ARTICLE OPEN ACCESS Transcriptome and Function of Novel Immunosuppressive Autoreactive Invariant Natural Killer T Cells That Are Absent in Progressive Multiple Sclerosis Belinda Carrión, MD, PhD, Yawei Liu, MD, PhD, Mahdieh Hadi, PhD, Jon Lundstrom, MSc, Correspondence Dr. Issazadeh-Navikas Jeppe Romme Christensen, MD, PhD, Cecilie Ammitzbøll, MD, PhD, Morten Hanefeld Dziegiel, MD, PhD, shohreh.issazadeh@bric.ku.dk Per Soelberg Sørensen, MD, Manuel Comabella, MD, PhD, Xavier Montalban, MD, PhD, Finn Sellebjerg, MD, PhD, and Shohreh Issazadeh-Navikas, PhD Neurol Neuroimmunol Neuroinflamm 2021;8:e1065. doi:10.1212/NXI.0000000000001065 Abstract Background and Objective The aim of this study was to determine whether natural killer T (NKT) cells, including invariant (i) NKT cells, have clinical value in preventing the progression of multiple sclerosis (MS) by examining the mechanisms by which a distinct self-peptide induces a novel, protective invariant natural killer T cell (iNKT cell) subset. Methods We performed a transcriptomic and functional analysis of iNKT cells that were reactive to a human collagen type II self-peptide, hCII707-721, measuring differentially induced genes, cytokines, and suppressive capacity. Results We report the first transcriptomic profile of human conventional vs novel hCII707- 721–reactive iNKT cells. We determined that hCII707-721 induces protective iNKT cells that are found in the blood of healthy individuals but not progressive patients with MS (PMS). By transcriptomic analysis, we observed that hCII707-721 promotes their development and proliferation, favoring the splicing of full-length AKT serine/threonine kinase 1 (AKT1) and effector function of this unique lineage by upregulating tumor necrosis factor (TNF)-related genes. Furthermore, hCII707-721–reactive iNKT cells did not upregulate interferon (IFN)-γ, interleukin (IL)-4, IL-10, IL-13, or IL-17 by RNA-seq or at the protein level, unlike the response to the glycolipid alpha-galactosylceramide. hCII707-721–reactive iNKT cells increased TNFα only at the protein level and suppressed autologous-activated T cells through FAS-FAS ligand (FAS-FASL) and TNFα-TNF receptor I signaling but not TNF receptor II. Discussion Based on their immunomodulatory properties, NKT cells have a potential value in the treat- ment of autoimmune diseases, such as MS. These significant findings suggest that endogenous peptide ligands can be used to expand iNKT cells, without causing a cytokine storm, consti- tuting a potential immunotherapy for autoimmune conditions, including PMS. From the Biotech Research and Innovation Centre (BRIC) (B.C., Y.L., M.H., J.L., S.I.-N.), University of Copenhagen; Danish Multiple Sclerosis Center (J.R.C., C.A., P.S.S.), University of Copenhagen and Department of Neurology, Rigshospitalet; Blood Bank (M.H.D.), Copenhagen University Hospital, Denmark; Centre d’Esclerosi Múltiple de Catalunya (M.C.), Cemcat, Unitat de Neuroimmunologia Clı́nica, Hospital Universitari Vall d´Hebron (HUVH) – Universitat Autònoma de Barcelona, Barcelona, Spain; and Centre d’Esclerosi Múltiple de Catalunya (X.M.), Cemcat, Unitat de Neuroimmunologia Clı́nica, Hospital Universitari Vall d´Hebron (HUVH) – Universitat Autònoma de Barcelona, Spain; Danish Multiple Sclerosis Center, University of Copenhagen and Department of Neurology, Rigshospitalet, Denmark. Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article. The Article Procesing Charge was funded by Lundbeck Foundation; Danish Multiple Sclerosis Society, Independent Research Fund Denmark-Medical and Health Sciences (DFF-M), and Foundation for Research in Neurology to SI-N. Belinda Carrion received a PhD fellowship from CONACYT-Mexico and the Lundbeck. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1 Glossary AKT1 = AKT serine/threonine kinase 1; cDMEM = complete Dulbecco’s Modified Eagle Medium; CFSE = carboxyfluorescein succinimidyl ester; convT = conventional CD4+ T cells; DMEM = Dulbecco’s Modified Eagle Medium; DN = double-negative; EAE = experimental autoimmune encephalomyelitis; FACS = fluorescence-activated cell sorting; FAS-FASL = FAS-FAS ligand; FDR = false discovery rate; GSEA = gene set enrichment analysis; HC = healthy control; HD = healthy donor; IFN = interferon; IL = interleukin; iNKT cell = invariant natural killer T cell; MS = multiple sclerosis; NKT = natural killer T cell; PMS = progressive MS, including PPMS and SPMS; PBMC = peripheral blood mononuclear cell; PPMS = primary progressive MS; RRMS = relapsing-remitting MS; SPMS = secondary progressive MS; TNF = tumor necrosis factor; TNFRI = TNF receptor I; TNFRII = TNF receptor II; α-Galcer = alpha-galactosylceramide; 7AAD = 7-aminoactinomycin D. Natural killer T (NKT) cells recognize lipid and glycolipid and secondary progressive MS [SPMS]) patients, classified per antigens on the major histocompatibility complex Class Lublin and Reingold,48 was obtained from the Danish Multiple I-related CD1d.1-5 In addition to their host defense6 and anti- Sclerosis Center and Multiple Sclerosis Center of Catalonia; tumor activities,7 NKT cells are linked to the prevention of blood from healthy donors (HDs) was acquired from Rig- inflammation in allergic and autoimmune diseases.8-10 Multiple shospitalet, Denmark. sclerosis (MS) is a chronic neuroinflammatory disease that is believed to be mediated by autoreactive lymphocytes, impli- Standard Protocol Approvals, Registrations, cating a defect in NKT cells.11,12 Invariant iNKT (iNKT) cell and Patient Consents levels are lower in the peripheral blood of patients with MS,13,14 The study was approved by the Scientific Ethics Committee and untreated patients with MS have more Th2-skewed for Copenhagen and Frederiksberg (protocol KF01-041/95) iNKT cells during remission.15 iNKT cells can be induced by and the Hospital Universitari Vall d’Hebron Ethics Com- the glycolipid alpha-galactosylceramide (α-GalCer) to modu- mittee [PR(AG)32/2008] and was conducted after individual late experimental autoimmune encephalomyelitis (EAE), a consent was obtained. model of MS, through Th2 cytokines.16,17 However, α-GalCer treatment for EAE in mice causes liver toxicity,18 induces Patients and Healthy Individuals abortions,19 and exacerbates atherogenesis20—possibly from a In this cross-sectional study, in total, 146 subjects were in- cytokine storm. Consistent with this model, on induction with cluded: 44 age- and gender-matched healthy controls (HCs), α-GalCer, NKT cells secrete high levels of IFN-γ and IL-4, 26 patients with RRMS, 34 patients with PPMS, and 42 pa- which can be harmful in MS.11 Thus, NKT cell responses to tients with SPMS. eTable 6 (links.lww.com/NXI/A555) de- antigen must be discerned to develop NKT cell-based thera- tails all the individuals’ information included age, EDSS score, peutics. We have found that self-antigenic peptide-reactive and disease duration. NKT cells maintain peripheral tolerance, demonstrating that an immunodominant epitope from mouse collagen type II Baseline CD4− and CD8− double-negative (DN) NKT cells (mCII707-721) induces NKT cells, in turn inhibiting arthritis, were studied in patients with MS and HCs who had been neuroinflammation, and lung inflammation in models of recruited to the Danish MS Center, comprising HC: 24 women rheumatoid arthritis, MS, and allergic asthma, respectively.21 (mean age 37 years; range 24–62) and 8 men (39 years; 24–63), patients with RRMS: 14 women (38 years; 21–55) and 5 men We report a novel self-peptide-reactive NKT cell pop- (34 years; 28–41), PPMS: 8 women (51 years; 43–59) and 12 ulation in human peripheral blood. On being challenged men (47 years; 28–59), and SPMS: 11 women (46 years; with hCII707-721, they are activated and immunosup- 33–60) and 9 men (46 years; 30–53). pressive in healthy individuals but hyporesponsive in pa- tients with MS. Furthermore, we present the first Baseline iNKT cells were analyzed in patients with MS and transcriptional profile of human iNKTs, based on the ef- HDs from the Danish MS Center; the blood bank at Rig- fects of hCII707-721. shospitalet, Denmark; and Universitari Vall d’Hebron, Spain. The study population comprised HD: 9 women (44 years; 26–58) and 3 men (30 years; 23–37), patients with RRMS: 4 women (30 years; 24–37) and 2 men (38 years; 36–41 years), Methods PPMS: 6 women (55 years; 52–60) and 6 men (51 years; Peripheral Blood Samples 31–62), and SPMS: 7 women (54 years; 53–64) and 8 men Peripheral blood was collected by standard venipuncture into (50 years; 36–62). vacuum tubes with ethylenediaminetetraacetic acid. Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll- The responsiveness to hCII707-721 was tested in vitro in isopaque density gradient centrifugation (Gibco-BRL). Pe- freshly isolated human PBMCs from patients with PMS from ripheral blood from untreated relapsing-remitting MS (RRMS) the Danish MS Center and HDs. The 39 PBMC samples and progressive MS (PMS) (primary progressive MS [PPMS] comprised 12 HDs, 12 PPMS, and 15 SPMS—HDs: 9 women 2 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN Table 1 Demography of Individual Subjects Included in the Study Gender and Mean and Healthy number of range in Flow cytometry study subject or MS samples years Clinic Baseline DN NKT HC Female n = 24 37 (24–62) Danish MS Clinic Male n = 8 39 (24–63) Danish MS Clinic RRMS Female n = 14 38 (21–55) Danish MS Clinic Male n = 5 34 (28–41) Danish MS Clinic PPMS Female n = 8 51 (43–59) Danish MS Clinic Male n = 12 47 (28–59) Danish MS Clinic SPMS Female n = 11 46 (33–60) Danish MS Clinic Male n = 9 46 (30–53) Danish MS Clinic Baseline iNKT HD Female n = 9 44 (26–58) Blood Bank Male n = 3 30 (23–37) Blood Bank RRMS Female n = 4 30 (24–37) Universitari Vall d’Hebron MS Clinic Male n = 2 38 (36, 41) Universitari Vall d’Hebron MS clinic PPMS Female n = 6 55 (52–60) Danish MS Clinic Male n = 6 51 (31–62) Danish MS Clinic SPMS Female n = 7 54 (53–64) Danish MS Clinic Male n = 8 50 (36–62) Danish MS Clinic In vitro hCII707-721 HD Female n = 9 44 (26–58) Blood Bank Reactivity Male n = 3 30 (23–37) Blood Bank PPMS Female n = 6 55 (52–60) Danish MS Clinic Male n = 6 51 (31–62) Danish MS Clinic SPMS Female n = 8 54 (36–64) Danish MS Clinic Male n = 7 50 (36–62) Danish MS Clinic iNKT HD Female n = 13 33 (19–57) Blood bank Suppressive assay and phenotyping Male n = 11 42 (21–59) Blood bank RNA-seq HD Female n = 3 30–40 Blood bank Abbreviations: DN = double-negative; HC = healthy control; HD = healthy donor; iNKT = invariant NKT; MS = multiple sclerosis; NKT = natural killer; PPMS = primary progressive MS; RRMS = relapsing-remitting MS; SPMS = secondary progressive MS. (44 years; 26–58) and 3 men (30 years; 23–37 years), PPMS: 6 Preparation and Treatment of Human PBMCs women (55 years; 52–60) and 6 men (51 years; 31–62), and Buffy coats of blood donors and blood (15 mL) from patients SPMS: 8 women (54 years; 36–64) and 7 men (50 years; with MS (10–12 mL) were used to prepare PBMCs using the 36–62). Ficoll-Paque Plus (7.5 mL) (GE Healthcare, 17-1440-02). The cells were then centrifuged at 400g for 30–40 minutes at In vitro phenotyping and suppression of iNKT cells were 18°C–20°C. The lymphocyte layer was collected, washed 3 analyzed in 24 HDs: 13 women (33 years; 19–57) and 11 men times with PBS, and centrifuged at 300g for 10 minutes. The (42 years; 21–59). Table 1 details the patients and HC in- cells were resuspended at 3 × 106 cells/mL in Dulbecco’s cluded in the various assays. Modified Eagle Medium (DMEM) with Glutamax-1 (Gibco), 10% fetal bovine serum (Sigma), 1% HEPES 1 M (Invitrogen), None of the patients have received immunosuppressive or 1% penicillin-streptomycin 10,000 U/mL (Gibco), and 0.001% immunomodulatory treatment before sample date. beta-mercaptoethanol 50 mM to make complete DMEM Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 3 (cDMEM). PBMCs were analyzed or stored in liquid nitrogen distinguish the hCII707-721-treated iNKT cells before purifi- until use. cation from nontreated iNKT cells, the hCII707-721–treated iNKT cell population is referred to as hCII707-721–reactive Human Collagen II Peptide 707-721–Activated iNKT cells. NKT Cells PBMCs were cultured in round-bottomed 96-well culture plates CFSE Staining (Nunc) in a humidified 37°C atmosphere, containing 5% CO2. PBMCs were labeled with fresh 1 μM carboxyfluorescein PBMCs were treated with 5 μg/mL α-GalCer (KRN7000 succinimidyl ester (CFSE; Molecular Probes) at 37°C for 2 Funakoshi Co, Ltd.), 100 μg/mL hCII707-721 (H- minutes and then at room temperature for 7 minutes in the PPGSNGNPGPPGPPG-OH Schafer-N), or collagen type dark. An equal volume of fetal bovine serum was added to halt I 707–721 as a negative control. Cells were stained with the reaction, and the cells were washed 3 times with cDMEM. antibodies after 48 hours for a fluorescence-activated cell sorting (FACS) analysis. In Vitro Suppression Assay PBMCs were labeled with CFSE and activated (2 × 105 cells/ FACS Staining well) with plate-bound anti-CD3 10 μg/mL (OKT3 16–003781 PBMCs were washed with PBS and 0.5% BSA (Sigma). Fc eBioscience) and human soluble anti-CD28 2 μg/mL (CD28.2 receptors (FcR) were blocked with antihuman FcR blocking 16–028981 eBioscience) for 48 hours. Autologous hCII707- reagent (Miltenyi Biotec 130-059-901) in PBS and 2% BSA 721–reactive iNKT cells from the same HD were purified by for 30 minutes at 4°C per the manufacturer’s instructions. All autoMACS and cocultured at 1:1 with CFSE+ PBMCs for an- antibodies were used at 1–5 μg/mL: TCR αβ anti-human other 24 hours. Pacific Blue (IP26 306716 Biolegend), CD4 antihuman APC- Cy7 (RPA-T4 300518 Biolegend), CD8 antihuman BV510 For certain experiments, 1 hour before coculture, PBMCs (RPA-T8 563526 BD Biosciences), CD56 antihuman BV605 were blocked with antibodies against human Fas (250 ng/mL, (HCD56 318334 Biolegend), CD161 antihuman FITC ZB4 Millipore 05–338), TNF receptor I (TNFRI) (10 μg/ (191B8 130-092-097 Miltenyi Biotec), CD94 antihuman mL, H398 In vivo Gen mab-htnfr1), and TNF receptor II APC (HP-3D9 17–5094 eBioscience) or Vα24Jα18 TCR (TNFRII) (0.5 μg/mL, 22221 R&D Systems MAB226). IgG antihuman APC (6B11 17–5806 eBioscience), CD178 (BD Pharmingen Purified Mouse IgG/IgG2a κ Isotype (FasL) antihuman PE (NOK-1 306407 Biolegend), CD3 564121/554126) was used as a negative control. antihuman APC- Cy7 (HIT3a 300318 Biolegend), CD4 Pe- Cy7 antihuman (RPA-T4 300512 Biolegend), CD8 antihu- After 24 hours of coculture, cells were harvested and sup- man eFluorNC 605 (RPA-T8 94–0088 eBioscience), CD56 plemented with 7ADD before FACS analysis of CFSE+ vs antihuman PerCP-Cy5.5 (HCD56 318322 Biolegend), and 7AAD+ cells. CD161 antihuman APC (DX12 V1NK12 BD Biosciences). 7-Aminoactinomycin D (7AAD) (BD Pharmingen 559925) ELISA was applied to exclude dead cells. Supernatants from in vitro PBMC cultures, treated with α-GalCer and hCII707-721 for 48 hours, were analyzed using Intracellular cytokine staining was performed in iNKT mag- the Granzyme B Platinum ELISA kit (Affymetrix eBioscience netic bead-enriched PBMCs after blockade with Brefeldin-A BMS2027) per the manufacturer’s instructions. 10 μg/mL (420601 Biolegend) for 6 hours, using IL-4 antihuman PE (8D4-8 500704 Biolegend), IFN-γ antihuman RNA-Seq Sample Preparation and Gene Alexa Fluor-700 (4S.B3 502520 Biolegend), TNFα antihuman Expression Analysis Alexa Fluor-488 (Mab11 55,722 BD Pharmingen), IL17F Total RNA was extracted from PBMCs samples (3 women antihuman PerCP efluor- 710 (SHLR17 46–7169 eBio- HDs, 30–40 years) using the RNeasy mini kit (Qiagen, science), granzyme B antihuman/mouse FITC (GB11 515403 74,104). The RNA-seq samples were composed of 3 biological Biolegend), perforin antihuman APC (DG9 308111 Biol- individual samples for each population, that is, 3 conventional egend), and TGF-β1 antihuman PE-Cy7 (TW4-2F8 349610 T cells, 3 conventional iNKT cells (without hCII707-721 ac- Biolegend). tivation), and 3 hCII707-721–reactive iNKT cells (with hCII707-721 activation). Messenger ribonucleic acid was pre- Cells were acquired on a BD LSR II using FACSDiva and pared using oligo-dT primers and reverse transcribed (KAPA analyzed in FlowJo (Tree), v8.8.7. Biosystems, KK2601). The double-stranded DNA was tag- mented (Illumina, FC-131–1096) and ligated with sequencing Purification of iNKT Cells adaptors (Illumina, FC-131–2002) as detailed.49 PBMCs (4–10 × 108) were cultured in cDMEM at 3 × 106 cells/mL with hCII707-721 (100 μg/mL) in T- 175 cm2 flasks Libraries were pooled (1.6 pM) and sequenced through (Nunc/Fischer). After activation with hCII707-721 for 48 paired-end 75-bp reads on an Illumina NextSeq500. Three hours, total iNKT cells were purified by autoMACS using biological replicates were used for RNA-seq for each condi- Anti-iNKT MicroBeads (130-094-842, Miltenyi Biotec). To tion (3 healthy women, 30–40 years). Quality control and 4 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN preprocessing steps were performed using FastQC and Trim- activation status, based on their clustering in principal compo- momatic. Gene expression was quantified by mapping the reads nent analysis plots and gene expression profiles (false discovery to the human genome (GRCh38.p13, GENCODE v32) using rate [FDR] < 0.05) (Figure 2A and eFigure 2A and B, links.lww. Salmon (v0.14.1). On average, there were 10.5 M mapped com/NXI/A548). reads per library. Pairwise comparison of biological replicates between conditions was performed using edgeR. Differential iNKT cells had a distinct transcriptional profile compared with isoform usage was analyzed using IsoformSwitchAnalyzeR.50 convT cells, wherein 1927 genes were differentially expressed Pathway and gene ontology enrichment analysis was per- (FDR < 0.05) (Figure 2A–B, eTable 1, links.lww.com/NXI/ formed in R using ReactomePA,51 CAMERA,52 and fgsea.53-55 A550), 139 of which were lncRNAs (Figure 2C, eTable 2, links. For the pathway enrichment analysis, we used annotated gene lww.com/NXI/A551). Of these 1,927 genes, 1,397 were signif- sets from the Molecular Signatures Database.54 icantly upregulated and 530 were downregulated (Figure 2B, D). Next, we annotated cell surface markers and transcription fac- Data Availability tors23 among these differential genes to identify the top surface All data generated or analyzed during this study are included in markers and a transcription factor network that were related to this published article (and its supplementary information files). iNKT function—constituting the gene signature of iNKTs (Figure 2E-F, eTable 1). The iNKT-specific transcription factor profile was used to perform STRING network analysis,24 in Results which overrepresented Reactome pathways were identified Reduced Circulating NKT Cells and a Lack of among the core factors (eFigure 2A–C, links.lww.com/NXI/ hCII707-721-Reactive iNKT Cells in PMS A548). Patients with PMS are often older than patients with RRMS, and because age could be a confounding factor affecting im- Next, we performed a Reactome pathway enrichment analysis mune cells, we investigated first if age affects the frequency of of the upregulated genes to identify the molecular pathways that NKT cells. The correlation analysis of age vs subset frequency were associated with iNKT differentiation and function in the HCs did not reveal a significant impact of the age (Figure 2G). The resulting pathways reflected the cytotoxic interval used in the study on the frequency of NKT cells nature of iNKT cells and included terms that are associated with (eFigure 1, links.lww.com/NXI/A547). important effector functions, such as degranulation (neutrophil degranulation), ROS production (ROS and RNA production in In examining whether PMS is associated with NKT subsets phagocytes), and interleukin signaling (interleukins 4, 13, and using established markers, we did not find any differences in 10). Finally, we conducted gene set enrichment analysis single positive CD3+ or CD4+ NKT cells between MS and HCs, (GSEA) of the iNKT profile vs convT, confirming the upre- nor did we find differences based on CD56+, CD95+, and gulation of iNKT effector functions using several pathway da- CD161+ triple-positive population (data not shown). However, tabases (Figure 2H). in support of previous studies,13,14 among double-negative cells (CD4−CD8− DN22), the percentages of CD4−CD8−CD161+ Transcriptomic Profiling Reveals hCII707- NKT cells and CD4−CD8−CD56+CD161+ NKT cells were 721–Induced Activation and Proliferation lower in patients with PPMS (Figure 1A–C). Based on the of iNKTs invariant TCR α-chain Vα24Jα18 (Figure 1D), we found a de- Compared with conventional nonactivated iNKT cells, crease in iNKT cells, albeit insignificant, in both PMS patient hCII707-721–reactive iNKT cells upregulated 115 genes groups but not RRMS patients (Figure 1E). and downregulated 59 genes (p < 0.05, |logFC| > 0.5) (Figure 3A–C, eFigure 2D, links.lww.com/NXI/A548, Healthy individuals harbored no hCII707-721–reactive CD4+ or eTable 3, links.lww.com/NXI/A552), 33 of which were DN NKT cells after a 48-hour challenge with this peptide (not lncRNAs (Figure 3D, eTable 4, links.lww.com/NXI/ shown) but had a significant population of a novel hCII707- A553)—the first indication that hCII707-721 activates 721–reactive iNKT subset (Figure 1F,G). Gating on Vα24Jα18+ human iNKT cells. iNKT cells (Figure 1F), hCII707-721, not α-GalCer, significantly increased the percentage of iNKTs in HDs (Figure 1G). Fur- By GSEA, genes in pathways that were related to iNKT ef- thermore, patients with PMS, particularly those with SPMS, had fector function were consistently upregulated. In 3 databases fewer hCII707-721–reactive iNKT cells vs HDs (Figure 1G). (GO, KEGG, and Reactome), proliferation-related pathways were the most highly upregulated (Figure 3E). Furthermore, Human iNKT-specific Transcriptomic Signature the appearance of hematopoietic cell lineage by KEGG (FDR No complete gene signature of human iNKT cells has been = 0.096) suggested that lineage commitment and differenti- reported. To address this shortcoming, we purified naı̈ve con- ation occurred on activation (Figure 3E–F). In addition, ventional CD4+ T cells (convT), conventional iNKT cells, and hCII707-721 enhanced the effector function of iNKTs, based hCII707-721–reactive iNKT cells from PBMCs and Vα24Jα18- on their upregulation of TNF-related genes (Reactome- expressing cells, respectively, and performed RNA sequencing. TNFs bind their physiologic receptors, FDR = 0.054) We considered all iNKTs as an entity, regardless of their (Figure 3E). Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 5 Figure 1 NKT Cell Levels are Reduced and hCII707-721-induced iNKT Cells Are Hyporesponsive in PPMS PBMCs were freshly isolated from each subject and stained with CD3, CD4, CD8, CD56, CD161, and Vα24Jα18. (A) FACS gating strategy of CD3+CD4−CD8− (DN) NKT cells and gating of DN CD56+CD161+ NKTs. (B) The percentage of DN NKT cells based on CD161 expression. (C) The percentage of DN NKT cells based on CD56 and CD161 expression. The bars indicate the mean ± SEM in HC (n = 32), patients with RRMS (n = 19), patients with PPMS (n = 20), and patients with SPMS (n = 20). (D) Gating of Vα24Jα18+ iNKT cells from single and live lymphocytes compared with isotype control-stained cells. (E) The percentage of Vα24Jα18+ iNKT cells. The bars indicate the mean ± SEM. HD (n = 12), RRMS (n = 6), PPMS (n = 10), and SPMS (n = 13). *p < 0.05 One-way ANOVA with the Tukey multiple comparison test was applied. (F–G) PBMCs were cultured in vitro with α- GalCer (5 μg/mL), hCII707-721 (100 μg/mL), or media for 48 hours. (F) Representative FACS dot plots showing gating of Vα24Jα18+ iNKT cells in the various treatments. (G) Ratio of Vα24Jα18+ iNKT cells. The value of Vα24Jα18+ iNKT cells in nontreated sample (media) = 1 for HD, PPMS, and SPMS (empty bar). Ratio is calculated as the percentage of treated Vα24Jα18+ iNKT cells/percentage of Vα24Jα18+ iNKT cells in media. The bars indicate the mean ± SEM, HD (n = 18), PPMS (n = 6–18), and SPMS (n = 15–18). Student t test was applied, *p < 0.05 between control and hCII707-721 in HDs. #p < 0.05 between HD and SPMS with hCII707-721 treatment. ANOVA = analysis of variance; CD = cluster of differentiation; DN = double-negative; FACS = fluorescence- activated cell sorting; HC = healthy control; HD = healthy donor; NKT = natural killer T; PBMC = peripheral blood mononuclear cell; PPMS = primary progressive MS; RRMS = relapsing-remitting MS; SEM = standard error of mean; SPMS = secondary progressive MS. 6 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN Figure 2 iNKT Cells Have a Distinct Gene Expression Profile Compared With Conventional T Cells Conventional T cells, conventional iNKT cells (without hCII707-721 activation), and hCII707-721–reactive iNKT cells (with hCII707-721 activation) purified from PBMCs were analyzed by RNA-seq. Conventional iNKT cells and hCII707-721–reactive iNKT cells were considered as a single group as iNKT cells in this analysis. n = 3/group. (A) Principal component analysis plot of convT (red) and iNKT cells (green). (B) Volcano plot of signature genes of iNKT cells vs conventional T cells (FDR < 0.05). (C) Volcano plot of differentially expressed lncRNAs between iNKT and conventional T cells (FDR < 0.05). (D) Heatmap of NKT signature genes. (E) A snake plot of cell surface markers of NKT cells highlighted the 15 upregulated and 15 downregulated genes. (F) A snake plot of transcription factor markers of NKT cells highlighted the 15 upregulated and 15 downregulated genes. (G) Top 50 enriched Reactome pathways among the genes that are upregulated in NKT cells vs convT. (H) GSEA showing the enriched pathways and molecular functions in iNKT cells per the GO, KEGG, and Reactome databases. convT = conventional CD4+ T cells; FDR = false discovery rate; GSEA = gene set enrichment analysis; iNKT = invariant NKT; NKT = natural killer T cell; PBMC = peripheral blood mononuclear cell. Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 7 Figure 3 hCII707-721 Induces Proliferation and Activation in hCII707-721–Reactive iNKT Cells Treated iNKT cells were analyzed by RNA-seq compared with iNKT cells and convT. (A) Principal component analysis plot of conventional T cells, conventional iNKT cells, and hCII707-721–reactive iNKT cells. (B) Heatmap of differentially expressed genes in hCII707-721–reactive iNKT compared with iNKT. (C) Volcano plot showing the signature genes of hCII707-721–reactive iNKT vs untreated iNKT cells (p < 0.05). (D) Volcano plot showing the lncRNAs differentially expressed between hCII707-721–reactive iNKT and conventional iNKT (p < 0.05). (E) GSEA of regulated GO, KEGG, and Reactome pathways in hCII707-721–reactive iNKT vs conventional iNKT cells. (F) Enrichment plot of KEGG-Hematopoietic cell lineage pathway in hCII707-721–reactive iNKT vs conventional iNKT cells (p = 0.0038, FDR = 0.0964, NES = 1.624). (G) Enrichment plot of KEGG-Spliceosome pathway in hCII707-721–reactive iNKT vs conventional iNKT cells (p = 0.00048, FDR = 0.022, NES = 1.693). (H) Differential isoform use of ARHGDIA in hCII707-721–reactive iNKT cells. (I) Differential isoform use of AKT1 in hCII707- 721–reactive iNKT cells. FDR = false discovery rate; GSEA = gene set enrichment analysis; iNKT = invariant NKT; NES = normalized enrichment scores. 8 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN Given the upregulation of the KEGG_SPLICEOSOME we hypothesized that hCII707-721–reactive iNKT cells sup- pathway (FDR = 0.022) (Figure 3G), we examined whether press by engaging the Fas- FasL pathway. Suppression assay, alternative splicing was induced by hCII707-721, finding 67 using a blocking antibody to Fas on activated T cells, revealed genes with significant differential isoform usage (eFigure 3A the importance of this pathway (Figure 5N–O). Because and B, links.lww.com/NXI/A549, eTable 5, links.lww.com/ hCII707-721–reactive iNKT cells have higher TNF levels and NXI/A554). Specifically, on stimulation, Rho GDP disso- because cytotoxic lymphocytes induce cell death through ciation inhibitor alpha (ARHGDIA) changed to a non- TNF/TNFR,31 we also blocked TNFRI and TNFRII on ac- functional isoform that lacked the functional domain, tivated T cells to determine the involvement of TNFRI in promoting Rho GTPase activation, which mediates T cell iNKT cell-mediated suppression (Figure 5N–O). Although activation, ROS production, and cell migration25,26 the secretion of other cytokines by NKT cells is important for (Figure 3H). Furthermore, hCII707-721 favored splicing of their function,3,6,7,32 we did not find any indication of their full-length AKT serine/threonine kinase 1 (AKT1), which association with reactivity to hCII707-721, consistent with the mediates cell growth and proliferation27 (Figure 3I). absence of cytokine-mediated suppression in murine CII707- 721–reactive NKT cells.21 Cytokine and Cytotoxic Molecule Profile of hCII707-721-Reactive iNKT Cells Blocking Fas-FasL and TNFRI abolished the suppressive function Given that cytokine pathways were among the top KEGG of hCII707-721–reactive iNKT cells against activated T cell pro- pathways in hCII707-721–reactive iNKT cells (Figure 3D), liferation, whereas blocking TNFRII had no such effect we examined NKT cell signature cytokines in our RNA-seq (Figure 5N–O). However, the percentage of cell death (7AAD data set. None of these cytokines was upregulated on stimu- staining) was not affected (data not shown). These data suggested lation with hCII707-721 (Figure 4A). The RNA-seq data were that both Fas-FasL and TNF−TNFRI signaling are involved in verified by intracellular cytokine measurements, merely hCII707-721–reactive iNKT cell-mediated suppression. showing significantly higher TNFα in hCII707-721–reactive iNKT cells vs conventional iNKT cells (gated on Vα24Jα18+ cells) (Figure 4B–F). By ELISA, IFN-γ, IL-4, and IL-17F Discussion levels were minor (data not shown). Notably, unlike the reported massive induction of cytokines on administration of The dysregulation of regulatory and suppressive immune cell α-GalCer,28 we found that hCII707-721 does not alter the types in autoimmune conditions33 has spawned efforts to identify expression or production of intracellular IFN-γ, TGF-β, or IL- new such populations, culminating in the description of murine 17F (Figure 4B). CII707-721–reactive NKT cells21 and circulating FoxA1+Treg cells in human blood.34 In response to self-antigens and foreign Perforin and granzyme B were expressed in nonactivated con- antigens, NKT cells activate rapidly and develop proinflammatory ventional iNKT cells (Figure 4G), consistent with previous and immunoregulatory functions, resulting in protective or reports.29,30 In addition, perforin expression did not differ be- harmful responses in many pathologies in mice and humans, tween hCII707-721–reactive and nonactivated conventional including autoimmune disease, allergy, and cancer.6,10,33,35,36 This iNKT cells (Figure 4G–I, M). Notably, granzyme B expression association underscores the importance of understanding the and secretion were reduced in hCII707-721–reactive iNKT cells function and regulation of NKT cells to exploit them in the vs conventional iNKT cells (Figure 4G, H, J, L, N). hCII707- management of these diseases.37 721–reactive iNKT cells also produced significantly less gran- zyme B than α-GalCer-reactive iNKT cells (Figure 4L). FasL NKT cells have been implicated in MS, causing motor deficits, was upregulated on activation of iNKT cells with hCII707-721 loss of coordination, cognitive and visual disturbances, and ulti- compared with nonactivated iNKT cells, albeit insignificantly mately severe disability.11 In keeping with their immunoregula- (Figure 4G-H, K, O). tory functions, lower levels of and functional defects in NKT cell subsets have been reported in patients with MS.13,14 Conclu- hCII707-721-Reactive iNKT Cells Suppress sively, previous studies on this topic have been contradictory,13- 15,22 Activated T Cells Through Fas/FasL and TNFα/ likely because different markers were used to study NKT cell TNFRI subsets in MS. Consistent with the evidence of a lower per- As shown in Figure 5A–D, by in vitro suppression assay, the centage of circulating NKT cells, in this study, we noted a lower proliferation of activated responder T cells declined significantly percentage of various NKT cell populations in patients with PMS, in the presence of enriched autologous hCII707-721–reactive including DN-NKT and iNKT cells. These findings indicate that iNKT cells, without significantly affecting cell death (by 7AAD+ NKT cells are protective in MS and that their absence is associ- staining) (Figure 5E–F), such as murine CII707-721–reactive ated with disease progression. NKT cells,21 whereas nonactivated conventional iNKT cells effected no significant suppression (Figure 5G–K). The major adverse effects of α-GalCer render it a suboptimal option for treating MS, perpetuating the unmet clinical need By transcriptomic analysis, hCII707-721–reactive iNKT cells for therapies that activate protective NKT cells. Given the upregulated Fas and TNF signaling (Figure 5L and M). Thus, therapeutic potential of NKT cells and our description of Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 9 Figure 4 Cytokine and Cytotoxic Molecule Profiles of hCII707-721–Reactive iNKT Cells Compared With Conventional iNKT Cells PBMCs were isolated from HDs and treated with or without hCII707-721 for 48 hours. Cells were gated for Vα24Jα18+ iNKT cells. (A) Normalized counts (from RNA- Seq) of IFN-γ+, IL-4+, TGFβ+, TNFα+, and IL-17F+ cells in iNKT cells. (B) The percentage of IFN-γ+, IL-4+, TGFβ+, TNFα +, and IL-17F+ cells in iNKT cells. Multiple t test, one per row was applied. *p < 0.05 between hCII707-721 and untreated conventional iNKTs in HDs. (C) Dot plots showing the back-gating strategy for lymphocyte population, single cells, and live cells. (D) Dot plots of isotype controls for various cytokine stains. (E) Representative dot plots of cytokine stains from untreated conventional iNKT cells. (F) Representative dot plots of cytokine stains from hCII707-721–reactive iNKT cells. (G) Representative dot plots of perforin+, granzyme B+, and FasL+ cells from gated nontreated conventional iNKT cells. (H) hCII707-721–reactive iNKT cells. (I) The percentage of perforin+ (J) and granzyme B+ cells in gated iNKT cells. (K) FasL+ cells in gated iNKT cells. (L) Granzyme B production by -GalCer, hCII707-721, and untreated conventional iNKT cells after 48 hours of culture, (M) normalized counts (log2-transformed normalized counts from DESeq2) of perforin expression by RNA-seq. (N) Normalized counts of granzyme B expression by RNA-seq. (O) Normalized counts of FasL by RNA-seq. The bars indicate the mean ± SEM; n = 3–4. HD = healthy donor; IFN = interferon; IL = interleukin; NKT = natural killer T; PBMC = peripheral blood mononuclear cell; SEM = standard error of mean; TGF = transforming growth factor; TNF = tumor necrosis factor. 10 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN Figure 5 hCII707-721–Reactive iNKT Cells Suppress Through Fas-FasL and TNFα-TNFRI CFSE+ PBMC cells from HDs were activated with plate-bound anti-CD3 (10 μg/mL) and soluble anti- CD28 (2 μg/mL) for 48 hours as responders (Resp) and cocultured with purified autologous hCII707-721–reactive Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells) for another 24 hours. (A) FACS gating strategy of back-gating and gating for Vα24Jα18+ iNKT cells. (B) FACS plot of gating strategy for responder CFSE + cells (hCII707-721–reactive Vα24Jα18+ iNKT cells as suppressors). (C) hCII707- 721–reactive Vα24Jα18+ iNKT cells from PBMCs of a HD were enriched with magnetic beads before coculture with autologous activated responder T lymphocytes for functional suppression assay. Vα24Jα18+ iNKT cells were obtained by treatment with hCII707-721 (100 μg/mL) for 48 hours (C–F) or from conventional Vα24Jα18+ iNKT (G–K). (D) Representative histograms of CFSE+ gating in Resp CFSE+ responder T cells (Resp) cocultured with reactive hCII707-721 Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells) for another 24 hours. (E) Representative histograms of 7AAD+ gating in Resp cocultured with reactive hCII707- 721 Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells). (F) The percentage of 7AAD+ dead cells in Resp CFSE+ responder T cells (Resp) were then cocultured with reactive hCII707-721 Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells) for another 24 hours. (G) FACS plot of gating strategy for responder CFSE+ cells (conventional Vα24Jα18+ iNKT cells as suppressors). (H) Conventional Vα24Jα18+ iNKT cells cocultured with autologous activated responder T lymphocytes for functional suppression assay. Representative histograms of CFSE+ gating in Resp CFSE+ responder T cells (Resp) cocultured with conventional Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells) for another 24 hours, with percentages in (I). (J) Representative histograms of 7AAD+ gating in Resp cocultured with conventional Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells). (K) The percentage of 7AAD+ dead cells in Resp CFSE+ responder T cells (Resp) cocultured with conventional Vα24Jα18+ iNKT cells or control cells (Vα24Jα18- cells) for another 24 hours. One-way ANOVA *p < 0.05. The bars indicate the mean ± SEM, n = 7. (L) TNF pathway (PID, MSidDB) enrichment in hCII707-721–reactive iNKT cells compared with conventional iNKT cells (FDR = 0.038). (M) FAS pathway (PID, MSidDB) enrichment in hCII707-721–reactive iNKT vs conventional iNKT cells (FDR = 0.14). (N) Gating strategy for responder CFSE + cells and proliferating CFSE+ cells. (O) The percentage of proliferating CFSE+ cells. Ratio of Resp-only (without neutralizing antibodies) set to 1; ratios calculated as percentage of proliferating CFSE+ cells in various cocultures/percentage of proliferating CFSE+ cells in Resp-only. The bars indicate the mean ± SEM; n = 6–8. Unpaired student t test was applied; *p < 0.05 between Resp + iNKT (Control) and Resp-only. One-way ANOVA and Dunnett multiple-comparisons test: #p < 0.05, ##p < 0.01 between groups with blocking antibodies and control group. ANOVA = analysis of variance; CFSE = carboxyfluorescein succinimidyl ester; FACS = fluorescence-activated cell sorting; FDR = false discovery rate; HD = healthy donor; MS = multiple sclerosis; NKT = natural killer T; PBMC = peripheral blood mononuclear cell; PID = The Pathway Interaction Database; SEM = standard error of mean; TNF = tumor necrosis factor; TNFRI = TNF receptor I; 7AAD = 7-Aminoactinomycin D. Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 11 murine CII707-721–reactive NKT cells,21 we examined antigens.33 Notably, certain NKT cells react with sulfatide, a whether similar cells with possible clinical benefit existed in self-glycolipid in the myelin sheaths that coats and insulates humans. This premise led to our identification of hCII707- nerve fibers in the CNS.44 CD1-dependent activation by 721, revealing the existence of a novel iNKT cell population in α-GalCer in mice induces all signature cytokines of various T healthy persons that is absent in patients with PMS and helper subsets, demonstrating the cytokine storm that is eli- prompting us to characterize their molecular and suppressive cited by this stimulation.45 In contrast, the self-peptide properties. We have identified the first human self-peptide hCII707-721 seems to be protective because of the conse- hCII707-721–reactive iNKT cell population in healthy hu- quent low cytokine output—except for TNFα, which initiates mans and determined their transcriptional profile, which was its TNFRI-mediated suppression—unlike with other CD1- borne out in their suppressive function. restricted NKT cells, such as α-GalCer-reactive cells. Effector function pathways are significantly enriched in iNKT In summary, we found iNKT cells to be hyporesponsive to vs convT cells, including NK cell-mediated immunity, che- hCII707-721 in patients with PMS, reflecting a potential state mokine and cytokine responses, signal transduction, and cell of anergy in circulating self-reactive iNKT cells in these pa- motility and survival.38 In our study, some such activities were tients. Determining how anergic iNKT cells can be activated validated in the general human iNKT cell signature, and in patients with PMS might be helpful in designing hCII707- others appeared only on activation of iNKT cells with 721–based therapies. Furthermore, direct clinical benefits hCII707-721. have been obtained by stimulating or adoptively transferring NKT cells to correct defects in the NKT cell pool.8,9,17,21,46 Previous studies on the suppressive functions of iNKT cells have Low production of IFN-γ, IL-4, and IL-17F by hCII707- focused on the involvement of cytokines, based on the ability of 721–reactive iNKT cells is a highly favorable attribute, NKT cells to robustly and rapidly produce cytokines that drive avoiding, for example, shifts in polarizing T helper subsets, Th1, Th2, and Th17 cell differentiation and function.3,6,7,32 which are associated with pathology in MS.47 Thus, based on Although the “cytokine-cytokine receptors” interaction was our novel identification of endogenous self-antigen hCII707- among the major KEGG pathways that were induced by 721–reactive iNKT cells, approaches that use hCII707-721 to hCII707-721 in iNKT cells, the activation of iNKT cells by expand and activate hCII707-721–reactive iNKT cells hCII707-721 was not associated with any specific cytokine have tremendous therapeutic potential to correct the defects pattern, except for higher TNFα levels. Consistent with this of NKT cells in patients with MS and prevent disease finding, our transcriptomic analysis of hCII707-721–reactive progression. iNKT vs conviNKT cells implicated the TNF pathway as being significantly enriched. Although blocking TNFRII on activated Study Funding T lymphocytes did not hinder the antiproliferative effects of This work was supported by grants from the Lundbeck hCII707-721–reactive iNKT cells, TNFα-mediated TNFRI Foundation, Danish MS Society, Independent Research Fund signaling—a pathway that induces cell death31—contributed to Denmark-Medical and Health Sciences (DFF-M), and their suppressive activity. Foundation for Research in Neurology (to S.I.-N.). B. Carrion received a PhD fellowship from CONACYT-Mexico and the We found that the suppressive function of hCII707- Lundbeck Foundation. M. Hadi was the recipient of a 721–reactive iNKT cells was also mediated by Fas-FasL sig- McDonald fellowship from the MSIF program. naling, as observed for murine CII707-721–reactive NKT cells.21 A similar mechanism was reported in the kill- Disclosure ing of tumor cells by GalCer-activated NKT cells,39 consistent The authors report no disclosures relevant to the manuscript. with the regulation of T cells by NKT cells through Fas/ Go to Neurology.org/NN for full disclosures. FasL.40,41 As our study has shown, cooperation between Fas/ FasL and TNFα/TNFRI is a potential suppressive mechanism Publication History of hCII707-721–reactive iNKT cells, as supported by the use Received by Neurology: Neuroimmunology & Neuroinflammation of cytotoxic mechanisms by human NKT cells against a February 4, 2021. Accepted in final form June 16, 2021. myelomonocytic leukemia cell line, including perforin/ granzyme-B, TNFα, FAS-L, and TRAIL.42 Although 7AAD staining of cell death was not shown in the differences, it does Appendix Authors not rule out a possibility that cell apoptosis could be involved Name Location Contribution in hCII707-721–reactive iNKT cell-mediated suppression. Belinda Biotech Research and Drafting/revision of the Carrión, MD, Innovation Centre (BRIC), manuscript for content, Although NKT cells primarily recognize and react to glyco- PhD University of Copenhagen, including medical writing Denmark for content; Major role in lipids in the context of CD1 molecules, they can react to the acquisition of data; peptide antigens.43 NKT cell regulation of immune homeo- Study concept or design; stasis might depend on the recognition of self-antigens under Analysis or interpretation of data stress, such as tissue damage, which causes the release of self- 12 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN 3. Kronenberg M. Toward an understanding of NKT cell biology: progress and para- doxes. Annu Rev Immunol. 2005;23:877-900. Appendix (continued) 4. Taniguchi M, Harada M, Kojo S, Nakayama T, Wakao H. The regulatory role of Valpha14 NKT cells in innate and acquired immune response. Annu Rev Immunol. Name Location Contribution 2003;21:483-513. 5. Van Kaer L, Parekh VV, Wu L. Invariant natural killer T cells: bridging innate and Yawei Liu, Biotech Research and Drafting/revision of the adaptive immunity. Cel Tissue Res. 2011;343(1):43-55. MD, PhD Innovation Centre (BRIC), manuscript for content, 6. Tupin E, Kinjo Y, Kronenberg M. The unique role of natural killer T cells in the University of Copenhagen, including medical writing response to microorganisms. Nat Rev Microbiol. 2007;5(6):405-417. Denmark for content; Major role in 7. Terabe M, Berzofsky JA. NKT cells in immunoregulation of tumor immunity: a new the acquisition of data; immunoregulatory axis. Trends Immunology. 2007;28(11):491-496. Analysis or interpretation 8. Teige A, Teige I, Lavasani S, et al. CD1-dependent regulation of chronic central of data nervous system inflammation in experimental autoimmune encephalomyelitis. J Immunol. 2004;172(1):186-194. Mahdieh Biotech Research and Major role in the 9. Teige A, Bockermann R, Hasan M, Olofsson KE, Liu Y, Issazadeh-Navikas S. CD1d- Hadi, PhD Innovation Centre (BRIC), acquisition of data; dependent NKT cells play a protective role in acute and chronic arthritis models by University of Copenhagen, Analysis or interpretation ameliorating antigen-specific Th1 responses. J Immunol. 2010;185(1):345-356. Denmark of data 10. Novak J, Lehuen A. Mechanism of regulation of autoimmunity by iNKT cells. Cy- tokine. 2011;53(3):263-270. Jon Biotech Research and Major role in the 11. Van Kaer L, Wu L, Parekh VV. Natural killer T cells in multiple sclerosis and its animal Lundstrom, Innovation Centre (BRIC), acquisition of data; model, experimental autoimmune encephalomyelitis. Immunology. 2015;146(1):1-10. MSc University of Copenhagen, Analysis or interpretation 12. Goverman J. Autoimmune T cell responses in the central nervous system. Nat Rev Denmark of data Immunol. 2009;9(6):393-407. 13. Illes Z, Kondo T, Newcombe J, Oka N, Tabira T, Yamamura T. Differential expression Jeppe Romme Danish Multiple Sclerosis Major role in the of NK T cell V alpha 24J alpha Q invariant TCR chain in the lesions of multiple Christensen, Center, University of acquisition of data; sclerosis and chronic inflammatory demyelinating polyneuropathy. J Immunol. 2000; MD PhD Copenhagen and Analysis or interpretation 164(8):4375-4381. Department of Neurology, of data 14. van der Vliet HJ, von Blomberg BM, Nishi N, et al. Circulating V(alpha24+) Rigshospitalet, Denmark; Vbeta11+ NKT cell numbers are decreased in a wide variety of diseases that are characterized by autoreactive tissue damage. Clin Immunol. 2001;100(2):144-148. Cecilie Danish Multiple Sclerosis Major role in the 15. Araki M, Kondo T, Gumperz JE, Brenner MB, Miyake S, Yamamura T. Th2 bias of Ammitzbøll, Center, University of acquisition of data CD4+ NKT cells derived from multiple sclerosis in remission. Int Immunol. 2003; MD, PhD Copenhagen and 15(2):279-288. Department of Neurology, 16. Pal E, Tabira T, Kawano T, Taniguchi M, Miyake S, Yamamura T. Costimulation- Rigshospitalet, Denmark dependent modulation of experimental autoimmune encephalomyelitis by ligand stimulation of V alpha 14 NK T cells. J Immunol. 2001;166(1):662-668. Morten Blood Bank, Copenhagen Major role in the 17. Singh AK, Wilson MT, Hong S, et al. Natural killer T cell activation protects mice against Hanefeld University Hospital, Denmark acquisition of data experimental autoimmune encephalomyelitis. J Exp Med. 2001;194(12):1801-1811. Dziegiel, MD, 18. Osman Y, Kawamura T, Naito T, et al. Activation of hepatic NKT cells and sub- PhD sequent liver injury following administration of alpha-galactosylceramide. Eur J Immunol. 2000;30(7):1919-1928. Per Soelberg Danish Multiple Sclerosis Major role in the 19. Ito K, Karasawa M, Kawano T, et al. Involvement of decidual Valpha14 NKT cells in Sørensen, MD Center, University of acquisition of data abortion. Proc Natl Acad Sci U S A. 2000;97(2):740-744. Copenhagen and 20. Whitman SC, Ramsamy TA. Participatory role of natural killer and natural killer Department of Neurology, T cells in atherosclerosis: lessons learned from in vivo mouse studies. Can J Physiol Rigshospitalet, Denmark Pharmacol. 2006;84(1):67-75. 21. Liu Y, Teige A, Mondoc E, Ibrahim S, Holmdahl R, Issazadeh-Navikas S. Endogenous Manuel Centre d’Esclerosi Múltiple de Major role in the collagen peptide activation of CD1d-restricted NKT cells ameliorates tissue-specific Comabella, Catalunya, Cemcat, Unitat de acquisition of data inflammation in mice. J Clin Invest. 2011;121(1):249-264. MD, PhD Neuroimmunologia Clı́nica, 22. O’Keeffe J, Gately CM, Counihan T, et al. T-cells expressing natural killer (NK) Hospital Universitari Vall d´ receptors are altered in multiple sclerosis and responses to alpha-galactosylceramide Hebron (HUVH) Universitat are impaired. J Neurol Sci. 2008;275(1–2):22-28. Autònoma de Barcelona, 23. Hu H, Miao YR, Jia LH, Yu QY, Zhang Q, Guo AY. AnimalTFDB 3.0: a compre- Barcelona, Spain hensive resource for annotation and prediction of animal transcription factors. Nucleic Acids Res. 2019;47(D1):D33-D38. Xavier Centre d Esclerosi Múltiple de Major role in the 24. Szklarczyk D, Gable AL, Lyon D, et al. STRING v11: protein-protein association Montalban, Catalunya, Cemcat, Unitat de acquisition of data networks with increased coverage, supporting functional discovery in genome-wide MD, PhD Neuroimmunologia Clı́nica, experimental datasets. Nucleic Acids Res. 2019;47(D1):D607-D613. Hospital Universitari Vall d´ 25. Lu W, Wang X, Liu J, et al. Downregulation of ARHGDIA contributes to human Hebron (HUVH) Universitat glioma progression through activation of Rho GTPase signaling pathway. Tumour Autònoma de Barcelona, Biol. 2016;37(12):15783-15793. Barcelona, Spain 26. Manresa-Arraut A, Johansen FF, Brakebusch C, Issazadeh-Navikas S, Hasseldam H. RhoA drives T-cell activation and encephalitogenic potential in an animal model of Finn Danish Multiple Sclerosis Major role in the multiple sclerosis. Front Immunol. 2018;9:1235. Sellebjerg, Center, University of acquisition of data 27. Duggal S, Jailkhani N, Midha MK, Agrawal N, Rao KVS, Kumar A. Defining the Akt1 MD, PhD Copenhagen and interactome and its role in regulating the cell cycle. Sci Rep. 2018;8(1):1303. Department of Neurology, 28. Wu L, Van Kaer L. Natural killer T cells and autoimmune disease. Curr Mol Med. Rigshospitalet, Denmark 2009;9(1):4-14. 29. Gumperz JE, Miyake S, Yamamura T, Brenner MB. Functionally distinct subsets of Shohreh Biotech Research and Drafting/revision of the CD1d-restricted natural killer T cells revealed by CD1d tetramer staining. J Exp Med. Issazadeh- Innovation Centre (BRIC), manuscript for content, 2002;195(5):625-636. Navikas, PhD University of Copenhagen, including medical writing 30. Lee PT, Benlagha K, Teyton L, Bendelac A. Distinct functional lineages of human Denmark for content; Study concept V(alpha)24 natural killer T cells. J Exp Med. 2002;195(5):637-641. or design; Analysis or 31. Zimmerer JM, Horne PH, Fiessinger LA, et al. Cytotoxic effector function of CD4- interpretation of data independent, CD8(+) T cells is mediated by TNF-alpha/TNFR. Transplantation. 2012;94(11):1103-1110. 32. Swann JB, Coquet JM, Smyth MJ, Godfrey DI. CD1-restricted T cells and tumor immunity. Curr Top Microbiol Immunol. 2007;314:293-323. References 33. Issazadeh-Navikas SNKT. Cell self-reactivity: evolutionary master key of immune 1. Bendelac A, Savage PB, Teyton L. The biology of NKT cells. Annu Rev Immunol. homeostasis? J Mol Cell Biol. 2012;4(2):70-78. 2007;25:297-336. 34. Liu Y, Carlsson R, Comabella M, et al. FoxA1 directs the lineage and immunosup- 2. Brigl M, Brenner MB. CD1: antigen presentation and T cell function. Annu Rev pressive properties of a novel regulatory T cell population in EAE and MS. Nat Med. Immunol. 2004;22:817-890. 2014;20(3):272-282. Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 13 35. Meyer EH, DeKruyff RH, Umetsu DT. iNKT cells in allergic disease. Curr Top helper type 1 (Th1), Th2 and Th17 immune responses. Immunology. 2008;123(1): Microbiol Immunol. 2007;314:269-291. 45-56. 36. Vivier E, Ugolini S, Blaise D, Chabannon C, Brossay L. Targeting natural killer cells 46. Jahng AW, Maricic I, Pedersen B, et al. Activation of natural killer T cells potentiates and natural killer T cells in cancer. Nat Rev Immunol. 2012;12(4):239-252. or prevents experimental autoimmune encephalomyelitis. J Exp Med. 2001;194(12): 37. Berzins SP, Ritchie DS. Natural killer T cells: drivers or passengers in preventing 1789-1799. human disease? Nat Rev Immunol. 2014;14(7):640-646. 47. Hirahara K, Nakayama T. CD4+ T-cell subsets in inflammatory diseases: beyond the 38. Cohen NR, Brennan PJ, Shay T, et al. Shared and distinct transcriptional programs Th1/Th2 paradigm. Int Immunol. 2016;28(4):163-171. underlie the hybrid nature of iNKT cells. Nat Immunol. 2013;14(1):90-99. 48. Lublin FD, Reingold SC. Defining the clinical course of multiple sclerosis: results of an 39. Kawano T, Cui J, Koezuka Y, et al. Natural killer-like nonspecific tumor cell lysis international survey. National Multiple Sclerosis Society (USA) Advisory Committee mediated by specific ligand-activated Valpha14 NKT cells. Proc Natl Acad Sci U S A. on Clinical Trials of New Agents in Multiple Sclerosis. Neurology. 1996;46(4): 1998;95(10):5690-5693. 907-911. 40. Arase N, Arase H, Good RA, Onoe K. Contribution of host radioresistant T cells to 49. Picelli S, Faridani OR, Bjorklund AK, Winberg G, Sagasser S, Sandberg R. Full-length the clonal elimination of minor lymphocyte stimulatory-1a reactive T cells in mouse RNA-seq from single cells using Smart-seq2. Nat Protoc. 2014;9(1):171-181. bone marrow chimeras. Cell Immunol. 1994;156(1):13-23. 50. Vitting-Seerup K, Sandelin A. The landscape of isoform Switches in human Cancers. 41. Moroda T, Iiai T, Suzuki S, et al. Autologous killing by a population of intermediate Mol Cancer Res. 2017;15(9):1206-1220. T-cell receptor cells and its NK1.1+ and NK1.1- subsets, using Fas ligand/Fas mol- 51. Yu G, He QY. ReactomePA: an R/Bioconductor package for reactome pathway ecules. Immunology. 1997;91(2):219-226. analysis and visualization. Mol Biosyst. 2016;12(2):477-479. 42. Metelitsa LS, Weinberg KI, Emanuel PD, Seeger RC. Expression of CD1d by mye- 52. Wu D, Smyth GK. Camera: a competitive gene set test accounting for inter-gene lomonocytic leukemias provides a target for cytotoxic NKT cells. Leukemia. 2003; correlation. Nucleic Acids Res. 2012;40(17):e133. 17(6):1068-1077. 53. Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a 43. Castano AR, Tangri S, Miller JE, et al. Peptide binding and presentation by mouse knowledge-based approach for interpreting genome-wide expression profiles. Proc CD1. Science. 1995;269(5221):223-226. Natl Acad Sci U S A. 2005;102(43):15545-15550. 44. Jahng A, Maricic I, Aguilera C, Cardell S, Halder RC, Kumar V. Prevention of au- 54. Liberzon A, Subramanian A, Pinchback R, Thorvaldsdottir H, Tamayo P, Mesirov JP. toimmunity by targeting a distinct, noninvariant CD1d-reactive T cell population Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011;27(12): reactive to sulfatide. J Exp Med. 2004;199(7):947-957. 1739-1740. 45. Niemeyer M, Darmoise A, Mollenkopf HJ, et al. Natural killer T-cell character- 55. Korotkevich T, Sukhov V, Budin N, et al. Fast gene set enrichment analysis. BioRxive. ization through gene expression profiling: an account of versatility bridging T 2019. doi: 10.1101/060012. 14 Neurology: Neuroimmunology & Neuroinflammation | Volume 8, Number 6 | November 2021 Neurology.org/NN