Vitamin D Regulates Mertk-Dependent Phagocytosis in Human Myeloid Cells Jelani Clarke, Moein Yaqubi, Naomi C

Vitamin D Regulates MerTK-Dependent Phagocytosis in Human Myeloid Cells Jelani Clarke, Moein Yaqubi, Naomi C. Futhey, Sara Sedaghat, Caroline Baufeld, Manon Blain, Sergio Baranzini, This information is current as Oleg Butovsky, Jack Antel, John H. White and Luke M. of October 4, 2021. Healy J Immunol published online 15 June 2020 http://www.jimmunol.org/content/early/2020/06/14/jimmun ol.2000129 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2020/06/14/jimmunol.200012 Material 9.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 4, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2020 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 15, 2020, doi:10.4049/jimmunol.2000129 The Journal of Immunology

Vitamin D Regulates MerTK-Dependent Phagocytosis in Human Myeloid Cells

Jelani Clarke,* Moein Yaqubi,* Naomi C. Futhey,* Sara Sedaghat,* Caroline Baufeld,† Manon Blain,* Sergio Baranzini,‡ Oleg Butovsky,† Jack Antel,* John H. White,x,{ and Luke M. Healy*

Vitamin D deﬁciency is a major environmental risk factor for the development of multiple sclerosis. The major circulating metabolite of vitamin D (25-hydroxyvitamin D) is converted to the active form (calcitriol) by the hydroxylase enzyme CYP27B1. In multiple sclerosis lesions, the tyrosine kinase MerTK expressed by myeloid cells regulates phagocytosis of myelin debris and apoptotic cells that can accumulate and inhibit tissue repair and remyelination. In this study, we explored the effect of calcitriol on homeostatic (M-CSF, TGF-b–treated) and proinﬂammatory (GM-CSF–treated) human monocyte-derived macrophages and

microglia using RNA sequencing. Transcriptomic analysis revealed significant calcitriol-mediated effects on both Ag presentation Downloaded from and phagocytosis pathways. Calcitriol downregulated MerTK mRNA and protein expression in both myeloid populations, resulting in reduced capacity of these cells to phagocytose myelin and apoptotic T cells. Proinflammatory myeloid cells expressed high levels of CYP27B1 compared with homeostatic myeloid cells. Only proinflammatory cells in the presence of TNF-a generated calcitriol from 25-hydroxyvitamin D, resulting in repression of MerTK expression and function. This selective production of calcitriol in proinflammatory myeloid cells has the potential to reduce the risk for autoantigen presentation while retaining the

phagocytic ability of homeostatic myeloid cells. The Journal of Immunology, 2020, 205: 000–000. http://www.jimmunol.org/

itamin D deficiency is a major environmental risk factor 25-hydroxyvitamin D (25OHD; calcifediol), the major circulating for the development of multiple sclerosis (MS) (1). Al- metabolite, and then to hormonally active 1,25-dihyroxyvitamin D V though widely prescribed for patients with MS, the (1,25(OH)2D; calcitriol) through sequential hydroxylation, catalyzed impact of vitamin D on disease course and severity, as well as its by 25-hydroxylases (CYP2R1, CYP27A1) and 25-hydroxyvitamin mechanisms of action, are poorly understood. Active vitamin D D3 1-a-hydroxylase (CYP27B1), respectively (2). Levels of 25OHD (calcitriol) is obtained from the cutaneous production of vitamin are used clinically to assess vitamin D status (3). Calcitriol functions D3 (cholecalciferol) in the presence of sufficient UV B irradiation, as a ligand for the vitamin D receptor, a member of the nuclear by guest on October 4, 2021 as well as limited dietary sources. Cholecalciferol is converted to receptor family of hormone-regulated transcription factors (3). Catabolism of 25OHD and calcitriol is initiated by the CYP24A1 enzyme, whose expression is tightly regulated by calcitriol in a *Neuroimmunology Unit, Department of Neurology and Neurosurgery, Montreal Neurological Institute, McGill University, Montreal, Quebec H3A 0G4, Canada; negative feedback loop. CYP27B1 is abundantly expressed in †Ann Romney Center for Neurologic Diseases, Department of Neurology, Brigham most biological systems, allowing for local calcitriol production ‡ and Women’s Hospital, Harvard Medical School, Boston, MA 02115; Department of in several tissues, including the CNS. Importantly, CYP27B1 Neurology, Weill Institute for Neurosciences, University of California-San Francisco, San Francisco, CA 94115; xDepartments of Physiology and Medicine, McGill University, expression is regulated by a complex cytokine network in im- Montreal, Quebec H3A 0G4, Canada; and {Evergrande Center for Immunologic mune cells, including cells of myeloid origin (4). Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA Cells of myeloid lineage, including endogenous microglia and 02115 infiltrating monocyte-derived macrophages (MDMs), are the ORCIDs: 0000-0001-7299-5930 (M.Y.); 0000-0002-0680-3141 (N.C.F.); 0000-0001- 7966-7534 (S.S.); 0000-0003-0067-194X (S.B.); 0000-0003-0186-8867 (O.B.); dominant cell population within active MS lesions (5). We have 0000-0002-4785-2687 (J.H.W.); 0000-0001-9496-2216 (L.M.H.). previously shown that the myeloid cell–mediated phagocytic Received for publication February 7, 2020. Accepted for publication May 17, 2020. clearance of myelin debris, a process required for efficient This work was supported by a Montreal Neurological Institute Start-up Fund and the remyelination, is regulatedbyMerTK,amemberoftheTAM Multiple Sclerosis International Federation (PA-1604-08459). family of receptor tyrosine kinases (6). MerTK deficiency results The RNA-sequencing data presented in this article have been submitted to the Gene in delayed remyelination in the cuprizone model of demyelination Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number (7). MDMs derived from MS patients show impaired ability to GSE148986. phagocytose myelin, a defect linked to a reduction in MerTK Address correspondence and reprint requests to Dr. Luke M. Healy, McGill Uni- versity, 3801 University Street, Montreal, QC H2V 4E5, Canada. E-mail address: expression (8). In addition to clearing myelin debris, MerTK [email protected] mediates the process of efferocytosis, the removal of dead/dying The online version of this article contains supplemental material. cells, which is important for autoreactive T cell fate determination Abbreviations used in this article: FSC, forward light scatter; GWAS, genome-wide in MS (9). The functions of myeloid cells are dependent on their association study; iCTNet, Integrated Complex Traits Network; IPA, ingenuity path- state of activation. TGF-b, a key cytokine involved in CNS ho- way analysis; MDM, monocyte-derived macrophage; MØ0, CNS homeostatic mac- meostasis, has been shown to maintain cells in a homeostatic state rophage differentiated using M-CSF+TGFb;MØGMcsf, proinflammatory macrophage differentiated using GM-CSF; MS, multiple sclerosis; 25OHD, 25-hydroxyvitamin characterized by high expression of MerTK, TREM2, CSF1R, and D; ORA, overrepresentation analysis; oRBC, opsonized RBC; PCA, principle com- MAFB (10). In contrast, MerTK expression is comparatively lower ponent analysis; RNAseq, RNA sequencing; SNP, single-nucleotide polymorphism. in proinflammatory myeloid cells, a population shown to contribute Copyright Ó 2020 by The American Association of Immunologists, Inc. 0022-1767/20/$37.50 to MS pathogenesis (6). Genome-wide association studies (GWAS)

www.jimmunol.org/cgi/doi/10.4049/jimmunol.2000129 2 VITAMIN D REGULATES MYELOID CELLS VIA THE MerTK PATHWAY have explained much of MS heritability. Single-nucleotide poly- marker (MerTK, CD80, CD86, HLA-DR/DP/DQ, HLA-ABC, CD40, CD274) morphisms (SNPs) in CYP24A1 and CYP27B1, which tightly test Abs. Cells were washed, and flow cytometry was performed using the regulate the intracellular levels of calcitriol, have been associated Attune NxT (Thermo Fisher Scientific). Myeloid cells were gated based on side scatter area and forward light scatter (FSC) area. Doublets were with an increased risk of MS (11–13). excluded using FSC area and FSC height. Live cells were gated based on In the current study, we investigated calcitriol-mediated tran- live-dead staining (Aqua; Life Technologies). scriptomic regulation of human MDMs and microglia. RNA se- Apoptosis assay quencing (RNAseq) revealed significant calcitriol-mediated negative regulation of both phagocytic and Ag-presenting pathways in these Isolated T cells were collected and resuspended to 1 3 106 cells per cell types. We demonstrate that calcitriol represses MerTK expres- milliliter in PBS. Cells were exposed to UV for 1 h. Following exposure, cells were collected, pelleted, and processed for phagocytosis, as previ- sion and phagocytic capacity of primary myeloid cells and signifi- ously described, for myelin. pHRodamine-dyed cells were inoculated into cantly downregulates components of the Ag presentation pathway. macrophage cultures at a density of 5:1 T/M and left to incubate for 1 h. Notably, proinflammatory myeloid cells expressing the lowest levels Assessment of apoptosis was done by flow cytometry using Alexa 488 of MerTK have the most active vitamin D metabolic processing Annexin V/Dead Cell Apoptosis Kit (Thermo Fisher Scientific). pathway and are, therefore, able to respond to the precursor 25OHD. RNAseq In contrast, lack of endogenous processing of 25OHD in homeostatic myeloid cells maintains high MerTK expression and, Control and calcitriol-treated MDMs and microglia were collected in TRIzol reagent (Invitrogen), and RNA was extracted according to the therefore, participation in the immunologically silent clearance manufacturer’s protocol (Qiagen). Smart-Seq2 libraries were prepared of myelin debris and apoptotic cells. by the Broad Technology Labs and sequenced by the Broad Genomics Platform. cDNA libraries were generated from the Smart-seq2 protocol (16). RNAseq was performed using Illumina NextSeq500 and a High Downloaded from Materials and Methods Output v2 Kit to generate 2 3 25 bp reads. Reads were aligned to the MDMs hg19 genome with STAR aligner and quantified by the Broad Technology Labs’ computational pipeline using Cuffquant version 2.2.1 (17, 18). Human PBMCs were isolated from healthy donors by Ficoll-Hypaque Raw counts were normalized using trimmed mean of maximum values density gradient centrifugation (GE Healthcare). Monocytes were isolated normalization and then log2-transformed. The read counts for each from PBMCs using magnetic CD14+ isolation beads (Miltenyi Biotec). sample were used for differential expression analysis with the edgeR Proinflammatory macrophages differentiated using GM-CSF (MØ ) package (19, 20). The differentially expressed genes were identified

GMcsf http://www.jimmunol.org/ and alternative (M2) macrophages were generated by differentiating using p value ,0.05 and log2 fold change .1. Principle component monocytes for 6 d in the presence of 25 ng/ml GM-CSF and M-CSF, analysis (PCA) was carried out using built-in R function, prcomp,and respectively. To generate CNS homeostatic macrophages differentiated visualized using gplot package. Heatmaps were created using ggplot2 using M-CSF+TGFb (MØ0), TGF-b (50 ng/ml) was added to the M2 M-CSF package in R. The full list of identified genes was used to generate volcano 2 culture conditions on days 3 and 6. A concentration of 10 7 M calcitriol plots in R. For PCA and heatmap graphs, variance of genes across all (Selleck Chemicals) was added to designated macrophages on day 1 of macrophage phenotypes was calculated and the top-500 highly variable culture and maintained throughout differentiation. Culture media was genes were used for further analysis. replenished every 2–3 d. Quantitative PCR Microglia and astrocytes Cells were lysed in TRIzol (Invitrogen). Total RNA extraction was per- Human adult microglia were isolated from brain tissue of patients under- formed using standard protocols followed by DNAse treatment according by guest on October 4, 2021 going brain surgery for intractable epilepsy. Cells were cultured in DMEM, to the manufacturer’s instructions (Qiagen). For gene expression analysis, 5% FBS, penicillin/streptomycin, and glutamine. Cell differentiation and random hexaprimers and Moloney murine leukemia virus reverse tran- calcitriol treatment was performed over 6 d as described above. Human fetal scriptase were used to perform standard reverse transcription. Analysis of astrocytes were isolated, as previously described (14), from human CNS individual gene expression was conducted using TaqMan probes to assess tissue from fetuses at 17–23 wk of gestation that were obtained from the expression relative to Gapdh. University of Washington Birth Defects Research Laboratory (project no. 5R24HD000836-51) following Canadian Institutes of Health Research– Study approval approved guidelines. All studies that were performed have been conducted according to Dec- Autologous T cells laration of Helsinki principles and with approval of the Research Ethics Office at McGill University. Human T cells were isolated from the same PBMC fraction as described for macrophages, using magnetic CD3+ isolation beads (Miltenyi Biotec). Statistics Proinflammatory cytokine assay Paired Student t test and one-way ANOVA were used to determine significance of results. Following differentiation, macrophage cultures were supplemented with 10 ng/ml TNF-a or IL-1b for 24 h. Cells were then treated with 1027 M Results 25OHD (Selleck Chemistry) for 48 h. Calcitriol mediates significant transcriptional changes in Phagocytosis assay human MDMs Human myelin was isolated as previously described (15). Myelin was We have previously identified MerTK as an important phagocytic found to be endotoxin-free using the Limulus amebocyte lysate test receptor for the immunologically silent clearance of myelin debris (Sigma-Aldrich). To evaluate myelin uptake, myelin was incubated (6). To identify compounds that are known to alter MERTK gene with a pH-sensitive dye (pHRodamine; Invitrogen) for 1 h in PBS (pH 8). expression, we used a data integration approach known as Inte- Dyed myelin was added to myeloid cells to a final concentration of 20 mg/ml and incubated for 1 h. Flow cytometry was performed using the FACS grated Complex Traits Network (iCTNet) (21). iCTNet retrieves Fortessa (BD Biosciences). Live cells were gated based on live-dead information from multiple databases and creates a single network staining, and doublets were excluded. with user-defined parameters for visualization. Calcitriol was revealed Flow cytometry as a regulatory factor upon visualization of a subset of Food and Drug Administration–approved compounds (gray) and diseases (pink) Human myeloid cells were detached gently using 2 mmol EDTA/PBS and related to MERTK (Fig. 1A). blocked-in FACS buffer supplemented with 10% normal human serum and normal mouse IgG (3 mg/ml). Cells were incubated at 4˚C for 15 min with To examine the effect of calcitriol on MDMs in different states Aqua viability dye (Life Technologies) and then subsequently incubated of polarization (Supplemental Fig. 1A, 1B), we analyzed the at 4˚C for 30 min with either control isotype Ab or appropriate surface transcriptomic profile of homeostatic (MØ0) and MØGMcsf MDMs The Journal of Immunology 3 Downloaded from http://www.jimmunol.org/

FIGURE 1. Calcitriol mediates signiﬁcant transcriptional changes in human MDMs. (A) iCTNet neighborhood visualization of MerTK including U.S. Food and Drug Administration–approved compounds and diseases associated with genetic variants or mutations in MerTK. Calcitriol is identiﬁed as an by guest on October 4, 2021

MerTK-interacting molecule. (B) PCA plot of MØ0 (n = 3), MØGMcsf (n = 3), and calcitriol-treated (n = 6) MDM samples shows separation along PC1 according to cellular phenotype and along PC2 in response to calcitriol treatment based on transcriptional proﬁle. (C) Unsupervised hierarchical clustering and heat map of control and treated MDMs shows that samples cluster according to calcitriol treatment and then according to their phenotype. Upregulated genes are shown in red, and downregulated genes are shown in green. Dendrogram provides a measure of the relatedness of gene expression in each sample

(top) and for each gene (left). (D) Volcano plots display comparison of gene expression between untreated and calcitriol-treated MØ0 and MØGMcsf cells. Genes with adjusted p value/FDR ,0.05 only are shown in red. Genes with log2 fold change .1 in orange, and if both requirements are met, genes appear in green. Genes of interest are marked, including genes CYP24A1 and CAMP, highlighting cellular response to calcitriol. (E) ORA networks display the most enriched biological processes. Differentially expressed genes (FDR , 0.05; log2 fold change .1) in response to calcitriol were used to generate networks. Set nodes represent biological processes, which are colored based on their FDR (most signiﬁcant nodes appear in red, followed by gold, orange, dark yellow, and light yellow, each with decreasing signiﬁcance). Size of the set nodes corresponds to the number of genes associated with that biological process. Smaller nodes represent individual genes, which are colored based on their fold change (upregulation is denoted by light and dark red; downregulation is denoted by light and dark green). FDR, false discovery rate.

generatedinvitroandsubjectedtobulkRNAseq.MØ0 show cells exposed to calcitriol (Fig. 1E). Set nodes represent biological high expression of CNS homeostatic myeloid markers such as processes colored based on p value (red to light yellow, most sig- TREM2, CSF1R, IL-10,andMAFB (Supplemental Fig. 1C). nificant to least significant). The size of the node corresponds to the Proinflammatory MØGMcsf cells expression signatures show number of genes associated with the biological process that corre- typical inflammatory markers such as IL-6, NLRP1, CCL22, MMP9, lates with the function of these genes. Smaller unlabeled nodes and ITGAX, as well as induction of inflammatory programs in- represent individual genes (red, upregulated; green, downregulated). volving the transcription factor BHLHE40, identified as part of the Downregulated genes of interest (MERTK and HLA-DRB1)with disease-associated transcriptomic signature (22). PCA (Fig. 1B) and their link to relevant biological processes (regulation of endocytosis heatmap (Fig. 1C) analyses showed that MØ0 and MØGMcsf cells and adaptive immune response) are highlighted. cluster separately based on their phenotypes with calcitriol-treated cells clustering together regardless of their starting phenotype (Fig. Calcitriol regulation of MerTK expression and function in 1B, 1C). Volcano plot analysis confirms this calcitriol-mediated human MDMs shift in the transcriptomic signature and highlights that both phe- Use of the ingenuity pathway analysis (IPA) bioinformatic tool notypes responded to calcitriol by upregulating known calcitriol highlighted “phagosome formation” as one of the top canonical target genes CYP24A1 and cathelicidin (CAMP) (Fig. 1D). Finally, pathways affected by calcitriol in MDMs (Supplemental Fig. 2). overrepresentation analysis (ORA) was carried out using signifi- Visualization of this pathway highlighted the downregulation of a cantly differentially expressed genes in both MØ0 and MØGMcsf number of phagocytic and immune-sensing receptors, including 4 VITAMIN D REGULATES MYELOID CELLS VIA THE MerTK PATHWAY complement receptors, Fc receptors, and integrins, suggesting that autologous apoptotic T cells, and opsonized RBCs (oRBCs). calcitriol may influence the cells’ ability to phagocytose a range of Calcitriol-treated MDMs displayed a reduced capacity to substrates (Fig. 2A). We identified a list of 30 genes associated phagocytose pHRhodamine-labeled human myelin regardless with phagocytosis by myeloid cells and assessed their ex- of cellular phenotype (Fig. 2E). pression in response to calcitriol in both MØ0 and MØGMcsf In addition to myelin, MerTK has been extensively characterized as cells. A total of seven genes were significantly downregulated in a mediator of apoptotic cell clearance (9). To investigate whether MØ0 and four genes in MØGMcsf in response to calcitriol treat- calcitriol also inhibited this process, pHRhodamine-labeled apoptotic ment. MERTK was the only gene significantly downregulated in T cells were incubated with autologous MDMs. We observed a both MØ0 and MØGMcsf cells (Fig. 2B). We validated this RNAseq significant inhibition of apoptotic T cell phagocytosis by calcitriol- finding by quantitative RT-PCR. Regardless of phenotype, calcitriol exposed MØ0 but not MØGMcsf cells. This is indicative of an significantly downregulated MERTK mRNA (Fig. 2C) and protein MØGMcsf–specific efferocytotic receptor that can compensate expression, as measured by flow cytometry (Fig. 2D). for the calcitriol-mediated downregulation of MerTK (Fig. 2F). To assess if reduced expression of MerTK would have a Finally, to validate the specificity of calcitriol in regulating MerTK- functional impact on the cells, we measured the ability of dependent phagocytosis, we assessed the uptake of oRBCs by both calcitriol-treated MDMs to phagocytose myelin debris, MDM phenotypes. Phagocytosis of oRBCs occurs through Fc Downloaded from http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 2. Calcitriol regulates MerTK expression and phagocytosis in human MDMs. (A) IPA of differentially expressed genes identifies “phagosome formation” as a significantly affected pathway. Visualization of this pathway highlights affected molecules (nodes) and relationships between nodes, which are denoted by lines (edges). Edges are supported by at least one reference in the Ingenuity Knowledge Base. The intensity of color in a node indicates the degree of downregulation (green). (B) Thirty phagocytosis-related genes are identified in RNAseq datasets. Direction of regulation is assessed in both MØ0 and MØGMcsf cells. MERTK is downregulated in both cellular phenotypes. (C) Exposure of MDMs to calcitriol (100 nM) downregulates MerTK mRNA and (D) protein expression in both MØ0 and MØGMcsf cells. (E) Both MØ0 and MØGMcsf cells are impaired in their ability to phagocytose myelin debris following treatment with calcitriol (100 nM) as compared with vehicle. Representative flow plot of myelin phagocytosis. (F)MØ0 cells, but not MØGMcsf cells, are impaired in their ability to phagocytose autologous apoptotic T cells. Representative flow plot of autologous apoptotic T cell phagocytosis. (G) There was no significant regulation on the ability of MDMs to phagocytose oRBCs, representative flow plot of oRBC phagocytosis. All data were analyzed using paired Student t test. **p , 0.01, ***p , 0.001, ****p , 0.0001. The Journal of Immunology 5 receptor–mediated endocytosis, an MerTK-independent pathway. (Fig. 4A, 4B). CYP27B1 expression is regulated by a complex In all cases, calcitriol had no influence on the ability of MDMs to network of cytokines (4); we therefore assessed the impact of phagocytose oRBCs, suggesting a specificity to the calcitriol- proinflammatory cytokines known to play a role in MS pa- mediated inhibition of phagocytosis by human MDMs (Fig. 2G). thology (TNF-a and IL-1b)onCYP27B1 expression, 25OHD metabolism, and MerTK expression (30). We observed that the Calcitriol downregulates the expression of Ag addition of TNF-a (and, to a lesser degree, IL-1b) enhanced presentation molecules the expression of CYP27B1 in MØGMcsf cells but not MØ0 Engagement of the adaptive immune system through the reac- (Fig. 4C). To assess the capacity of the vitamin D metabolic tivation of antimyelin T cell responses in the CNS acts as a key pathway to regulate MerTK expression, cells were treated with pathogenic step in the initiation and exacerbation of MS (23). the major circulating metabolite 25OHD. Despite the increased + + Activation of CD8 and CD4 T cells requires recognition of basal expression of CYP27B1 in MØGMcsf cells, exposure to cognate Ags loaded on the surface of APCs. The strongest MS risk 25OHD did not significantly alter MerTK expression (Fig. 4D). loci maps to the HLA region, which is a gene complex encoding However, combinatorial treatment of MDMs with 25OHD and the major histocompatibility family of proteins (MHC). GWAS TNF-a (and, to a lesser degree, IL-1b) selectively and sig- has identified the HLA-DRB1 as the strongest risk locus, confer- nificantly downregulated MerTK expression in MØGMcsf cells ring a 3-fold–increased MS risk (24). Activation of T cells re- to a similar degree as calcitriol (Fig. 4D). TNF-a alone did not quires expression of MHC class molecules by APCs (signal one) change MerTK expression. Altogether, we show that proinflammatory in addition to a “second” signal in the form of expression of co- MØGMcsf cells are the only cells capable of converting 25OHD to stimulatory molecules such as CD40 and CD86, both also iden- active calcitriol, leading to the downregulation of the myelin- tified as MS risk loci (15, 25, 26). IPA analysis of our sequencing phagocytic receptor MerTK. Downloaded from results highlights the “Ag presentation pathway” as a significantly Calcitriol regulation of MerTK expression in primary affected pathway (Supplemental Fig. 2), with downregulation of human glia both MHC class I and MHC class II molecules as indicated using the pathway visualization tool (Fig. 3A). We identified a list of 24 In addition to recruited MDMs, both resident microglia and as- genes associated with Ag presentation in our dataset and assessed trocyte populations take part in the neuroinflammatory process and http://www.jimmunol.org/ expression in response to calcitriol in MØ0 and MØGMcsf cells the phagocytic clearance of myelin debris. We therefore examined (Fig. 3B). Expression of a large number of HLA/MHC genes were the effect of calcitriol on human microglia isolated from resected downregulated by calcitriol treatment in both cellular phenotypes, brain tissue and astrocytes derived from the fetal human CNS. including the major MS risk gene HLA-DRB1. We validated these Microglia were polarized to CNS homeostatic (MG0) and proin- sequencing findings by measuring protein expression using flow flammatory (MGGMcsf) phenotypes. Similar to MDMs, cells were cytometry. Protein expression of both MHC class I (HLA-ABC) exposed to M-CSF (MG0)orGM-CSF(MGGMcsf) over a 6- and MHC class II (HLA-DR/DP/DQ) molecules were downreg- d period, with homeostatic cells receiving additional TGF-b. ulated by calcitriol treatment in both MØ0 and MØGMcsf cells Confirmation of these phenotypes is highlighted by expression of (Fig. 3C). Expression of costimulatory molecules CD86 and established CNS homeostatic markers, including microglia-specific by guest on October 4, 2021 CD40 were also significantly reduced in response to calcitriol markers TMEM119, SALL1,andOLFML3 (Supplemental Fig. 1C). (Fig.3D).Interestingly,weobserved increased expression of Proinflammatory microglia are characterized by high expression of immune checkpoint molecule CD274(PD-L1) both at the mRNA canonical inflammatory myeloid markers including genes that show (Fig. 3B) and protein (Fig. 3E) level following treatment with relative specificity to microglia, CCL17 and IL-1a (Supplemental calcitriol. CD274(PD-L1) suppresses the adaptive immune re- Fig. 1D). Bulk RNAseq was carried out on calcitriol-treated MG0 sponse by inducing apoptosis in CD279-expressing T cells (27). and MGGMcsf cells. PCA of these samples showed that, similar to Moreover, previous work has shown that the human CD274(PD-L1) MDMs, microglia cluster along the first principal component based gene is a direct target of the 1,25(OH)2D–regulated vitamin D on their cellular phenotype (MG0 and MGGMcsf) and along the receptor (28). Finally, a “third” signal in the form of proin- second principal component based on treatment with calcitriol (Fig. flammatory cytokine release from the APC is suggested to be 5A). ORA carried out on differentially expressed genes in both necessary for the induction of T cell proliferation. IL-6 is a phenotypes exposed to calcitriol show a similar pattern of calcitriol- cytokine that, when released from APCs, can promote the dif- responsive biological processes, including “inflammatory response” ferentiation of IL-17–producing Th-17 cells, known to be highly and “cytokine production/secretion” (Fig. 5B). These transcriptomic pathogenic in MS (29). We observed a significant decrease in IL-6 results were validated in vitro whereby calcitriol downregulated mRNA and protein release by ELISA in response to calcitriol MERTK mRNA and MerTK protein in human microglia (Fig. 5C). (Fig. 3F) in MØGMcsf cells. Finally, calcitriol had no influence on MerTK mRNA or protein expression in human fetal astrocytes (Fig. 5D), indicating that the Endogenous production of calcitriol inhibits MerTK selectively regulation of MerTK expression by calcitriol is specific to cells of in proinflammatory MDMs the myeloid lineage. The in vivo–circulating concentrations of calcitriol (40–100 pM) are much lower than those of 25OHD (20–150 nM). It is therefore Discussion important to determine whether there is sufficient intracellular The link between vitamin D and MS risk and the overrepresentation metabolism of 25OHD to calcitriol within MDMs to affect MerTK of genes involved in vitamin D metabolism as part of the genetic expression. As shown in Fig. 4A, proinflammatory MØGMcsf cells architecture of MS highlights the need for understanding the exhibited the highest expression of the calcitriol-producing enzyme, functional pathways under the control of vitamin D. In this study, CYP27B1 (Fig. 4E). This high level of CYP27B1 expression neg- we explored the response of human myeloid populations to cal- atively correlated with MERTK expression. Cells that expressed the citriol (1,25(OH)2D) on the level of the whole transcriptome. lowest levels of MERTK (MØGMcsf) expressed the highest levels of Using network-based analysis, we observed significant modulation CYP27B1 and, conversely, cells (MØ0) that expressed the highest of both Ag presentation and phagocytosis pathways in MDMs and levels of MERTK displayed the lowest expression of CYP27B1 primary human microglia. We report that calcitriol controls 6 VITAMIN D REGULATES MYELOID CELLS VIA THE MerTK PATHWAY Downloaded from http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 3. Calcitriol regulates the Ag presentation pathway in human MDMs. (A) IPA of differentially expressed genes identifies “Ag presentation” as a significantly affected pathway. Visualization of this pathway highlights affected molecules (nodes) and relationships between nodes, which are denoted by lines (edges). Edges are supported by at least one reference in the Ingenuity Knowledge Base. The intensity of color in a node indicates the degree of downregulation (green). (B) Twenty-four Ag presentation genes are identified in RNAseq datasets. Direction of regulation is assessed in both MØ0 and MØGMcsf cells with HLA genes significantly downregulated and immune checkpoint molecules upregulated in both cellular phenotypes. (C) Exposure of MDMs to calcitriol (100 nM) downregulates protein expression of HLA-ABC and HLA-DR/DP/DQ as measured by flow cytometry (D) Calcitriol treatment downregulates protein expression of costimulatory molecules CD86 and CD40 in both MØ0 and MØGMcsf cells. (E) Both MØ0 and MØGMcsf cells up- regulate CD274 (PD-L1) protein expression following treatment with calcitriol (100 nM). (F) IL-6 mRNA and protein release (ELISA) are downregulated by calcitriol treatment in MØGMcsf cells. All data were analyzed using paired Student t test. *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001. expression of the phagocytic receptor MerTK and subsequent levels. Cells exposed to inflammatory cytokines expressed the uptake of myelin debris and apoptotic cells. Calcitriol also es- highest levels of CYP27B1 and had an enhanced ability to respond tablishes an immune-regulatory phenotype in these myeloid to 25OHD. Based on our findings, we would predict that circu- cells, significantly reducing expression of inflammatory medi- lating 25OHD may not be as important as the CYP27B1-mediated ators and Ag presentation machinery while increasing the ex- production of intracellular calcitriol and subsequent transcrip- pression of immune checkpoint molecules. tional regulation of cellular function, particularly in cells of the Mendelian randomization studies have shown that genetically innate immune system. Therefore, supplementation that in- determined variations in 25OHD serum levels play a causal role in creases serum 25OHD levels may not be targeting the cellular MS (31, 32). Clinical studies are ongoing (vitamin D to ameliorate functions relevant to the pathogenesis of MS. Based on our MS and efficacy of vitamin D supplementation in MS), yet a results, we would propose that an individual’s ability to respond reproducible benefit of vitamin D supplementation has not been to vitamin D supplementation may fluctuate with time based on evident thus far. Standard of care preparations of vitamin D are their inflammatory status and their cells’ abilities to produce comprised of precursors such as 25OHD, the major circulating active calcitriol from circulating 25OHD. form of vitamin D. 25OHD is processed locally to biologically- Myelin clearance through myeloid cell–mediated phagocytosis active calcitriol yet, it is 25OHD levels that define an individ- is an essential process that allows for efficient remyelination and ual’s vitamin D “status” (3). Both systemic and intracellular CNS repair (33). We and others have reported reduced MerTK conversion of 25OHD to calcitriol is dependent on sufficient expression and phagocytic capacity in myeloid cells of MS pa- expression of the enzyme CYP27B1 (2). Our study demonstrates a tients (8). Expression of both membrane-bound and -soluble forms significant effect of the activation state of the cell on CYP27B1 of MerTK are elevated in MS lesional tissues (34). In animal The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 4. 25OHD selectively downregulates MerTK in proinflammatory MDMs. (A and B)MØGMcsf cells express high levels of CYP27B1 (i.e., low cycle threshold [Ct] values by quantitative PCR) and express low levels of MERTK. In contrast, MØ0 cells express the highest levels of MERTK and low levels of CYP27B1.**p , 0.01, ****p , 0.0001. Paired Student t test. (C) Exposure of MDMs to TNF-a and IL-1b selectively upregulates CYP27B1 expression in MØGMcsf cells but not in MØ0 cells. **p , 0.01. One-way ANOVA. (D) Combinatorial treatment of TNF-a + 25OHD selectively reduces MERTK expression to a similar level as calcitriol in MØGMcsf cells only. ***p , 0.001. One-way ANOVA. (E) Schematic representation of data shows high expression of MerTK and myelin-phagocytic function in homeostatic MØ0 cells. These cells are unable to convert 25OHD to calcitriol because of the low expression of CYP27B1 and, therefore, maintain MerTK expression and function. However, proinflammatory MØGMcsf cells express high levels of CYP27B1 and are thus able to produce calcitriol from its precursor, downregulate MerTK (molecules associated with Ag presentation), and inhibit phagocytosis. models, MerTK and its cognate ligand Gas6 have shown to play This particular SNP (rs7422195) displays discordant association protective roles, particularly in the cuprizone toxin model in depending on the individual’s HLA-DRB1*15:01 status, conferring which Gas6-knockout mice develop a more severe level of de- increased risk, but converting to a protective effect on an HLA- myelination coupled with a delayed remyelination process (7). DRB1*15:01 homozygous background. The stratification of risk Experimental evidence strongly supports a functional role for based on DRB1 status is strongly suggestive of a functional inter- MerTK in inflammation resolution, debris clearance, and repair play or crosstalk between phagocytosis and Ag presentation in (35). GWAS has identified several SNPs in the MERTK gene as cells capable of carrying out such functions. The beneficial role independently associated with the risk of developing MS (12, 25). of high MerTK expression is dependent on the underlying pa- Fine-mapping of the MERTK locus identifies a risk variant that thology, the phase of the disease, and the activation status of the operates in trans with the HLA-DRB1 locus and is associated cell in which it is expressed. A recent study has shown poly- with higher expression of MerTK in MS patient monocytes (13). morphisms in the MERTK gene that drive low expression of the 8 VITAMIN D REGULATES MYELOID CELLS VIA THE MerTK PATHWAY Downloaded from http://www.jimmunol.org/

FIGURE 5. Calcitriol selectively downregulates MerTK in microglia in the brain. (A) Transcriptomic changes in primary human microglia (MG0 and by guest on October 4, 2021 MGGMcsf) treated with calcitriol are visualized on a PCA plot. Microglia separate along PC1 according to cellular phenotype and along PC2 in response to calcitriol treatment. (B) ORA networks display the most enriched biological processes. Differentially expressed genes (FDR , 0.05; log2 fold change .1) in response to calcitriol treatment were used to generate networks. Set nodes represent biological processes, which are colored based on their FDR (most significant nodes appear in red, followed by gold, orange, dark yellow, and light yellow, each with decreasing significance). Size of the set nodes corresponds to the number of genes associated with that biological process. Smaller nodes represent individual genes, which are colored based on their fold change (upregulation is denoted by light and dark red; downregulation is denoted by light and dark green). (C) Exposure of primary human microglia to calcitriol downregulates MerTK mRNA and protein expression. **p , 0.01. Paired Student t test (D) Calcitriol does not modulate MerTK mRNA or protein expression in human fetal astrocytes. Representative flow plot of MerTK expression. Paired Student t test. ns, not significant. protein in Kupffer cells to protect against the development of Calcitriol also inhibited IL-6 expression and release. These liver fibrosis in nonalcoholic steatohepatitis (36). In addition to combined data highlight the ability of calcitriol to modulate the genomic determinants of MerTK expression and function, both the ingestion of material and the expression of molecular our study highlights how environmental factors can also influ- machinery involved in Ag presentation, potentially lowering ence expression of this key phagocytic and immunomodulatory the risk of autoantigen presentation to the adaptive immune receptor. system. In addition to myelin debris, impaired clearance of cells un- In summary our data demonstrates that calcitriol down- dergoing apoptosis leads to sustained proinflammatory responses as regulates MerTK expression and MerTK-mediated phagocytosis cells progress to secondary necrosis (37). Digestion of phagocy- in primary human myeloid cells. Intracellular production of tosed substrate and presentation as Ags loaded on MHC molecules active calcitriol from its precursor and resultant repression of (signal 1) coupled with costimulation (signal 2) and secretion MerTK is limited to proinflammatory myeloid cells (due to high of inflammatory cytokines (signal 3) from APCs play a critical expression of CYP27B1). This proinflammatory-specific effect role in stimulating the adaptive immune response (38). GWAS may underlie a beneficial mechanism of vitamin D in MS. has identified an extended HLA haplotype, HLA DRB1*15:01, Proinflammatory myeloid cells are potent Ag presenters; selective DQA1*0102, DQB1*0602, within the MHC class II region that inhibition of myelin uptake by these cells may lower the risk of is strongly associated with MS risk. In accordance with previous myelin Ag presentation to infiltrating T cells. In contrast, main- reports, we observed that calcitriol downregulated the expression tenance of MerTK, and therefore phagocytic function in homeo- of both MHC class I and II molecules on the surface of myeloid static myeloid populations (due to low expression of CYP27B1), cells including HLA DRB1/DQA1/DQB1. Calcitriol downregulated the would allow these cells to maintain clearance of myelin debris and expression of major costimulatory molecules and upregulated immune contribute to the process of repair. Overall, we uncover a functional checkpoint molecule CD274 (PD-L1), as previously reported (28). interaction between one of the strongest environmental modulators The Journal of Immunology 9 of MS risk (vitamin D) and the MerTK pathway that is selective to expression analysis of RNA-seq experiments with TopHat and Cufflinks. [Pub- lished erratum appears in 2014 Nat. Protoc. 9: 2513.] Nat. Protoc. 7: 562–578. disease-relevant populations of primary human myeloid cells. 19. Robinson, M. D., D. J. McCarthy, and G. K. Smyth. 2010. edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26: 139–140. Disclosures 20. McCarthy, D. J., Y. Chen, and G. K. Smyth. 2012. Differential expression The authors have no financial conflicts of interest. analysis of multifactor RNA-seq experiments with respect to biological varia- tion. Nucleic Acids Res. 40: 4288–4297. 21. Wang, L., D. S. Himmelstein, A. Santaniello, M. Parvin, and S. E. Baranzini. 2015. iCTNet2: integrating heterogeneous biological interactions to understand References complex traits. F1000Res. 4: 485. 1. Zhang, H. L., and J. Wu. 2010. Role of vitamin D in immune responses and 22. Krasemann, S., C. Madore, R. Cialic, C. Baufeld, N. Calcagno, R. El Fatimy, autoimmune diseases, with emphasis on its role in multiple sclerosis. Neurosci. L. Beckers, E. O’Loughlin, Y. Xu, Z. Fanek, et al. 2017. The TREM2-APOE Bull. 26: 445–454. pathway drives the transcriptional phenotype of dysfunctional microglia in 2. Prosser, D. E., and G. Jones. 2004. Enzymes involved in the activation and in- neurodegenerative diseases. Immunity 47: 566–581.e9. activation of vitamin D. Trends Biochem. Sci. 29: 664–673. 23. Lopes Pinheiro, M. A., A. Kamermans, J. J. Garcia-Vallejo, B. van Het Hof, 3. Wierzbicka, J., A. Piotrowska, and M. A. Zmijewski._ 2014. The renaissance of L. Wierts, T. O’Toole, D. Boeve, M. Verstege, S. M. van der Pol, Y. van Kooyk, vitamin D. Acta Biochim. Pol. 61: 679–686. et al. 2016. Internalization and presentation of myelin antigens by the brain 4. Noyola-Martıńez, N., L. Dıáz, V. Zaga-Clavellina, E. Avila, A. Halhali, endothelium guides antigen-specific T cell migration. eLife 5: e13149. F. Larrea, and D. Barrera. 2014. Regulation of CYP27B1 and CYP24A1 gene 24. Baranzini, S. E., and J. R. Oksenberg. 2017. The genetics of multiple sclerosis: expression by recombinant pro-inflammatory cytokines in cultured human tro- from 0 to 200 in 50 years. Trends Genet. 33: 960–970. phoblasts. J. Steroid Biochem. Mol. Biol. 144: 106–109. 25. International Multiple Sclerosis Genetics ConsortiumWellcome Trust Case 5. Croxford, A. L., S. Spath, and B. Becher. 2015. GM-CSF in neuroinflammation: Control Consortium 2. 2011. Genetic risk and a primary role for cell-mediated licensing myeloid cells for tissue damage. Trends Immunol. 36: 651–662. immune mechanisms in multiple sclerosis. Nature 476: 214–219. 6. Healy, L. M., G. Perron, S.-Y. Won, M. A. Michell-Robinson, A. Rezk, 26. Sawcer, S., R. J. M. Franklin, and M. Ban. 2014. Multiple sclerosis genetics. S. K. Ludwin, C. S. Moore, J. A. Hall, A. Bar-Or, and J. P. Antel. 2016. MerTK Lancet Neurol. 13: 700–709. is a functional regulator of myelin phagocytosis by human myeloid cells. J. 27. Freeman, G. J., A. J. Long, Y. Iwai, K. Bourque, T. Chernova, H. Nishimura, Downloaded from Immunol. 196: 3375–3384. L. J. Fitz, N. Malenkovich, T. Okazaki, M. C. Byrne, et al. 2000. Engage- 7. Binder, M. D., H. S. Cate, A. L. Prieto, D. Kemper, H. Butzkueven, ment of the PD-1 immunoinhibitory receptor by a novel B7 family member M. M. Gresle, T. Cipriani, V. G. Jokubaitis, P. Carmeliet, and T. J. Kilpatrick. leads to negative regulation of lymphocyte activation. J. Exp. Med. 192: 2008. Gas6 deficiency increases oligodendrocyte loss and microglial activation 1027–1034. in response to cuprizone-induced demyelination. J. Neurosci. 28: 5195–5206. 28. Dimitrov, V., M. Bouttier, G. Boukhaled, R. Salehi-Tabar, R. G. Avramescu, 8. Healy, L. M., J. H. Jang, S.-Y. Won, Y. H. Lin, H. Touil, S. Aljarallah, A. Bar-Or, B.Memari,B.Hasaj,G.L.Lukacs,C.M.Krawczyk,andJ.H.White.2017. and J. P. Antel. 2017. MerTK-mediated regulation of myelin phagocytosis by Hormonal vitamin D up-regulates tissue-specific PD-L1 and PD-L2 surface

macrophages generated from patients with MS. Neurol. Neuroimmunol. Neuro- glycoprotein expression in humans but not mice. J. Biol. Chem. 292: 20657– http://www.jimmunol.org/ inflamm. 4: e402. 20668. 9. Scott, R. S., E. J. McMahon, S. M. Pop, E. A. Reap, R. Caricchio, P. L. Cohen, 29. Bettelli, E., Y. Carrier, W. Gao, T. Korn, T. B. Strom, M. Oukka, H. L. Weiner, H. S. Earp, and G. K. Matsushima. 2001. Phagocytosis and clearance of apo- and V. K. Kuchroo. 2006. Reciprocal developmental pathways for the generation ptotic cells is mediated by MER. Nature 411: 207–211. of pathogenic effector TH17 and regulatory T cells. Nature 441: 235–238. 10. Butovsky, O., M. P. Jedrychowski, C. S. Moore, R. Cialic, A. J. Lanser, 30. Kemanetzoglou, E., and E. Andreadou. 2017. CNS demyelination with TNF-a G. Gabriely, T. Koeglsperger, B. Dake, P. M. Wu, C. E. Doykan, et al. 2014. blockers. Curr. Neurol. Neurosci. Rep. 17: 36. Identification of a unique TGF-b-dependent molecular and functional signature 31. Mokry, L. E., S. Ross, O. S. Ahmad, V. Forgetta, G. D. Smith, D. Goltzman, in microglia. [Published erratum appears in 2014 Nat. Neurosci. 17: 1286.] Nat. A. Leong, C. M. Greenwood, G. Thanassoulis, and J. B. Richards. 2015. Vitamin Neurosci. 17: 131–143. D and risk of multiple sclerosis: a mendelian randomization study. [Published 11. Rivera, C. R., J. M. Kollman, J. K. Polka, D. A. Agard, and R. D. Mullins. 2011. erratum appears in 2016 PLoS Med. 13: e1001981.] PLoS Med. 12: e1001866. Architecture and assembly of a divergent member of the ParM family of bac- 32. Rhead, B., M. Baä¨rnhielm, M. Gianfrancesco, A. Mok, X. Shao, H. Quach,

terial actin-like proteins. J. Biol. Chem. 286: 14282–14290. L. Shen, C. Schaefer, J. Link, A. Gyllenberg, et al. 2016. Mendelian randomi- by guest on October 4, 2021 12. Australia and New Zealand Multiple Sclerosis Genetics Consortium (ANZgene). zation shows a causal effect of low vitamin D on multiple sclerosis risk. Neurol. 2011. Polymorphisms in the receptor tyrosine kinase MERTK gene are associ- Genet. 2: e97. ated with multiple sclerosis susceptibility. PLoS One 6: e16964. 33. Lampron, A., A. Larochelle, N. Laflamme, P. Pre´fontaine, M. M. Plante, 13. Binder, M. D., A. D. Fox, D. Merlo, L. J. Johnson, L. Giuffrida, S. E. Calvert, M. G. Sańchez, V. W. Yong, P. K. Stys, M. E. Tremblay, and S. Rivest. 2015. R. Akkermann, G. Z. M. Ma, A. A. Perera, M. M. Gresle, et al; ANZgene. 2016. Inefficient clearance of myelin debris by microglia impairs remyelinating Common and low frequency variants in MERTK are independently associated processes. J. Exp. Med. 212: 481–495. with multiple sclerosis susceptibility with discordant association dependent upon 34. Weinger, J. G., K. M. Omari, K. Marsden, C. S. Raine, and B. Shafit-Zagardo. HLA-DRB1*15:01 status. PLoS Genet. 12: e1005853. 2009. Up-regulation of soluble Axl and Mer receptor tyrosine kinases negatively 14. Jack, C. S., N. Arbour, J. Manusow, V. Montgrain, M. Blain, E. McCrea, correlates with Gas6 in established multiple sclerosis lesions. Am. J. Pathol. 175: A. Shapiro, and J. P. Antel. 2005. TLR signaling tailors innate immune responses 283–293. in human microglia and astrocytes. J. Immunol. 175: 4320–4330. 35. Shafit-Zagardo, B., R. C. Gruber, and J. C. DuBois. 2018. The role of TAM 15. Ji, Q., L. Castelli, and J. M. Goverman. 2013. MHC class I-restricted myelin family receptors and ligands in the nervous system: from development to epitopes are cross-presented by tip-DCs that promote determinant spreading to pathobiology. Pharmacol. Ther. 188: 97–117. CD8+ T cells. Nat. Immunol. 14: 254–261. 36.Cai,B.,P.Dongiovanni,K.E.Corey,X.Wang,I.O.Shmarakov,Z.Zheng, 16. Picelli, S., A. K. Bjo¨rklund, O. R. Faridani, S. Sagasser, G. Winberg, and C. Kasikara, V. Davra, M. Meroni, R. T. Chung, et al. 2020. Macrophage R. Sandberg. 2013. Smart-seq2 for sensitive full-length transcriptome profiling MerTK promotes liver fibrosis in nonalcoholic steatohepatitis. Cell Metab. 31: in single cells. Nat. Methods 10: 1096–1098. 406–421.e7. 17. Dobin, A., C. A. Davis, F. Schlesinger, J. Drenkow, C. Zaleski, S. Jha, P. Batut, 37. Szondy, Z., Z. Sarang, B. Kiss, E´ . Garabuczi, and K. Ko¨ro¨skeńyi. 2017. Anti- M. Chaisson, and T. R. Gingeras. 2013. STAR: ultrafast universal RNA-seq inflammatory mechanisms triggered by apoptotic cells during their clearance. aligner. Bioinformatics 29: 15–21. Front. Immunol. 8: 909. 18. Trapnell, C., A. Roberts, L. Goff, G. Pertea, D. Kim, D. R. Kelley, H. Pimentel, 38. Blander, J. M. 2008. Phagocytosis and antigen presentation: a partnership ini- S. L. Salzberg, J. L. Rinn, and L. Pachter. 2012. Differential gene and transcript tiated by toll-like receptors. Ann. Rheum. Dis. 67(Suppl. 3): iii44–iii49. A. B. MDM MG

MØ0 MØGMcsf

150µM 150µM MØGMcsf MØ0 MGGMcsf MG0

C. TREM2 CSF1R IGF1 IL10 Mafb TMEM119 SALL1 OLFML3

50 1200 300 120 25 ✱ 60 6 500 ✱✱✱ ✱✱ ✱✱✱✱ ✱ ✱ 1000 100 20 40 50 400 800 200 80 40 4 15 30 300

TPM 600 60 30 10 20 200 400 100 40 20 2 5 10 100 200 20 10 n.e n.e n.e n.e 0 0 0 0 0 0 0 0

MG MG MG MG MG MG MG MG MG MG MG MG MG MG MG MG MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM D. IL6 NLRP1 ITGAX BHLHE40 MMP9 CCL22 CCL17 IL1α

✱ 40 8 ✱✱✱ 400 ✱✱ 100 600 2500 80 15 ✱ ✱ ✱✱ ✱✱ ✱✱✱ 80 500 2000 30 6 300 60 400 10 ✱ 60 1500 20 4 200 300 40 200 TPM 40 200 150 5 10 2 100 20 100 20 100 50 0 0 0 0 0 0 0

MG MG MG MG MG MG MG MG MG MG MG MG MG MG MG MG MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM MDM M M M M M M M M M M M M M M M M 0 GMcsf 0 GMcsf 0 GMcsf 0 GMcsf 0 GMcsf 0 GMcsf 0 GMcsf 0 GMcsf

SUPPLEMENTARY FIGURE 1. Confirmation of myeloid cell phenotypes. MDMs and microglia were polarized in homeostatic (MØ0 and MG0) and proinflammatory (MØGMcsf and MGGMcsf) phenotypes. (A) shows typical morphological difference between the phenotypes. MØ0 display a bipolar morphology, MØGMcsf have a more ameboid and activated morphology. (B) Heat map of RNAseq results of polarized MDMs and microglia shows significant transcriptional differences between the two phenotypes, in both cell types. (C) Homeostatic myeloid cells show higher expression of known brain homeostatic markers TREM2, CSF1R, IL10 and Mafb. Homeostatic microglia show increased or exclusive expression of specific homeostatic microglia markers such as TMEM119, SALL1 and OLFML3. (D) Proinflammatory myeloid cells show higher expression of markers of inflammation IL6, NLRP1, ITGAX, BHLHE40 and MMP9. Some inflammatory markers are enriched in microglia populations (CCL17 and IL1a) with others enriched in MDMs (CCL22)

MG 0 MGGMcsf

SUPPLEMENTARY FIGURE 2. Pathway analysis of calcitriol-treated myeloid cells (A) Pathway analysis using IPA was carried out on common significantly- and differentially-expressed genes in response to calcitriol treatment. (B) Fluorescent images of human microglia phagocytosing labelled myelin debris.