Six- to eight-week-old C57BL/6 wild-type, C57BL/6 x BALB/c (F1), Foxp3/GFP and RAG-KO mice were used (both under C57BL/6 background). All mice were maintained under pathogen free conditions, with a 12 h light/dark cycle and food and water ad libitum. This study was carried out after the revision and approval of its experimental protocol in accordance with the recommendations of bioethical Committee guidelines from the Faculty of Medicine, Universidad de los Andes, and the Comisión Nacional de Ciencia y Tecnología, CONICYT.

For skin transplantation experiments, mice were anesthetized with a mixture of ketamine at 27% and xylazine at 3% via intraperitoneal (i.p.) injection. Tail skin (~1 cm²) from C57BL/6 (syngeneic) or F1 (allogeneic) donors was transplanted onto the dorsal area of C57BL/6 or RAG^−/− recipient mice. l-Ascorbic acid (AA, Sigma-Aldrich, MI, USA) was added daily in the water at 0.86 mg/mL, starting the day of surgery until the end of the experiment. Survival of skin allografts was evaluated twice per week and grafts were considered rejected when 80% of the original graft had disappeared or become necrotic.

Dendritic cells were purified from wild-type C57BL/6 mouse spleen, digested in the presence of 1 mg/mL of collagenase D (Roche, Germany) and 2 µg/mL of DNAse I (Roche, Germany) in RPMI medium (Thermo Scientific, NH, USA) supplemented with 10% FBS (Gibco, USA), 1% penicillin/streptomycin (Corning, USA), HEPES (Gibco, Great Britain), β-mercaptoethanol (Sigma, MO, USA) at 50 µM. The 1 mL of this mixture was used to perfuse the organ, followed by an incubation of 45 min at 37°C in waterbath. Undigested material was filtered through a cell strainer (Fisher Scientific, NH, USA) and CD11c⁺ cells were isolated from the obtained cell suspension by using mouse CD11c MicroBeads UltraPure (MACS, Miltenyi, Germany), according to the manufacture instructions. For lymphocytes, spleens from Foxp3/GFP mice were collected in PBS 1× + 5% of FBS and mechanically disaggregated through a cell strainer. Red blood cells were lysed using RBC lysis buffer 10× (Biolegend, CA, USA) and total CD4⁺ T cells were purified using mouse CD4⁺ T cell isolation kit (MACS Miltenyi Biotec, Germany), according to the manufacture instructions. B cells and CD8⁺ T cells were sorted from the CD4⁻ fraction based on the expression of CD8 and CD19, while CD4⁺ T conventional and CD4⁺ T regulatory (Tregs) cells were sorted from the CD4⁺ fraction based on the expression of CD4 and Foxp3/GFP using a BD FACSAria III (Franklin Lakes, NJ, USA).

Total CD4⁺ T cells were isolated from the spleens of Foxp3/GFP mice, using mouse CD4⁺ T cell isolation kit (MACS Miltenyi Biotec, Germany), according to the manufacture instructions. 2 × 10⁵ cells were cultured in 24-well plates, previously coated with anti-CD3 (clone 2c11, at 10 µg/mL) and anti-CD28 (clone PV-1, at 1 µg/mL) for 6 days. Culture media contained RPMI medium (Thermo Scientific, NH, USA) supplemented with 5% FBS (Gibco, USA), HEPES (Gibco, Great Britain), 1% penicillin/streptomycin, and β-mercaptoethanol 50 µM, in the presence of IL-2 (at 100 U/mL) and human TGF-β (at 20 ng/mL), both from Peprotech (NJ, USA). At days 1 and 3, AA (Sigma-Aldrich, MI, USA) was added at a final concentration of 200 µM, and at day 6 cells were harvested for phenotype analysis. All cell cultures were maintained at 37°C and 5% CO₂.

To evaluate in vivo the direct effect of VitC on naive CD4⁺ T cells, these were isolated from a Foxp3/GFP reporter mouse spleen and sorted based on CD4⁺CD62L^hiCD44^loFoxp3/GFP⁻ phenotype, and intravenously (i.v.) injected into RAG-KO mice a day before skin surgery (4 × 10⁵ cells per mouse). To test the effect of VitC on both natural and iTreg, conventional CD4⁺ Foxp3/GFP⁻ T cells were sorted from a Ly5.1 Foxp3/GFP reporter mouse, and nTreg were obtained from a Ly5.2 Foxp3/GFP mouse using a mouse CD4⁺ CD25⁺ Regulatory T Cell isolation kit (MACS Miltenyi, Germany). A day before surgery, RAG-KO recipient mice received intravenously 1:1 mixture of CD4⁺ conventional T cells and Treg. Grafting was performed as previously described and AA (Sigma-Aldrich, MI, USA) was delivered daily in the water at 0.86 mg/mL from the day of surgery until the end of the experiment.

For in vitro assay, CD4⁺ Treg cells were sorted from the spleen of a Ly5.1 Foxp3/GFP reporter mouse, as described above, and incubated with VitC (at 50 µM) for 1 h before coculture with antigen presenting cells (APCs) and CD4 conventional T cells. 1 × 10⁵ mitomycin C-treated APCs (FACS sorted from the spleen of a Ly5.2 Foxp3/GFP mouse, based on CD4⁻ CD8⁻ MHC-II⁺ phenotype), 5 × 10⁴ CFSE-labeled CD4⁺ conventional T cells and Tregs (at 1:1, 1:2, and 1:4 Treg:CD4⁺ T effector ratio) were cultured in 96 round well-bottom plates with complete RPMI medium, in the presence of soluble anti-CD3 (clone 2c11, 5 µg/mL) for 3 days. Mitomycin C was used at 50 µg/mL and CFSE at 5 µM (Molecular Probes, Invitrogen, USA). For ex vivo experiments, Tregs were isolated at day 20 from draining lymph nodes (DLNs) of RAG^−/− transplanted mice, treated or not with VitC, using a mouse CD4⁺ CD25⁺ Regulatory T Cell isolation kit (MACS Miltenyi, Miltenyi). The assay was set up as described above. Cell division was analyzed by flow cytometry to determine responder T cell proliferation (by CFSE dilution) and FlowJo used to determine division index of responder cells. Suppression was calculated with the formula% Suppression = (1 − DI_Treg/DI_Tresp) ×100% (where DI_Treg stands for the division index of responder cells with Tregs, and DI_Tresp stands for the division index of responder cells activated without Tregs).

Cell samples were stained with anti-CD4 (clone RM4-5), anti-CD8 (clone 53-6.7), anti-CD44 (clone IM7), anti-CD62L (clone MEL-14), anti-CD19 (clone 6D5), anti-CD25 (clone PC61), anti-Foxp3 (clone FJK-16s), anti-CD45.1 (clone A20), anti-CD45.2 (clone 104), and anti-CD11c (clone N418) (all from BioLegend, CA, USA, except Foxp3 from EBioscience, San Diego, CA, USA). All antibodies were conjugated with FITC, phycoerythrin, peridinin chlorophyll protein, or allophycocyanin. All flow cytometry was performed using a BD FACS Canto II cytometer (BD Biosciences, CA, USA), and data were analyzed using FlowJo software (Tree Star, OH, USA).

Supernatants from Treg induction assays were collected at day 6 and stored at 80°C until cytokine quantification by sandwich ELISA. Briefly, 96 well flat-bottomed plates were coated overnight with anti-IFNγ, anti-IL-17, or anti-IL-10 capture antibodies at a final concentration of 1 µg/mL. After several washing steps with PBS 1× + 0.05% Tween 20 and a blocking step with PBS 1× + 10% FBS, samples and standard curve were incubated at room temperature for 2 h. Biotinylated antibodies in conjunction with HRP–avidin were used for detecting the immobilized cytokine and tetramethylbenzidine (TMB, ThermoFisher) substrate solution was added to detect HRP activity. The reaction was stopped by adding sulfuric acid 0.16 M and absorbance measured at 450 nm using a Tecan absorbance microplate reader.

RNA from spleen/DLNs sorted-populations or in vitro cultured cells were treated for RNA isolation using TRIzol reagent (ThermoFisher, NH, USA), according to manufacture instructions. cDNA was prepared using iScript cDNA synthesis kit (Bio-Rad, CA, USA) and the expression of SVCT2 and housekeeping gene 18S was measured by real-time PCR, using HOT FIREPol EvaGreen qPCR Supermix (Solis BioDyne, Tartu, Estonia) with Mx3000P quantitative PCR system (Agilent Technologies, CA, USA). Primer sequences are showed in Table 1.

The 500 µl of blood was collected by cardiac puncture from naive and skin transplanted C57BL/6 mice in tubes coated with heparin. Samples were centrifuged at 3,000 rpm for 15 min and plasma collected and stored at −80°C until analysis. VitC measurements were performed using the methylene blue assay under acidic conditions described by Dilgin et al. (41).

Statistical significance was determined by Student’s t-test, using GraphPad Prism (GraphPad Software, CA, USA).

Due to the existence of few studies considering VitC as a potential mediator of immune tolerance, we sought to evaluate SVCT2 on different leukocyte subsets. For this purpose, we FACS sorted total splenic APCs (CD3⁻MHC-II⁺), conventional DCs (CD11c⁺ CD11b⁺), B cells (CD19⁺ CD3⁻), total CD8⁺ and CD4⁺ T cells, conventional CD4⁺ T cells (CD4⁺ Foxp3/GFP⁻), and Tregs (CD4⁺ Foxp3/GFP⁺), and evaluated SVCT2 expression by qRT-PCR. SVCT2 mRNA was detected in all cell subsets (Figure 1A), but Tregs and CD8⁺ T cells showed the highest levels in comparison with total CD4⁺ T cells, ~2.5-fold increase (Figure 1B). Interestingly, SVCT2 mRNA levels were dramatically upregulated on Tregs upon in vitro polyclonal activation (~25-fold increase, compared with CD4⁺ Foxp3/GFP⁻ T cells) (Figure 1C). Thus, Tregs correspond to one of the main leukocyte subsets expressing SVCT2, suggesting the ability to sense VitC, which may be higher under T cell activation conditions.

Next, we tested VitC influence on the suppression capacity of nTreg cells. For this purpose, we isolated splenic nTregs from a Foxp3/GFP reporter mouse, and pretreated them with VitC before coculture with CFSE-labeled effector CD4⁺ T cells and APCs, FACS-sorted from the spleen of a Foxp3/GFP reporter mouse. After three days of polyclonal activation with soluble anti-CD3, we measured CFSE dilution on effector CD4⁺ T cells by flow cytometry. Our results indicate that VitC pretreated Tregs were less suppressive than non-treated Tregs, detecting lower number of divisions of effector CD4⁺ T cells in the latter (Figures 2A,B). Decreased function, which can also can be expressed in terms of percentage of suppression, was observed when 1:1, 1:2, and 1:4 Tregs:effector CD4⁺ T cells ratios were tested (Figure 2C). In addition, no significant changes on Foxp3⁺ Tregs frequency nor Foxp3 expression levels were observed (Figure 2D), suggesting that the poor suppressive capacity observed on VitC pretreated nTregs was not due to the loss of Foxp3 expression.

In order to define the potential mechanism targeted by VitC in Tregs, we evaluated gene expression by qRT-PCR of pro-inflammatory and Tregs signature cytokines aside from other relevant proteins. In these experiments, we did not detect major changes in the expression of genes involved in metabolic disruption (CD39 and CD73), lytic activity (Granzyme and Perforin), proinflammatory cytokines (IFN-γ and IL-17), or suppressive cytokines (IL-10 and TGF-β) (Figure 2E). Altogether, our findings indicate that VitC affects Tregs function without impacting Foxp3 expression (or other major Treg-associated genes), cell proliferation, or survival.

Recent studies have shown that VitC supplementation during induction of in vitro Tregs with TGF-β increases Foxp3 expression (35, 36). In addition, it has been shown that VitC promotes demethylation of the conserved CpG-rich noncoding sequence 2, a regulatory region within the first intron of the Foxp3 gene that is crucial for maintaining stable expression of Foxp3 (35, 36). This increased Foxp3 stability has also been detected after in vitro restimulation and after in vivo transfer of these cells in unchallenged mice (36). However, it is not known whether VitC could induce the differentiation of Tregs in vivo, or whether these in vivo-generated iTregs are functional. To answer these questions, we recapitulated Treg induction in vitro experiments in the presence of VitC, which demonstrated that total CD4⁺ T cells cultured in the presence of IL-2 and VitC do not upregulate Foxp3 expression (~0.9% of CD4⁺ Foxp/GFP⁺ T cells versus ~1.9% in IL-2 alone) (Figure 3A, top plot). As expected, ~35% of total CD4⁺ T cells cultured with IL-2 and TGF-β differentiated toward iTregs, but the inclusion of VitC enhanced iTregs conversion to ~45% and their number (Figure 3A, last two dot plots and 3B, top left column graph). Furthermore, we observed that iTregs differentiated without VitC contain a population of Tregs with intermediate levels of Foxp3 (Foxp3^int), representing ~70% of total Tregs (Figure 3B). Interestingly, the inclusion of VitC in iTregs differentiation conditions shifted this Foxp3^int population toward a population with higher levels of Foxp3 expression (Foxp3^hi), increasing from 10.6 ± 2.5% in control cultures to 34.4 ± 3.3% (Figure 3B). Furthermore, these VitC-iTregs are able to suppress in vitro, in the absence of retinoic acid (RA), as previously reported (36), and as shown in Figure S1D in Supplementary Material, in which the % of proliferating cells is depicted when cocultured with nTreg, iTreg generated in the absence of VitC (iTreg control) and iTreg generated in the presence of VitC (iTreg⁺ VitC); each condition at 1:1, 1:2, and 1:4 proliferating:Treg cell ratios. More clearly, Figure S1E in Supplementary Material displays the % of suppression under the mentioned culture conditions (~50–80% for nTreg, ~75–80% for iTreg control, and >85% for iTreg⁺ VitC). The same data set demonstrates the inversely correlated frequencies of both effector T (A) and Treg (B) cells under the coculture conditions, and the unchanged Foxp3 expression on Treg cells in the cultures (C). Furthermore, Figure S3 in Supplementary Material displays the post-sorting purity of naïve CD4⁺ T cells used in the previously described in vitro assay, and Figure S4 in Supplementary Material demonstrates that only signal from proliferating effector CD4⁺ T cells is detected, excluding potential staining of proliferating Treg cells. In addition, cytokine expression and secretion was evaluated using quantitative RT-PCR and ELISA, respectively. As observed in Figure 3C, IL-17 production was negatively regulated by VitC in both control (IL-2) and iTregs (IL-2 and TGF-β) conditions (from 1104 ± 255 pg/mL in control Tregs versus 58 ± 35 pg/mL in VitC-treated Tregs with IL-2; and from 137 ± 54 pg/mL in control Tregs versus 36 ± 74 pg/mL in VitC-treated Tregs with IL-2 and TGF-β). Additionally, VitC affected neither IL-10 nor IFN-γ secretion. The same findings were obtained when TGF-β and SVCT2 expression were analyzed by qRT-PCR (Figure 3D). Thus, VitC may inhibit IL-17 secretion from in vitro activated CD4⁺ T cells, favoring a milieu for iTreg differentiation and/or stabilization.

Next, we assessed the effect of VitC on both Treg generation and function in vivo using a previously standardized model to study immune tolerance based on allogeneic skin transplantation in RAG-KO mice, receiving adoptive transfer of naive CD4⁺ Foxp3/GFP⁻ T cells the day before surgery (6). For in vivo VitC supplementation, skin transplanted mice received a daily mega-dose of VitC (0.86 mg/mL) in the drinking water, for 20 days (Figure 4A), which was corroborated by measuring VitC levels in their serum (Figure S2 in Supplementary Material). During this time, skin transplants were monitored for signs of rejection, observing that VitC-supplemented mice reject at the same kinetic than control mice (Figure 4B). Analysis of CD4⁺ T cells from skin transplant-DLNs showed no changes in the frequencies of total CD4⁺ T cells, CD4⁺ Foxp3⁻CD44^hi effector T cells nor total CD4⁺ Foxp3/GFP⁺ Tregs (Figures 4C,D). However, similar to what we observed during in vitro iTregs differentiation, a population of iTregs with high expression of Foxp3 was significantly increased in mice treated with VitC, representing ~60% of total Tregs (Figure 4D). In terms of cellularity, VitC significantly increased the number of total iTregs (Figure 4D, bottom graphs). These results show that VitC also induces the differentiation of iTregs in vivo; however, this increase does not translate into graft tolerance, suggesting that VitC may impact Tregs subsets, effector CD4⁺ T cells or both.

In order to define whether VitC treatment can also expand the nTreg population and alter their function in vivo, we performed adoptive transfer of a mixture of T cells a day before skin transplantation: Treg from a Ly5.2⁺ Foxp3/GFP⁺ mouse and conventional CD4⁺ Foxp3/GFP⁻ T cells from a Ly5.1⁺ Foxp3/GFP mouse (Figure 5A). As before, VitC was added daily into drinking water and DLNs were analyzed by flow cytometry at day 20 postsurgery. Kaplan–Meier plots show, as expected, that grafted mice without cell transfer do not reject (Figure 5B, black line), while mice receiving only effector CD4⁺ T cells start rejection after two weeks post-surgery (Figure 5B, gray line). As it has been shown in other studies, adoptive cotransfer of Treg allows for graft acceptance (Figure 5B, red line), however, some of the mice treated with VitC rejected the graft despite receiving Treg (Figure 5B, blue line).

Discriminating by congenic marker and Foxp3/GFP expression, we studied total CD4⁺ T cells, nTregs (Foxp3/GFP⁺ Ly5.2⁺) and iTregs (Foxp3/GFP⁺ Ly5.1⁺) in response to VitC (Figure 5C), finding that VitC did not alter effector T cell frequency, but exhibit a tendency to decrease the frequency of total, nTreg and iTreg, by 50% (Figures 5D,E). However, when cell number was determined, we observed a tendency to augment the number of all Treg cell populations (with no significant difference), but effector CD4⁺ T cells were enriched in response to VitC supplementation (Figures 5D,E), suggesting that VitC may promote effector T cell expansion (Figure 5D).

Regardless of VitC augmenting the cell number of CD4⁺ T cell populations, and favoring iTreg differentiation, its impact on in vivo Tregs functionality appears unaltered. In order to corroborate this, we performed syngeneic and allogeneic skin transplants on wild-type C57/BL6 mice (receiving or not VitC in drinking water), and then isolated sufficient number of Treg to further test their suppressive activity ex vivo (Figure 6A). First, we administered VitC in water the same day mice were transplanted (day 0) obtaining no changes on Treg cells (data not shown); then, we decided to deliver VitC in water from day 4 aiming at “priming” the immune system before the surgery. Graft survival analysis showed that syngeneic transplanted mice (both control, black line, and VitC-treated, gray line) do not reject their graft, while allogeneic transplanted mice show rejection signs between 7 and 8 days of surgery, without detecting differences between control, red line and VitC-treated mice, blue line (Figure 6B). After 10 days postsurgery, DLNs quantification analysis showed no changes in the frequencies of CD4⁺ Foxp3/GFP⁻ CD44^hi activated T cells or total CD4⁺ Foxp3/GFP⁺ Treg in VitC-treated mice compared with non-treated controls (Figure 6C, top graphs). As in our previous results, VitC treatment increases the number of CD4⁺ Foxp3/GFP⁺ Treg in the allogeneic group (from ~20,000 Tregs in untreated mice to ~60,000 Tregs in VitC-treated mice), as displayed in Figure 6C. Ex vivo suppression assays using Treg isolated from dLN of naive, syngeneic and allogeneic mice (obtained from controls and VitC-treated animals) confirmed that VitC does not improve Treg-mediated suppression, on the contrary, Treg isolated from VitC-treated naive and allogeneic mice are less suppressive than controls (Figures 6D,E). This decreased suppression capacity was also detected when a lower ratio of Treg:effector T cells (1:4) was used in the assay (Figures 6D,E). Intriguingly, Treg cells isolated from VitC-treated syngeneic animals were more suppressive than untreated animals, at both Treg:T effector cell ratios tested.

Overall, our work indicates that Treg express the SVCT2 transporter and increase the level of Foxp3 expression in response to VitC. Despite our in vitro results, in vivo VitC supplementation enriches for Treg, but this does not translate into allograft acceptance. Even more, when VitC-treated Treg are tested for suppressive function ex vivo, their regulatory activity was indeed decreased.

Even though recent literature involving VitC as an enhancer of in vitro-differentiated Treg is promising, our data highlight a differential behavior of in vivo-expanded Treg in a VitC supplemented state.