Blood from healthy donors was collected after informed consent in heparin tubes (Greiner Bio-One, Alphen a/d Rijn, The Netherlands) for peripheral blood mononuclear cell (PBMC) and neutrophil isolation by density gradient centrifugation on Lymphoprep (d = 1.077 ± 0.001 g/ml; Axis-Shield, Oslo, Norway). Subsequently, PBMCs were separated into monocyte and peripheral blood lymphocyte (PBL) fractions by density gradient centrifugation on Percoll (GE Healthcare, Hoevelaken, The Netherlands). Monocyte-derived DCs were generated from the monocyte fraction, as described previously (9), and the purity of DCs was 95% or higher. CD4⁺CD45RA⁺ naive cells were subsequently purified by negative selection and CD4⁺CD45RO⁺ memory cells by positive selection from the PBL fraction by magnetic cell separation, as described previously (9). The purity of obtained naive T cells always exceeded 98%, and T cells were stored in liquid nitrogen until use in co-culture. Neutrophils were isolated fresh on the day of co-culture initiation from the erythrocyte pellet of Lymphoprep density centrifugation, as described previously (9). The purity of neutrophils was always more than 98%. The purity of all cell types was analyzed by flow cytometry and is shown in Supplementary Figure 1 .

Cultures of C. albicans (clinical isolate) were maintained at the Department of Experimental Microbiology, Amsterdam University Medical Center, Amsterdam. Yeast particles were resuspended in Iscove’s modified Dulbecco’s medium (IMDM; Thermo Scientific, Gibco, Waltham, MA, USA) supplemented with 10% heat-inactivated fetal bovine serum (Sigma-Aldrich, St. Louis, MO, USA) w/o gentamycin and plated at 10,000 particles per well on flat-bottom 96-well plates (Costar, Corning Inc, Corning, NY, USA). For the scraping of hyphae, 60,000 particles were plated in 24-well plates (Costar). (Pseudo-)hyphae were formed during 4 h of incubation at 37°C, 5% CO₂, followed by heat killing for 2 h at 70°C.

On day 6 of DC generation, immature DCs were primed for 2 h with 2.5 µM of VD3 metabolite calcitriol 1,25(OH)2D3 purchased from Sigma-Aldrich in IMDM/5% FCS prior to harvesting. Titrations of VD3 were done to determine the optimal concentration for use, and we primed for only 2 h given that 30% of cholecalciferol, a highly lipophilic compound, was taken up within 1 h of culture of Caco-2 cells (32). DCs were extensively washed prior to use in culture since VD3 has been demonstrated to have a profound direct effect on T cells (33). Co-cultures were done in IMDM supplemented with 5% heat-inactivated human serum (Lonza, Verviers, Belgium) and gentamycin (86 μg/ml, Duchefa Biochemie B.V., Haarlem, The Netherlands). Culturing was done initially in 96-well C. albicans hyphae-coated plates with 50,000 DCs and 50,000 autologous naive CD4⁺ T cells, in the presence or absence of 100,000 autologous neutrophils. These culture conditions for Th17 cell development were optimized previously (9). In indicated experiments, recombinant human IL-23 (R&D Systems, Minneapolis, MN, USA) or epacadostat (Selleck Chemicals, Houston, TX, USA) was used (epacadostat pretreatment of DCs for 30 min at 37°C, no washing). Cells were transferred after 4 days to 48-well plates (Costar) and refreshed with a medium containing recombinant human IL-2 (Novartis AG, Basel, Switzerland) at a final concentration of 10 U/ml every 2 days. Dependent on the growth of the culture, cells were transferred to 24-well plates 2 to 4 days later and cultured until resting on days 11–13 of culture when cytokine production was determined either by flow cytometry (IL-17 and IFN-γ) or by ELISA for IL-10, and cells were stained for Treg markers by flow cytometry. T cells (100,000) were restimulated with anti-CD3 (clone 1XE, 0.15 µg/ml) and anti-CD28 (15E8, 1 µg/ml) (both purchased from Sanquin, Amsterdam, The Netherlands) in IMDM 5% heat-inactivated human serum on flat-bottom 96-well plates, and 24-h supernatants were collected to measure IL-10 by ELISA. To assess the stimulatory capacity of DCs, 30,000 DCs were co-cultured with 50,000 human naive CD4⁺ T cells in the presence or absence of 60,000 neutrophils (2:1 ratio neutrophil:DC) on C. albicans hyphae-coated plates. On day 5 of culture, the percentage of proliferating T cells was assessed by measuring the incorporation of EdU into dividing cells by flow cytometry using the Click-iT™ EdU kit (C10424, Invitrogen, Waltham, MA, USA) according to the manufacturer’s instructions.

T cells that were co-cultured for 8 days with either control or VD3-primed DCs with or without neutrophils (“test cells”) were incubated for 2 h at 37°C with 100 µg/ml of mitomycin C (Sigma), to prevent expansion, and then harvested, extensively washed, and counted. Autologous memory T cells were labeled with 5,6-carboxyfluorescein diacetate succinimidyl ester (0.5 mM; Molecular Probes, Eugene, OR, USA) and subsequently used as bystander target cells. Test cells (50,000) were co-cultured in a round-bottom 96-well plate (Costar) with 25,000 target cells and 500 beads/well Dynabeads™ CD3/CD28 (Gibco). After 3 to 4 days, the proliferation of the target T cells was determined by flow cytometry.

After 2 h of priming of immature DCs in a 24-well plate with calcitriol (2.5 µM) in IMDM/5% FCS on days 5 to 7 of DC generation, DCs were washed and matured in IMDM/5% FCS containing 60,000 scraped C. albicans hyphae per well in the presence of 500 U/ml recombinant human GM-CSF (Schering-Plough B.V., Brussels, Belgium), with or without the presence of 600,000 neutrophils (both hyphae and neutrophils had a similar ratio to the surface area as in the 96-well plate). After 48 h, DCs were harvested and stained with a fixable viability dye eFluor 780 (eBioscience, San Diego, CA, USA), and surface marker or intracellular indoleamine-2,3-dioxygenase (IDO) expression was determined by flow cytometry.

Cells were restimulated for the determination of cytokine production by T cells as described previously (9). On the same day as the restimulation (days 11–13 after the start of culture), cells were stained for CD25 and CD127, and subsequently, cells were fixated and permeabilized with a Transcription Factor Buffer Set (BD Biosciences, San Jose, CA, USA) and stained for FoxP3. A total of 10,000 cells were acquired in the live gate on a FACSCanto machine (BD Biosciences). Alternatively, cells were stained for CD25, CD39, CD49b, CD69, CD127, ICOS, PD1, TIGIT, and TIM3 prior to fixing + permeabilization with the Transcription Factor Buffer Set and staining for CTLA4 and FoxP3. Cells were stained for GARP after 24 h of stimulation with αCD3 and αCD28. A total of 20,000 cells were acquired in the live gate on the SP6800 Spectral Analyzer (Sony, San Jose, CA, USA). The following antibodies were used: αCD3-FITC, αCD4-APC, αCD45RA-FITC, αIFNγ-FITC, αCD25-FITC, αCD39-BV510, αCD49b-BV605, αLAG-3-PE-CF594, αCD69-APC-R700, αPD1-BB515, αTIGIT-BB700, αTIM3-BV480, αGARP-BV750, and αICOSL-PE-CF954 (all from BD Biosciences); αCD127-PE, αCTLA4-PE-Cy5, αCD11c-APC, αCD83-PE-Cy5, αCD86-APC, αHLA-DR-PerCP, and αHLA-DR-BV421 (all from BD Pharmingen, San Diego, CA, USA); αFoxP3-AF647, αCD25-PE-Cy7, αICOS-BV421, and αCD66b-PE (all from Biolegend, San Diego, CA, USA); αCD45RO-PE (DAKO); αIL-17A-eFluor660 (eBioscience); and αIDO-PE (R&D Systems). Data were analyzed using FlowJo™ software (for Windows, Version 10.6.2., Ashland, OR, USA). Heatmaps were generated using Tercen™ (Waterford, Ireland). Cells stained for CD25 and CD127 were sorted with the 4-laser FACSAria IIu SORP (BD Biosciences) prior to restimulation for the assessment of IL-17 expression.

DCs were primed with VD3 as done for the co-cultures, and subsequently, 30,000 DCs were matured in flat-bottom 96-wells with 10,000 C. albicans hyphae per well and 30,000 CD40 ligand-expressing murine plasmacytoma cells (J558 cells; a gift from Dr. P. Lane, University of Birmingham, Birmingham, United Kingdom), with or without 60,000 neutrophils (similar ratio as the T-cell co-culture). Sandwich ELISAs were performed on 24-h culture supernatants to determine the concentrations of CXCL8 (Invitrogen), IL-1β (U-CyTech, Utrecht, The Netherlands), IL-10 (BD Biosciences), IL-12_p70 (own culture), and IL-23 (U-CyTech). For the analysis of IL-12_p70, 1,000 U/ml of IFN-γ (U-CyTech) was added to the cultures and TGF-β ELISA of eBioscience was used. Without additional DC stimulation, values were below the detection level (20 pg/ml) for IL-12_p70. Briefly, after blocking overnight the antibody-coated plates in PBS/0.1% Tween 20/1% BSA (PTB), they were incubated with supernatants for 1 h at RT diluted in PTB, and washing was done in PBS/0.1% Tween 20 (PT). Detection antibody with biotin was added for 1 h at RT, and after washing, poly-streptavidin-horseradish peroxidase conjugate (Sanquin) diluted in PTB with 2% Protifar (Nutricia, Utrecht, The Netherlands) was added and incubated for 30 min at RT. Finally, the plates were washed with PT and developed with 3,3′,5,5′-tetramethylbenzidine (TMB, Merck, Germany). The conversion was stopped by adding 1 M of H₂SO₄. The absorbance was measured at 450 nm with reference at 655 nm using VersaMax™ (Molecular Devices, Wokingham, UK).

Statistical analyses were performed using GraphPad Prism software (La Jolla, CA, USA, version 8.3.0 for Windows). Repeated-measures one-way ANOVA with Holm–Sidak’s multiple comparisons test was used. The Friedman test with Dunn’s multiple comparisons test, a non-parametric test, was applied if the distribution of data was not normal. The Shapiro–Wilk test was performed to test the normality of data. All statistics used were indicated in the figures and legends. p-values of 0.05 or less were considered significant.

Since the effect of VD3 on neutrophil-assisted DC-driven Th17 cell development from human naive T cells is unknown, we investigated the effect of VD3 priming of DCs on T-cell development from naive T cells. In the absence of neutrophils, Th17 cell differentiation was negligible, whereas neutrophil presence elicited the capacity of control DCs to promote Th17 cell differentiation ( Figures 1A, B ), as published before (9). Priming of DCs with VD3 significantly impedes Th17 cell development ( Figure 1B ), with a mean of 6.1% IL-17⁺ cells driven by control DCs versus 1.7% by VD3-primed DCs in the presence of neutrophils. Additionally, Th1 cell development was reduced by VD3 priming of DCs ( Figure 1C ), which is in line with the expectation as VD3 is known to suppress IL-12 expression (23, 24). The potency of VD3 to restrict Th1 cell development was neutrophil-independent.

Furthermore, we observed that the proliferation of T cells cultured with VD3-primed DCs was reduced. Therefore, we assessed the T-cell stimulatory capacity of DCs on day 5 of the culture ( Figure 1D ). VD3 priming of DCs inhibited DC-driven proliferation of naive CD4⁺ T cells (mean control DCs vs. mean VD3-primed DCs without neutrophils: 39% vs. 21%). Moreover, in most donors, we observed that neutrophil addition slightly reduced the proliferation of T cells, although this was not significant either in control DCs or VD3-primed DCs. In the absence of DCs, neutrophils cultured with C. albicans hyphae did not stimulate naive T-cell proliferation, as shown in our previous work (9). Taken together, VD3 priming of DCs reduced the capability of DCs to stimulate T-cell proliferation and restricted Th17 and Th1 cell development.

Since we observed a strong reduction in T-cell stimulatory capacity by VD3 priming of DCs ( Figure 1D ), we hypothesized that increased development of Tregs in these conditions could underlie this observation. VD3 is well known to endow DCs with tolerance-inducing properties, resulting in elevated Treg development (11, 14–16). VD3 priming of DCs significantly increased the development of FoxP3⁺CD127_lowCD25⁺ cells, as shown in representative flow cytometry plots ( Figure 2A ) and as combined data from 22 independent experiments and donors ( Figure 2B ). FoxP3⁺CD127_low cells were always CD25⁺ ( Supplementary Figure 2A ). Interestingly, we observed a trend of increased Treg development in the presence of neutrophils; however, this effect was not significant. In these Th17-polarizing conditions, VD3 priming of DCs significantly heightened FoxP3⁺ Treg development.

In addition to the distinction of Tregs by FoxP3, CD127, and CD25 expression, IL-10 production is characteristic for various Treg types, including type 1 regulatory (Tr1) cells, ICOS⁺ Tregs, and CD69⁺FoxP3⁺ Tregs, and IL-10 has been proven crucial for their suppressive capacity (14, 19, 20, 34). The combined data of 19 donors showed a significant induction of IL-10 production upon VD3 priming of DCs, also under Th17-polarizing conditions ( Figure 2C ). Again, as seen for FoxP3⁺CD127_low T cells, a slight increase in IL-10-producing T cells was observed in the presence of autologous neutrophils. Although FoxP3⁺ Tregs can produce IL-10, not necessarily both FoxP3⁺ Tregs and IL-10-producing T cells are induced simultaneously, as shown by van der Aar et al. (14). In some donors, an increase of FoxP3⁺CD127_low Tregs was observed, while in other donors under similar circumstances, IL-10-producing cells were primarily induced ( Figure 2D ). When we further analyzed Treg induction of at least two-fold in the presence of neutrophils, VD3 priming of DCs resulted in the polarization to either FoxP3⁺ Tregs (32% of donors), both FoxP3⁺ and IL-10-secreting T cells (58% of donors), or IL-10-producing T cells (10% of donors). When the induction of FoxP3⁺ and IL-10-secreting T cells was pooled as a Treg measure, neutrophils significantly enhanced Treg development driven by either control DCs or VD3-primed DCs ( Figure 2E ).

Most importantly, we tested the capacity of induced T cells to inhibit the proliferation of autologous bystander-activated CD4⁺ T cells, the ultimate feature of Tregs. T cells cultured with control DCs and neutrophils slightly suppressed the proliferation of bystander-activated T cells compared to T cells without neutrophils ( Figure 2F ). The proliferation of bystander T cells was significantly inhibited by T cells polarized by VD3-primed DCs compared to control DCs, and this was not altered by neutrophil presence. Collectively, the results show that DC-driven T-cell polarization is shifted from pro-inflammatory Th17 cell development toward Treg development upon exposure of DCs to VD3, even in inflammatory conditions where neutrophils are present.

Since we observed increased induction of Tregs under Th17-polarizing conditions (neutrophil presence), we wondered whether these Tregs were identical to the IL-17⁺ subpopulation of T cells. We sorted resting T cells by flow cytometry on the expression of CD25 and CD127, which allowed us to gate for Tregs (CD25^highCD127_low cells) versus presumed T effector cells (Teff; CD25_lowCD127^high cells) using the flow cytometry gating strategy in Supplementary Figure 2B , prior to restimulation of T cells for the analysis of IL-17 expression ( Supplementary Figure 2C ). In the CD25_lowCD127^high T effector cell population, most IL-17⁺ cells were observed in control DCs with neutrophils when compared to the CD25^highCD127_low population of Tregs. Albeit the percentage of CD25^highCD127_low cells was increased in VD3-primed DCs with neutrophil conditions, Th17 cell development was low, as seen in Figure 1B and Supplementary Figure 2C . Taken together, the CD25^highCD127_low Treg subsets induced by VD3 priming of DCs in the presence of neutrophils are functional suppressor cells with low IL-17 expression.

Since different subsets of Tregs, characterized by the expression of specific molecules, have been described, we investigated the expression of functional and Treg subset markers ( Figure 3 ). Immune checkpoint cytotoxic T-lymphocyte-associated protein (CTLA)-4 was expressed on both ICOS⁺ Tregs and on natural or induced FoxP3⁺CD127_lowCD25⁺ Tregs (19). The frequency of ICOS⁺ Tregs (ICOS⁺CTLA4⁺FoxP3⁺) and FoxP3⁺CD127_lowCD25⁺CTLA4⁺ Tregs was significantly increased by VD3 priming of DCs, also in the presence of neutrophils ( Figures 3A–C ). Moreover, C-type lectin receptor CD69 and T-cell immunoglobulin and ITIM domain (TIGIT) expression was elevated both on FoxP3⁺ and FoxP3^– cells by VD3 priming, while the expression of ectonucleotidase CD39 and T-cell immunoglobulin and mucin domain 3 (TIM3) was specifically induced on FoxP3⁺ T cells and reduced on FoxP3^– T cells ( Figure 3A ). In line with the increased polarization toward functional suppressive Tregs, CD69⁺FoxP3⁺ Tregs ( Figure 3D ) and TIGIT⁺FoxP3⁺ cells ( Figure 3E ) were significantly induced by VD3 priming of DCs, even in the presence of neutrophils, while no additional effect of neutrophils was seen. While the average frequency of CD39⁺FoxP3⁺ and TIM3⁺FoxP3⁺ T cells was elevated in VD3-primed DCs, enhanced expression was not seen in all donors and the effect was not significant ( Supplementary Figures 3A–C ).

Furthermore, we measured the expression of glycoprotein A repetitions predominant (GARP), involved in Treg function and homeostasis via TGF-β (35), after αCD3 and αCD28 restimulation, since we could not detect GARP expression on day 11 of T-cell culture without restimulation. GARP expression was significantly elevated on T cells polarized by VD3-primed DCs, and the presence of neutrophils did not affect this ( Figure 3F ). The percentage of GARP⁺ T cells that express FoxP3 was significantly increased by VD3 priming as well ( Figure 3G ). We were not able to assess the frequency of the IL-10-producing Tr1 subset, since lymphocyte-activation protein 3 (LAG-3) staining was low, while LAG-3 and CD49b are the most specific Tr1 markers (36). CD49b expression was not affected by VD3 priming or neutrophil presence ( Supplementary Figure 3D ). Overall, these data suggest that under Th17 cell-polarizing conditions (presence of neutrophils), DCs primed with VD3 are still capable of inducing a functional Treg response, with the polarization of naive T cells toward suppressor cells that express a variety of Treg markers, including CD69, ICOS, CTLA4, TIGIT, and GARP.

To understand the mechanism of altered T-cell development, we examined whether VD3 priming and the presence of neutrophils influenced the expression of cell surface molecules associated with DC maturation and the production of key cytokines in response to C. albicans hyphae. We used a viability dye to exclude dead cells (both DCs and neutrophils). Ninety percent of neutrophils cultured with C. albicans hyphae died overnight ( Supplementary Figure 4A ). VD3 priming of DCs inhibited the induction of CD83 expression by C. albicans hyphae stimulation of DCs ( Figures 4A, B ). Furthermore, neutrophils enhanced the expression of CD83 on both control and VD3-primed DCs. This is in line with the ability of lactoferrin, a neutrophil granule protein, to upregulate the expression of CD83 on immature DCs (37, 38). Furthermore, lactoferrin was shown to induce CD86 and HLA-DR expression (38). However, we did not observe enhanced expression of CD86 or HLA-DR by neutrophils nor an effect by the 2-h VD3 priming of DCs ( Supplementary Figure 4B ).

Next, we examined IL-1β and IL-23 release by DCs, known to be involved in the induction and maintenance of a Th17 cell phenotype (6–8). Since C. albicans hyphae alone did not induce substantial cytokine release by DCs, costimulation with CD40L was applied, mimicking DC–T-cell interactions, although TGF-β levels in the supernatant were still below the detection level. As expected, VD3 priming inhibited the release of IL-23 versus control DCs without neutrophils (mean ± SD = 122 ± 74 ng/ml), while IL-1β secretion (control DC – neutro: mean ± SD = 0.38 ± 0.31 ng/ml) was not affected by VD3 ( Figure 4C ). To determine whether the reduction in IL-23 was involved in the reduced Th17 cell development observed upon VD3 priming, we added recombinant IL-23 and found a 1.5-fold increase in Th17 cell development in three out of seven donors, a slight induction in three donors, and a reduction in one donor ( Supplementary Figure 5 ). Overall, no significant effect of IL-23 was found. Neutrophils did not influence IL-23 secretion, while a reduction of IL-1β release was observed in the presence of neutrophils, which was significant in VD3-primed DCs. Moreover, CXCL8 (77), of which a processed form is required for Th17 cell development from naive CD4⁺ T cells (9), was slightly reduced by VD3 priming of DCs, albeit not significantly ( Figure 4C ). Neutrophils significantly reduced CXCL8 release from DCs in control DCs (control DC – neutro: mean ± SD = 0.64 ± 0.26 µg/ml).

The 2-h VD3 priming of DCs strongly inhibited the secretion of IL-12_p70 (control DC – neutro: mean ± SD = 6.98 ± 3.76 ng/ml; Figure 4D ), as expected since the vitamin D receptor (VDR) downregulates both the expression of p35 and p40 subunits (24). Neutrophils reduced IL-12_p70 release by VD3-primed DCs in each donor, although the effect is not significant. Intriguingly, although the generation of DCs in the presence of VD3 was repeatedly shown to induce IL-10 production in DCs (15, 22), we observed an inhibition in IL-10 release by DCs (control DC – neutro: mean ± SD = 65.68 ± 39.14 ng/ml; Figure 4D ) that were treated with VD3 for only 2 h prior to C. albicans stimulation. Moreover, the presence of neutrophils reduced IL-10 secretion in both control and VD3-primed DCs. Taken together, the 2-h VD3 priming of DCs significantly reduced the secretion of specific T-cell-polarizing cytokines, while neutrophils lowered the release of IL-1β, CXCL8, and IL-10.

One of the mechanisms via which DCs induce Tregs is via the activity of the immunosuppressive enzyme IDO, which degrades tryptophan. Downstream metabolites of tryptophan suppress T effector cell responses (39). We investigated whether VD3 priming and neutrophil presence affected the intracellular expression of IDO in DCs. Stimulation with C. albicans hyphae significantly increased the expression of IDO compared to unstimulated DCs ( Figures 5A, B ). VD3 priming reduced IDO expression, while neutrophils increased the expression in both control and VD3-primed DCs ( Figures 5A, B ). Consequently, we wondered whether this enhanced IDO expression in DCs could be involved in the increased Treg development by neutrophil presence. Inhibition of IDO was achieved by pretreatment of DCs with epacadostat, an IDO selective inhibitor (39), prior to culture with naive T cells and neutrophils. IDO inhibition did not significantly affect Treg development (data not shown).

In addition to IDO expression in DCs, ICOS ligand (ICOSL) expression is a tolerogenic feature that DCs could employ in order to induce Tregs (20, 40). Therefore, we assessed the expression of ICOSL on control and VD3-primed DCs after stimulation with C. albicans hyphae in the absence or presence of neutrophils. Neutrophils significantly increased the expression of ICOSL on DCs and VD3 priming of DCs reduced the expression of ICOSL ( Figures 5A, C ). Hence, both of these tolerogenic DC features of IDO and ICOSL expression follow a similar pattern, with reduced IDO and ICOSL expression in VD3-primed conditions, while neutrophils enhance the expression.