ORIGINAL RESEARCH published: 19 June 2019 doi: 10.3389/fimmu.2019.01251 MAP7 and MUCL1 Are Biomarkers of Vitamin D3-Induced Tolerogenic Dendritic Cells in Multiple Sclerosis Patients Juan Navarro-Barriuso 1,2 , María José Mansilla 1,2 , Bibiana Quirant-Sánchez 1,2 , Alicia Ardiaca-Martínez 1,2 , Aina Teniente-Serra 1,2 , Silvia Presas-Rodríguez 3,4 , Anja ten Brinke 5 , Cristina Ramo-Tello 3 and Eva M. Martínez-Cáceres 1,2* 1 Division of Immunology, LCMN, Germans Trias i Pujol University Hospital and Research Institute, Barcelona, Spain, 2 Department of Cellular Biology, Physiology and Immunology, Universitat Autònoma de Barcelona, Bellaterra, Spain, 3 Multiple Sclerosis Unit, Department of Neurosciences, Germans Trias i Pujol University Hospital, Barcelona, Spain, 4 Department of Medicine, Universitat Autònoma de Barcelona, Bellaterra, Spain, 5 Department of Immunopathology, Sanquin Research, and Landsteiner Laboratory, Amsterdam UMC, University of Amsterdam, Amsterdam, Netherlands Edited by: Piotr Trzonkowski, The administration of autologous tolerogenic dendritic cells (tolDC) has become a Medical University of Gdansk, Poland promising alternative for the treatment of autoimmune diseases, such as multiple Reviewed by: sclerosis (MS). Specifically, the use of vitamin D3 for the generation of tolDC (vitD3-tolDC) Aurelie Moreau, Institut National de la Santé et de la constitutes one of the most widely studied approaches, as it has evidenced significant Recherche Médicale immune regulatory properties, both in vitro and in vivo. In this article, we generated (INSERM), France human vitD3-tolDC from monocytes from healthy donors and MS patients, characterized Claudia Volpi, University of Perugia, Italy in both cases by a semi-mature phenotype, secretion of IL-10 and inhibition of allogeneic *Correspondence: lymphocyte proliferation. Additionally, we studied their transcriptomic profile and selected Eva M. Martínez-Cáceres a number of differentially expressed genes compared to control mature and immature emmartinez.germanstrias@gencat.cat dendritic cells for their analysis. Among them, qPCR results validated CYP24A1, MAP7 Specialty section: and MUCL1 genes as biomarkers of vitD3-tolDC in both healthy donors and MS patients. This article was submitted to Furthermore, we constructed a network of protein interactions based on the literature, Immunological Tolerance and Regulation, which manifested that MAP7 and MUCL1 genes are both closely connected between a section of the journal them and involved in immune-related functions. In conclusion, this study evidences Frontiers in Immunology that MAP7 and MUCL1 constitute robust and potentially functional biomarkers of the Received: 18 December 2018 generation of vitD3-tolDC, opening the window for their use as quality controls in clinical Accepted: 16 May 2019 Published: 19 June 2019 trials for MS. Citation: Keywords: tolerogenic dendritic cells, multiple sclerosis, biomarkers, vitamin D3, immune tolerance Navarro-Barriuso J, Mansilla MJ, Quirant-Sánchez B, Ardiaca-Martínez A, Teniente-Serra A, INTRODUCTION Presas-Rodríguez S, ten Brinke A, Ramo-Tello C and The role of dendritic cells (DC) within the immune system is crucial, since they are in charge of Martínez-Cáceres EM (2019) MAP7 orchestrating immune responses and maintaining the homeostasis between immunogenicity and and MUCL1 Are Biomarkers of Vitamin D3-Induced Tolerogenic tolerance. Under normal conditions, DC remain in an immature status (iDC), characterized by Dendritic Cells in Multiple Sclerosis their ability to capture and present antigens and other signals in their environment. However, these Patients. Front. Immunol. 10:1251. cells are not stable, and when exposed to a danger signal, iDC become activated and differentiate doi: 10.3389/fimmu.2019.01251 into professional, antigen-presenting mature DC (mDC). This pro-inflammatory status is Frontiers in Immunology | www.frontiersin.org 1 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers characterized by an increase in the expression of MHC and co- availability and the glycolytic activity mediated through mTOR stimulatory molecules, thus enabling mDC to induce an efficient signaling are crucial for the induction and maintenance of their and potent immunogenic response (1–4). tolerogenic function (27). However, despite the identification of If the immune homeostasis is lost, and a breach of tolerance several pathways involved in the anti-inflammatory role of vitD3- causes mDC to recognize and present specific self-antigens to T tolDC, the specific mechanisms for the induction of immune cells, different autoimmune pathologies may develop depending tolerance by these cells have not been clearly identified yet. on which protein or tissue is targeted, such as multiple sclerosis Previously, our group has successfully generated human (MS), rheumatoid arthritis or type 1 diabetes. These complex vitD3-tolDC —demonstrating their tolerogenic properties in disorders involve many innate and adaptive mechanisms of the vitro using cells generated from both healthy donors and MS immune system, and their etiology still remains unknown. For patient samples—, and has studied their transcriptomic profile this reason, a cure has not been found yet, and the currently compared to other tolDC protocols (12, 25, 26, 33). Additionally, available treatments consist in strong immunomodulatory or in further in vivo studies, we also reported a positive and immunosuppressive drugs. In general, these are focused on beneficial effect of antigen-specific vitD3-tolDC treatment over addressing the symptoms in a poorly effective and unspecific the course of the murine model of MS, the experimental manner, with potentially severe side effects. Hence, there is an autoimmune encephalomyelitis (30, 31). Altogether, these results increasing need for new, more specific and effective therapies. have led to the development of an ongoing Phase I clinical trial Over the last years, tolerogenic DC (tolDC) have been postulated in MS patients with peptide-loaded vitD3-tolDC (http://www. as a novel and promising alternative to treat these disorders (5). clinicaltrials.gov, NCT02903537). In fact, several approaches have already been tested in Phase However, for the full translation of an autologous, antigen- I clinical trials for autoimmune diseases, as reviewed by ten specific, tolerogenic cell therapy into the common clinical Brinke et al. (6), and other clinical studies are still ongoing. In practice, several additional steps yet need to be taken. Among all cases, these treatments have demonstrated to be safe, with them, the definition of adequate, robust and objective biomarkers no relevant side effects on the patients. Consequently, many constitutes one of the priorities. These markers would, on the initiatives are now focused on assessing the actual efficacy of one hand, guarantee the proper generation and functionality of tolerogenic cell therapies. tolDC, without compromising the safety for the patients. On In general, tolDC are defined as a stable and semi-mature the other hand, these biomarkers would enable the comparison subset of DC with the potential to restore immune tolerance of results with other research groups, hereby accelerating the in an antigen-specific manner if loaded with one or more translation of tolDC therapies into the clinic in a collaborative determined peptides, thus not compromising the protective endeavor (6). However, although many efforts have been made in immunity of the patients. Compared to mDC, these cells are this regard, and despite several genes and molecules have been typically characterized by their low —or lower— expression identified for a variety of tolDC protocols separately, such as of MHC and co-stimulatory molecules (such as CD40, CD80, IDO1, GILZ, or ANXA1, the definition of universal biomarkers of CD83 or CD86), and by their reduced or null secretion of tolerance-inducing cell products has not been possible so far, and IL-12, IFN-γ and other pro-inflammatory cytokines, combined it seems unlikely provided the wide heterogenicity of approaches with an increment in the production of IL-10 or TGF-β. These being used (34). For this reason, in this study we have analyzed features confer to tolDC a reduced capability to induce T the transcriptomic profile of vitD3-tolDC in order to select and cell proliferation and the possibility to prime regulatory T cell validate several differentially expressed genes (DEG) that may be (Treg) responses, thus potentially directing the immune response used as transcriptomic biomarkers of these cells in a clinical trial toward a regulatory context (7–9). However, these characteristics for MS patients. can sometimes be very variable, since there is a wide variety of protocols to generate tolDC in vitro from human peripheral blood monocytes. These approaches include the use of several MATERIALS AND METHODS compounds, cytokines and immunomodulatory drugs such as IL-10 (10, 11), dexamethasone (11–15) or rapamycin (11, 12, 16), Sample Collection as well as different genetic engineering techniques (17, 18). Buffy coat samples from 24 randomized healthy donors were Among all of them, the use of 1,25-dihydroxyvitamin D3 obtained from the Banc de Sang i Teixits (Barcelona, Spain), —the active form of vitamin D3— to generate tolDC is following the institutional Standard Operating Procedures for one of the most widely established protocols. Specifically, blood donation, which included a signed informed consent. vitamin D3-induced tolDC (vitD3-tolDC) present a semi-mature Whole blood samples from 10 MS patients were collected by profile, accompanied by an ability to inhibit allogenic T cell standard venipuncture in lithium heparin tubes. Patients did proliferation and to polarize the immune response toward an not receive any corticoid or disease-modifying therapy during at anti-inflammatory TH 2 profile (12, 19–28). Furthermore, several least the previous 2 months, and both relapsing and progressive studies using animal models of autoimmune diseases have forms of the disease were considered. The same procedure was demonstrated their functionality in vivo (29–31). In general, followed for whole blood obtention from 34 healthy donors for these cells are characterized by the suppression of the NF-κB the functional assays (see below). This study was approved by pathway (21, 32), accompanied by an increased activity of the Germans Trias i Pujol Hospital ethical committee, and all the oxidative metabolism of glucose, and indeed the glucose patients and healthy controls signed an informed consent. Frontiers in Immunology | www.frontiersin.org 2 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers Complementary DNA (cDNA) of paired IL-10-induced Phenotype Analysis tolDC (IL10-tolDC) and mDC samples from 5 healthy donors Surface protein expression of CD11c, CD14, CD25, CD83, CD86 were obtained from Sanquin Bloodbank (Amsterdam, The and HLA-DR of iDC, mDC and vitD3-tolDC was determined Netherlands) after informed consent. These samples were by flow cytometry. For each measurement, DC suspensions were generated as described in Boks et al. (11). incubated for 20 min, protected from light, with the adequate amounts of monoclonal antibodies anti-: CD11c PE-Cyanine dye 7 (PE-Cy7), CD14 Violet 450 (V450), CD25 allophycocyanin Monocyte Isolation (APC), CD83 APC, CD86 fluorescein isothiocyanate (FITC) and Samples from healthy donors were processed first depleting HLA-DR Violet 500 (V500) (all of them from BD Biosciences). CD3+ cells using the RosetteSep R Human Monocyte Afterwards, at least 10,000 CD11c+ events of each sample were Enrichment Cocktail kit (StemCell Technologies, Vancouver, acquired in a FACSCanto II flow cytometer and analyzed using Canada) prior to a ficoll-hypaque (Rafer, Zaragoza, Spain) FACSDiva software. density gradient separation. Subsequently, CD14+ cells were isolated by positive selection using the EasySep R Human CD14 Functionality Test Positive Selection Kit (StemCell) following manufacturer’s Allogeneic peripheral blood mononuclear cells (PBMC) from instructions. For the isolation of monocytes from MS patients, whole blood of healthy donors were isolated by ficoll-hypaque peripheral blood monocytes were isolated from 50 mL of whole density gradient separation. Cells were washed twice afterwards, blood by ficoll-hypaque density gradient separation, followed and their absolute number and viability was determined as by the abovementioned CD14 positive selection step. The initial shown above. CD3+ cells depletion step was not performed due to the limited Subsequently, a proliferation assay was performed in 96- amount of blood. Samples were acquired on a FACSCanto II flow well round bottom plates with co-cultures of 105 allogeneic cytometer (BD Biosciences, Franklin Lakes, NJ, USA), and cell PBMC and 5,000 iDC, mDC, or vitD3-tolDC (1:20 ratio) in a viability was determined using 7-amino-actinomycin D (7-AAD) total volume of 200 µL of supplemented X-VIVO 15 medium. (BD Biosciences) and phycoerythrin (PE)-conjugated annexin The same number of PBMC cultured in the presence of either V (Immunotools, Friesoythe, Germany) staining for 20 min at supplemented X-VIVO 15 medium or 50 ng/mL phorbol 12- 4◦ C, protected from light. Cell counts were quantified using myristate-13-acetate (PMA) and 500 ng/mL ionomycin were PerfectCount microspheres (Cytognos, Salamanca, Spain) and used as negative and positive controls, respectively. Six replicates monocyte purity was determined using forward and side scatter of each condition were performed. Cells were incubated for 4 gating strategies on FACSDiva software (BD Biosciences). days at 37◦ C in a 5% CO2 atmosphere. Finally, 1 µCi [3 H]-thymidine (PerkinElmer, Waltham, MA, USA) was added to each well and the plate was incubated VitD3-tolDC Generation for further 18 h under the same conditions. Cells were then Monocytes from both healthy donors and MS patients were collected using a HARVESTER96 2M cell harvester (Tomtec Inc, cultured at a density of 1 × 106 cells/mL in 24-well plates at Hamden, CT, USA) and read on a 1450 MicroBeta TriLux liquid 37◦ C and a 5% CO2 atmosphere for 6 days in 1 mL X-VIVO scintillation counter (Wallac, Turku, Finland). 15 medium (Lonza, Basel, Switzerland), supplemented with 2% heat-inactivated human AB serum, 2 mM L-glutamine (Sigma- Cytokine and Soluble Protein Production Aldrich, St. Louis, MO, USA), 100 U/mL penicillin (Reig Jofre, The production of granzyme B (GZMB) and vascular endothelial Sant Joan Despí, Spain) and 100 µg/mL streptomycin (Normon, growth factor (VEGF), as well as IL-10, IFN-γ and IL-12p70 Tres Cantos, Spain). For the generation of iDC, monocytes cytokines, was quantified in the culture supernatants of tolDC were differentiated in the presence of 200 U/mL granulocyte using the Human Soluble Protein CBA Flex Set (BD biosciences) macrophage colony-stimulating factor (GM-CSF) and 250 U/mL according to manufacturer’s instructions. Samples were acquired IL-4 (both from Peprotech, London, UK). Whole volume of fresh on an LSR Fortessa flow cytometer (BD Biosciences) and medium and cytokines was replenished on day 4. In addition, a analyzed using FACSDiva software. maturation cocktail containing 1,000 U/mL IL-1β, 1,000 U/mL The production of TGF-β was determined using the TNF-α (both from Peprotech) and 1 µM prostaglandin E2 Human/Mouse TGF beta 1 Uncoated ELISA kit (Invitrogen) in (PGE2) (Pfizer, New York, NY, USA) was added on day 4 to both 100 µL of supernatant samples, again following manufacturer’s mDC and vitD3-tolDC conditions. Finally, for the induction of instructions. The optical density of each well was measured at vitD3-tolDC, these cells were besides treated with 1 nM vitamin 450 nm, and the optical density at 570 nm was then substracted D3 (Calcijex, Abbott, Chicago, IL, USA) on days 0 and 4. On day as background signal, using a Varioskan Flash Multimode Reader 6, all three conditions were harvested following an incubation (Thermo Fisher Scientific, Waltham, MA, USA). with accutase (Invitrogen, Carlsbad, CA, USA) for 30 min at 37◦ C to detach the cells from the plate, and washed twice. Cell counts Differential Expressed Genes Selection and and viability were determined by flow cytometry, as shown above, Generation of a Network of Protein and after the phenotypical and functional characterization, dry Interactions pellets of each condition were generated by centrifugation and The data of a comparative transcriptomic analysis of vitD3- stored at −80◦ C. tolDC, iDC and mDC from 5 healthy donors, previously Frontiers in Immunology | www.frontiersin.org 3 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers performed by our group in a microarray study (33), was vs. both iDC and mDC when either logFC vitD3-tolDC vs. mDC > used to select several DEG. For that, the mean difference 0.5 —first criterium—, while logFC iDC vs. mDC < 0.5, —second of expression (MeanDiff) of each gene was evaluated, and criterium—; or logFC vitD3-tolDC vs. mDC < −0.5, while logFC only those genes that were specifically induced or repressed iDC vs. mDC > −0.5, if statistical significance was reached for the in vitD3-tolDC vs. both iDC and mDC conditions (either first criterium (p < 0.05). MeanDiff vitD3-tolDC vs. mDC > 0.5 —first criterium—, while MeanDiff iDC vs. mDC < 0.5, —second criterium—; or MeanDiff Immunocytochemistry Validation vitD3-tolDC vs. mDC < −0.5, while MeanDiff iDC vs. mDC > −0.5), In order to confirm the qPCR results and analyze the protein with a statistically significant differential expression in the first expression and distribution of MAP7 and MUCL1 molecules, criterium (p < 0.01), were selected in order to validate them. an indirect immunocytochemistry (ICC) staining was carried Unlike in our previous study, B-statistic —an indicator of the out in vitD3-tolDC, mDC and iDC from healthy donors and likelihood of the results— was not considered for the selection MS patient samples. Cell culture differentiations of vitD3- of candidate genes this time. Microarray raw and processed tolDC, mDC and iDC were performed in 24-well plates, data were deposited in the ArrayExpress database at EMBL- following the same protocol described above, over 12 mm round EBI (www.ebi.ac.uk/arrayexpress) under accession number coverslips. After 6 days of culture, cells were fixed with 4% E-MTAB-6937. paraformaldehyde, permeabilized with 0.5% TWEEN20 (Sigma- Additionally, a bioinformatic analysis was performed in Aldrich) and subsequently blocked with 10% goat serum for order to characterize the molecular mechanisms involved in 15 min. Afterwards, samples were incubated for 1 h at room the tolerogenicity of vitD3-tolDC. Briefly, the Therapeutic temperature or overnight at 4◦ C with the primary antibodies Performance Mapping System technology (35, 36) was used to mouse anti-human α-tubulin (Invitrogen) and either rabbit generate a mathematical model of protein interactions from our anti-human MAP7 (Invitrogen) or rabbit anti-human MUCL1 transcriptomic microarray data, based on an effectors database of (Sigma-Aldrich). Next, cells were washed and later incubated tolDC biology and functionality (Anaxomics, Barcelona, Spain). with AlexaFluor (AF) 488 goat anti-mouse IgG and AF 594 This database was generated using information available in the goat anti-rabbit IgG secondary antibodies (both from Jackson literature as well as public and private repositories. From this ImmunoResearch Laboratories Inc, West Grove, PA, USA) model, a network of protein interactions between all the effectors for 30 min at room temperature, protected from light. Cells found in our microarray data was built using an Artificial were washed again, and the coverslips were finally mounted Neural Networks analysis and represented using Cytoscape using ProLong R Gold Antifade Mountant medium with DAPI 3.6.1 software. (Invitrogen) for nucleus staining. Samples were analyzed on an Axio Observer Z1 fluorescence microscope (Zeiss, Oberkochen, RNA Extraction and qPCR Validation Germany) with a 63x objective, using ZEN software (Zeiss), Total RNA was isolated from dry pellets using the RNeasy Mini and the expression of MAP7 and MUCL1 was calculated as the Kit (Qiagen, Hilden, Germany) along with a complementary corrected total cell fluorescence (CTCF) of each protein (38, 39) DNAse treatment with the RNAse-free DNAse Set (Qiagen), using the FIJI distribution of ImageJ software (40, 41). following manufacturer’s instructions. Quantity and purity of the samples was then determined using a Nanodrop ND- Statistical Analysis 1000 spectrophotometer (Thermo Fisher Scientific), and the All the statistical analyses were performed with either parametric RNA was subsequently retrotranscribed into cDNA using or non-parametric tests depending on the normality of each the High Capacity cDNA Reverse Transcription Kit (Applied compared data set, as determined by the D’Agostino & Pearson Biosystems, Foster City, CA, USA). Finally, 250 ng cDNA were test, using Prism 6.0 software (GraphPad, La Jolla, CA, USA). preamplified using the TaqManTM PreAmp Master Mix Kit For multiple comparisons, either the one-way ANOVA test with (Applied Biosystems). Geisser-Greenhouse correction or the non-parametric Friedman The expression of genes CA2, CAMP, CLEC5A, CYP24A1, test with Dunn’s correction were used depending on the DHRS9, GAPDH, GZMB, IL1R1, MAP7, MUCL1, OS9, PPIA, normality of the groups. Analogously, for comparisons between SNORD30, SPARC, ST6GAL1, TBP and THBS1 was determined two groups, either the t-Student or the Wilcoxon test were used by quantitative PCR (qPCR) using the respective TaqMan if the samples were normally distributed or not, respectively. Gene Expression Assays (Applied Biosystems) shown in When N ≤ 6, parametric tests were used anyway due to the small Supplementary Table 1, following the instructions provided by sample size (42). Results were expressed as mean ± standard the manufacturer, in a LightCycler 480 System thermocycler deviation (SD), unless noted otherwise, and they were considered (Roche, Basel, Switzerland). Housekeeping genes CYPA, TBP and statistically significant when p < 0.05. GAPDH were used as controls. The quantitative expression of each gene was calculated based on the 2−1Cp method (37), using RESULTS the mean Cp values of the 3 housekeeping genes. The decimal logarithm of fold change (logFC) expression values of each gene VitD3-tolDC Show Phenotypical were considered for the definition of validation criteria. Similar Characteristics of tolDC to the MeanDiff parameter from the microarray, in this case TolDC were generated from samples of 24 healthy donors and genes were considered as differentially expressed in vitD3-tolDC 10 MS patients with 83.1 ± 0.01% purity of monocytes and 98.4 ± Frontiers in Immunology | www.frontiersin.org 4 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers 0.03% viability after CD14+ cells positive selection. After in vitro and mDC from 5 healthy donor samples (33), we applied differentiation, DC were harvested, and their phenotype, purity the filtering criteria described in the methods section to and viability were characterized by flow cytometry following the data results (either MeanDiff vitD3-tolDC vs. mDC > 0.5 the gating strategy depicted in Supplementary Figure 1a. As —first criterium—, while MeanDiff iDC vs. mDC < 0.5, —second determined by the percentage of CD11c+ cells, purity of criterium—; or MeanDiff vitD3-tolDC vs. mDC < −0.5, while DC was always >90%, with 91.1 ± 0.04% viability in both MeanDiff iDC vs. mDC > −0.5, with a statistically significant healthy donor and MS patient cells (Supplementary Figure 1b). differential expression in the first criterium). Briefly, these When we analyzed the phenotype of the cells, the median parameters spotted those genes that were specifically induced by fluorescence intensity (MFI) values for CD83, CD86 and the effect of vitamin D3 over DC according to our microarray HLA-DR on each condition were considered. As displayed in analysis, since they were differentially expressed in vitD3-tolDC Supplementary Figures 1c–e, vitD3-tolDC from healthy donors compared to both mDC and iDC, and therefore could be showed reductions of 83 ± 34%, 59 ± 15% and 71 ± 15% in the considered as potential transcriptomic biomarkers of vitD3- expression of CD83, CD86 and HLA-DR, respectively, compared tolDC (Table 1). As a result, we selected CA2, CAMP, CLEC5A, to mDC, and similar to the expression levels showed by iDC. In CYP24A1, GZMB, IL1R1, MAP7, MUCL1 and SNORD30 genes the case of MS patient-derived vitD3-tolDC, a similar behavior for validation. was observed, with reductions of 77 ± 30%, 60 ± 13% and 61 The subsequent qPCR analysis of the actual expression of ± 14% in CD83, CD86 and HLA-DR expression, respectively, these genes in healthy donors evidenced that all of them compared to mDC, and similar to iDC. All the results reached showed an expression pattern compliant with the expression statistical significance for both healthy donors and MS patients thresholds that were established for the validation —analogously (p < 0.05). to those from the microarray but depicted in logFC, as defined in the methods section—, except for CA2 (uncompliant with VitD3-tolDC Induce Allogeneic the second criterium), GZMB (uncompliant with the second Hyporresponsiveness and Produce criterium) and SNORD30 (uncompliant with the first criterium), as shown in Table 1 and Figure 1A. Consequently, these 3 genes Anti-inflammatory Cytokines were discarded from further analysis and the expression of the The functional assay results evidenced that vitD3-tolDC remaining genes was tested on MS patient samples. In this case, significantly inhibited the proliferation of allogeneic PBMC. As Table 1 and Figure 1B show that, however, neither CAMP, nor shown in Supplementary Figure 2a, a 73.6 ± 16.6% reduction CLEC5A, nor IL1R1 fulfilled our expression criteria. As a result, compared to mDC was observed in healthy donors, similar to the only CYP24A1, MAP7 and MUCL1 could be validated as DEG 84.1 ± 10.7% reduction exhibited by iDC. Analogously, although for our vitD3-tolDC product —reaching statistical significance in a less strong degree, for MS patient-derived DC, a 47.9 ± 25.6% for the first criterium, as required, in all cases (p < 0.05)—, and a 46.3 ± 28.9% reduction in the proliferation induction was and therefore were selected as transcriptomic biomarkers of our observed for vitD3-tolDC and iDC, respectively, in comparison tolDC-inducing protocol. to mDC. In all four comparisons, statistical significance was reached (p < 0.05). When the cytokine secretion profile of vitD3-tolDC was MUCL1 Is Also Induced in tolDC Generated compared to mDC and iDC, an increase in the production of the With IL-10 anti-inflammatory cytokines IL-10 (Supplementary Figure 2b) Given the positive results, we therefore intended to test and TGF-β (Supplementary Figure 2c) was detected in vitD3- the expression of some of these genes in 5 samples of tolDC differentiated from healthy donors (IL-10 vitD3-tolDC : IL10-tolDC differentiated from healthy donors, in order 166.0 ± 287.7 pg/mL vs. IL-10 mDC : 44.0 ± 51.4 pg/mL; p to assess their potential value as biomarkers of a different = 0.003; and TGF-β vitD3-tolDC : 306.5 ± 159.5 pg/mL vs. tolDC-inducing protocol. While the expression of CA2, TGF-β mDC : 188.1 ± 165.3 pg/mL; p = 0.046), but only CAMP, CLEC5A, IL1R1 (data not shown) and MAP7 of IL-10 in the case of vitD3-tolDC generated from MS (Figure 1C) was not altered in these cells, we observed patients (IL-10 vitD3-tolDC : 148.6 ± 141.7 pg/mL vs. IL-10 the up-modulation of MUCL1 gene expression in IL10- mDC : 62.1 ± 54.1 pg/mL; p = 0.043). Furthermore, IL-12 tolDC compared to mDC (logFC IL10-tolDC vs. mDC = 0.960 production could not be detected in any condition (data not ± 0.395; p = 0.034), similar to that observed for vitD3- shown). Finally, no statistically significant changes were found tolDC (Figure 1C). The expression of CYP24A1 was not in the production of GZMB (Supplementary Figure 2d), analyzed, since this gene is directly related to the response nor IFN-γ (Supplementary Figure 2e), between the of the cells to vitamin D3, which was not present in these different conditions. specific cultures. CYP24A1, MAP7 and MUCL1 Genes Are MAP7 and MUCL1 Are Functionally Induced in vitD3-tolDC From Healthy Related in vitD3-tolDC Through a Network Donors and MS Patients of Protein Interactions Following the study previously performed by our group In order to find potentially common genes related to CYP24A1, in a comparative microarray study of vitD3-tolDC, iDC MAP7 and MUCL1 that could provide a mechanistic insight Frontiers in Immunology | www.frontiersin.org 5 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers TABLE 1 | Expression by microarray and qPCR of the selected genes in dendritic cells differentiated from healthy donors and multiple sclerosis patients. MICROARRAY DATA FROM HEALTHY DONORS [FROM NAVARRO-BARRIUSO ET AL. (33)] iDC vs. mDC vitD3-tolDC vs. mDC Gene MeanDiff p-value MeanDiff p-value CA2 1.044 0.097 1.680 0.010 CAMP 0.337 0.279 1.351 <0.001 CLEC5A 0.015 0.975 1.573 0.003 CYP24A1 −0.585 0.264 2.271 <0.001 GZMB 0.097 0.582 0.734 <0.001 MAP7 0.075 0.663 0.880 <0.001 MUCL1 0.084 0.857 2.132 <0.001 SNORD30 −0.040 0.949 1.616 0.016 HEALTHY DONORS iDC vs. mDC vitD3-tolDC vs. mDC vitD3-tolDC vs. iDC Gene logFC ± SD p-value logFC ± SD p-value logFC ± SD p-value CA2 0.566 ± 0.391 0.001 0.858 ± 0.445 <0.001 0.292 ± 0.276 0.042 CAMP 0.290 ± 0.397 0.042 0.957 ± 0.447 <0.001 0.668 ± 0.433 0.003 CLEC5A −0.254 ± 0.371 0.063 0.602 ± 0.324 0.001 0.856 ± 0.346 <0.001 CYP24A1 −1.680 ± 0.422 0.008 1.532 ± 0.637 0.004 3.212 ± 0.914 0.004 GZMB 0.588 ± 0.426 0.001 0.607 ± 0.554 0.001 0.019 ± 0.515 >0.999 IL1R1 −0.133 ± 0.259 0.250 −0.662 ± 0.353 <0.001 −0.528 ± 0.293 <0.001 MAP7 0.473 ± 0.353 0.002 1.015 ± 0.220 <0.001 0.542 ± 0.308 0.007 MUCL1 −0.318 ± 0.453 0.130 1.511 ± 0.419 <0.001 1.829 ± 0.405 <0.001 SNORD30 −0.025 ± 0.204 0.056 −0.001 ± 0.207 0.233 0.024 ± 0.278 >0.999 MULTIPLE SCLEROSIS PATIENTS iDC vs. mDC vitD3-tolDC vs. mDC vitD3-tolDC vs. iDC Gene logFC ± SD p-value logFC ± SD p-value logFC ± SD p-value CAMP 0.431 ± 0.342 0.042 0.441 ± 0.331 0.076 0.010 ± 0.459 >0.999 CLEC5A −0.837 ± 0.539 0.002 0.111 ± 0.249 0.668 0.947 ± 0.490 0.002 CYP24A1 −0.762 ± 0.739 0.160 1.465 ± 0.908 <0.001 2.227 ± 1.448 <0.001 IL1R1 0.143 ± 0.158 0.103 −0.307 ± 0.179 0.002 −0.450 ± 0.214 0.008 MAP7 0.320 ± 0.296 0.081 0.506 ± 0.297 0.003 0.187 ± 0.292 0.174 MUCL1 −0.283 ± 1.209 0.221 0.790 ± 0.325 0.042 1.073 ± 1.362 <0.001 Gene expression values from a microarray study with healthy donors (n = 5) or from qPCR analysis from healthy donors (n = 24, except for CYP24A1, in which n = 10, and GZMB and SNORD30, in which n = 20) and multiple sclerosis patients (n = 10). Data presented as either the mean difference of expression (MeanDiff) for the microarray data or as the mean decimal logarithm of fold change (logFC) ± SD for the qPCR results. Housekeeping genes GAPDH, TBP and CYPA were used as controls. One qPCR experiment was performed for each donor or patient, with triplicated measurements for each sample. Microarray p-values < 0.01 and qPCR p-values < 0.05 are highlighted in bold. iDC, immature dendritic cells; mDC, mature dendritic cells; vitD3-tolDC, vitamin D3-induced tolerogenic dendritic cells. Friedman test with Dunn’s correction or one-way ANOVA test with Geisser-Greenhouse correction. in their functionality and the metabolic pathways triggered in —such as HLA class II antigen presentation and different anti- vitD3-tolDC, we constructed a network of protein interactions. inflammatory mediators— and other cellular processes and Briefly, previously reported data from the literature and several pathways —highlighting the metabolism of retinoic acid and public and private databases were crossed with our microarray the oxidative metabolism—, as shown in Figure 2. Furthermore, results in order to link the function of each gene of our CAMP, CLEC5A, GZMB and IL1R1 genes appeared closely transcriptomic study between them. This approach would allow related to MAP7 and MUCL1 in our network. us to reveal potential interactions between our biomarkers that To validate these results and confirm the close functional might explain their involvement in immune regulation. Our relation between MUCL1 and MAP7, we studied the expression study revealed that, while CYP24A1 seemed to be functionally of DHRS9, OS9, SPARC, ST6GAL1 and THBS1 genes. As shown separated from the rest of the genes, MAP7 and MUCL1 were in Table 2, their expression in our microarray indicated an closely related through common immune-related mechanisms overexpression of DHRS9, SPARC and ST6GAL in vitD3-tolDC Frontiers in Immunology | www.frontiersin.org 6 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers FIGURE 1 | Expression of the selected genes as candidate biomarkers of vitD3-tolDC. (A) Expression of CA2, GZMB and SNORD30 in healthy donors by microarray (n = 5) and quantitative PCR (qPCR; n = 20, except CA2, in which n = 24) in immature dendritic cells (iDC), mature DC (mDC) and vitD3-tolDC. (B) Expression of CAMP, CLEC5A, CYP24A1, MAP7 and MUCL1 in healthy donors both by microarray (n = 5) and qPCR analysis (n = 24, except CYP24A1, in which n = 10) in iDC, mDC and vitD3-tolDC, and in multiple sclerosis (MS) patients by qPCR only (n = 10). (C) Expression of MAP7 and MUCL1 in vitD3-tolDC (n = 24) and in IL10-tolDC (n = 5) by qPCR. Data presented as the mean difference of expression (MeanDiff) or the decimal logarithm of fold change (logFC) expression for the microarray and qPCR results, respectively, in both cases normalized to mDC expression. Housekeeping genes GAPDH, TBP and CYPA were used as controls. One qPCR experiment was performed for each donor or patient, with triplicated measurements for each sample. Error bars corresponding to SEM. Dotted lines represent the logFC = 0.5 or −0.5 expression threshold. *p < 0.05. Friedman test with Dunn’s correction, one-way ANOVA test with Geisser-Greenhouse correction or paired t-test. Frontiers in Immunology | www.frontiersin.org 7 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers FIGURE 2 | Network of protein interactions between CYP24A1, MAP7 and MUCL1. The protein interactions network was built based in both our microarray results and previously reported data. Each node represents a different protein encoding gene. The color of the border of each node indicates the level of expression of each gene in vitD3-tolDC compared to immature dendritic cells (iDC) and the color of the body of each node indicates the expression of each gene compared to mature dendritic cells (mDC). The color scale indicates the level of mean difference expression (MeanDiff) of each gene according to our microarray study, from green (MeanDiff ≤ −1.35) to red (MeanDiff ≥ 1.35). Green and blue lines indicate protein interactions that are related to MAP7 and MUCL1 genes (respectively) in 3 or less parenthood levels, and the red lines indicate interactions that are shared by both genes in <6 parenthood levels. Genes have been grouped in functional clusters, as indicated by the colored bubbles. Arrows indicate those genes that were selected for further validation studies. vs. mDC, but not iDC, while THBS1 and OS9 appeared induced was evidenced in vitD3-tolDC from healthy donors compared and slightly down-modulated, respectively, in vitD3-tolDC to iDC —since it could not be detected in this condition—, compared to iDC. We, therefore, validated their expression in 10 but no statistical significance could be reached in the reduction samples from healthy donors and 10 samples from MS patients, found compared to mDC (VEGF vitD3-tolDC : 205.5 ± 276.9 and observed that the expression pattern of these 5 genes in pg/mL; VEGF mDC : 422.2 ± 497.6 pg/mL). Furthermore, similar healthy donor samples was in accordance to that shown in results were evidenced in samples from MS patients, and our microarray data (Table 2). Furthermore, and with the only VEGF production could not be detected on iDC either (VEGF exception of DHRS9 and OS9, these expression patterns were also vitD3-tolDC : 331.8 ± 321.5 pg/mL; VEGF mDC : 369.2 ± 243.1 confirmed in DC samples obtained from MS patients, as shown pg/mL). On the other hand, the study of the expression of in Table 2. CD25 in healthy donor samples from the microarray evidenced Additionally, VEGFA and IL2RA were also studied due to a strong up-modulation of the IL2RA gene in vitD3-tolDC their relevance in our network, but their expression was validated compared to iDC only. The expression of this gene in vitD3- at the protein level. Their gene expression in our microarray tolDC compared to mDC, however, did not reach statistical study is shown in Table 2. VEGF production (encoded by the significance (Table 2). Accordingly, our results in samples from VEGFA gene) was analyzed by cytometric bead array (CBA) 6 healthy donor DC cultures showed a statistically significant in supernatant samples from 10 healthy donors and 8 MS increase (p = 0.040) in the MFI values of this marker in vitD3- patients, and the surface expression of CD25 (encoded by the tolDC compared to iDC, but not mDC, probably due to the high IL2RA gene) was assessed by flow cytometry in 6 healthy variability observed in this specific condition (CD25 vitD3-tolDC : donor samples. First, in accordance with the results from the 433.8 ± 194.9; CD25 mDC : 796.3 ± 621.2; CD25 iDC : 175.0 microarray (Table 2), an increase in the production of VEGF ± 16.8). Frontiers in Immunology | www.frontiersin.org 8 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers TABLE 2 | Expression by microarray and qPCR of the selected genes from the protein interaction network in dendritic cells differentiated from healthy donors and multiple sclerosis patients. MICROARRAY DATA FROM HEALTHY DONORS [FROM NAVARRO-BARRIUSO ET Al.(33)] vitD3-tolDC vs. mDC vitD3-tolDC vs. iDC Gene MeanDiff p-value MeanDiff p-value DHRS9 1.160 0.007 −0.161 0.685 IL2RA −1.539 0.040 2.297 0.003 OS9 −0.014 0.926 −0.478 0.004 SPARC 1.347 <0.001 0.506 0.101 ST6GAL1 1.564 <0.001 0.189 0.570 THBS1 −1.052 0.048 1.944 <0.001 VEGFA 0.200 0.509 0.696 0.028 qPCR DATA FROM HEALTHY DONORS vitD3-tolDC vs. mDC vitD3-tolDC vs. iDC Gene logFC ± SD p-value logFC ± SD p-value DHRS9 0.683 ± 0.387 0.002 n/a OS9 −0.011 ± 0.172 0.917 −0.265 ± 0.182 0.015 SPARC 1.210 ± 0.346 0.003 n/a ST6GAL1 0.571 ± 0.223 0.002 n/a THBS1 −0.023 ± 0.285 0.694 2.508 ± 0.303 0.004 qPCR DATA FROM MULTIPLE SCLEROSIS PATIENTS vitD3-tolDC vs. mDC vitD3-tolDC vs. iDC Gene logFC ± SD p-value logFC ± SD p-value DHRS9 0.264 ± 0.358 0.557 n/a OS9 −0.036 ± 0.152 >0.999 −0.190 ± 0.222 0.076 SPARC 0.552 ± 0.194 0.011 n/a ST6GAL1 0.403 ± 0.331 0.001 n/a THBS1 −0.027 ± 0.216 0.420 2.391 ± 0.381 0.002 Gene expression values from a microarray study with healthy donors (n = 5) or from qPCR analysis from healthy donors (n = 10) and multiple sclerosis patients (n = 10). Data presented as either the mean difference of expression (MeanDiff) for the microarray data or as the mean decimal logarithm of fold change (logFC) ± SD for the qPCR results. Housekeeping genes GAPDH, TBP and CYPA were used as controls for the qPCR experiments. One qPCR experiment was performed for each donor or patient, with triplicated measurements for each sample. Microarray p-values < 0.01 and qPCR p-values < 0.05 are highlighted in bold. iDC, immature dendritic cells; mDC, mature dendritic cells; vitD3-tolDC, vitamin D3-induced tolerogenic dendritic cells. Friedman test with Dunn’s correction, one-way ANOVA test with Geisser-Greenhouse correction, Wilcoxon matched-pairs signed rank test or paired t-test. Protein Expression of Both MAP7 and × 103 ; MAP7 mDC = 49.58 ± 17.16 × 103 ; MAP7 iDC = MUCL1 Is Strongly Induced in vitD3-tolDC 70.54 ± 36.39 × 103 ), reaching statistical significance in both Since MAP7 and MUCL1 proved themselves as transcriptomic cases (p = 0.016; p = 0.007, respectively). The induction of the biomarkers of vitD3-tolDC in both healthy donors and MS MUCL1-encoded protein was even stronger, as evidenced by the patient samples and, additionally, they were closely related 6.72-fold and 13.02-fold CTCF expression in vitD3-tolDC vs. between them in our functional network, we further analyzed iDC and mDC, respectively (MUCL1 vitD3-tolDC = 1204.85 ± them and determined the actual expression of their respective 509.91 × 103 ; MUCL1 mDC = 92.54 ± 94.19 × 103 ; MUCL1 3 encoded proteins in order to provide more reliability to our iDC = 179.19 ± 32.93 × 10 ). In this case, however, statistical qPCR results. The ICC analysis in 4 healthy donor samples significance was only reached in the comparison vs. mDC evidenced that, in fact, a strong up-modulation of these proteins (p = 0.024), probably due to the small sample size (Figure 3B). was observed in vitD3-tolDC compared to both iDC and mDC Representative microscopy pictures of the expression of both conditions. As shown in Figure 3A, the microtubule-associated proteins in healthy donor samples are shown in Figures 3C,D. protein 7, encoded by MAP7, showed a 2.81-fold and a 4.00- Furthermore, the ICC analysis in 3 MS patient samples presented fold higher CTCF expression in vitD3-tolDC compared to iDC a similar tendency. On the one hand, as shown in Figure 3A, and mDC, respectively (MAP7 vitD3-tolDC = 198.35 ± 40.45 the study of MAP7 expression presented a 1.57-fold and a Frontiers in Immunology | www.frontiersin.org 9 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers 2.14-fold higher CTCF expression in vitD3-tolDC compared study (33), the current report constitutes their first validation to iDC and mDC, respectively (MAP7 vitD3-tolDC = 132.68 as such. ± 91.61 × 103 ; MAP7 mDC = 61.74 ± 31.52 × 103 ; MAP7 Several years ago, mucin-like 1, the protein encoded by 3 iDC = 84.32 ± 66.48 × 10 ). On the other hand, and as MUCL1 gene, was initially identified as a breast-specific gene shown in Figure 3B, the expression of MUCL1 was, again, even expressed in more than 90% of breast cancer cell lines, developing stronger, with a 2.76-fold and 4.49-fold CTCF expression in an important role in the proliferation of these tumor cells (45– vitD3-tolDC compared to iDC and mDC, respectively (MUCL1 48). However, no other specific role has been reported for it. In 3 vitD3-tolDC = 132.34 ± 46.44 × 10 ; MUCL1 mDC = 29.47 fact, its name comes given by the structural analogy of MUCL1 ± 11.66 × 10 ; MUCL1 iDC = 47.97 ± 24.65 × 103 ). 3 protein with mucin proteins, characterized by regions of high However, statistical significance for MS patient samples could tandem repeated serine and threonine content with extensive O- only be reached for the expression of MUCL1 in vitD3-tolDC glycosylation of these residues (46). On the other hand, we have compared to mDC (p = 0.049), again, probably due to the MAP7. This gene encodes the microtubule-associated protein 7, small sample size. Representative microscopy pictures of the a cytoskeleton component that has been mainly related with the expression of both proteins in MS patient samples are shown in development of collateral axon morphogenesis and development Figures 3C,D. in neurons (49, 50). For both MUCL1 and MAP7, a potential role in DC function has not been reported yet. However, interestingly, microtubule associated proteins have been described to be DISCUSSION O-glycosylated —just like MUCL1—, although the functional significance of this modification remains to be determined (51). The identification of biomarkers is a key point for the In the case of vitD3-tolDC, our results have demonstrated a translation of tolDC into the clinic. In this article, we have strong induction of gene and protein expression of both MUCL1 evidenced that CYP24A1, MAP7 and MUCL1 genes appear and MAP7, as already hinted in our previous microarray study, strongly induced in vitD3-tolDC, both in healthy donors thus suggesting that they might be developing an important role and MS patients. Therefore, the differential expression of in the functionality of these cells. However, further functional these genes in our tolerogenic cell product gives them the studies should be performed first in order to elucidate to what potential to unequivocally identify vitD3-tolDC with a simple extent these genes might be involved in the mechanisms of qPCR analysis, without ambiguity, thus ensuring that they are immune tolerance. In any case, the fact that our results also not immunogenic (not mDC), nor susceptible of maturation evidence a strong induction of MUCL1 in IL10-tolDC indicates (not iDC), and consequently characterizing their tolerogenic that this gene might constitute a potential biomarker of the potential. Additionally, the study of the protein expression of regulatory function of tolDC, and it would be interesting to test MAP7 and MUCL1 further supported these results. The study its expression in other protocols generating tolerance-inducing of these genes in the context of the whole transcriptome of cell products. Nevertheless, on the other hand, it is also true the vitD3-tolDC has also elucidated that the role of MAP7 that our above-mentioned preliminary microarray results did and MUCL1 —but not of CYP24A1— seems to be closely not show the overexpression of MUCL1 in dexamethasone nor related with important and widely described immune- and non- rapamycin-induced tolDC, although proper PCR validations immune-related pathways, which correlates with many of the should be performed in order to confirm this. results that we have obtained at the phenotypical (reduction of In concordance with previous reported studies by our co-stimulatory molecules and HLA-DR expression), functional own group (26), and with the objective to finally develop (increased secretion of IL-10 and TGF-β; non detectable an autologous tolerogenic cell-therapy for MS, our results production of IL-12) and transcriptomic levels (reduction of demonstrated that the generation of vitD3-tolDC in both IL1R1 gene expression). Furthermore, many of the interactions healthy donors and MS patients did not significantly differ suggested by our network have also been confirmed in both regarding their phenotypic and functional characteristics —with healthy donor and MS patient cells. Therefore, even though the exception of an increase on TGF-β production in healthy the specific role of MAP7 and MUCL1 in the tolerogenic donor samples only—. In all cases, a consistent decrease in the functionalities of vitD3-tolDC is not fully clear, our results surface expression of CD86 and HLA-DR, as well as in the manifested that they are at least having an effect on several induction of allogeneic proliferation, was observed, accompanied relevant immune regulatory mechanisms and different metabolic by a significant increase in the production of IL-10 and no IL-12 pathways. On the other hand, CYP24A1 gene, encoding the detection. However, the tolerogenic potential of vitD3-tolDC vitamin D3 24-hydroxylase, might well serve as a strong and generated from monocytes of MS patients, although sufficient, robust biomarker of vitD3-tolDC, although it seems to have did not seem as robust as in healthy donors, provided that the little to no influence in the actual regulatory properties of changes in their functional and phenotypic markers were not these cells. After all, this gene is directly involved in the as pronounced. Interestingly, though, the study of CYP24A1 metabolism of vitamin D3 (43, 44), and thus could constitute expression raised another concern in this regard. Even though a robust indicator of the response of the cell product to the this gene constitutes a robust biomarker of the generation of treatment with this compound. Even though we previously vitD3-tolDC, as discussed above, its expression also manifested suggested the potential of CYP24A1, MAP7 and MUCL1 as important differences in iDC from healthy donors and MS candidate biomarkers of vitD3-tolDC in a previous microarray patients; compared to mDC, a strong repression of CYP24A1 was Frontiers in Immunology | www.frontiersin.org 10 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers FIGURE 3 | Immunocytochemistry study of MAP7 and MUCL1 protein expression in dendritic cells. (A) Relative levels of expression of microtubule-associated protein 7 (MAP7) in dendritic cells differentiated from healthy donor samples (n = 4) and MS patient samples (n = 3). (B) Relative levels of expression of mucin-like 1 (MUCL1) in dendritic cells differentiated from healthy donor samples (n = 4) and MS patient samples (n = 3). Results are calculated as percentage (%) of corrected total cell fluorescence (CTCF) in immature dendritic cells (iDC) and mature DC (mDC) compared to vitD3-tolDC. One single immunocytochemistry experiment was performed for each sample. Error bars corresponding to SEM. *p < 0.05. One-way ANOVA test with Geisser-Greenhouse correction or paired t-test. Representative pictures of the expression of (C) MAP7 and (D) MUCL1 in iDC, mDC and vitD3-tolDC from healthy donor and MS patient samples. α-tubulin staining is shown in green; either MAP7 or MUCL1 staining is shown in red; nuclei staining is shown in blue. The immunocytochemistry analysis was performed on a fluorescence microscope using a 63x objective. Immunocytochemistry primary staining was performed using mouse anti-human α-tubulin and either rabbit anti-human MAP7 or rabbit anti-human MUCL1 antibodies. Secondary stainings were performed using AlexaFluor (AF) 488 goat anti-mouse IgG and AF 594 goat anti-rabbit IgG antibodies. Nuclei staining was performed using DAPI. Frontiers in Immunology | www.frontiersin.org 11 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers observed in healthy donor samples, but the expression of this for the use of these genes as a reliable quality control in gene in iDC from MS patients was similar to mDC. Therefore, clinical trials with a simple qPCR analysis, before administering this result implicates that the transcriptomic profile of DC in MS the cell product into the patients. Furthermore, the role of patients is already different before receiving the same treatment MAP7 and MUCL1, but not of CYP24A1, seems to be strongly with vitamin D3, thus suggesting a more pro-inflammatory related to important immune-related functions. Specifically, the baseline status in these cells. However, the strong induction of case of MUCL1 is of significant relevance, since this gene CYP24A1 observed in vitD3-tolDC still indicates that these cells has also demonstrated an interesting potential as a broad- are adequately responding to the effect of vitamin D3, suggesting use biomarker of tolerance, based on its validation both in that the problem has to be downstream. vitD3-tolDC and in IL10-tolDC. Consequently, MUCL1 sets These differences between healthy donors and MS patient an interesting path for future experiments with the objective samples also forced us to discard another interesting candidate to validate the role of this gene as a potential biomarker of transcriptomic biomarker, DHRS9, the gene encoding the other tolDC protocols, and thus, hopefully as a wide biomarker dehydrogenase/reductase 9, involved in the metabolism of of tolerogenic cell products and their mechanisms of immune retinoic acid (52). The relevance of this gene comes given by its tolerance induction. potential as a broad-use biomarker of immune tolerance, since it has been already described as differentially induced in vitamin ETHICS STATEMENT D3 + dexamethasone-induced tolDC (53) and, especially, in regulatory macrophages (54). It is also worth noting that, even This study was approved by the Germans Trias i Pujol Hospital though DHRS9 could not be confirmed as a biomarker of vitD3- ethical committee, and all patients and healthy controls signed an tolDC from MS patients, this does not mean that it could not informed consent. still be useful in cells generated from patients with a different autoimmune disease, such as rheumatoid arthritis or type 1 AUTHOR CONTRIBUTIONS diabetes. In this regard, our results already allowed us to point out some transcriptomic differences between healthy donors and EM-C, JN-B, and MM conceived the experiments. CR-T MS patients. We observed that CLEC5A could not be validated and SP-R obtained the patient samples. AtB provided cDNA in MS patient samples either, and when we studied our network, samples from IL10-tolDC. AA-M, BQ-S, JN-B, and MM we noticed that both CLEC5A and DHRS9 were closely related performed the cell cultures. AA-M, BQ-S, JN-B, and MM to MUCL1 through GALNT2, GALNT7 and GALNT11 genes. performed the cell characterization analyses. JN-B and MM Specifically, these 3 genes have been described to be in charge of analyzed the results. EM-C, JN-B, and MM interpreted the the processes of O-linked glycosylation. Consequently, this might results. JN-B wrote the manuscript. AA-M, AtB, AT-S, BQ-S, be an indicator of a potential misfunction in the glycosylation CR-T, EM-C, JN-B, MM, and SP-R reviewed the manuscript. mechanisms of MS patients, that might be indirectly affecting the expression of CLEC5A, DHRS9 and probably many other ACKNOWLEDGMENTS genes. In fact, several glycosylation defects have already been related with the pathogenesis of MS, as reviewed by Grigorian This work was partially supported by projects PI14/01175, et al. (55). However, MUCL1 gene expression would also be PI16/01737 and PI17/01521, integrated in the Plan Nacional de expected to be affected by a defect in these genes, but our results I+D+I and co-supported by the ISCIII-Subdirección General showed that this was not the case. As a possible explanation, de Evaluación and the Fondo Europeo de Desarrollo Regional the glycosylation of MUCL1 might be sufficiently mediated by (FEDER), and by grant 140191 from IWT-TBM (Belgium). ST6GAL1 gene alone —whose expression still remains slightly JN-B was beneficiary of a FI-DGR Ph.D. contract from induced in vitD3-tolDC from MS patient samples—. In any case, AGAUR (Agència de Gestió d’Ajuts Universitaris i de Recerca), our results have raised the need to explore the mechanisms that supported by the Government of Catalonia (02/2015-01/2018). might be different between healthy donors and MS patients. The contracts of AA-M, JN-B, and MM have been partially On the other hand, most of the genes closely related to supported by grant 140191 from IWT-TBM. The authors are MAP7 showed a similar behavior in both healthy donor and members of a consolidated group recognized by AGAUR (2017 MS patient vitD3-tolDC, such as SPARC, THBS1 or, specially, SGR 103), supported by the Government of Catalonia. This VEGFA, which has been reported to develop a role in the work has been supported by positive discussion through A recruitment, inhibition of maturation and IDO1 induction in FACTT network (COST Action BM1305: www.afactt.eu) and DC (56–58). Project 779316 - ReSToRe (www.h2020restore.eu). COST and In conclusion, our results evidenced that, despite not having ReSToRe are supported by the EU Framework Programme an obvious involvement in the tolerogenic functionality of the Horizon 2020. cells in all cases, several genes have shown a strong differential expression in vitD3-tolDC from healthy donors, compared to SUPPLEMENTARY MATERIAL both mature and immature control conditions. Among them, CYP24A1, MAP7 and MUCL1 have also been validated as robust The Supplementary Material for this article can be found transcriptomic biomarkers of vitD3-tolDC generated from MS online at: https://www.frontiersin.org/articles/10.3389/fimmu. patient samples. Thus, this finding opens a promising window 2019.01251/full#supplementary-material Frontiers in Immunology | www.frontiersin.org 12 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers REFERENCES 22. Pedersen AW, Holmstrøm K, Jensen SS, Fuchs D, Rasmussen S, Kvistborg P, et al. Phenotypic and functional markers for 1α,25-dihydroxyvitamin 1. Banchereau J, Briere F, Caux C, Davoust J, Lebecque S, Liu Y-J, et al. D3-modified regulatory dendritic cells. Clin Exp Immunol. (2009) 157:48– Immunobiology of dendritic cells. Annu Rev Immunol. (2000) 18:767– 59. doi: 10.1111/j.1365-2249.2009.03961.x 811. doi: 10.1146/annurev.immunol.18.1.767 23. Széles L, Keresztes G, Töröcsik D, Balajthy Z, Krenács L, Póliska S, et al. 2. Mueller DL. Mechanisms maintaining peripheral tolerance. Nat Immunol. 1,25-dihydroxyvitamin D3 is an autonomous regulator of the transcriptional (2010) 11:21–7. doi: 10.1038/ni.1817 changes leading to a tolerogenic dendritic cell phenotype. J Immunol Baltim 3. Ganguly D, Haak S, Sisirak V, Reizis B. The role of dendritic cells in Md 1950. (2009) 182:2074–83. doi: 10.4049/jimmunol.0803345 autoimmunity. Nat Rev Immunol. (2013) 13:566–77. doi: 10.1038/nri3477 24. Ferreira GB, Kleijwegt FS, Waelkens E, Lage K, Nikolic T, Hansen DA, 4. Hopp A-K, Rupp A, Lukacs-Kornek V. Self-antigen presentation et al. Differential protein pathways in 1,25-dihydroxyvitamin d(3) and by dendritic cells in autoimmunity. Front Immunol. (2014) dexamethasone modulated tolerogenic human dendritic cells. J Proteome Res. 5:55. doi: 10.3389/fimmu.2014.00055 (2012) 11:941–71. doi: 10.1021/pr200724e 5. Hilkens CMU, Isaacs JD, Thomson AW. Development of dendritic cell- 25. Raïch-Regué D, Naranjo-Gómez M, Grau-López L, Ramo C, Pujol-Borrell based immunotherapy for autoimmunity. Int Rev Immunol. (2010) 29:156– R, Martínez-Cáceres E, et al. Differential effects of monophosphoryl lipid 83. doi: 10.3109/08830180903281193 A and cytokine cocktail as maturation stimuli of immunogenic and 6. Ten Brinke A, Hilkens CMU, Cools N, Geissler EK, Hutchinson JA, tolerogenic dendritic cells for immunotherapy. Vaccine. (2012) 30:378– Lombardi G, et al. Clinical use of tolerogenic dendritic cells-harmonization 87. doi: 10.1016/j.vaccine.2011.10.081 approach in European collaborative effort. Mediators Inflamm. (2015) 26. Raïch-Regué D, Grau-López L, Naranjo-Gómez M, Ramo-Tello 2015:471719. doi: 10.1155/2015/471719 C, Pujol-Borrell R, Martínez-Cáceres E, et al. Stable antigen- 7. Adler HS, Steinbrink K. Tolerogenic dendritic cells in health and disease: specific T-cell hyporesponsiveness induced by tolerogenic dendritic friend and foe! Eur J Dermatol. (2007) 17:476–91. doi: 10.1684/ejd.2007.0262 cells from multiple sclerosis patients. Eur J Immunol. (2012) 8. Morelli AE, Thomson AW. Tolerogenic dendritic cells and the quest for 42:771–82. doi: 10.1002/eji.201141835 transplant tolerance. Nat Rev Immunol. (2007) 7:610–21. doi: 10.1038/nri2132 27. Ferreira GB, Vanherwegen A-S, Eelen G, Gutiérrez ACF, Van Lommel L, 9. Suwandi JS, Nikolic T, Roep BO. Translating mechanism of regulatory action Marchal K, et al. Vitamin D3 induces tolerance in human dendritic cells by of tolerogenic dendritic cells to monitoring endpoints in clinical trials. Front activation of intracellular metabolic pathways. Cell Rep. (2015) 10:711–25. Immunol. (2017) 8:01598. doi: 10.3389/fimmu.2017.01598 doi: 10.1016/j.celrep.2015.01.013 10. Steinbrink K, Wölfl M, Jonuleit H, Knop J, Enk AH. Induction of tolerance by 28. Malaguarnera L, Marsullo A, Zorena K, Musumeci G, Di Rosa M. Vitamin IL-10-treated dendritic cells. J Immunol. (1997) 159:4772–80. D3 regulates LAMP3 expression in monocyte derived dendritic cells. Cell 11. Boks MA, Kager-Groenland JR, Haasjes MSP, Zwaginga JJ, van Ham SM, Immunol. (2017) 311:13–21. doi: 10.1016/j.cellimm.2016.09.013 ten Brinke A. IL-10-generated tolerogenic dendritic cells are optimal for 29. Farias AS, Spagnol GS, Bordeaux-Rego P, Oliveira COF, Fontana AGM, de functional regulatory T cell induction–a comparative study of human Paula RFO, et al. Vitamin D3 induces IDO+ tolerogenic DCs and enhances clinical-applicable DC. Clin Immunol Orlando Fla. (2012) 142:332– Treg, reducing the severity of EAE. CNS Neurosci Ther. (2013) 19:269– 42. doi: 10.1016/j.clim.2011.11.011 77. doi: 10.1111/cns.12071 12. Naranjo-Gómez M, Raïch-Regué D, Oñate C, Grau-López L, Ramo-Tello C, 30. Mansilla MJ, Sellès-Moreno C, Fàbregas-Puig S, Amoedo J, Navarro-Barriuso Pujol-Borrell R, et al. Comparative study of clinical grade human tolerogenic J, Teniente-Serra A, et al. Beneficial effect of tolerogenic dendritic cells pulsed dendritic cells. J Transl Med. (2011) 9:89. doi: 10.1186/1479-5876-9-89 with MOG autoantigen in experimental autoimmune encephalomyelitis. CNS 13. Xia C-Q, Peng R, Beato F, Clare-Salzler MJ. Dexamethasone induces IL-10- Neurosci Ther. (2015) 21:222–30. doi: 10.1111/cns.12342 producing monocyte-derived dendritic cells with durable immaturity. Scand J 31. Mansilla MJ, Contreras-Cardone R, Navarro-Barriuso J, Cools N, Berneman Immunol. (2005) 62:45–54. doi: 10.1111/j.1365-3083.2005.01640.x Z, Ramo-Tello C, et al. Cryopreserved vitamin D3-tolerogenic dendritic cells 14. Cabezón R, Carrera-Silva EA, Flórez-Grau G, Errasti AE, Calderón-Gómez E, pulsed with autoantigens as a potential therapy for multiple sclerosis patients. Lozano JJ, et al. MERTK as negative regulator of human T cell activation. J J Neuroinflammation. (2016) 13:113. doi: 10.1186/s12974-016-0584-9 Leukoc Biol. (2015) 97:751–60. doi: 10.1189/jlb.3A0714-334R 32. Xing N, L Maldonado ML, Bachman LA, McKean DJ, Kumar R, 15. Flórez-Grau G, Cabezón R, Borgman KJE, España C, Lozano JJ, Garcia- Griffin MD. Distinctive dendritic cell modulation by vitamin D(3) and Parajo MF, et al. Up-regulation of EP2 and EP3 receptors in human glucocorticoid pathways. Biochem Biophys Res Commun. (2002) 297:645– tolerogenic dendritic cells boost the immunosuppressive activity of PGE2. J 52. doi: 10.1016/S0006-291X(02)02262-3 Leukoc Biol. (2017) 102:881–95. doi: 10.1189/jlb.2A1216-526R 33. Navarro-Barriuso J, Mansilla MJ, Naranjo-Gómez M, Sánchez-Pla A, Quirant- 16. Fedoric B, Krishnan R. Rapamycin downregulates the inhibitory receptors Sánchez B, Teniente-Serra A, et al. Comparative transcriptomic profile of ILT2, ILT3, ILT4 on human dendritic cells and yet induces T cell tolerogenic dendritic cells differentiated with vitamin D3, dexamethasone hyporesponsiveness independent of FoxP3 induction. Immunol Lett. (2008) and rapamycin. Sci Rep. (2018) 8:14985. doi: 10.1038/s41598-01 120:49–56. doi: 10.1016/j.imlet.2008.06.009 8-33248-7 17. Giannoukakis N, Phillips B, Finegold D, Harnaha J, Trucco M. Phase I (safety) 34. Navarro-Barriuso J, Mansilla MJ, Martínez-Cáceres EM. Searching study of autologous tolerogenic dendritic cells in type 1 diabetic patients. for the transcriptomic signature of immune tolerance induction— Diabetes Care. (2011) 34:2026–32. doi: 10.2337/dc11-0472 biomarkers of safety and functionality for tolerogenic dendritic 18. Lu L, Lee WC, Takayama T, Qian S, Gambotto A, Robbins PD, et al. Genetic cells and regulatory macrophages. Front Immunol. (2018) engineering of dendritic cells to express immunosuppressive molecules 9:02062. doi: 10.3389/fimmu.2018.02062 (viral IL-10, TGF-beta, and CTLA4Ig). J Leukoc Biol. (1999) 66:293– 35. Pujol A, Mosca R, Farrés J, Aloy P. Unveiling the role of network and 6. doi: 10.1002/jlb.66.2.293 systems biology in drug discovery. Trends Pharmacol Sci. (2010) 31:115– 19. Adorini L, Penna G, Giarratana N, Roncari A, Amuchastegui S, Daniel 123. doi: 10.1016/j.tips.2009.11.006 KC, et al. Dendritic cells as key targets for immunomodulation by 36. Perera S, Artigas L, Mulet R, Mas JM, Sardón T. Systems biology applied Vitamin D receptor ligands. J Steroid Biochem Mol Biol. (2004) 89–90:437– to non-alcoholic fatty liver disease (NAFLD): treatment selection based 41. doi: 10.1016/j.jsbmb.2004.03.013 on the mechanism of action of nutraceuticals. Nutrafoods. (2014) 13:61– 20. Penna G, Roncari A, Amuchastegui S, Daniel KC, Berti E, Colonna M, et al. 8. doi: 10.1007/s13749-014-0022-5 Expression of the inhibitory receptor ILT3 on dendritic cells is dispensable for 37. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real- induction of CD4+Foxp3+ regulatory T cells by 1,25-dihydroxyvitamin D3. time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods San Blood. (2005) 106:3490–7. doi: 10.1182/blood-2005-05-2044 Diego Calif. (2001) 25:402–8. doi: 10.1006/meth.2001.1262 21. Adorini L, Penna G. Induction of tolerogenic dendritic cells 38. McCloy RA, Rogers S, Caldon CE, Lorca T, Castro A, Burgess A. Partial by vitamin D receptor agonists. Handb Exp Pharmacol. (2009) inhibition of Cdk1 in G 2 phase overrides the SAC and decouples mitotic 188:251–73. doi: 10.1007/978-3-540-71029-5_12 events. Cell Cycle Georget Tex. (2014) 13:1400–12. doi: 10.4161/cc.28401 Frontiers in Immunology | www.frontiersin.org 13 June 2019 | Volume 10 | Article 1251 Navarro-Barriuso et al. MAP7 and MUCL1, vitD3-tolDC Biomarkers 39. Burgess A, Vigneron S, Brioudes E, Labbé J-C, Lorca T, Castro A. Loss of 51. Ding M, Vandré DD. High molecular weight microtubule-associated proteins human Greatwall results in G2 arrest and multiple mitotic defects due to contain O-linked-N-acetylglucosamine. J Biol Chem. (1996) 271:12555– deregulation of the cyclin B-Cdc2/PP2A balance. Proc Natl Acad Sci USA. 61. doi: 10.1074/jbc.271.21.12555 (2010) 107:12564–9. doi: 10.1073/pnas.0914191107 52. Napoli JL. Physiological insights into all-trans-retinoic acid biosynthesis. 40. Abramoff MD, Magalhães PJ, Ram SJ. Image processing with ImageJ. Biochim Biophys Acta. (2012) 1821:152–67. doi: 10.1016/j.bbalip.2011. Biophotonics Int. (2004) Available at: http://dspace.library.uu.nl/handle/1874/ 05.004 204900 (accessed January 23, 2018). 53. Malinarich F, Duan K, Hamid RA, Bijin A, Lin WX, Poidinger M, et al. 41. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, High mitochondrial respiration and glycolytic capacity represent a metabolic et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. phenotype of human tolerogenic dendritic cells. J Immunol. (2015) 194:5174– (2012) 9:676. doi: 10.1038/nmeth.2019 86. doi: 10.4049/jimmunol.1303316 42. Bland JM, Altman DG. Analysis of continuous data from small samples. BMJ. 54. Riquelme P, Amodio G, Macedo C, Moreau A, Obermajer N, Brochhausen (2009) 338:a3166. doi: 10.1136/bmj.a3166 C, et al. DHRS9 is a stable marker of human regulatory macrophages. 43. Sakaki T, Sawada N, Komai K, Shiozawa S, Yamada S, Yamamoto K, et al. Dual Transplantation. (2017) 101:2731–38. doi: 10.1097/TP.0000000000 metabolic pathway of 25-hydroxyvitamin D3 catalyzed by human CYP24. Eur 001814 J Biochem. (2000) 267:6158–65. doi: 10.1046/j.1432-1327.2000.01680.x 55. Grigorian A, Mkhikian H, Li CF, Newton BL, Zhou RW, Demetriou 44. Masuda S, Kaufmann M, Byford V, Gao M, St-Arnaud R, Arabian A, M. Pathogenesis of multiple sclerosis via environmental and genetic et al. Insights into Vitamin D metabolism using cyp24 over-expression dysregulation of N-glycosylation. Semin Immunopathol. (2012) 34:415– and knockout systems in conjunction with liquid chromatography/mass 24. doi: 10.1007/s00281-012-0307-y spectrometry (LC/MS). J Steroid Biochem Mol Biol. (2004) 89–90:149– 56. Ellis LM, Hicklin DJ. VEGF-targeted therapy: mechanisms of anti-tumour 153. doi: 10.1016/j.jsbmb.2004.03.094 activity. Nat Rev Cancer. (2008) 8:579–91. doi: 10.1038/nrc2403 45. Houghton RL, Dillon DC, Molesh DA, Zehentner BK, Xu J, Jiang J, et al. 57. Marti LC, Pavon L, Severino P, Sibov T, Guilhen D, Moreira-Filho Transcriptional complementarity in breast cancer: application to detection of CA. Vascular endothelial growth factor-A enhances indoleamine circulating tumor cells. Mol Diagn J Devoted Underst Hum Dis Clin Appl Mol 2,3-dioxygenase expression by dendritic cells and subsequently Biol. (2001) 6:79–91. doi: 10.2165/00066982-200106020-00003 impacts lymphocyte proliferation. Mem Inst Oswaldo Cruz. (2014) 46. Colpitts TL, Billing P, Granados E, Hayden M, Hodges S, Roberts L, et al. 109:70–9. doi: 10.1590/0074-0276130252 Identification and immunohistochemical characterization of a mucin-like 58. Rodriguez-Fernandez S, Pujol-Autonell I, Brianso F, Perna-Barrull D, Cano- glycoprotein expressed in early stage breast carcinoma. Tumour Biol J Sarabia M, Garcia-Jimeno S, et al. Phosphatidylserine-liposomes promote Int Soc Oncodevelopmental Biol Med. (2002) 23:263–78. doi: 10.1159/0000 tolerogenic features on dendritic cells in human type 1 diabetes by apoptotic 68566 mimicry. Front Immunol. (2018) 9:253. doi: 10.3389/fimmu.2018.00253 47. Miksicek RJ, Myal Y, Watson PH, Walker C, Murphy LC, Leygue E. Identification of a novel breast- and salivary gland-specific, mucin-like gene Conflict of Interest Statement: The authors declare that the research was strongly expressed in normal and tumor human mammary epithelium. Cancer conducted in the absence of any commercial or financial relationships that could Res. (2002) 62:2736–40. be construed as a potential conflict of interest. 48. Conley SJ, Bosco EE, Tice DA, Hollingsworth RE, Herbst R, Xiao Z. HER2 drives Mucin-like 1 to control proliferation in breast cancer cells. Oncogene. Copyright © 2019 Navarro-Barriuso, Mansilla, Quirant-Sánchez, Ardiaca-Martínez, (2016) 35:4225–34. doi: 10.1038/onc.2015.487 Teniente-Serra, Presas-Rodríguez, ten Brinke, Ramo-Tello and Martínez-Cáceres. 49. Tymanskyj SR, Yang B, Falnikar A, Lepore AC, Ma L. MAP7 Regulates axon This is an open-access article distributed under the terms of the Creative Commons collateral branch development in dorsal root ganglion neurons. J Neurosci. Attribution License (CC BY). The use, distribution or reproduction in other forums (2017) 37:1648–61. doi: 10.1523/JNEUROSCI.3260-16.2017 is permitted, provided the original author(s) and the copyright owner(s) are credited 50. Tymanskyj SR, Yang BH, Verhey KJ, Ma L. MAP7 regulates axon and that the original publication in this journal is cited, in accordance with accepted morphogenesis by recruiting kinesin-1 to microtubules and modulating academic practice. No use, distribution or reproduction is permitted which does not organelle transport. eLife. (2018) 7:e36374. doi: 10.7554/eLife.36374 comply with these terms. Frontiers in Immunology | www.frontiersin.org 14 June 2019 | Volume 10 | Article 1251