Stromal Interaction Molecules 1 and 2 Are Key Regulators of Autoreactive Activation in Murine Autoimmune Central Nervous System Inflammation This information is current as of September 28, 2021. Michael K. Schuhmann, David Stegner, Alejandro Berna-Erro, Stefan Bittner, Attila Braun, Christoph Kleinschnitz, Guido Stoll, Heinz Wiendl, Sven G. Meuth and Bernhard Nieswandt

J Immunol 2010; 184:1536-1542; Prepublished online 18 Downloaded from December 2009; doi: 10.4049/jimmunol.0902161 http://www.jimmunol.org/content/184/3/1536 http://www.jimmunol.org/

Supplementary http://www.jimmunol.org/content/suppl/2009/12/18/jimmunol.090216 Material 1.DC1 References This article cites 27 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/184/3/1536.full#ref-list-1 by guest on September 28, 2021 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

*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 © 2010 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Stromal Interaction Molecules 1 and 2 Are Key Regulators of Autoreactive T Cell Activation in Murine Autoimmune Central Nervous System Inflammation

Michael K. Schuhmann,*,1 David Stegner,†,1 Alejandro Berna-Erro,† Stefan Bittner,* Attila Braun,† Christoph Kleinschnitz,* Guido Stoll,* Heinz Wiendl,* Sven G. Meuth,* and Bernhard Nieswandt†

Calcium (Ca2+) signaling in T lymphocytes is essential for a variety of functions, including the regulation of differentiation, transcription, and effector functions. A major Ca2+ entry pathway in nonexcitable cells, including T cells, is store-operated Ca2+ entry (SOCE), wherein depletion of intracellular Ca2+ stores upon receptor stimulation causes subsequent influx of extracellular Ca2+ across the plasma membrane. Stromal interaction molecule (STIM) 1 is the Ca2+ sensor in the , which Downloaded from controls this process, whereas the other STIM isoform, STIM2, coregulates SOCE. Although the contribution of STIM molecules and SOCE to T lymphocyte function is well studied in vitro, their significance for immune processes in vivo has remained largely elusive. In this study, we studied T cell function in mice lacking STIM1 or STIM2 in a model of myelin-oligodendrocyte glycoprotein

(MOG35–55)-induced experimental autoimmune encephalomyelitis (EAE). We found that STIM1 deficiency significantly impaired the generation of neuroantigen-specific T cell responses in vivo with reduced Th1/Th17 responses, resulting in complete protection from EAE. Mice lacking STIM2 developed EAE, but the disease course was ameliorated. This was associated with a reduced clinical http://www.jimmunol.org/ peak of disease. Deficiency of STIM2 was associated with an overall reduced proliferative capacity of lymphocytes and a reduction of IFN-g/IL-17 production by neuroantigen-specific T cells. Neither STIM1 nor STIM2 deficiency altered the phenotype or function of APCs. These findings reveal a crucial role of STIM-dependent pathways for T cell function and activation under autoimmune inflammatory conditions, establishing them as attractive new molecular therapeutic targets for the treatment of inflammatory and autoimmune disorders. The Journal of Immunology, 2010, 184: 1536–1542.

cells orchestrate inflammation in multiple sclerosis, an diacylglycerol and 1,4,5-inositol trisphosphate, which triggers Ca2+ inflammatory disorder of the CNS resulting in demye- release from the endoplasmic reticulum through 1,4,5-inositol by guest on September 28, 2021 lination, oligodendrocyte loss, and axonal damage (1, 2). trisphosphate receptor channels and leads to a transient elevation T 2+ 2+ Calcium signals control proliferation, differentiation, apoptosis, of the cytosolic Ca concentration (7). This Ca store depletion is and a variety of transcriptional programs and effector functions of followed by Ca2+ entry through specialized Ca2+ release-activated T cells (3–5). Conversely, disturbed Ca2+ signaling and dysregu- calcium channels in the plasma membrane, resulting in longer- lation of the intracellular Ca2+ network in T lymphocytes have lasting (∼1 h) elevated cytosolic Ca2+ concentration, which is been associated with inflammatory and autoimmune disorders, mandatory for transcription-dependent steps of T cell activation including multiple sclerosis (6). (8, 9). This so-called store-operated calcium entry (SOCE) occurs Signaling cascades after TCR stimulation involve phospholipase through the four-transmembrane channel, Orai1, (6, 10, 11) and is C-mediated cleavage of phosphotidylinositol 4,5-bisphosphate into regulated by the endoplasmic reticulum-resident calcium sensor molecule stromal interaction molecule (STIM) 1, which connects *Department of Neurology and †Rudolf Virchow Center, Deutsche Forschungsge- store depletion to the opening of the channel (12). The closely meinschaft Research Center for Experimental Biomedicine, University of Wu¨rzburg, related STIM2 has been shown to regulate basal cytosolic and Wu¨rzburg, Germany store Ca2+ concentrations in different cell types (13). The recent 1 M.K.S. and D.S. contributed equally to this work. generation of mice lacking STIM molecules in T cells confirmed Received for publication July 7, 2009. Accepted for publication November 20, 2009. the essential function of STIM1 for SOCE (12, 14) and also re- This work was supported by the Rudolf Virchow Center and the Deutsche For- vealed the contribution of STIM2 to this process (15). The sig- schungsgemeinschaft (Sonderforschungsbereiche Grants 487 and 688 to B.N. and nificance of SOCE for T cell function in vivo, however, has not Grant 581 to S.G.M.). D.S. was supported by a grant of the German Excellence Initiative to the Graduate School of Life Sciences, University of Wu¨rzburg. been definitively addressed to date despite the recent description Address correspondence and reprint requests to Dr. Sven G. Meuth, Department of of human immunodeficiency syndromes associated with defi- Neurology, Josef-Schneider-Straße 11, 97080 Wu¨rzburg, Germany, or Dr. Bernhard ciencies in STIM1 function (16). A first study has revealed that Nieswandt, Rudolf-Virchow Zentrum, Josef-Schneider-Straße 2, 97080 Wu¨rzburg STIM1-deficient T cells display defects in homeostatic T cell Germany. E-mail addresses: [email protected] and bernhard.nieswandt @virchow.uni-wuerzburg.de proliferation but, despite abolished SOCE, are able to support The online version of this paper contains supplemental material. humoral immune responses after vaccination and can induce acute Abbreviations used in this paper: aGVHD, acute graft-versus-host disease; bd, below graft-versus-host disease following adoptive transfer into alloge- detection level; BM, bone marrow; DC, dendritic cell; EAE, experimental autoimmune neic hosts (12). In contrast, the significance of STIM-dependent encephalomyelitis; LFB, Luxol fast blue; MOG, myelin-oligodendrocyte glycoprotein; SOCE for inflammatory reactions in vivo has remained elusive. SOCE, store-operated Ca2+ entry; STIM, stromal interaction molecule; wt, wild-type. Inthisstudy,weinvestigatethecontributionofSTIM1andSTIM2in Copyright Ó 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00 a model of autoimmune CNS inflammation, myelin oligodendrocyte www.jimmunol.org/cgi/doi/10.4049/jimmunol.0902161 The Journal of Immunology 1537

glycoprotein (MOG35–55) -induced experimental autoimmune rat anti-mouse CD8a-PE (no. 553033), rat anti-mouse CD44-FITC (no. encephalomyelitis (EAE). We show that STIM1 deficiency abrogates 553133), rat anti-mouse CD62L-APC (no. 553152), rat anti-mouse CD11b- disease development, whereas STIM2 deficiency significantly delays PerCP (no. 350993), hamster anti-mouse CD11c-APC (no. 550261), mouse anti-mouse MCHII-FITC (no. MCA1501F; Serotec, Du¨sseldorf, Germany), onset and attenuates the clinical course of MOG EAE. These findings rat anti-mouse CD86-FITC (no. 553691), hamster anti-mouse CD80-FITC provide a strong rationale for considering STIM molecules as po- (no. 553768), and hamster anti-mouse CD40-FITC (no. 553723). 2 2 tential new pharmacological targets in T cell-mediated disorders. For functional analysis, dentritic cells (DCs) from wt or Stim2 / mice were prepared from BM cells flushed from femur and tibia bones as de- scribed before (19). BM-derived DCs were harvested on day 8, and 2 3 106 Materials and Methods cells in 1 ml splenocyte complete medium were incubated with 50 mg/ml Mice 5 MOG35–55 for 6 h. Thereafter MOG35–55-loaded DCs (1 3 10 ) were co- + 3 5 The generation of Stim12/2 bone marrow (BM) chimeric mice has been cultured with MACS isolated CD4 T cells (5 10 ) (Miltenyi Biotec, described previously (17). Stim12/2 BM chimeric mice or equally treated Bergisch Gladbach, Germany) for 3 d, and supernatants were assessed for wild-type (wt) mice were used 11–20 wk after transplantation. The gener- IFN-g and IL-17 levels by ELISA. ation of Stim22/2 mice has been reported recently (18). Genotypes were determined by PCR analysis using the following primer pairs: Stim2 wt Blood values allele forward primer 59-CCCATATGTAGATGTGTTCAG-39; reverse At the end of the EAE experiment (day 50), whole blood was taken from primer 59-GAGTGTTGTTC-CCTTCACAT-39; Stim2 knockout allele for- three animals of each of the four experimental groups. Blood was centri- ward primer 59-TTATCGATGAGCGTGGTGGTTATGC-39; reverse primer fuged, and sera were analyzed for aspartate aminotransferase (GOT), al- 59-GCGCGTACATCGGGCAAATAATATC-39. All animal experiments anine transaminase (GPT), glucose, and urea by spectrophotometric assays were approved by local authorities and conducted according to the German (Cobas Integra 800 auto-analyzer; Roche, Mannheim, Germany). law of animal protection. Downloaded from Induction and evaluation of EAE Immunohistochemistry EAE was induced by immunization of female Stim12/2 BM chimeras, Paraformaldehyde-fixed spinal cord cryosections were stained with pri- Stim22/2 mice, and corresponding wt littermates with 200 mg myelin oli- mary Ab (rat anti-CD3, 1:100, Serotec), secondary Ab (Cy3-labeled goat anti-rat 1:300; Dianova, Hamburg, Germany) and 0.5 mg/ml DAPI (Merck, godendrocyte glycoprotein (MOG35–55) peptide (Biotrend, Cologne, Ger- many). CFA was supplemented with MOG to obtain a 1 mg/ml emulsion Darmstadt, Germany). Pictures were collected by immunofluorescence and 23 100 ml were injected s.c. at two different sites of the flank of deeply microscopy (Axiophot; Zeiss, Jena, Germany). H&E and Luxol fast blue anesthesized mice. Pertussis toxin was injected on the day of immunization (LFB) stainings were done according to standard procedures. http://www.jimmunol.org/ and 2 d later (400 ng, Alexis, San Diego, CA). All animals were kept under 2+ standard conditions and had access to water and food ad libidum. Ca measurements in primary T cells The clinical course of EAE was monitored using following score system: CD4+ T cells were loaded with 5 mM Fura-2/AM (Invitrogen, Karlsruhe, grade 0, no abnormality; grade 1, limp tail tip; grade 2, limp tail; grade 3, Germany) in the presence of 0.2 mg/ml Pluronic F-127 (Invitrogen, moderate hind limb weakness; grade 4, complete hind limb weakness; grade Karlsruhe, Germany) for 30 min at 37˚C. After loading, T cells were washed 5, mild paraparesis; grade 6, paraparesis; grade 7, heavy paraparesis or twice and resuspended in calcium-free Tyrode’s-HEPES buffer (134 mM paraplegie; grade 8, tetraparesis; grade 9, quadriplegia or premoribund state; NaCl, 0.34 mM Na2HPO4, 2.9 mM KCl, 12 mM NaHCO3, 20 mM HEPES grade 10, death. [N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid], pH 7.0) contain- Western blot analysis of STIM expression in lymph node cells ing 5 mM glucose, 0.35% BSA, and 2 mM CaCl2). Stirred T cells were

activated after 50 s equilibration time with 5 mM (Invitrogen) by guest on September 28, 2021 Whole lysates were separated by SDS-PAGE gel and blotted onto or, for TCR stimulation, cross-linking of anti-CD3e Abs (T cells were a polyvinylidene difluoride membrane. Anti-STIM1 Ab from Abnova preincubated for 15 min with 10 mg/ml anti-CD3e Abs) with 10 mg/ml (H00006786-M01; Abnova, Taipei, Taiwan), and anti-STIM2 (CT) Ab goat anti-hamster IgG Ab (Jackson ImmunoResearch, West Grove, PA) from ProSci (catalog no. 4123; ProSci, Poway, CA) were used. STIM was used, and fluorescence was measured with a fluorimeter (LS 55; expression was normalized according to a-tubulin expression using rat PerkinElmer). Excitation was alternated between 340 and 380 nm, and anti-mouse a-tubulin Ab (MAB1864; Chemicon, Hofheim, Germany). emission was measured at 509 nm. Each measurement was calibrated using Triton X-100 and EGTA. Proliferation assay—[3H]thymidine uptake Splenocytes were isolated by standard procedures. Inbrief, 1 3105 splenocytes Statistical analysis 2/2 2/2 from wt, Stim1 chimeric, and Stim2 mice were cultured for 3 d after All results are presented as mean 6 SEM. Statistical analysis was per- isolation and stimulated with CD3/CD28 beads (cell to bead ratio 2:1; Dynal 3 formed using a modified Student t test for parametric data (20) or a Mann- Biotech, Hamburg, Germany). A total of 1 mCi of [ H]thymidine (Amersham, Whitney U test for nonparametric datasets. p values , 0.05 were con- Piscataway,NJ)wasadded for the final 14 h, and radioactivity wasmeasuredon sidered statistically significant. pp , 0.05; ppp , 0.01; pppp , 0.001. a b-scintillation counter (TopCount NXT; PerkinElmer, Rodgau-Ju¨gesheim, Germany). Experiments were performed in quintuplicates. Results Cytokine production To assess the role of the two STIM isoforms in EAE, we analyzed 2 2 2 2 Splenocytes from wt, Stim1 / chimeric, and Stim2 / mice were isolated mice lacking either STIM1 or STIM2. Because STIM1 deficiency is 3 6 from naive mice and at the disease maximum (day 15), and 3 10 sple- associated with ∼70% perinatal lethality (17), wt mice were trans- nocytes in 1 ml medium were plated for 48 h in DMEM containing 10 mM 2 2 planted with Stim1 / BM and analyzed after 11 wk. In contrast, HEPES, 25 mg/ml gentamicin, 50 mM mercaptoethanol, 5% FCS, 2 mM 2/2 glutamine, and 1% nonessential amino acids (Cambrex, Verviers, Belgium). Stim2 mice were born apparently healthy, developed normally to Splenocytes were either left unstimulated or were stimulated with MOG adulthood, and were fertile. Starting at the age of 8 wk after birth, peptide (10 mg/ml) or CD3/CD28 beads, and supernatants were assessed for however, sudden death of Stim22/2 mice was observed, and only IFN-g, IL-2, IL-4, IL-5, IL-10, and IL-17 protein levels by ELISA (R&D 10% of the mice reached the age of 7 mo (18). Spontaneous death of Systems, Wiesbaden, Germany) or by cytokine bead array (Bender Med- 2/2 Systems, Vienna, Austria) according to the manufacturer’s instructions. For Stim2 mice has also recently been reported by others (15), al- ELISPOT assays, 1 3 105 splenocytes per well were cultured with 20 mg/ml though in this study, mice died by the age of 4 to 5 wk most likely due MOG peptide in 96-well plates according to the manufacturer’s instructions to variability in the genetic background of both strains. The absence (MABTECH, Hamburg, Germany). Spots were evaluated by CTL Europe of the targeted STIM isoforms in lymph nodes was confirmed by (Bonn, Germany). Western blot analysis (Fig. 1B, inset). In agreement with previous Immunotyping and analysis of Ag presentation reports (12, 15), SOCE elicited by the sarcoplasmic/endoplasmic reticulum calcium ATPase inhibitor thapsigargin or anti-CD3 Analysis of T cell subtype distribution and APCs by flow cytometry was done 2/2 as previously described (19) using appropriate Abs or isotype controls (all by stimulation was virtually abolished in Stim1 T cells and signif- 2 2 BD Biosciences, San Jose, CA): rat anti-mouse CD4-PerCP (no. 553052), icantly reduced in Stim2 / T cells (Supplemental Fig. 1). 1538 STIM1 AND STIM2 IN EAE

A B

C D Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 1. Mice lacking STIM1 or STIM2 show a significantly altered disease course and histopathology after EAE induction. A and B, EAE was induced by MOG immunization (200 mg) in Stim12/2 chimeric (A,grayline,n =9),Stim22/2 mice (B, gray line, n = 10) and wt littermates (black line, n = 9/10). Inset: Western blot analyses of lymph node lysates from 5-wk-old micewith the indicated genotypes (1, wt; 2, Stim22/2;3,Stim12/2). C, H&E staining of spinal cord sections from wt (n =3)andStim12/2 chimeric mice (upper panel, n = 3; at day 15 and 50); white arrows indicate the lesions. Bar graph representation of inflammatory lesion load (number of lesions/250 randomly chosen sections) in wt and Stim12/2 chimeric animals at day 15 and 50. Scale bar (100 mm) accounts for all images. D,Im- munohistochemistry of spinal cord lesions revealed different lesion patterns (marked with arrows) in Stim22/2 mice (n = 4) compared with wt (n = 4) controls. H&E staining indicates a comparable amount of infiltrating cells in lesions from wt animals and Stim22/2 mice. LFB showed no significant differences in demyelination between wt and Stim22/2 mice. CD3 indicates the number of infiltrating T lymphocytes. Scale bar (100 mm) accounts for all images. pp , 0.05; pppp , 0.001.

EAE is significantly ameliorated in Stim22/2 mice and 5.9 6 0.8; Fig. 1A, Table I). Immunohistopathological analyses abrogated in the absence of STIM1 revealed no inflammatory lesions in the spinal cord of the mutant In a first set of experiments, we examined the role of STIM1 in chimeras, whereas profound infiltrates and demyelination were seen 6 2/2 , MOG35–55 peptide-induced EAE (Fig. 1A,1C). Although eight of in wt littermates (wt: 80.4 0.02%; Stim1 : 0%; p 0.001; n =4) nine Stim12/2 chimeric animals (∼90%) were resistant to EAE, as identified by H&E and LFB staining for demyelination (Fig. 1C). only one animal showed weak disease signs and was therefore ex- A different picture emerged when Stim22/2 mice were subjected cluded from further analysis (score at disease maximum for wt mice: to the MOG EAE model. These animals displayed a significantly The Journal of Immunology 1539

Table I. Summary of EAE incidence, day of disease onset, maximum clinical score, and EAE score at day 30 and day 50

Mouse genotype Immunization Incidence Day of Disease Onset Score at Disease Maximum Score at Day 30 Score at Day 50 wt (Stim1) 200 mg MOG 9/9 13.5 6 1.5 5.9 6 0.8 5.1 6 0.8 5.9 6 0.8 Stim12/2 200 mg MOG 1/9 n.a. n.a. n.a. n.a. wt (Stim2) 200 mg MOG 9/10 13.8 6 0.8 6.4 6 0.8 4.6 6 0.8 4.8 6 1.3 Stim22/2 200 mg MOG 10/10 15.2 6 0.6* 4.6 6 1.2** 5.0 6 0.3*** 4.6 6 0.8**** pp . 0.05; ppp , 0.001; pppp = 0.68; ppppp = 0.47 n.a., not applicable. reduced disease maximum (wt: day 13.8; score: 6.4 6 0.8; Stim22/2: mice compared with wt littermates by flow cytometry using CD4, day 15.2; score: 4.6 6 1.2; p , 0.05), whereas disease incidence, CD8, CD44, and CD62L (21) as markers (Fig. 2A,2B). Naive T cells onset, overall disease course, and disease scores on day 50 were express a CD62LhighCD44low phenotype, whereas memory T cells 2/2 unaltered (Fig. 1B, Table I). Because Stim2 mice showed sig- exhibit a CD62LlowCD44high phenotype. Analysis of CD62L and nificant clinical differences only at the disease maximum, we fo- CD44 expression in naive splenocytes showed an increase in acti- + 2/2 cused on the lesion pattern in the spinal cord at this time point. vated CD4 CD62LlowCD44high T cells in Stim1 chimeric mice 2/2 Stim2 mice showed a nonsignificant tendency to reduced de- (wt: 30.9 6 3.46, n =3;Stim12/2: 58.9 6 5.47, n =3,p , 0.05; Fig. 6 2/2 6 myelination (wt: 1.25 0.18%, Stim2 : 0.97 0.18%, p = 0.295; 2A; Supplemental Table II) as described previously (12). To address n = 3 each), but a comparable number of infiltrating T cells (Fig. 1D). the global functionality of knockout and wt immune cells, we Downloaded from Of note, routine laboratory parameters including GOT, GPT, stimulated isolated murine splenocytes with CD3/CD28 beads, glucose, and urea were evaluated at the end of the experiment and 2/2 2/2 a widely accepted method for polyclonal stimulation of lympho- showed no differences between Stim1 chimeric, Stim2 , and 2/2 cytes. Naive Stim1 chimeric mice showed a significantly reduced control animals (Supplemental Table I). secretion of the Th1 cytokines IFN-g (wt: 15,869 6 500 pg/ml, n =8; Absence of STIM1 or STIM2 influences cytokine secretion and Stim12/2: 7,756 6 553 pg/ml, n =9;p , 0.01) and IL-2 (wt: 8,756 6 proliferation under primary stimulation conditions 612 pg/ml, n =8;Stim12/2: 4,150 6 803 pg/ml, n =8;p , 0.01, Fig. http://www.jimmunol.org/ Based on the above results, we assessed the influence of STIM de- 2C). No change was observed for IL-17 (wt: 7,333 6 1464 pg/ml, 2 2 ficiency on several immunological parameters. First, we examined T n =8;Stim1 / : 8,231 6 688 pg/ml, n =9;p . 0.05) and IL-10, lymphocyte subset distribution in Stim12/2 chimeras and Stim22/2 whereas Stim12/2 cells showed significantly higher levels of the

A B by guest on September 28, 2021

FIGURE 2. Immunophenotyping, splenocyte cyto- kine production, and T cell proliferation in naive 2/2 2/2 D Stim1 chimeric, Stim2 , and wt mice. A and B, C Immunofluorescent analysis of naive wt, Stim12/2 chimeric, and Stim22/2 (n = 4) animals concerning the cell surface markers CD4+, CD8+, CD44, and CD62L. C and D, IFN-g, IL-2, IL-4, IL-5, IL-10, and IL-17 pro- duction upon CD3/CD28 bead (48 h) stimulation of splenocytes from control animals, Stim12/2 chimeric, and Stim22/2 mice (n = 8). E and F, Proliferation of splenocytes from wt (black column, n = 5), Stim12/2 chimeric (E, gray column, n = 5), and Stim22/2 mice (F, gray column, n = 5) assessed by thymidine uptake upon CD3/CD28 bead stimulation. pp , 0.05; ppp , 0.01. E F 1540 STIM1 AND STIM2 IN EAE

Th2 cytokines IL-4 and IL-5. This shift in the cytokine pattern was Ag-specific IFN-g production was detectable after MOG recall, accompanied by an unaltered proliferation rate of naive splenocytes indicating that the lack of this Ca2+ sensor almost completely (wt: 10,338 6 1,300 cpm, n =5;Stim12/2: 9,373 6 1,263 cpm, n = abolishes the generation of Ag-specific T cells. Comparable re- 5; p . 0.05, Fig. 2E). sults were found for IL-2 and IL-17 (wt: 2,104 6 63 pg/ml, n =5; No differences were observed between Stim22/2 and control Stim12/2: 698 6 100 pg/ml, n =5;p , 0.01). The latter stands in mice (Fig. 2B), both showing a large naive population interesting contrast to the unaltered activation-induced production (CD62LhighCD44low). STIM2-deficient lymphocytes displayed no of IL-17 by naive CD3/CD28 activated cells, emphasizing a role significant differences in the production of IFN-g and IL-2, but the of STIM1 for the generation of Ag-specific T cell responses in secretion of IL-17 was markedly reduced compared with wt cells vivo (Fig. 3C). Stim22/2 cells showed a significantly reduced Ag- (wt: 10,042 6 3,187 pg/ml; Stim22/2: 3,805 6 905 pg/ml, n =8; specific production of IFN-g and IL-17 and no significant change p , 0.05; Fig. 2D). Additionally, IL-5 levels were increased sig- for IL-2 production (Fig. 3D). In parallel to the deviated cytokine nificantly in Stim22/2 lymphocytes, whereas no difference was pattern, MOG-specific proliferation was reduced for both Stim12/2 found for IL-4 and IL-10. This was accompanied by a significantly and Stim22/2 splenocytes (Fig. 3C,3D, insets). reduced T cell proliferation rate (wt: 8,964 6 767 cpm, n =5; IL-17 ELISPOT assays were performed to investigate the Th17 Stim22/2: 5,438 6 1,105 cpm, n =5;p , 0.05, Fig. 2F). The re- response of Stim22/2 cells in EAE in more detail. Although Stim22/2 duced ability of Stim22/2 cells to produce IL-17 (Fig. 2D) may be cells showed reduced levels of IL-17 production, the general number particularly relevant in the setting of EAE (Fig. 1B), as Th17 cells of IL-17–producing cells was not significantly altered (p = 0.61; Fig. are considered key pathogenic players in MOG-induced EAE. 3E). However, the IL-17 output on the cellular level was significantly

reduced in the absence of STIM2 (230%; p = 0.004; Fig. 3F). Downloaded from Lack of STIM1 or STIM2 affects the generation of Altogether, these results indicate that STIM-deficient mice neuroantigen-specific T cell responses undergo important changes in their cytokine phenotype upon Peripheral T lymphocyte subset distribution in EAE-induced immunization. Mice lacking STIM1 are virtually unable to raise 2 2 2 2 Stim1 / chimeric and Stim2 / mice compared with wt litter- a relevant Ag-specific T cell response against MOG. Both Th1 mates was assessed by flow cytometry using CD4, CD8, CD44, cytokines (IFN-g, IL-2) and, most importantly, IL-17 are strongly

and CD62L as markers. No significant differences were observed diminished, reflecting the resistance to EAE. In contrast, although http://www.jimmunol.org/ (Fig. 3A,3B). We next assessed the quantity and quality of the showing the same tendencies on the cytokine level, Stim22/2 mice T cell response following MOG immunization in mice. Spleno- are able to generate more robust neuroantigen-specific T cell re- 2 2 cytes were isolated from Stim1 / chimeric mice at disease sponses. This response, however, is attenuated because individual maximum and stimulated with MOG. In this study, virtually no levels of IL-17 (and IFN-g) are significantly reduced.

A B by guest on September 28, 2021

FIGURE 3. Immunophenotyping and assessment of neuroantigen-specific T cells in Stim12/2 chimeric, Stim22/2, and wt mice at the maximum of EAE. A and B, Flow cytometry analysis of wt (n = 3), Stim12/2 CD chimeric (n = 3), and Stim22/2 (n = 3) animal T cell subsets using CD4, CD8, CD44, and CD62L as cell surface markers at EAE disease maximum. C and D, MOG recall assays of splenocytes from control (n = 5), Stim12/2 chimeric (n = 5), and Stim22/2 mice were analyzed for IL-2, IFN-g, and IL-17 cytokine levels. Insets show proliferation of T lymphocytes from wt (black columns, n = 5), Stim12/2 chimeric, and Stim22/2 mice (gray columns, n = 5) upon MOG restimulation. E and F, MOG recall IL-17 ELISPOT assay of splenocytes from control (n =4) 2 2 F and Stim2 / (n = 5) mice at disease maximum. Data E are presented as spot counts (E) and mean spot size (F) normalized on wt values. pp , 0.05; ppp , 0.01. The Journal of Immunology 1541

The lack of STIM1 or STIM2 has no impact on the phenotype Although the effect of STIM2 deficiency on T cell responses was and function of APCs moderate compared with STIM1 deficiency, the lack of STIM2 Peripheral APCs, identified by the expression of CD11b and CD11c influenced basal T cell-derived IL-17 production and the quanti- from Stim12/2 chimeric, Stim22/2, and control mice did not differ tative release of cytokines by individual Ag-specific T cells. in markers of maturation and function (MHCII, CD80, CD86, We have previously shown that serum cytokine levels in naive wt 2/2 CD40; Fig. 4A: Stim12/2,4B: Stim22/2). To investigate the po- and Stim1 mice are not significantly different (12), whereas tential influence of STIM deficiency on Ag presentation, we in- polyclonal in vitro stimulation (CD3/CD28), which results in cubated cultured DCs from Stim12/2 chimeric, Stim22/2,or much higher cytokine levels, revealed differences in T cell ef- 2/2 control animals with MOG peptide (10 mg/ml) and CD4+ T fector function caused by STIM deficiency. Although Stim1 lymphocytes from Stim12/2 chimeric, Stim22/2, or wt mice (Fig. chimeric mice are characterized by a lymphoproliferative pheno- 4C–F). This criss-cross approach clearly showed that neither type despite normal thymic T cell maturation, the lack of SOCE STIM1 nor STIM2 deficiency had an effect on DC function, results in significantly reduced cytokine production when T cells whereas significantly lower IFN-g and IL-17 production was de- are stimulated with CD3/CD28 beads in vitro and an impaired tected in STIM1- and STIM2-deficient T cells compared with the cytokine production and proliferation when Ag-specific T cells are wt controls. restimulated with their cognate Ag (Figs. 2C,2D,3C,3D). In our hands, Stim12/2 cells display unaltered proliferation capacities Discussion when stimulated maximally both via the TCR and CD28. It seems plausible that the effect of STIM deficiency depends on the

We investigated the role of STIM1 and STIM2 for T cell function Downloaded from and the generation of Ag-specific immune responses in a model of strength of the activation signal and that it may be overcome by 2+ inflammation. We demonstrate that lack of STIM1-dependent the Ca -independent costimulatory CD28 pathway (22). 2+ SOCE severely impairs the generation of Ag-specific T cell Accordingly, despite the severe impairment of Ca signaling, 2 2 responses in a MOG-induced EAE, resulting in resistance to Stim1 / immune cells can generate acute graft-versus-host dis- disease development. The influence of STIM1 deficiency for T cell ease (aGVHD) (12). aGVHD is primarily due to recipient-specific responses was demonstrated both on a qualitative as well as on cytotoxic donor T lymphocytes, and the underlying cellular http://www.jimmunol.org/ a quantitative level. Lack of STIM2 resulted in a decreased, but not mechanisms differ from chronic GVHD, which has been shown to absent, SOCE in T cells and a significantly attenuated EAE course. be caused by autoreactive helper T lymphocytes (23). Thus, acute

A B by guest on September 28, 2021

C D

FIGURE 4. Characterization of APCs in wt, Stim12/2 chimeric, and Stim22/2 mice after EAE induction (day 15). A and B, Analysis of APCs expressing CD11b+/CD11c+ by flow cytometry for the surface markers MHCII, CD80, CD86, and CD40 revealed no significant differences between naive control and knockout mice (n = 3). C–F, Criss-cross experiment of DCs and T lymphocytes from wt (n = 4), Stim12/2 chimeric (n = 4), and Stim22/2 (n = 4) animals and vice versa. Co- cultures were assessed for IL-17 and IFN-g pro- duction after MOG stimulation. pp , 0.05; ppp , 0.01. bd, below detection level. E F 1542 STIM1 AND STIM2 IN EAE and chronic non-MHC GVHD are caused by two different cellular novel interesting molecular target structures for the treatment of T mechanisms stimulated by different Ags (i.e., non–H-2 Ags in cell-mediated autoimmune disorders. aGVHD and autologous Ia Ag in chronic GVHD). The EAE model is probably better comparable with chronic GVHD because Disclosures of the relevance of autoreactive T cells resulting in chronic The authors have no financial conflicts of interest. autoimmune-like symptoms (23, 24). We therefore hypothesize that the capability of T cells to raise an adaptive immune response in the absence of STIM1 under certain conditions depends on References the strength of the MHC-dependent TCR signal (possibly in 1. Sospedra, M., and R. Martin. 2005. Immunology of multiple sclerosis. Annu. combination with the comodulatory influence of accessory mol- Rev. Immunol. 23: 683–747. 2. Frohman, E. M., M. K. Racke, and C. S. Raine. 2006. Multiple sclerosis—the ecules and cytokine signals). Upon maximal in vitro stimulation plaque and its pathogenesis. N. Engl. J. Med. 354: 942–955. with CD3/CD28 beads, we found an intact IL-17 response of 3. Gomes, B., M. Savignac, M. Moreau, C. Leclerc, P. Lory, J. C. Guery, and Stim12/2 immune cells (Fig. 2C), whereas T cell cytokine pro- L. Pelletier. 2004. Lymphocyte calcium signaling involves dihydropyridine- sensitive L-type calcium channels: facts and controversies. Crit. Rev. Immunol. duction was clearly impaired in Ag-specific T cells under recall 24: 425–447. conditions (Fig. 3C). This residual competence of strongly optimal 4. Quintana, A., D. Griesemer, E. C. Schwarz, and M. Hoth. 2005. Calcium- activated Stim12/2 immune cells might explain their relatively dependent activation of T-lymphocytes. Pflugers Arch. 450: 1–12. 5. Savignac, M., B. Mellstro¨m, and J. R. Naranjo. 2007. Calcium-dependent normal response in the aGVHD model. transcription of cytokine in T lymphocytes. Pflugers Arch. 454: 523–533. 2 2 Stim1 / chimeric mice show severely impaired generation of 6. Feske, S. 2007. Calcium signalling in lymphocyte activation and disease. Nat. Rev. Immunol. 7: 690–702. Downloaded from Ag-specific responses in vivo with compromised functions of both 7. Panyi, G., G. Va´mosi, A. Bodna´r, R. Ga´spa´r, and S. Damjanovich. 2004. Th1 and Th17 cells. The generation of sizeable MOG-specific T Looking through ion channels: recharged concepts in T-cell signaling. Trends cell responses is strongly diminished in Stim12/2 chimeric mice. Immunol. 25: 565–569. 8. Prakriya, M., and R. S. Lewis. 2003. CRAC channels: activation, permeation, In contrast, the consequences of STIM2 deficiency for the gen- and the search for a molecular identity. Cell Calcium 33: 311–321. eration of adaptive immune responses are rather mild, as revealed 9. Negulescu, P. A., N. Shastri, and M. D. Cahalan. 1994. Intracellular calcium by the robust neuroantigen-specific T cell responses in Stim22/2 dependence of gene expression in single T lymphocytes. Proc. Natl. Acad. Sci. USA 91: 2873–2877. mice (Fig. 1B). However, the strength (or quality) of this MOG- 10. Penna, A., A. Demuro, A. V. Yeromin, S. L. Zhang, O. Safrina, I. Parker, and http://www.jimmunol.org/ specific immune response is strongly diminished in comparison M. D. Cahalan. 2008. The CRAC channel consists of a tetramer formed by Stim- induced dimerization of Orai dimers. Nature 456: 116–120. with wt at disease maximum (Fig. 3D,3F) underlining the im- 11. Firth, A. L. 2009. Fine tuning I(CRAC): the interactions of STIM-1 and Orai. paired immune activation upon immunization at this time point. J. Physiol. 587: 15–16. Furthermore, we found significantly reduced IL-17 production of 12. Beyersdorf, N., A. Braun, T. Vo¨gtle, D. Varga-Szabo, R. R. Galdos, S. Kissler, 2/2 T. Kerkau, and B. Nieswandt. 2009. STIM1-independent T cell development and Stim2 cells compared with wt control in response to both effector function in vivo. J. Immunol. 182: 3390–3397. polyclonal stimulation with CD3/CD28 beads as well as Ag re- 13. Brandman, O., J. Liou, W. S. Park, and T. Meyer. 2007. STIM2 is a feedback stimulation. In contrast, STIM2 deficiency had less influence on regulator that stabilizes basal cytosolic and endoplasmic reticulum Ca2+ levels. Cell 131: 1327–1339. Th1 cytokines, albeit significantly lower IFN-g levels were ob- 14. Oh-hora, M., and A. Rao. 2008. Calcium signaling in lymphocytes. Curr. Opin. served at disease maximum. It is generally accepted that both Th1 Immunol. 20: 250–258. by guest on September 28, 2021 15. Oh-Hora, M., M. Yamashita, P. G. Hogan, S. Sharma, E. Lamperti, W. Chung, and Th17 cells contribute to the development of EAE with Th17 M. Prakriya, S. Feske, and A. Rao. 2008. Dual functions for the endoplasmic cells being the first effector cells that invade the CNS during au- reticulum calcium sensors STIM1 and STIM2 in T cell activation and tolerance. toimmune inflammation (25, 26). This might explain why the Nat. Immunol. 9: 432–443. 16. Picard, C., C. A. McCarl, A. Papolos, S. Khalil, K. Lu¨thy, C. Hivroz, F. LeDeist, effect of STIM2 deficiency in the model of MOG-induced CNS F. Rieux-Laucat, G. Rechavi, A. Rao, et al. 2009. STIM1 mutation associated inflammation was evident, especially at early stages of the disease. with a syndrome of immunodeficiency and autoimmunity. N. Engl. J. Med. 360: Thus, our data suggest that STIM2 deficiency affects the genera- 1971–1980. 17. Varga-Szabo, D., A. Braun, C. Kleinschnitz, M. Bender, I. Pleines, M. Pham, tion of adaptive T cell responses in vivo, which is most likely T. Renne´, G. Stoll, and B. Nieswandt. 2008. The calcium sensor STIM1 is an based on changes in cytokine patterns and levels. It is currently essential mediator of arterial thrombosis and ischemic brain infarction. J. Exp. Med. 205: 1583–1591. not clear how STIM2 function is mechanistically linked to the 18. Berna-Erro, A., A. Braun, R. Kraft, C. Kleinschnitz, M. K. Schuhmann, ability of T cells to secrete IL-17, but it is known that IL-17 ex- D. Stegner, T. Wultsch, J. Eilers, S. G. Meuth, G. Stoll, and B. Nieswandt. 2009. pression depends especially on the calcium-regulated transcription STIM2 regulates capacitive Ca2+ entry in and plays a key role in hypoxic neuronal cell death. Sci. Signal. 2: ra67. factor NFAT (27), whereas other cytokines might be less sensitive 19. Melzer, N., S. G. Meuth, D. Torres-Salazar, S. Bittner, A. L. Zozulya, to subtle changes in calcium levels because of redundant pathways C. Weidenfeller, A. Kotsiari, M. Stangel, C. Fahlke, and H. Wiendl. 2008. A like MAPK or NF-kB. Of note, our data do not support a major b-lactam antibiotic dampens excitotoxic inflammatory CNS damage in a mouse model of multiple sclerosis. PLoS One 3: e3149. role of STIM1 or STIM2 for the phenotype and Ag-presenting 20. Meuth, S. G., S. Bittner, P. Meuth, O. J. Simon, T. Budde, and H. Wiendl. 2008. function of APCs (Fig. 4). Thus, the observed effects in vivo are TWIK-related acid-sensitive K+ channel 1 (TASK1) and TASK3 critically in- fluence T lymphocyte effector functions. J. Biol. Chem. 283: 14559–14570. most likely explained by a specific effect of STIM deficiency on T 21. Sallusto, F., J. Geginat, and A. Lanzavecchia. 2004. Central memory and effector cell function. It remains to be demonstrated, however, whether memory T cell subsets: function, generation, and maintenance. Annu. Rev. Im- STIM molecules are involved in other mechanisms of T cell munol. 22: 745–763. 22. Rudd, C. E., and M. Raab. 2003. Independent CD28 signaling via VAVand SLP- function beyond proliferation or cytokine secretion, such as mi- 76: a model for in trans costimulation. Immunol. Rev. 192: 32–41. gration of T cell subtypes to the CNS. The use of mice with T cell- 23. Parkman, R. 1986. Clonal analysis of murine graft-vs-host disease. I. Phenotypic specific STIM deficiencies may provide further detailed insights and functional analysis of T lymphocyte clones. J. Immunol. 136: 3543–3548. 24. Iwasaki, T. 2004. Recent advances in the treatment of graft-versus-host disease. into these open questions. Clin. Med. Res. 2: 243–252. Taken together, we have shown that the lack of STIM1 in immune 25. Korn, T., A. C. Anderson, E. Bettelli, and M. Oukka. 2007. The dynamics of effector T cells and Foxp3+ regulatory T cells in the promotion and regulation of cells protects mice from the development of neuroantigen-specific T autoimmune encephalomyelitis. J. Neuroimmunol. 191: 51–60. cell responses and therefore renders them resistant to EAE. In 26. Korn, T., M. Oukka, V. Kuchroo, and E. Bettelli. 2007. Th17 cells: effector contrast, the lack of STIM2 leads to a reduced peak severity of EAE, T cells with inflammatory properties. Semin. Immunol. 19: 362–371. 27. Liu, X. K., X. Lin, and S. L. Gaffen. 2004. Crucial role for nuclear factor of possibly due to an effect of STIM2 on the individual IL-17 pro- activated T cells in T cell receptor-mediated regulation of human interleukin-17. duction by T cells. Therefore, STIM-regulated pathways represent J. Biol. Chem. 279: 52762–52771.