Store-Operated Ca²⁺ Entry through ORAI1 Is Critical for T Cell-Mediated Autoimmunity and Allograft Rejection

Activation of T cells in response to TCR stimulation requires Ca²⁺ influx, which mediates important Ca²⁺-dependent signaling events (1). Arguably the most important Ca²⁺ influx pathway in T cells is store-operated Ca²⁺ entry (SOCE), so named because it is activated by depletion of endoplasmic reticulum (ER) Ca²⁺ stores after activation of phospholipase C γ1 and production of inositol 1,4,5-triphosphate (InsP3). InsP3 binds to and opens InsP3 receptor ion channels allowing Ca²⁺ to efflux from the ER. Ca²⁺ release results in a transient increase in the intracellular Ca²⁺ concentration [Ca²⁺]_i and activation of Ca²⁺ release-activated Ca²⁺ (CRAC) channels in the plasma membrane (2). The CRAC channel is the prototypical SOC channel and possesses unique electrophysiological properties including high Ca²⁺ selectivity and a very low single-channel conductance (<1 pS) (3). The CRAC channel is composed of ORAI1 (or CRACM1), a tetra-spanning plasma-membrane protein that is structurally unrelated to other ion channels and that serves as the pore-forming subunit of the channel (2, 4). ORAI1-CRAC channels are activated through direct physical interaction with the ER Ca²⁺ sensors stromal interaction molecule (STIM) 1 and 2, transmembrane proteins located in the membrane of the ER (5).

CRAC channel activation and the resulting SOCE are required for various T cell functions including cytokine gene expression in vitro (1, 6, 7). In humans, the importance of ORAI1 and STIM1 in vivo is illustrated by a rare group of immunodeficiency diseases affecting patients with mutations in ORAI1 and STIM1 that abolish SOCE in T cells and other cell types (8–11). Whereas some mutations abolish ORAI1 or STIM1 protein expression, a missense mutation in ORAI1 results in a single amino acid substitution (R91W) located at the interface of the cytoplasmic N terminus with the first transmembrane domain of ORAI1. The mutant ORAI1-R91W protein is expressed but its CRAC channel function is abolished (12, 13). Clinically, all ORAI1- and STIM1-deficient patients suffer from immunodeficiency with an increased susceptibility to infections. The latter has been attributed to defects in the activation of the patients' T cells from in vitro studies (14). In addition, nonimmunological symptoms, such as congenital muscular hypotonia and anhydrotic ectodermal dysplasia, are present in ORAI1- and STIM1-deficient patients but are not life threatening (10, 11).

Mice lacking Orai1 expression have been generated and, in contrast with human patients, die neonatally even under specific pathogen-free conditions (15–18). A minority of surviving mice was severely runted indicating that ORAI1 serves critical functions outside the immune system. Hematopoietic lineage cells, such as T and B cells (16), mast cells (17), and platelets (15), isolated from surviving Orai1^−/− mice showed a defect in SOCE and impaired cell function. In one study (17), however, SOCE in ORAI1-deficient T cells was normal and T cell function only modestly impaired raising questions about the precise contribution of ORAI1 to SOCE in murine T cells. Because naive T cells also express ORAI2 and ORAI3, two highly conserved paralogues of ORAI1 (16, 17), it was suggested that these molecules, especially ORAI2, could be responsible for the residual SOCE observed in ORAI1-deficient T cells (17). It is noteworthy that both ORAI2 and ORAI3—when overexpressed together with STIM1—are able to form functional Ca²⁺ channels with properties similar to those of the native CRAC channel (19, 20), but the contribution of endogenously expressed ORAI2 and ORAI3 to SOCE and CRAC channel function has yet to be established.

Given the controversial role of ORAI1 for SOCE in murine T cell and B cell function, and because a systematic analysis of ORAI1-dependent, T cell-mediated immune responses in vivo has been lacking, we generated Orai1 knock-in (Orai1^KI/KI) mice expressing a mutant ORAI1-R93W protein that is equivalent to the nonfunctional ORAI1-R91W mutant found in human SCID patients (12, 18). Orai1^KI/KI mice die neonatally, but fetal liver chimeric mice that homozygously express ORAI1-R93W in hematopoietic cells develop normally. T cells from chimeric Orai1^KI/KI mice have a profound defect in SOCE and CRAC channel function that results in severely compromised T cell function in vitro and in vivo. Orai1^KI/KI mice expressing a nonfunctional ORAI1 channel protein provide a useful tool to study the role of ORAI1 for immune responses in vivo, and their phenotype can be directly compared with that of Orai1^−/− mice and patients homozygous for the ORAI1-R91W mutation.

Gene targeting of the Orai1 gene was performed by homologous recombination in B6/3 embryonic stem cells derived from C57BL/6 mice (TaconicArtemis GmbH, Koln, Germany) as described (18). Briefly, Orai1^KI/KI mice were generated by replacing codon 93 (CGG encoding R93) in exon 1 of the Orai1 gene with TGG (encoding W93). Chimeric mice with targeted Orai1 alleles were generated by blastocyst injection of heterozygous Orai1^neo/+ embryonic stem cell clones, the Neo cassette was deleted by Cre expression under the control of the testis-specific ACE promoter during spermatogenesis, and founder Orai1^neo/+ chimeric mice were bred to C57BL/6 mice to establish Orai1^+/KI mice. Successful gene targeting was confirmed by two PCR approaches, detecting (1) the remaining loxP site in the Orai1 locus after excision of the Neo^r/ACE-Cre cassette and (2) the mutated codon 93, which creates a recognition site for the restriction endonuclease PvuII (CAGCTG). Primers used for detection of the loxP site were as follows: forward 5′-ATTTCCCAATACGTTCCACCTCCC-3′; reverse 5′-TCGTACCACCTTCTTGGGACTTGA-3′. Primers used for PCR amplification of Orai1 exon 1 (followed by PvuII digest) were as follows: forward 5′-TGGATCGGCCAGAGTTACTCC-3′; reverse 5′-GATTACATGCAGGGCAGACTTCTTA-3′. Stim1^fl/fl Cd4-Cre mice were generated as described (21).

To generate fetal liver chimeras (FLCs), fetal liver cells were obtained from E14.5 mouse embryos derived from matings of Orai1^KI/+ mice on the C57BL/6 background. Orai1^KI/KI and Orai1^+/+ littermate control cells were injected intravenously into sublethally irradiated (4.5 + 4.5 Gy) Rag2^−/−, cγ^−/− mice (Taconic, Hudson, NY) or Rag2^−/− mice (Taconic). Reconstituted mice were sacrificed 5–6 wk after transplantation to harvest blood, spleens, lymph nodes, and thymi for isolation of lymphocytes. To generate homozygous Orai1^KI/KI mice on a mixed genetic background, heterozygous Orai1^KI/+ mice on the C57BL/6 background were mated for 3–4 generations with mice of the outbred ICR strain (Taconic), after which Orai1^KI/+ mice were intercrossed to generate Orai1^KI/KI N3f1 and Orai1^KI/KI N4f1 mice, respectively. All mice were maintained in specific pathogen-free barrier facilities at Harvard Medical School and New York University Langone Medical Center and were used in accordance with protocols approved by the institutional animal care and use committees at both institutions.

CD4⁺ and CD8⁺ T cells and B220⁺ B cells were isolated from thymi, spleens, and lymph nodes of Orai1^KI/KI and Orai1^+/+control fetal liver chimeric mice by Dynal magnetic bead separation (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. CD4⁺ and CD8⁺ T cells were differentiated into ThN, Th1, and CTLs by stimulation with plate-bound anti-CD3 and anti-CD28 Abs for 2 d, followed by culture in DMEM medium containing either 20 U/ml IL-2 alone (ThN, nonpolarizing conditions), or 20 U/ml IL-2 plus IL-12 and anti–IL-4 (Th1 conditions), or 100 U/ml IL-2 (CTL) for an additional 4 d. HEK293 cells were cultured in DMEM (Mediatech, Manassas, VA) at 37°C, 10% CO₂.

Total RNA was extracted from T cells, skeletal muscle, and colon using TRIzol reagent (Invitrogen). cDNA was synthesized from total RNA using oligo(dT) primers and the Superscript II First-Strand kit (Invitrogen). Real-time PCR was performed using an iCycler system (BioRad, Hercules, CA) and SYBR Green dye (Applied Biosystems, Foster City, CA) with the following primer pairs (sense, antisense): Orai1 (NM_175423): 5′-CCAAGCTCAAAGCTTCCAGC-3′, 5′-GCACTAAAGACGATGAGCAACC-3′; Orai2 (NM_178751): 5′-TCCTCAGACACACCAAGG-3′, 5′-CAAAAGACACCAATCATGTTCTGC-3′; Orai3 (NM_198424): 5′-GGATCCTGGGTTAAATGAGAG-3′, 5′-TTGAGGACAGTTGTGCAGAC-3′; Ppia (AK028210): 5′-AGCTCTGAGCACTGGAGAGA-3′, 5′-GCCAGGACCTGTATGCTTTA-3′; Gadp (BC083080.1): 5′-CTGGAGAAACCTGCCAAGTA-3′, 5′-TGTTGCTGTAGCCGTATTCA-3′; Atf3 (NM_007498): 5′-CCAGGTCTCTGCCTCAGAAG-3′, 5′-CATCTCCAGGGGTCTGTTGT-3′; Ddit3 (Chop10, NM_007837): 5′-CATACACCACCACACCTGAAAG-3′, 5′-CCGTTTCCTAGTTCTTCCTTGC-3′; Bip (NM_022310): 5′-GAAAGGATGGTTAATGATGCTGAG-3′, 5′-GTCTTCAATGTCCGCATCCTG-3′; Il17a (NM_010552.3): 5′-CTCCAGAAGGCCCTCAGACTAC-3′, 5′-AGCTTTCCCTCCGCATTGACACAG-3′; Ifng (NM_008337.3): 5′-GATGCATTCATGAGTATTGCCAAGT-3′, 5′-GTGGACCACTCGGATGAGCTC-3′. SYBR Green signals were captured at the end of each polymerization step (72°C), a threshold was set in the linear part of the amplification curve, and threshold cycles (C_T) for genes of interest were normalized to the C_T for Ppia (Fig. 4D) and Gadp (Fig. 6H, Supplemental Fig. 2C) housekeeping gene control obtaining Δ^C_T. Expression data were plotted as 0.5^ΔC_T or percentage of Gadp expression. The specificity of PCR amplification was verified by melt curve analysis and agarose gel electrophoresis of PCR products. The efficiency of PCR amplification was tested by serial dilutions and analyzed using iCycler software. mOrai1: 112.7 ± 12.1% (n = 6 repeat experiments); mOrai2: 96.7 ± 8.0% (n = 6); mOrai3: 121.4 ± 10.5% (n = 6); Ppia 93.3 ± 4.2% (n = 6).

For flow cytometry, single-cell suspensions prepared from thymi, lymph nodes, and spleens of Orai1^KI/KI and Orai1^+/+ control mice were preincubated with anti-CD16/32 Ab (clone 93; eBioscience, San Diego, CA) followed by incubation with the following Abs (all from eBioscience): anti-B220 (clone RA3-6B2), anti–CD-4 (clone GK1.5), anti-CD8 (clone 53-6.7), anti-CD25 (clone PC61.5), anti-CD44 (clone IM7), anti-TCRβ (clone H57-597), anti-CD69 (clone H1.2F3), anti-AA4.1 (clone AA4.1), anti-IgD (clone 11-26), anti–IL-2 (clone JES6-5H4), anti–IFN-γ (clone XMG1.2), anti–IL-4 (clone 11B11), anti–IL-10 (clone JES5-16E3), anti-granzyme B (clone 16G6), anti-Foxp3 (clone NRRF-30); anti-IgM (clone R6-60.2) Ab was purchased from BD Biosciences (San Jose, CA). For intracellular cytokine staining, ThN cells or CTLs were restimulated with 10 nM PMA (Sigma) and 1 μM ionomycin (Sigma, St. Louis, MO) for 6 h in the presence or absence of 2 μM cyclosporin A (EMD Biosciences, Gibbstown, NJ). Brefeldin A (5 μM) was added during the last 2 h of stimulation. At the end of the stimulation, cells were incubated with anti-CD16/32 for 20 min on ice to block Fc receptors followed by incubation with Abs to surface proteins for 20 min on ice. Cells were washed in FACS buffer (1% BSA in 1× PBS), fixed with 4% paraformaldehyde, and incubated with anti-cytokine Abs in permeabilization buffer (0.5% saponin in 1× PBS and 1% BSA). Fluorescence was measured using an LSRII flow cytometer (BD Biosciences), and data were analyzed with FlowJo software (Tree Star, Ashland, OR).

CD4⁺ and CD8⁺ T cells isolated from lymph nodes and spleen of Orai1^KI/KI and Orai1^+/+control mice were loaded with 4 μM CFSE directly following magnetic bead separation and stimulated with plate-bound anti-CD3 and anti-CD28 for 72 h. Alternatively, cells were first differentiated into ThN cells (see earlier) for 3 d, then loaded with CFSE on day 3 and stimulated with anti-CD3 and anti-CD28 for 72 h. Cells were stimulated for proliferation assays in the absence of exogenous IL-2. The number of cell divisions was analyzed using an LSRII flow cytometer.

Regulatory T cell (Treg)-mediated suppression was analyzed as described (21). Briefly, CD4⁺ CD25⁺ T cells were isolated from peripheral lymph nodes and spleens of wild-type and Orai1^KI/KI chimeric mice by cell sorting using an iCyt Reflection parallel cell sorter (iCyt, Champaign, IL). CD4⁺ CD25⁻ T cells were isolated from peripheral lymph nodes by negative selection using CD25 MACS MicroBead Kit (Miltenyi Biotec, Bergisch-Gladbach, Germany) followed by labeling with 2 μM CFSE (Invitrogen) at 37°C for 3 min according to the manufacturer’s instructions. CD4⁺ CD25⁻ T cells were stimulated with 0.3 μg/ml anti-CD3 (2C11) in the presence of 5 × 10⁴ T cell-depleted splenocytes (pretreated with 40 μg/ml mitomycin C [Sigma] at 37°C for 30 min) and cocultured with CD4⁺ CD25⁺ Tregs at various ratios for 72 h at 37°C in 96-well round-bottom plates.

Tail-skin grafts from BALB/c mice (H-2^d) were harvested and transplanted onto the backs of Orai1^KI/KI and Orai1^+/+control fetal liver chimeric mice (C57BL/6 background, H-2^b). Donor skin from the base of the tail of BALB/c mice was removed and cut to the size of 8 × 8 mm for each skin graft. The hair on the left side of the thorax of each recipient mouse was shaved and the epidermal and dermal layers of the skin removed to expose the underlying panniculus carnosus creating a graft bed of ∼9 × 9 mm, slightly bigger than the size of the skin graft. Skin allografts were fixed on graft-beds with nylon black monofilament (6-0; Tyco Healthcare, Norwalk, CT) at the corners. Recipients of skin allograft transplants were monitored daily for size and integrity of the skin graft. The graft was considered rejected when the size was <20% of the original transplanted skin.

Orai1^KI/KI and Orai1^+/+control mice were sensitized to hapten 5–6 wk after transfer of fetal liver cells by epicutaneous application of 200 μl 0.5% FITC (Sigma) diluted 1:1 in acetone/dibutyl phthalate (vehicle) onto the shaved abdomen or left unsensitized. Six days later, baseline thickness of both ears was measured with a dial thickness gauge (GWJ Co., Hacienda Heights, CA). Each side of the right and left ear was then treated with 10 μl 0.5% FITC (dissolved in acetone/dibutyl phthalate) and vehicle alone, respectively. Thickness of both ears was measured 24 h later in sensitized and nonsensitized animals. The change in ear thickness (ΔT) was calculated as ΔT = (ear thickness 24 h after elicitation) – (baseline ear thickness) as described (22).

CD4⁺ CD25^– CD45RB^hi T cells were isolated from peripheral lymph nodes and spleens of wild-type and Orai1^KI/KI chimeric mice by staining with anti-CD4 (clone GK1.5), anti-CD25 (clone PC61.5), and anti-CD45RB (clone C363.16A; all from eBioscience) Abs and cell sorting using an iCyt Reflection parallel cell sorter (iCyt). A >95% purity of the sorted CD4⁺CD25^–CD45RB^hi T cells was confirmed by flow cytometry using an LSRII flow cytometer, and 5 × 10⁵ cells were injected i.p. into 3- to 4-mo-old syngeneic Rag2^−/− mice. Recipient mice were weighed and assessed for disease symptoms weekly. Mice were sacrificed and organs harvested 12 wk after adoptive cell transfer.

Colons of mice were excised and flushed with PBS, cut longitudinally, and rolled as described (23). For histology, proximal and distal colon sections were fixed in 4% paraformaldehyde, paraffin-embedded, and stained with H&E using standard protocols. Colon histology was scored in a blinded fashion by two individuals using a grading system from 0 to 5 as described (24): 0, no changes; 1, minimal scattered mucosal inflammatory cell infiltrates, with or without epithelial hyperplasia; 2, mild scattered to diffuse inflammatory cell infiltrates, sometimes extending into the submucosa and associated with erosions, with minimal to mild epithelial hyperplasia and minimal to mild mucin depletion from goblet cells; 3, mild to moderate inflammatory cell infiltrates that were sometimes transmural, often associated with ulceration, with moderate epithelial hyperplasia and mucin depletion; 4, marked inflammatory cell infiltrates that were often transmural and associated with ulceration, with marked epithelial hyperplasia and mucin depletion; 5, marked transmural inflammation with severe ulceration and loss of intestinal glands.

Cytokine levels were measured in cell culture supernatants of cells isolated from mesenteric lymph nodes and spleens that were stimulated for 48 h with plate-bound anti-CD3ε Ab (5 μg/ml). Supernatants were harvested, centrifuged to remove cell debris, and stored at –20°C. IL-17A, IFN-γ, and TNF-α were detected in triplicate measurements by Ready-SET-Go ELISA kits (eBioscience) following the manufacturer's instructions. Absorbance was measured on a DTX880 Multimode Detector (Beckman Coulter, Brea, CA).

CD4⁺ T, CD8⁺ T, and CD19⁺ B cells were isolated from lymph nodes of Orai1^KI/KI and littermate control mice and either used for calcium measurements directly or differentiated in vitro for 6 d (see earlier). For time-lapse Ca²⁺ imaging, cells were loaded with 1 μM fura 2-AM (Invitrogen) for 30 min at 22–25°C at a concentration of 1 × 10⁶ cells/ml and attached to poly-l-lysine–coated coverslips for 15 min. T cells were stimulated by passive store depletion with 1 μM thapsigargin or anti-CD3 cross-linking by incubating cells with 5 μg/ml biotin-conjugated anti-CD3 mAb (clone 2C11; Invitrogen) for 15 min at 22–25°C followed by perfusion with 10 μg/ml streptavidin (Pierce, Rockford, IL). B cells were stimulated with 1 μM thapsigargin. Measurements of [Ca²⁺]_i were conducted at 22–25°C using a Zeiss (Thornwood, NY) Axiovert S200 epifluorescence microscope and OpenLab imaging software (Improvision, Waltham, MA) as described (13). Approximately 100 cells per experiment were analyzed using Igor Pro analysis software (Wavemetrics, Lake Oswego, OR). For calcium measurements by flow cytometry, cells isolated from lymph nodes of Orai1^KI/KI and littermate control mice were simultaneously incubated with 1 μM fluo-4 acetoxymethyl ester (Invitrogen) and Abs to CD4, CD44, and CD62L for 15 min at 22–25°C. Cells were stimulated with 1 μM thapsigargin and calcium levels analyzed on an LSRII flow cytometer. Data were analyzed using FlowJo software (Tree Star).

Fluorescence resonance energy transfer (FRET) experiments were conducted and analyzed as described previously (25). Briefly, HEK293 cells were transfected with fusion proteins of wild-type Orai–cyan fluorescent protein and wild-type Orai1–yellow fluorescent protein or wild-type Orai1–cyan fluorescent protein and R91W-ORAI1–yellow fluorescent protein. FRET was measured in nonstimulated cells in 2 mM extracellular Ca²⁺ followed by stimulation with 1 μM thapsigargin in 0 mM extracellular Ca²⁺.

Sacrificed mice were fixed by cardiac perfusion with 4% paraformaldehyde/1% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.3. After dissection, fixation was allowed to continue at room temperature for 90 min or 4°C overnight. Cardiac and skeletal muscle tissue was washed with 100 mM Tris (pH 7.2) and 160 mM sucrose for 30 min followed by washes in iso-osmotic phosphate buffer (150 mM sodium chloride, 5 mM potassium chloride, 10 mM sodium phosphate, pH 7.3) for 30 min twice. After treatment with 1% osmium tetroxide in 140 mM sodium phosphate (pH 7.3) for 1 h, muscle tissue was washed (twice for 1 h) in water. The muscle was stained en bloc with saturated uranyl acetate for 1 h, dehydrated in ethanol, and embedded in Epon (Electron Microscopy Sciences, Hatfield, PA). Semithin sections were cut at 1 μm and stained with 1% toluidine blue to evaluate the quality of preservation. Ultrathin sections (60 nm) were cut and stained with uranyl acetate and lead citrate by standard methods. Stained grids were examined using a Philips CM-12 electron microscope (FEI; Eindhoven, The Netherlands) and photographed with a Gatan Erlangshen ES1000W digital camera (Model 785, 4 k × 2.7 k; Gatan, Pleasanton, CA).

CD4⁺ T cells were cultured in ThN conditions and harvested at days 5–6. Patch-clamp recordings were performed using an Axopatch 200 amplifier (Axon Instruments, Foster City, CA) interfaced to an ITC-18 input/output board (Instrutech, Port Washington, NY) and an iMac G5 computer. Currents were filtered at 1 kHz with a 4-pole Bessel filter and sampled at 5 kHz. Recording electrodes were pulled from 100-μl pipettes, coated with Sylgard (Dow Corning, Midland, MI), and fire-polished to a final resistance of 2–5 MΩ. Stimulation and data acquisition and analysis were performed using in-house routines developed on the Igor Pro platform (Wavemetrics). All data were corrected for the liquid junction potential of the pipette solution relative to Ringer’s solution in the bath (−10 mV) and for leak currents collected in 20 mM [Ca²⁺]_o + 25 μM La³⁺. The standard extracellular Ringer’s solution contained (in mM): 130 NaCl, 4.5 KCl, 20 CaCl₂, 1 MgCl₂, 10 d-glucose, and 5 Na-HEPES (pH 7.4). The standard divalent-free (DVF) Ringer’s solutions contained (in mM): 150 NaCl, 10 hydroxyethyl EDTA, 1 EDTA, and 10 HEPES (pH 7.4). Charybdotoxin, 25 nM (Sigma), was added to all extracellular solutions to eliminate contamination from Kv1.3 channels. The standard internal solution contained (in mM): 145 Cs aspartate, 8 mM MgCl₂, 10 BAPTA, and 10 Cs-HEPES (pH 7.2). Averaged results are presented as the mean value ± SEM.

Unless indicated otherwise, the unpaired, two-tailed Student t test was used for all statistical analyses. Differences were considered significant when p values were at least <0.05. The log-rank test was used to compare the survival of skin allografts.

Orai1^KI/KI mice were generated by replacing exon 1 of the Orai1 gene by homologous recombination with a mutant allele carrying a C > T missense mutation resulting in an R→W single amino acid substitution at position 93 (Supplemental Fig. 1) (18). Mice heterozygous for the Orai1-R93W mutation (Orai1^KI/+) developed normally on a C57BL/6 background with survival rates comparable with those of wild-type littermate controls (Fig. 1A). By contrast, the numbers of homozygous Orai1^KI/KI mice were reduced in late gestation at embryonic day 18.5 (Fig. 1B) and at birth relative to the expected Mendelian distribution and died within 12 h post partum (Fig. 1A). At birth, Orai1^KI/KI pups were of normal size, motile, and lacked obvious signs of paralysis or hypoxemia during the first hours postpartum. Survival of <30% of Orai1^KI/KI pups could be rescued for up to 9 d by crossing Orai1^KI/+ mice to the outbred ICR strain for 3–4 generations followed by intercrossing of heterozygotes (Fig. 1A). Nevertheless, the majority of Orai1^KI/KI mice on this background died on the day of birth; all remaining mice died within 9 d. Newborn Orai1^KI/KI pups lacked visible milk patches (Supplemental Fig. 2A) and showed signs of severe dehydration. The few mice that survived until days 5–9 after birth were severely runted. The poor survival of Orai1^KI/KI pups is in contrast with an ∼50% survival rate of Orai1^−/− mice on an ICR background (16) and greater survival of Orai1-deficient mice generated by insertional mutagenesis on a mixed genetic background (15, 17).

Orai1^KI/KI embryos (E16.5) and newborn pups lacked gross morphological abnormalities that could account for the perinatal lethality (data not shown). In particular, no histological abnormalities in skeletal muscle fibers from leg, intercostal, and diaphragm muscles were detected. This is in contrast with atrophic type II muscle fibers in an ORAI1-R91W mutant patient suffering from congenital myopathy (10) and Stim1^−/− mice, which showed a marked reduction in muscle cross-sectional area compared with that of wild-type mice. Ultrastructural analysis by transmission electron microscopy revealed largely intact myofibril structure and cell organelles in skeletal muscle from Orai1^KI/KI mice (Fig. 1C, 1D). A small fraction (<5%) of skeletal myofibers, however, showed markedly swollen mitochondria with abnormal cristae structure (Fig. 1E). Comparable morphological abnormalities were absent in skeletal muscle samples from wild-type littermate controls (data not shown). The mitochondrial changes in Orai1^KI/KI mice, which were also observed in cardiomyocytes (Supplemental Fig. 2B), are reminiscent of a similar, although more pronounced, mitochondriopathy in Stim1-deficient mice, which have been suggested to die perinatally from a skeletal myopathy (26). In addition, we observed dilated ER structures and nuclear envelopes in a small minority (<5%) of cells present in skeletal muscles of Orai1^KI/KI mice suggestive of ER stress. mRNA levels of ER stress-associated genes, however, were not increased in muscle from Orai1^KI/KI mice compared with those in wild-type mice (Supplemental Fig. 2C). Taken together, Orai1^KI/KI mice show severely impaired perinatal survival indicative of an important role for ORAI1 in tissues outside the immune system despite the lack of overt signs of myopathy in Orai1^KI/KI mice.

To study the role of ORAI1 for lymphocyte development and immune responses in vivo, we generated fetal liver chimeric mice by transferring stem cells harvested from the livers of Orai1^KI/KI or Orai1^+/+ embryos on day E14.5 into sublethally irradiated Rag2^−/− or Rag2^−/−, cγ^−/− C57BL/6 mice. Orai1^KI/KI FLCs were viable, and their T cells, isolated 5 to 8 wk after stem cell transfer, were homozygous for the targeted Orai1 locus (Fig. 1F). Lymphocyte numbers in the spleens and lymph nodes, including those of T, B, NK, NKT, and dendritic cells, were comparable in Orai1^KI/KI and Orai1^+/+ control mice (Fig. 1G). This finding suggested that overall lymphocyte development is normal, enabling us to analyze T cell function in Orai1^KI/KI mice.

Human patients homozygous for the ORAI1-R91W mutation or mutations that abolish ORAI1 expression lack CRAC channel function and SOCE in T cells (8–10, 12). In mice, the role of ORAI1 for CRAC channel function in T cells has been debated as SOCE is impaired in T cells from Orai1^−/− mice (16), whereas T cells from another strain of Orai1-deficient mice have normal Ca²⁺ influx (17). To investigate the role of ORAI1 for SOCE in mouse T cells further, we analyzed Ca²⁺ influx in Orai1^KI/KI T cells at various stages of differentiation. SOCE was measured in CD4⁺ T cells from Orai1^KI/KI mice that were cultured in vitro under nonpolarizing (ThN) or Th1-polarizing conditions and CD8⁺ T cells differentiated in vitro into CTLs. Ca²⁺ influx in response to passive store depletion with the sarco/endoplasmic reticulum Ca²⁺ ATPase inhibitor thapsigargin was severely compromised in CD4⁺ Th1 and ThN cells and CD8⁺ T cells apparent in strongly reduced peak Ca²⁺ levels and initial rates of Ca²⁺ influx (Fig. 2A–C, Supplemental Fig. 3A–C). This finding is consistent with the complete lack of SOCE in T cell lines from human patients homozygous for the ORAI1-R91W mutation, although in contrast with patient T cells, a small residual Ca²⁺ influx was observed in mouse Th1, ThN, and CD8⁺ T cells (13).

The small residual Ca²⁺ influx observed in T cells from Orai1^KI/KI mice could be due to vestigial function of the mutant ORAI1^R93W protein or, alternatively, another Ca²⁺ channel with properties distinct from ORAI1–CRAC channels, such as transient receptor potential or other nonselective cation channels. To test these hypotheses, we analyzed CRAC channel currents in CD4⁺ T cells from Orai1^KI/KI mice differentiated into ThN cells in vitro. A small residual CRAC channel current was observed in Orai1^KI/KI T cells after passive store depletion with thapsigargin in the presence of either 20 mM extracellular Ca²⁺ (Ca²⁺_o) or an Na⁺-containing DVF solution. I_CRAC amplitudes in Orai1^KI/KI T cells were ∼15% (20 mM Ca²⁺) and ∼13% (DVF) of those in control T cells (Fig. 2D–F). The electrophysiological properties of the residual current in T cells from Orai1^KI/KI mice were indistinguishable from I_CRAC in T cells from littermate controls or human T cells (13) in the following respects: 1) block of Ca²⁺ current by La³⁺ (Fig. 2D, 2E, left panels); 2) depotentiation of Na⁺ current in DVF solution over tens of seconds (Supplemental Fig. 4); 3) block of Na⁺ current by micromolar concentrations of Ca²⁺ (data not shown); and 4) fast inactivation of Ca²⁺ current (data not shown). Taken together, these data show that the weak Ca²⁺ current and SOCE in T cells from Orai1^KI/KI mice are due to a calcium channel with properties indistinguishable from I_CRAC. Vestigial function of the mutant ORAI1^R93W protein, however, is unlikely because a residual Ca²⁺ current of comparable magnitude had been observed in T cells from Orai1^−/− mice (16) and because T cells from human patients homozygous for the ORAI1-R91W mutation had no recordable I_CRAC (12, 13). Instead, residual Ca²⁺ currents in Orai1^KI/KI T cells may result from another SOC channel, for instance ORAI2 or ORAI3. Residual Ca²⁺ currents in Orai1^KI/KI T cells did not, however, show any of the hallmarks reported for ORAI2 and ORAI3—when these molecules were overexpressed together with STIM1—such as direct activation by 50 μM 2-aminoethoxydiphenyl borate (2-APB) 2-APB or Ca²⁺ dependent reactivation (not shown) (19, 20). Taken together, our data show that ORAI1 is critical for CRAC channel function in differentiated mouse T cells and accounts for the majority of store-operated Ca²⁺ influx.

ORAI2, but not ORAI1, was suggested to be responsible for SOCE in naive T cells because naive T cells from one ORAI1-deficient mouse strain had normal SOCE and expressed very high levels of ORAI2 mRNA compared with levels of ORAI1 and ORAI3 mRNA (17). These findings are in contrast with reduced SOCE in CRAC currents in naive T cells of another Orai1^−/− strain (16). To analyze the role of ORAI1 in naive T and B cells, we measured SOCE in CD4⁺ and CD8⁺ T cells and B220⁺ B cells freshly isolated from Orai1^KI/KI mice. We observed an almost complete defect in Ca²⁺ influx in splenic B220⁺ B cells from Orai1^KI/KI mice after thapsigargin stimulation (Fig. 3A) consistent with similar findings in B cells from Orai1^−/− mice (16). Importantly, a pronounced, albeit incomplete, defect in SOCE was observed in naive CD4⁺ and CD8⁺ T cells from Orai1^KI/KI mice, which is apparent in both reduced peak Ca²⁺ levels as well as attenuated rates of Ca²⁺ influx (the latter being an indirect readout for Ca²⁺ channel activity) (Fig. 3B–D). This finding indicates that ORAI1 is required for CRAC channel function and SOCE in naive T cells. It is of note that the SOCE defect in naive CD4⁺ and CD8⁺ T cells from Orai1^KI/KI mice is less pronounced than the defect in T cells differentiated in vitro (Fig. 3E, Supplemental Fig. 3D) suggesting that the molecular composition of the CRAC channel may change during T cell differentiation.

To test this hypothesis and to compare the role of ORAI1 in naive and Ag-experienced memory T cells in vivo, we measured SOCE in CD62L⁺CD44^lo naive and CD62L^–CD44^hi memory CD4⁺ T cells freshly isolated from Orai1^KI/KI and wild-type control mice. Thapsigargin-induced SOCE, measured either as the peak of the Ca²⁺ response or the rate of Ca²⁺ influx, was significantly impaired in both naive and memory T cells from Orai1^KI/KI mice compared with that of wild-type controls (Fig. 4A–C). The peak and steady-state Ca²⁺ levels are important determinants of Ca²⁺-dependent cellular responses, and the Ca²⁺ influx rate provides an indirect readout for CRAC channel function in the absence of direct electrophysiological measurements. Significantly reduced Ca²⁺ influx rates in T cells from Orai1^KI/KI mice indicate that ORAI1 is a major SOCE channel in both naive and memory T cells in vivo. Other Ca²⁺ channels, such as ORAI2 and ORAI3, are likely to contribute to SOCE as well given the residual Ca²⁺ influx in naive and memory T cells from Orai1^KI/KI mice. We therefore compared mRNA expression levels of ORAI1, ORAI2, and ORAI3 using quantitative real-time PCR in CD62L⁺CD44^lo naive and CD62L^–CD44^hi memory CD4⁺ and CD8⁺ T cells from wild-type C57BL/6 mice. ORAI1 and ORAI3 transcript levels were moderately higher in memory compared with naive CD4⁺ T cells (Fig. 4D). Importantly, the relative expression levels of all three ORAI paralogues were similar in naive and memory T cells. This is consistent with microarray gene expression data from C57BL/6 mice showing comparable expression of ORAI1, ORAI2, and ORAI3 (14, 27). ORAI2, ORAI3, or both may, therefore, contribute to SOCE in T cells. Defining the precise contribution of ORAI2 and ORAI3 to Ca²⁺ influx in naive T cells from electrophysiological measurements alone is, however, difficult because the biophysical properties of all three ORAI isoforms are very similar (19). Collectively, our data show that ORAI1 is essential for SOCE and CRAC channel function in T cells.

To investigate the role of ORAI1 for T cell function directly in vivo, we tested Orai1^KI/KI mice in several models of T cell-mediated immunity. We assessed the ability of Orai1^KI/KI mice to reject skin allografts from MHC class I and class II mismatched mice, which is generally considered to be mediated by both CD4⁺ and CD8⁺ T cells (28). Tail-skin grafts from BALB/c (H-2^d) mice were transplanted onto the backs of Orai1^KI/KI or Orai1^+/+ chimeric mice on the C57BL/6 background (H-2^b). Transplantation of tail skin to wild-type mice resulted in rejection of skin grafts with a median graft survival time of 13.7 d after skin transplantation (Fig. 5A). By contrast, Orai1^KI/KI mice showed significantly prolonged skin allograft survival with a median graft survival of 21.1 d. Five days after transplantation, skin grafted onto wild-type C57BL/6 mice in general showed more pronounced perivascular leukocyte infiltration and occlusion of blood vessels—criteria for rejection of skin allografts (29)—compared with Orai1^KI/KI recipient mice (Fig. 5B). Although both wild-type and Orai1^KI/KI mice eventually rejected the allograft, the prolonged graft survival in Orai1^KI/KI mice indicates that CD4⁺ and CD8⁺ T function in vivo is significantly impaired in the absence of functional ORAI1.

Th2 cells mediate contact hypersensitivity reactions in response to exposure of epidermal skin cells to contact allergens (30). We tested whether Orai1^KI/KI mice can mount a delayed-type hypersensitivity response after sensitization with FITC. Six days after Orai1^KI/KI and control mice had been sensitized epicutaneously with FITC, their ears were painted with FITC or vehicle alone to elicit an immune response. Without prior sensitization, neither control nor Orai1^KI/KI mice showed a contact hypersensitivity response upon exposure to FITC. Sensitized wild-type mice showed a robust increase in ear thickness 24 h after FITC treatment relative to the ear thickness before FITC treatment (Fig. 5C). By contrast, no ear swelling was detectable in FITC-sensitized Orai1^KI/KI mice in response to FITC. These findings indicate that T cell-mediated delayed-type hypersensitivity responses and Th2 cell function in vivo are significantly impaired in Orai1^KI/KI mice.

Autoreactive T cells mediate inflammation and destruction of target tissues in numerous forms of autoimmune disease including colitis. Transfer of naive CD4⁺CD25^–CD45RB⁺ T cells into lymphopenic mice causes severe inflammatory bowel disease (IBD) in recipient animals. IBD develops in response to activation of naive T cells to stimulation by the gut microflora and their differentiation into proinflammatory Th1 and Th17 cells that goes unopposed due to the lack of Foxp3⁺ Tregs (31–33). We tested whether T cells from Orai1^KI/KI mice are able to induce IBD. CD4⁺CD25^–CD45RB⁺ T cells (5 × 10⁵) from Orai1^KI/KI and wild-type Orai1^+/+ mice were transferred into syngeneic Rag2^−/− mice. Mice that had received naive T cells from Orai1^+/+ mice progressively lost weight in contrast with recipients of Orai1^KI/KI T cells, who actually gained weight in the 12 wk after adoptive T cell transfer (Fig. 6A). Whereas colons of wild-type recipients showed severe signs of inflammation on macroscopic and histological inspection, the colons of mice that had received Orai1^KI/KI T cells were only moderately inflamed and showed few signs of epithelial hyperplasia, loss of goblet cells, or ulceration (Fig. 6B–D). A similar resistance to autoimmune colitis was observed in recipient mice that had received CD4⁺CD25^–CD45RB⁺ T cells from Stim1^fl/fl Cd4Cre mice that lack STIM1 expression and SOCE in T cells (Fig. 6E, 6F) (21) suggesting that STIM1 and ORAI1 are similarly important for the ability of T cells to induce inflammation in vivo. IFN-γ and IL-17 produced by CD4⁺ T cells in the lamina propria of mice with IBD are critical mediators of inflammation (31). We therefore analyzed proinflammatory cytokine gene expression in cells isolated from mesenteric lymph nodes and colon explants of Rag2^−/− recipients. Levels of IFN-γ, IL-17A, and TNF-α were strongly reduced in anti-CD3–stimulated cells isolated from mesenteric lymph nodes of mice that had received T cells from Orai1^KI/KI mice compared with the levels in mice that had received Orai1^+/+ T cells (Fig. 6G). In addition, IFN-γ and IL-17A mRNA expression levels in homogenates of resected colons from mice transferred with Orai1^KI/KI T cells were barely detectable in contrast with those of mice that had received wild-type T cells (Fig. 6H). Taken together, the inability of T cells from Orai1^KI/KI mice to cause autoimmune colitis is likely due to their failure to express proinflammatory cytokines. Defects in the expansion and/or homeostasis of ORAI1-deficient proinflammatory T cells may also contribute to the reduced levels of proinflammatory cytokines and protection from colitis as indicated by the reduced numbers of CD4⁺ T cells in the mesenteric lymph nodes of mice that had received Orai1^KI/KI T cells compared with those in mice that had received wild-type T cells (Fig. 6I). The majority of Orai1^KI/KI and wild-type T cells in lymph nodes showed a memory phenotype (CD44⁺CD62L^–) indicating that they had been activated in vivo. Collectively, these findings show that ORAI1 is critical for T cell activation in vivo and T cell-mediated immune responses.

To assess the role of ORAI1 for T cell function in more detail, we measured expression of cytokines in CD4⁺ and CD8⁺ T cells isolated from Orai1^KI/KI mice that were grown in vitro under nonpolarizing conditions (ThN). T cells restimulated with PMA and ionomycin showed a severe defect in IL-2, IFN-γ, and IL-4 production, whereas expression of IL-10 was only partially impaired (Fig. 7A, 7B). The almost complete lack of IL-4 expression in ORAI1-deficient T cells is consistent with impaired Th2-mediated contact hypersensitivity responses in Orai1^KI/KI mice. In addition to CD4⁺ T cells, CD8⁺ T cells from Orai1^KI/KI mice that were differentiated into CTLs in vitro also showed a severe defect in IFN-γ production upon restimulation (Fig. 7C). It is noteworthy that the level of cytokine production in CD4⁺ and CD8⁺ T cells from Orai1^KI/KI mice was comparable with that observed in Orai1^+/+ control T cells stimulated in the presence of cyclosporin A, an inhibitor of the Ca²⁺/calmodulin-dependent phosphatase calcineurin (Fig. 7B, 7C), suggesting that SOCE present in Orai1^KI/KI T cells is below the threshold for Ca²⁺-dependent activation of calcineurin. The dramatic defect in cytokine production observed in ORAI1-deficient T cells differentiated in vitro is consistent with the almost complete lack of detectable IFN-γ and IL-17 production in Orai1^KI/KI T cells isolated from Rag2^−/− mice and the absence of colitis in these mice (Fig. 6).

T cells from human patients homozygous for the ORAI1-R91W mutation showed a similar defect in cytokine gene expression as that in T cells from Orai1^KI/KI mice and failed to proliferate in vitro (6, 12, 34). In the colitis experiments described earlier, we had observed reduced numbers of CD4⁺ T cells in mesenteric lymph nodes of Orai1^KI/KI T cell-transferred mice, potentially due to a defect in T cell proliferation in vivo (Fig. 6I). To test this hypothesis, we isolated CD4⁺ and CD8⁺ T cells from lymph nodes of Orai1^KI/KI and Orai1^+/+ control mice and restimulated them in vitro for 72 h with αCD3 and αCD28. Orai1^KI/KI T cells, however, proliferated as vigorously as T cells from littermate controls (Fig. 7D). Only when CD4⁺ and CD8⁺ T cells were first differentiated in vitro for 3 d and subsequently restimulated for 72 h with αCD3 and αCD28 did Orai1^KI/KI T cells failed to expand as robustly as wild-type control T cells (Fig. 7E, 7F, left panels). Impaired proliferation of in vitro-differentiated compared with naive Orai1^KI/KI T cells is consistent with their more pronounced defect in SOCE (Fig. 3E). Treatment of wild-type and Orai1^KI/KI CD4⁺ and CD8⁺ T cells with cyclosporin A completely blocked their proliferation (Fig. 7E, 7F, right panels). Taken together, these findings suggest that residual SOCE in T cells from Orai1^KI/KI mice is sufficient to induce Ca²⁺/calcineurin-dependent cell proliferation but not cytokine gene expression.

As indicated earlier, the numbers of T and B cells in spleens and lymph nodes of Orai1^KI/KI chimeric mice were comparable with those in wild-type Orai1^+/+ mice (Fig. 1G, Supplemental Fig. 7A), suggesting that gross T and B cell development is unperturbed in ORAI1-deficient mice. When we analyzed populations of immature T cells in the thymus directly, we found comparable percentages of double-negative (CD4⁻CD8⁻), double-positive (CD4⁺CD8⁺), and single-positive (CD4⁺ or CD8⁺) thymocytes in Orai1^KI/KI and Orai1^+/+ wild-type mice (Fig. 8A). Orai1^KI/KI thymocytes expressed similar levels of the TCRβ-chain and the activation marker CD69 compared with levels in control mice (Supplemental Fig. 6B). In addition, B cell maturation appeared normal in ORAI1-deficient mice as the numbers of mature AA4.1^– and IgM^–IgD⁺ B cells in the spleens of Orai1^KI/KI and Orai1^+/+ mice were similar (Supplemental Fig. 7B–E). Collectively, these findings indicate that ORAI1 function is dispensable for lymphocyte development consistent with similar findings in mice lacking ORAI1 and STIM1 expression (16, 17, 21, 35) and normal numbers of PBLs in ORAI1- and STIM1-deficient human patients (10, 11, 14).

In contrast with conventional T cells, the development of Tregs appears to be more dependent on SOCE. Whereas Orai1^−/− and Stim1^−/− mice had normal development of conventional T cells and Tregs (16, 17, 35), mice with T cell-specific deletion of both STIM1 and STIM2 had severely reduced numbers of Tregs but not conventional T cells (21) suggesting that SOCE is required for Treg development. Orai1^KI/KI mice had numbers of CD25⁺Foxp3⁺ Tregs in thymus, lymph nodes, and spleen that were comparable with those in Orai1^+/+ controls consistent with the incomplete SOCE defect in naive Orai1^KI/KI CD4⁺ T cells (Fig. 8B). Orai1^KI/KI Tregs did, however, show a reduced ability to suppress proliferation of conventional T cells in vitro compared with Tregs from Orai1^+/+ control mice (Fig. 8C). This defect in Treg-mediated suppression in Orai1^KI/KI T cells is less pronounced than that observed in STIM1/STIM2-deficient Tregs (21) indicating that Treg function is modulated by the strength of the Ca²⁺ signal provided by SOCE.

We generated ORAI1^R93W homozygous knock-in mice to mimic an inherited human ORAI1 mutation that abolishes CRAC channel function and causes severe immunodeficiency to gain a detailed understanding of the role of ORAI1 for immune responses in vivo (10, 12). In this study, we show that ORAI1 is critically important for CRAC channel function and store-operated Ca²⁺ influx in T cells, that T cells from Orai1^KI/KI mice have functional defects including impaired cytokine expression and Treg-mediated suppression, and, most importantly, that ORAI1 is required for cell-mediated immune responses in vivo, such as T cell-dependent allograft rejection, hypersensitivity responses, and autoimmunity.

Lack of ORAI1 function in human T cells results in life-threatening immunodeficiency and recurrent infections due to impaired SOCE in T cells and potentially other immune cells (36). A role for ORAI1 in other T cell-mediated immune responses in vivo has not been defined. We show that Orai1^KI/KI mice tolerate allogeneic skin grafts significantly longer than wild-type mice. Rejection of MHC class I and class II mismatched skin allografts is mediated by both CD4⁺ and CD8⁺ T cells (28) indicating that in Orai1^KI/KI mice, CD4⁺ and CD8⁺ T cell function is severely compromised. Because CD4⁺ T cells alone are able to reject MHC class I and class II mismatched skin grafts in mice lacking CD8⁺ T cells (37, 38), attenuated allograft rejection in Orai1^KI/KI mice furthermore suggests an important role for ORAI1 in CD4⁺ T cell function in vivo. In addition, Th2-mediated skin contact hypersensitivity responses were absent in Orai1^KI/KI mice consistent with impaired IL-4 production in T cells from Orai1^KI/KI mice in vitro.

Orai1^KI/KI mice showed severely compromised proinflammatory Th1 and Th17 responses in vivo, apparent in the inability of T cells from ORAI1-deficient mice to induce colitis in an adoptive transfer model of IBD. Orai1^KI/KI T cells found in mesenteric lymph nodes and the colon failed to express IL-17A and IFN-γ, which are required to induce colitis (31, 39). In mice that had received wild-type CD4⁺CD45RB⁺ T cells, the IFN-γ secreted by Th1 cells present in the inflamed colon results in activation of macrophages that in turn release proinflammatory IL-12 and TNF-α (39). A similar role for Th17 cells and the retinoid-related orphan receptor γt-dependent production of IL-17A and IL-17F was recently described in the pathogenesis of intestinal inflammation following transfer of CD4⁺CD45RB⁺ naive T cells (31). Treatment of mice with anti–IFN-γ Abs or IL-10 and neutralizing anti–IL-17A Abs, respectively, inhibited disease (31, 39). The lack of IFN-γ and IL-17A expression in Orai1^KI/KI T cells is very likely to be responsible for their failure to induce colitis in the adoptive transfer model we used to evaluate the role of ORAI1 for proinflammatory T cell function in vivo. In addition, numbers of transferred CD4⁺ Orai1^KI/KI T cells were reduced in the mesenteric lymph nodes of recipient mice. Impaired ORAI1 function does not interfere with proliferation of naive mouse T cells in vitro (Fig. 7), which is consistent with similar findings in Orai1^−/− (16) and Stim1^−/− (35) mice. We recently found, however, that STIM1-deficient Th17 cells failed to expand in vitro and in vivo and are compromised in the expression of cytokines and cytokine receptors required for Th17 differentiation (40). A similar defect in the homeostasis of Orai1^KI/KI Th17 cells may contribute to their inability to induce colitis after adoptive transfer. Taken together, our findings demonstrate that ORAI- mediated Ca²⁺ influx is required for the function of several T cell subsets in vivo in the context of T cell-mediated hypersensitivity, autoimmunity, and allograft rejection. These findings raise the possibility that many other aspects of T cell-mediated immunity, such as antiviral and antitumor responses are also compromised in ORAI1-deficient mice, an issue that will have to be investigated in future studies.

The inability of T cells from Orai1^KI/KI mice to mediate classical T cell-dependent immune responses in vivo is explained to a large degree by their defective production of a number of key cytokines, such as the proinflammatory IFN-γ, IL-17, and TNF-α and the Th2 cytokine IL-4. The important role of ORAI1 for cytokine gene expression is emphasized by the almost complete absence of these cytokines upon restimulation of Orai1^KI/KI T cells in vitro and the fact that stimulation of wild-type T cells in the presence of the calcineurin inhibitor cyclosporin A resulted in a comparable inhibition of cytokine expression. This finding suggests that SOCE in Orai1^KI/KI T cells is not sufficient for the Ca²⁺-dependent activation of calcineurin, which appears to depend entirely on Ca²⁺ influx through ORAI1. Other factors that contribute to the defect in cell-mediated immunity in Orai1^KI/KI mice could be related to a role of ORAI1 in degranulation of CTLs and NK cells, as well as a potential role for SOCE in the differentiation and homeostasis of Th17 cells (40).

Surprisingly, however, T cells isolated from spleens and lymph nodes of Orai1^KI/KI mice showed normal proliferative responses upon TCR stimulation. This is in contrast with impaired proliferation of T cells isolated from peripheral blood of ORAI1-deficient patients (8–10, 34). The discrepancy could be due to differential Ca²⁺ signaling requirements in mouse and human T cells with regard to proliferation as T cells from STIM1-deficient mice proliferated normally in vitro (21, 35), whereas those of STIM1-deficient patients did not (11). Proliferation of Orai1^KI/KI T cells is unlikely to result from their residual SOCE and the potential contribution of other ORAI isoforms to SOCE in mouse T cells as STIM1-deficient murine T cells proliferated normally despite the absence of measurable Ca²⁺ influx in these cells (21, 35). Nonetheless, although murine ORAI1-deficient T cells proliferated normally in vitro, we cannot currently exclude that T cell homeostasis in vivo is compromised in Orai1^KI/KI mice given that numbers of Orai1^KI/KI CD4⁺ T cells were reduced in mesenteric lymph nodes of mice in which colitis had been induced by adoptive transfer of naive T cells.

In contrast with impaired T cell function, T and B cell development appeared normal in Orai1^KI/KI mice, which is consistent with normal lymphocyte development in Orai1^−/− (16, 17), and Stim1^−/− mice (21, 35), as well as human ORAI1- (8–10, 34) and STIM1- (11) deficient patients, who present with normal lymphocyte numbers and subsets in their peripheral blood. It should be pointed out that normal T cell development in Orai1^KI/KI mice as described in this study refers to numbers of T cells and T cell subsets only as we have not analyzed TCR repertoire selection in these mice. We cannot currently exclude that thymic selection of Orai1^KI/KI T cells is altered due to impaired Ca²⁺ signaling. Several indirect lines of evidence suggest that Ca²⁺ signals are required for proper T cell development in the thymus including Ca²⁺ oscillations observed in T cells undergoing positive selection (41), defects in positive selection of thymocytes in mice lacking the calcineurin B1 regulatory subunit (42), and the fact that ORAI1 is expressed at all stages during thymocyte development (14). Given the normal T cell development in ORAI1- (16, 17) and STIM1- (11, 21, 35) deficient mice and human patients, it appears, however, that the STIM1/ORAI1 pathway is dispensable for lymphocyte development and raises the possibility that other Ca²⁺ entry channels mediate this process (for a more detailed discussion, refer to Ref. 43).

The importance of ORAI1 for CRAC channel function in lymphocytes is evident from the clinical immunodeficiency and severely impaired SOCE in T and B cells from patients with mutations in ORAI1 that result in either nonfunctional channel protein (12) or abolished ORAI1 expression (10). Deletion of Orai1 in mice has led to conflicting results as to the role of ORAI1 for SOCE in mouse T cells. One study showed impaired SOCE and CRAC channel currents in T cells from Orai1^−/− mice resulting in impaired cytokine gene expression (16), whereas another study found that deletion of Orai1 did not affect SOCE in T cells and only moderately impaired cytokine production (17). Analyzing naive and differentiated T cells from Orai1^KI/KI mice, we found that ORAI1 is required for SOCE and CRAC channel function in mouse T cells as naive CD4⁺ and CD8⁺ T cells from Orai1^KI/KI FLCs showed impaired SOCE compared with that in wild-type controls. An even more pronounced defect with very low residual SOCE and CRAC channel currents was observed in CD4⁺ and CD8⁺ T cells differentiated in vitro. This finding suggested that the configuration of the CRAC channel complex may change during T cell differentiation from naive to Ag-experienced effector T cells with ORAI1 being the predominant ORAI isoform in effector T cells. We did observe, however, a comparable SOCE defect in naive (CD62L⁺CD44^lo) and memory (CD62L⁻CD44^hi) CD4⁺ T cells isolated from Orai1^KI/KI mice. It remains to be determined whether the composition of CRAC channel alters between distinct lymphocyte subsets.

The residual SOCE and Ca²⁺ current in T cells from Orai1^KI/KI mice we observed were not due to vestigial function of the mutant ORAI1-R93W protein because residual I_CRAC in Orai1^KI/KI T cells was comparable with that observed in T cells from Orai1^−/− mice (16) and because T cells from patients homozygous for the equivalent ORAI1-R91W mutation lack I_CRAC completely (12, 13). Thus, residual SOCE in naive Orai1^KI/KI T cells is likely to be due to—as has been suggested—ORAI2 or ORAI3 function, consistent with the observed expression of ORAI2 and ORAI3 mRNA expression in naive CD4⁺ T cells (16, 17). ORAI2 has been proposed to be the dominant ORAI isoform in naive T cells as ORAI2 mRNA levels in these cells were reported to vastly exceed those of ORAI1 or ORAI3 (17). By contrast, our quantitative mRNA expression analysis indicates that all three ORAI isoforms are expressed at roughly comparable levels. This is consistent with Northern blot (16) and microarray (14, 27) data from naive T cells showing comparable ORAI1 and ORAI2 mRNA levels that were moderately higher than those for ORAI3.

It proved to be difficult to evaluate by electrophysiological means whether residual SOCE and Ca²⁺ currents in Orai1^KI/KI T cells are due to ORAI2 or ORAI3 function because the biophysical properties of all three ORAI isoforms are very similar under overexpression conditions (19, 20) and are insufficiently defined for endogenous ORAI channels. The biophysical properties of residual Ca²⁺ currents in T cells from Orai1^KI/KI mice were indistinguishable from native CRAC channel currents and those in cells ectopically expressing ORAI1. Currents did not show any of the properties observed in cells ectopically expressing ORAI2, such as lack of slow inactivation of Ca²⁺ currents (19, 20), or ORAI3, such as direct activation of Ca²⁺ currents with 50 μM 2-APB (19). Although direct activation of native ORAI3 currents by 2-APB was recently described in breast cancer cells (44), the properties of overexpressed ORAI2 and ORAI3 channels may not accurately mimic those of endogenous ORAI2 and ORAI3 CRAC currents in T cells. In summary, I_CRAC in naive and differentiated murine T cells is in large part due to ORAI1. Although ORAI2 and ORAI3 are expressed in these cells, their contribution to CRAC channel function and T cell-mediated immune responses remains to be determined.

A surprising difference between human patients and mice homozygous for the equivalent ORAI1-R91W and -R93W mutations, respectively, is the neonatal lethality observed in Orai1^KI/KI mice but not human patients (10). ORAI1-deficient patients generally succumb to immunodeficiency in their first year of life but do survive after treatment by hematopoietic stem cell transplantation. Neonatal lethality in Stim1^−/− mice has been attributed to a myopathy characterized by morphological and functional abnormalities in skeletal myotubes (26). This finding is intriguing because human ORAI1- and STIM1-deficient patients suffer from congenital muscular hypotonia that in a patient with ORAI1-R91W mutation was associated with the histological observation of atrophic type II muscle fibers (10, 11). Newborn Orai1^KI/KI pups, however, lacked obvious signs of gross muscular dysfunction apart from their inability to feed and death due to dehydration within ∼12 h post partum. Structurally, their skeletal muscles appeared normal although a small fraction (<5%) of muscle fibers contained “swollen” mitochondria in transmission electron microscopy reminiscent of the mitochondriopathy observed in Stim1-deficient mice (26). It is unlikely, however, that the ultrastructural abnormalities observed in a small minority of muscle cells account for the neonatal lethality of Orai1^KI/KI mice, which is consistent with the observation that transgenic mice with skeletal muscle-specific expression of a dominant-negative mutant of ORAI1 are viable (45).

It is noteworthy that the phenotype of Orai1^KI/KI mice is more severe than that observed in Orai1^−/− mice in several respects raising the possibility that the mutant ORAI1-R93W protein exerts an inhibitory effect on other Ca²⁺ channels, such as ORAI2 and ORAI3. The mutant ORAI1-R91W protein is expressed in cells and can interact with wild-type ORAI1 [assessed by coimmunoprecipitation and FRET experiments (Supplemental Fig. 5)] in what is likely to be a tetrameric CRAC channel complex composed of four ORAI1 subunits (4). Alternatively, ORAI1 might form a tetrameric complex with other ORAI proteins as it was shown to be able to physically interact with ORAI2 and ORAI3 (19, 46) and to form heteromeric Ca²⁺ channels when coexpressed with ORAI3 (47). The following observations suggest that the mutant ORAI1-R93W protein may exert an additional inhibitory effect on SOCE compared with cells “simply” lacking ORAI1 expression. i) Orai1^KI/KI mice on a mixed genetic background (ICR) show a more severe, essentially nonrescuable neonatal lethality (Fig. 1) compared with Orai1^−/− (ICR) mice (16) and Orai1^−/− mice that were generated on a mixed genetic background (15, 17). ii) Cytokine production, especially that of IL-2, is more severely impaired in CD4⁺ T cells from Orai1^KI/KI mice than in those from Orai1^−/− mice (Fig. 7) (16, 17). iii) Residual SOCE in naive CD4⁺ and CD8⁺ T cells from Orai1^KI/KI mice is lower than that in Orai1^−/− T cells (Fig. 3) (16). By contrast, in vitro differentiated ThN cells from Orai1^KI/KI and Orai1^−/− mice had comparable residual I_CRAC amplitudes as expected for cells in which ORAI1 is the predominant ORAI1 isoform (16). iv) Orai1^KI/+ T cells heterozygous for the R93W mutation show CRAC currents (in 20 mM extracellular Ca²⁺) that are ∼30% of those in wild-type T cells (Supplemental Fig. 4B–D), slightly less than expected for a “simple” loss-of-function mutant. A dominant-negative effect of mutant ORAI1-R91W on CRAC channel function was postulated by Thompson et al. (48) who showed that incorporating one mutant ORAI1-R91W subunit together with three wild-type subunits in a concatenated tetramer reduced I_CRAC by ∼50%; two mutant subunits decreased I_CRAC by >90%. We speculate that nonfunctional ORAI1-R93W (or R91W in human) might be incorporated into heteromeric CRAC channel complexes containing ORAI2 and ORAI3 (although these have not been demonstrated to exist except under overexpression conditions), thereby preventing ORAI2 or ORAI3 from rescuing SOCE.

We show that ORAI1 is essential for SOCE and CRAC channel currents in T cells and that in the absence of ORAI1 function in Orai1^KI/KI mice, T cell-mediated immunity is severely impaired both in vitro and in vivo. Interfering with Ca²⁺ influx through inhibition of ORAI1 function may therefore be useful to suppress unwanted immune responses in the context of transplant rejection, inflammation, and autoimmunity.

We thank Dr. Klaus Rajewsky and laboratory for help with the generation of Orai1 knock-in mice. We thank Dr. Fengxia Liang and Tim Macatee for technical assistance with transmission electron microscopy and histological analysis, respectively.

Disclosures S.F. is a co-founder of Calcimedica, Inc. and member of its scientific advisory board.