples or repeated measures (RM) one-way analysis of variance (ANOVA) or ordinary one-way ANOVA with Tukey’s multiple comparison test for multiple samples) tests were performed as appropriate. Correlation analysis were done using the Pearson's correlation test. For survival analysis, patients were grouped into two groups according to the median (ie, grouped into below or above the median of the total group for each parameter), after which survival was tested using Kaplan-Meier method, and statistical significance of the survival distribution was analyzed by log-rank testing. All statistical tests were performed at the 0.05 significance level, and differences were considered significant when p<0.05, as indicated with an asterisk (*p<0.05, **p<0.01, ***p<0.001 and ****p<0.0001). Statistical analyses were performed using GraphPad Prism V.8.2.1 (San Diego, USA).

To evaluate the intraepithelial and stromal myeloid cell infiltration, tumors of 20 HPV16⁺ patients with OPSCC (online supplementary table 1) were analyzed for cells expressing CD14, CD33 and/or CD163 by triple immunofluorescent confocal microscopy (figure 1A; online supplementary table 2). Clear differences were observed in the number and type of infiltrating myeloid cells between individual tumors and between epithelium and stroma (figure 1B). In general, the stroma was more densely infiltrated with myeloid cells. The most abundant myeloid cells within the tumor epithelium were CD14⁺CD33 ^‒ CD163⁺, CD14⁺CD33⁺CD163⁺ or CD14 ^‒ CD33 ^‒ CD163⁺ cells. While CD14⁺CD33 ^‒ CD163⁺ and CD14⁺CD33⁺CD163⁺ cells could be detected in the tumor stroma at similar levels to the epithelium, CD14 ^‒ CD33⁺CD163 ^‒ immature myeloid cells and in particular CD14 ^‒ CD33 ^‒ CD163⁺ cells dominated the tumor stroma. Furthermore, the numbers of intratumoral and stromal CD14 ^‒ CD33 ^‒ CD163⁺ cells were highly correlated, while this was less the case for the other subpopulations (online supplementary figure 1A).

To gain insight in the impact of these cells in the TME, we assessed the relation between the number of infiltrating myeloid cells and a type 1-oriented tumor immune infiltrate. Positive correlations were found between the number of CD14 ^‒ CD33 ^‒ CD163⁺ myeloid cells and total CD8⁺, CD8⁺Foxp3 ^‒ Tbet⁺ and CD8 ^‒ Foxp3 ^‒ Tbet⁺ (CD4) cells, both in the tumor epithelium and in the stroma (figure 1C and online supplementary figure 1B). This was not found for the other myeloid cell subsets (online supplementary figure 1C), indicating that only the CD14^‒CD33^‒CD163⁺ cells accumulate at type 1-inflamed sites where potentially a tumor-specific immune response has developed. Notably, we have shown before that approximately 99% of the CD3⁺CD8^– T cells express CD4.23 As the reactivity of the TILs against the tumor-specific viral oncoproteins E6 and E7 of HPV16 was already known,3 8 9 we were able to assess whether these CD14^‒CD33^‒CD163⁺ cells preferentially accumulated in tumors with a concomitant tumor-specific type 1 T cell reaction. Indeed, higher numbers of intraepithelial and stromal CD14^‒CD33^‒CD163⁺ cells were found in OPSCC containing HPV16-specific TIL (HPV16⁺IR⁺) when compared with the group of OPSCC in which no intratumoral HPV16-specific TILs (HPV16⁺IR^-) were detected (figure 1D and online supplementary figure 2). Interestingly, relatively high numbers of tumor-infiltrating CD14^‒CD33^‒CD163⁺ cells also correlated with improved overall survival (figure 1E). These two relationships did not exist for the other populations of myeloid cells analyzed (online supplementary figures 2 and 3), suggesting that specifically the presence of CD14^‒CD33^‒CD163⁺ cells correlated with an ongoing effective antitumor response.

The positive association between intratumoral CD14^‒CD33^‒CD163⁺ cells and tumor immunity prompted the in-depth characterization of this cell population. Therefore, 10 HPV16⁺IR⁺ OPSCC tumors with high CD14 ^‒ CD33 ^‒ CD163⁺ cell content were stained with a multiplex immunofluorescent VECTRA panel consisting of the myeloid markers CD11c, HLA-DR, CD14, CD68, and CD163 (figure 2A). This analysis confirmed our findings on the existence of a CD14^‒CD163⁺ population, which constituted approximately 30% of the identified CD11c⁺ cells in these tumors. The absence of the markers CD14 and CD68 refutes that they are monocytes or macrophages, while the coexpression of CD11c and HLA-DR (figure 2B) confirmed their myeloid origin and suggested that these cells were a population of DCs. Notably, the relative contribution of the CD14^‒CD68^‒CD163⁺ cells to the total CD163⁺ cells is 40%, whereas that of CD14^‒CD163⁺ cells to the total CD163⁺ cells detected by three-color immunofluorescence was 67%, indicating that the originally identified CD14^‒CD33^‒CD163⁺ cell population also included a fraction of CD14^‒CD68⁺CD163⁺ M2 macrophages (figure 2C).

Recently a novel CD163⁺CD1c⁺ conventional type 2 DC subset (CD163⁺ cDC2; also known as DC314 15 was identified in the blood, the percentage of which was increased in the blood of patients with systemic lupus erythematosus (SLE) and correlated with the SLE disease activity index,14 suggesting that these DCs promoted inflammation. Therefore, a series of HPV16⁺ OPSCC single cell digests was used to delineate whether the observed tumor-infiltrating CD14^‒CD33^‒CD163⁺ cells resemble these DC3s, using phenotypic analysis by flow cytometry. Nine freshly digested HPV16⁺ OPSCC samples (n=3 IR^‒ and n=6 IR⁺) were stained with fluorescently labeled antibodies (online supplementary table 3) and analyzed by flow cytometry according to the gating strategy shown in figure 3A. Higher frequencies of total CD11c⁺CD163^‒ and CD11c⁺CD163⁺ cells were detected in the TME of HPV16⁺IR⁺ OPSCC (figure 3B). Two main subsets were found, namely CD14⁺CD1c^‒ monocytes (Mo) and CD14^‒CD1c⁺ cDC2. Whereas CD163⁺ Mo and CD163⁺ cDC2 were significantly increased in HPV16⁺IR⁺ tumors, this was not the case for their CD163^‒ counterparts (figure 3B). These data clearly indicate that CD14^‒CD33^‒CD163⁺ cDC2 in the TME of HPV16⁺ OPSCC phenotypically resemble blood DC3.

It is well established that DCs are potent stimulators of CD4⁺ and CD8⁺ T cell responses. Hence, we sorted CD14⁺CD163^‒ and CD14⁺CD163⁺ monocytes, and CD14^‒CD163^‒ and CD14^‒CD163⁺ DCs from peripheral blood (online supplementary figure 4A) and assessed their T cell stimulatory potential and cytokine secretion in vitro. Analysis of the sorted populations revealed that CD14^‒CD163⁺ DCs displayed the highest expression of HLA-DR and consisted of a pure population of CD1c⁺ DCs, whereas the CD14^‒CD163^‒ population consisted of a mixture of CD1c⁺ and CD1c^‒ DCs (online supplementary figure 4B–D). In addition, the CD14^‒CD163⁺ cells expressed typical DC markers such as CD40, CD54, CD80 and CD86 (online supplementary figure 4E,F). Unlike monocytes, both types of DCs were capable of stimulating T cell proliferation and production of the type 1 cytokines IFNγ and IL-22 (figure 4A–D). Interestingly, while IL-2, IL-3, IL-4, IL-5 and IL-17A levels were undetectable in all conditions, T cells stimulated with CD14^‒CD163^‒ DCs produced also much more of the type 2 cytokines IL-9 and IL-13 than CD14^‒CD163⁺ DC-stimulated T cells (figure 4D, online supplementary figure 5). This suggests that CD14^‒CD163⁺ cDC2 predominantly stimulate type 1 T cell responses. Activation of the monocyte and DC subsets with LPS, R848, and poly(I:C) revealed that the CD14^‒CD163⁺ DCs displayed a strong capacity to produce proinflammatory mediators IL-12p70, IL-18, GM-CSF, IL-10, IL-23, IL-6, IL-1α, IL-1β, MIP1α, MIP3α, TGFα and TNFα (figure 4E,F and online supplementary figure 6). Finally, the DC signature24 of CD14^‒CD163⁺ myeloid cells was confirmed by transcriptome analysis showing the expression of CLEC4A, CLEC10A, IRF4, ID2, FCER1G, KLF4 and ZEB2 (figure 5A). Consistent with their DC gene signature, these cells also expressed high levels of HLA class I and II molecules, which are important for T cell stimulation, and expressed toll-like receptor (TLR)2, 4–6 and 8, making them sensitive for TLR-mediated activation (figure 5). Interestingly, the CD14^‒CD163⁺ DCs also expressed IL-1β, IL-15, IL-16, IL-18, CCL3–5, CXCL16, cytokines and chemokines (figure 5A), which are important for attraction and activation of immune cells in the TME.25–31 Finally, we compared the gene expression profile of CD14^‒CD163⁺ DC with CD14⁺CD163⁺ cells. This revealed several subset-specific upregulated genes, among which TLR3, CXCR2 and CXCR4 may be utilized as therapeutic targets for specific activation or recruitment of the CD14^‒CD163⁺ DC (figure 5B and online supplementary table 5). Moreover, pathway analysis revealed a strong activation (positive Z score) of the Th1 pathway and an inhibition (negative Z score) of the programmed death (PD)-1/programmed death ligand 1 (PD-L1) cancer immunotherapy and inducible T cell costimulator (ICOS)/ICOS ligand (ICOSL) signaling pathways (figure 5C), highlighting the T cell stimulatory profile of the CD14^‒CD163⁺ DC.

TLR-mediated activation of CD14^–CD163⁺ cDC2 showed that they produced high amounts of IL-12p70 and IL-18 (figure 4E and online supplementary figure 6). These two cytokines have been described as important bridges between the innate and adaptive immune system, and to play a key role in T cell activation and driving Th1 differentiation by augmenting IFNγ production.32–34 Moreover, it has been described that intratumoral injection of DC engineered to secrete IL-12 and IL-18 leads to accelerated tumor rejection by the activation of a strong and diverse type 1 immune response in tumor-bearing mice.35 Therefore, we examined whether these cytokines also played a role in the type 1 T cell polarization by CD14^–CD163⁺ cDC2. IL-18 receptor expression analyses showed its expression by approximately 20% of CD4⁺ and 40% of CD8⁺ peripheral blood T cells (figure 6A). Stimulation of T cells with either IL-12 or IL-18 alone failed to induce IFNγ production, but their combination synergized to stimulate the production of IFNγ by CD4⁺ and CD8⁺ memory T cells (figure 6B and online supplementary figure 7A), confirming previous studies.36 37 This effect was specific for IFNγ, as production of TNFα, IL-2 and IL-17A was not induced (online supplementary figure 7B).

Next, the role of IL-12 and IL-18 on the T cell stimulatory capacity of FACS-sorted CD14^‒CD163^‒ DC and CD14^‒CD163⁺ cDC2 was determined by stimulating allogeneic CD4⁺ and CD8⁺ T cells in the presence or absence of neutralizing IL-12 and IL-18 antibodies. Neutralization of IL-12 and IL-18 had only minimal effect on the proliferation of CD4⁺ and CD8⁺ memory T cells (figure 6C). However, the production of the type 1 cytokines IFNγ and IL-22 by these T cells was clearly reduced on neutralization of IL-12 and IL-18 (figure 6D), while the production of the Th2 cytokine IL-13 was increased (figure 6 and online supplementary figure 8). The changes in cytokine production were on average stronger in T cells stimulated with CD14^–CD163⁺ DCs (figure 6D). These data suggest an additional role for these CD14^–CD163⁺ cDC2 in dampening Th2 differentiation via IL-12 and IL-18 production. The levels of IL-2, IL-3, IL-4, IL-5, IL-9 and IL-17A were low or undetectable and remained unaffected on neutralization of IL-12 and IL-18 (online supplementary figure 8).

Next, we studied the effects of IL-12 and IL-18 on the reactivity of tumor-specific T cells, which are present in HPV16⁺IR⁺ OPSCC where CD14^‒CD163⁺ cDC2 were found. On stimulation of 18 different HPV16-specific CD4⁺ helper T cell clones with increasing concentrations of cognate HPV16 antigen, in the presence or absence of recombinant IL-12 and IL-18, it became clear that the percentage of T cells producing IFNγ was significantly higher when IL-12 and IL-18 were present (figure 7A,B). Moreover, IFNγ-producing cells were detected already at much lower antigen concentrations (figure 7A,B). This effect was already observed at IL-12/IL-18 concentrations of 500 pg/mL and was specific for IFNγ, as this was not observed for GM-CSF and for TNFα only at the highest antigen and cytokine concentrations. Importantly, the increased IFNγ production by the helper T cell clones in the presence of IL-12 and IL-18 was truly a qualitative difference and not a quantitative difference, as expression of the activation marker CD154 on T cells was similar in the presence and absence of IL-12 and IL-18 at all of the antigen concentrations used to stimulate the T cells (figure 7C).

Finally, we also stimulated two different HPV16-specific Th clones with peptide-loaded, HLA-DPB1-matched CD14⁺CD163^‒, CD14⁺CD163⁺, CD14^‒CD163^‒ and CD14^‒CD163⁺ myeloid cells in the presence or absence of IL-12 and IL-18 neutralizing antibodies. CD14^‒CD163⁺ cDC2 stimulated the IFNγ production by HPV16-reactive Th clones, and they do this as efficient as CD14^‒CD163^‒ DCs (figure 7D). Moreover, IL-12 and IL-18 neutralization resulted in a clear reduction in the percentage of IFNγ-producing cells, whereas the level of CD154 activation was not affected (figure 7E). Importantly, the IL-12/IL-18 blockade had a significantly stronger effect on Th cells stimulated with CD14^‒CD163⁺ cDC2, in line with the fact that they produced the highest amounts of IL-12 and IL-18.