Loss of T cell CD98hc specifically ablates T cell clonal expansion and protects from autoimmunity

Abstract

CD98hc (CD98 heavy chain, 4F2 antigen, Slc3a2) was discovered as a lymphocyte activation antigen. Deletion of CD98hc in B cells leads to complete failure of B cell proliferation, plasma cell formation, and antibody secretion. Here we examined the role of T cell CD98 in cell-mediated immunity and autoimmune disease pathogenesis by specifically deleting it in murine T cells. Deletion of T cell CD98 prevented experimental autoimmune diabetes associated with dramatically reduced T cell clonal expansion. Nevertheless initial T cell homing to pancreatic islets was unimpaired. In sharp contrast to B cells, CD98-null T cells showed only modestly impaired antigen-driven proliferation and nearly normal homeostatic proliferation. Furthermore, these cells were activated by antigen leading to cytokine production (CD4) and efficient cytolytic killing of targets (CD8). The integrin binding domain of CD98 was necessary and sufficient for full clonal expansion, pointing to a role for adhesive signaling in T cell proliferation and autoimmune disease. When we expanded CD98-null T cells in vitro, they adoptively transferred diabetes, establishing that impaired clonal expansion was responsible for protection from disease. Thus the integrin binding domain of CD98 is required for antigen-driven T cell clonal expansion in the pathogenesis of an autoimmune disease and may represent a useful therapeutic target.

Introduction

T cell-dependent autoimmune diseases such as multiple sclerosis (MS) and Type I Diabetes (T1D) result from tissue injury by autoreactive T lymphocytes (1, 2). Autoimmune and normal T cell responses rely on rapid clonal expansion of antigen-specific naïve T lymphocytes to generate a large population of effector cells (3–7) (8, 9). In protective immunity, this expanded population of effectors efficiently kills targets, provides cytokine “help” to B cells, and stimulates macrophages. Vertebrates are unique in using antigen-driven clonal expansion to enable adaptive immunity (10). CD98hc³ (4F2 antigen, Slc3a2), a Type II transmembrane protein, is also a unique feature of vertebrates (11). CD98hc was first identified as a T cell activation marker more than 25 years ago (12); however, its role in T cell function and pathogenesis of autoimmune disease is uncertain. Treatment of human peripheral blood T cells with anti-CD98hc antibody in vitro augmented anti-CD3-induced proliferation (13), suggesting that CD98hc might act as a co-stimulator of T cell division(14). However, due to the embryonic lethal phenotype of a CD98 KO mouse (15), definitive in vivo loss-of-function studies have been lacking.

To investigate the role of CD98hc in pathogenic and protective T-dependent immunity, we genetically deleted CD98hc in single-positive T cells, crossing an Slc3a2^fl/fl mouse³ with one carrying the Cre³ recombinase under the control of the T-cell specific distal Lck (dLck) promoter (16). Slc3a2^fl/fldLck-Cre+ mice were completely protected from autoimmune disease in an animal model for Type I Diabetes (T1D). CD98hc-null T cells populated peripheral lymphoid organs normally, formed immune synapses with antigen-presenting cells (APC), and displayed activation markers in response to antigen receptor stimulation; however, these cells exhibited impaired proliferation in vitro and in vivo in response to T cell receptor engagement. CD4 and CD8 T cell effector functions were intact in single cells but drastically reduced at the population level due to a critical lack of clonal expansion. Pre-expansion/differentiation of CD98-null T cells in vitro prior to adoptive transfer restored population effector functions and diabetogenicity, confirming that the mechanism by which CD98hc deletion protected from T1D was a block of clonal expansion. These results demonstrate an in vivo role for CD98hc in T dependent immune responses and identify it as a potential target in therapy of T cell-mediated diseases.

Materials and Methods

Mice

Slc3a2-floxed mice were generated by flanking exons 1 and 2, which encode the transmembrane portion of CD98hc, with loxP sites (17). The neomycin selection cassette used to select ES clones positive for a Slc3a2-floxed allele was flanked by Flp sites and thus was excised when Slc3a2-floxed mice were bred with human β-actin FLPe deleter mice (Jackson Laboratories). Slc3a2^f/fdLck-Cre⁺ mice were the result of crossing Slc3a2^f/f with the dLck-Cre⁺ strain (16). Slc3a2^f/+dLck-Cre⁺ offspring were identified by PCR and backcrossed with Slc3a2^f/f mice to create mice heterozygous for dLck-Cre and homozygous for the Slc3a2 floxed allele (Slc3a2^flf). Wherever possible, littermate comparisons were made. OT-1 mice have been described previously (18). All mice were housed at the University of California San Diego animal facility, and all experiments were approved by the Institutional Animal Care and Use Committee (IACUC).

Staining and flow cytometry of peripheral T cells

Spleen, lymph node (mesenteric), or thymic single-cell suspensions were prepared by dissociating whole spleens using a 7ml tissue grinder (Kontes), and lysing erythrocytes. After counting, spleen, lymph node, or thymus cells were stained in 100 μl of staining buffer (PBS, 0.5% BSA) containing fluorochrome-conjugated antibodies (BD Biosciences) against mouse CD3, CD4, CD8, CD25, and CD98hc, at optimal concentrations. After incubation on ice for 30–45 min, followed by three washes in staining buffer, subsets were analyzed by flow cytometry using a FACSCalibur cytometer (BD Biosciences). For analysis of natural Treg populations, surface staining for CD4/CD25 was following by fixation and intracellular staining for Foxp3 using the cytofix/cytoperm kit (eBioscience).

T cell purification and proliferation analysis in vitro

T cells were purified from spleens of Slc3a2^fl/fl, dlck-Cre+ mice or control Slc3a2^fl/fl littermates by negative depletion using the Invitrogen (Dynal) magnetic bead kit. dlck-Cre+ samples included anti-CD98hc (Biolegend) antibody to deplete the remaining 5–10% CD98+ T cells present in these mice. Purified T cells (routinely 93–98% CD3+ by flow cytometry) were labeled with CFSE (Invitrogen) at 4 μM for 10 min. at 37C. For stimulation of polyclonal populations, T cells were incubated for 3 days (as indicated) in complete medium (DMEM, supplemented with l-glutamine, pen/strep antibiotics, and 100 μM 2-ME) and recombinant human IL-2 (NCI Preclinical Repository) on tissue culture plates that had been coated with 5 μg ml each of anti CD3 and anti-CD28 antibodies, or at a 1:1 ratio with 4.5 um latex beads (Polysciences) that had been coated with anti-CD3 antibody (2.5 μg/ml) overnight at 4C and blocked with BSA for 2h at 37C. To analyze proliferation in response to antigen and antigen-presenting cells (APC), T cells were purified from halves of spleens from OT-1+ Slc3a2^fl/f, dlck-Cre+ mice or control OT-1+ Slc3a2^fl/fl littermates. The other spleen half was irradiated (3500 rads) for use as APC. Purified OT-1+ T cells and irradiated autologous splenocytes were recombined, labeled with CFSE, and plated with 100ng/ml SIINFEKL for 5 days. All in vitro proliferation assays were analyzed by flow cytometry after additional staining with antibodies for T cell markers (CD4 or CD8) and CD98hc.

Conjugate formation assay

Purified T cells from spleens of OT-2+, Slc3a2^fl/fl, dlck-Cre+ mice or control OT-2+, Slc3a2^fl/fl littermates were labeled with CFSE and cultured with OVA_323–339 peptide and DiD-labeled purified B cells that had been activated for 24h with 20 μg/ml LPS (Invivogen). At the indicated time points, T+B+peptide mixtures were fixed with 4% paraformaldehyde, washed in PBS, 0.5% BSA and analyzed by flow cytometry for APC (DiD+CFSE+) double positive cells (conjugates).

Visualization of T cell: APC immune synapses by confocal microscopy

To generate splenic dendritic cells as APC, C57/BL6 mice were injected s.c. with 5 × 10⁶ Flt3L-expressing B16 cells (gift of G. Dranoff). Seven days later, mice were sacrificed and dendritic cells were isolated from splenocytes using CD11c microbeads (Miltenyi). Purified dendritic cells were labelled with the fluorescent dye CellTracker Blue (Invitrogen), pulsed with SIINFEKL peptide (1 nM) for 1h at 37°C, and washed to remove peptide. Dendritic cells were combined with naïve OT-1 TCR transgenic CD8+ T cells or OT-1 TCR transgenic CTLs at a 1:1 ratio, and incubated in serum-free media for 20 minutes at 37°C. Cells were harvested, allowed to adhere briefly to coverslips coated with poly-L-lysine (Sigma), and fixed with 4% paraformaldehyde. Staining was performed with the following antibodies: anti-α-tubulin (Sigma), anti-CD11a (eBioscience), and talin (8D4, Sigma). DAPI (Invitrogen) was used to detect DNA. T lymphocytes interacting with dendritic cells were selected for imaging, and image stacks were acquired at 0.5 mm intervals using an Olympus FV1000 Confocal Microscope. Image stacks were processed for display using iVision (BioVision) and Adobe Photoshop CS4 (Adobe).

Homeostatic proliferation

T cells (1×10⁶) were purified from spleens of OT-1+, Slc3a2^fl/fl, dlck-Cre+ mice or control OT-1+, Slc3a2^fl/fl littermates, labeled with CFSE, and injected into sub-lethally irradiated (600rads) wildtype BL6 mice. Seven days later, recipient mice were sacrificed, and spleen cells stained for TCR Vα1/Vβ5+ (OT-1) T cells. Flow cytometric analysis of CFSE dilution was performed on TCR Vα1/Vβ5+ (OT-1) T cells.

Analysis of Clonal Expansion in vivo

For OT-1 clonal expansion, T cells (2×10⁶) were purified from spleens of OT-1+, Slc3a2^fl/fl, dlck-Cre+ mice or control OT-1+, Slc3a2^fl/fl littermates, labeled with CFSE, and injected into wildtype BL6 mice. One day later, recipient mice containing “parked” OT-1 T cells were immunized with 75 μg SIINFEKL peptide in Complete Freund's Adjuvant (CFA) i.p. Two days after immunization, mice were sacrificed, and splenic single-cell suspensions were stained with antibodies for TCR Vα2 and CD98hc. Cell division was analyzed by CFSE dilution as measured by flow cytometry. To analyze clonal expansion on a polyclonal TCR background, Slc3a2^fl/fl, dlck-Cre+ mice or control Slc3a2^fl/fl littermates were immunized with a combination of 100 μg poly I:C (Invivogen), 50 μg anti-CD40 Ab (Biolegend), and 500 μg whole ovalbumin protein (Sigma) in PBS i.p (19). Six days later, mice were sacrificed and spleen cells stained using anti-CD8+ antibody and a H-2K^b-SIINFEKL(PE) tetramer (Coulter). Total tetramer+ cells were calculated from %tetramer+CD8+ cells multiplied by total cell number.

Retroviral CD98hc-chimera reconstitution and BM transplant

Donor mice were treated with 4 μg per mouse of 5-Fluorouracil (Adrucil, from Sicor pharmaceuticals) in 200 μl PBS, i.p. Three days later, ecopack293 packaging cells (Imgenex) were transfected with pCl-Eco (Imgenex) and with one of 4 retroviral constructs (C₉₈T₉₈E₉₈,C₉₈T₆₉E₉₈,C₉₈T₉₈E₆₉, or C₆₉T₆₉E₆₉ in an MSCV-IRES-Thy1.1 backbone) (Addgene) and cultured for 48 hours. On day 4, donor mice were sacrificed, and BM cultured overnight in complete medium containing IL-3 (25 ng/ml), IL-6 (50 ng/ml), and SCF (50 ng/ml) (all from Peprotech). Viral supernatants were collected from packaging cells on day 5 and used to spin-fect BM at 2800 rpm, room temperature for 90 min on days 5, 6, and 7. BM cells were harvested on day 8, washed, and injected (250,000 per mouse) in 100 μl PBS i.v. Small samples were retained and stained for Sca-1 (eBioscience, clone D7), a marker of stem cells. In all samples, >90% of Sca-1⁺ populations were Thy1.1+ before injection. 8–10 wk after transplant, (mouse) CD98hc-deficient T cells were purified from spleens of recipient mice, labeled with CFSE, and stimulated with platebound anti-CD3, anti-CD28, and soluble IL-2. Proliferation was assessed as described above.

CD8+ expansion, differentiation to CTL, and effector function

To generate functional CTL, splenocytes from OT-1+, Slc3a2^fl/fl, dlck-Cre+ mice or control OT-1+, Slc3a2^fl/fl littermates were cultured at 4×10⁷ per well in a 6-well plate with 1 μg/ml SIINFEKL and 100U/ml IL-2 (NCI). On day 2, live cells were purified by running over a Lympholyte M gradient, washed to remove SIINFEKL peptide, and expanded for 4 days in 100U/ml IL-2 alone. CTL were harvested, counted, and cultured with SIINFEKL pulsed (1 μg/ml for 1–2h) or unpulsed splenocytes as targets in a 96-well plate at various effector to target ratios in the presence of anti-LAMP-1-PE antibodies (eBioscience) for 2.5 h at 37C. Effector/target cultures were harvested, stained with antibodies to CD98hc and CD8, and analyzed by flow cytometry. For target lysis in vitro, CTL were generated as above and cultured with a ~50/50 mixture of peptide-pulsed (CFSElo) and unpulsed (CFSEhi) splenocyte targets overnight at the indicated effector-target ratios as described (20). Specific lysis was calculated as the %decrease in the percentage of the peptide-pulsed peak between CTL-containing and no-CTL control cultures. For in vivo target cell lysis, differentially labeled splenocyte targets were prepared as for in vitro experiments and injected i.v. into mice that had been immunized or had received OT-1 CTL 24h prior. Four to six hours late, mice were sacrificed and splenocytes analyzed by flow cytometry for loss of peptide-pulsed CFSE peak. Specific lysis was calculated as for in vitro experiments, comparing with mice that had not been immunized or had not received OT-1 CTL.

CD4+ differentiation/cytokine secretion in vitro

Splenocytes (3×10⁶) from adult Slc3a2^f/fdLck-Cre⁺ or control littermate mice were cultured in 24-well plates with soluble anti-CD3 and anti-CD28 antibodies (2 μg/ml each) with the following cytokine combinations to promote differentiation to different T helper subsets: Th1 = IL-2(50 U/ml), IL-12(10 ng/ml), and anti-IL-4(10 μg/ml); Treg = IL-2(100 U/ml), TGF-β(10 ng/ml); Th17 = TGF-β(10 ng/ml), anti-IL-4(10 μg/ml), anti-IFN-γ, and IL-6(10 ng/ml). After 4 days, splenocyte cultures were stimulated with 100 ng/ml PMA (Sigma) and 1 μg/ml Ionomycin (Sigma) for 4–6 hours in the presence of Brefeldin A (eBioscience). To detect cytokines, cells were stained with antibodies for surface markers (CD4 and CD98hc), fixed with Cytofix/Cytoperm kit (eBioscience), stained for intracellular IFN-γ or IL-17A, and analyzed by flow cytometry.

CD4+ effector function analyses in vivo

For antigen-specific antibody responses, adult Slc3a2^f/fdLck-Cre⁺ and control littermate mice were injected i.p. with 50 μg TNP-LPS (Sigma) in 250 μl PBS (T cell-independent antigen), or 100 μg TNP-KLH (Biosearch) emulsified in 250 μl CFA (T cell-dependent antigen). Blood serum was collected by centrifugation of tail vein bleed (100–200 μl with 1–2 mM EDTA solution as an anticoagulant) before (pre-immune) and 1 week after immunization. TNP-specific antibody concentrations in blood sera were assessed by direct ELISA with TNP-OVA as the coating antigen and AP-conjugated polyclonal anti-mouse IgG (Sigma) or anti-mouse IgM (Sigma) as the detection antibody.

Type I Diabetes disease model

Autoimmune diabetes transfer with OT-1+ T cells into RIPmOVA recipients has been described (21). Naïve OT-1 T cells were purified from spleens of OT-1+, Slc3a2^fl/fl, dlck-Cre+ mice or control OT-1+, Slc3a2^fl/fl littermates by negative depletion using the Invitrogen (Dynal) magnetic bead kit. dlck-Cre+ samples included anti-CD98hc antibody to deplete the remaining 5–10% CD98+ T cells present in these mice. T cells (7.5×10⁷) were transferred by i.v. injection into adult RIPmOVA recipients. Daily blood glucose readings were made, and mice were considered diabetic with >250 mg/dl glucose for 2 consecutive days. For transfer of pre-expanded OT-1 CTL (Fig. 7), CTL were generated from OT-1+ Slc3a2^fl/fl, dlck-Cre+ mice or control OT-1+, Slc3a2^fl/fl littermate spleens as for CD8+ effector function assays and injected into RIPmOVA recipients. For histological analyses, pancreas was removed upon sacrifice of RIPmOVA recipients, fixed in formalin, and sectioned. Five μm sections were cleared with orange oil, dehydrated, and stained with hematoxylin and eosin (H&E). For immunofluorescence tracking, pancreas was removed and frozen in O.C.T. solution. Five μm sections were fixed in acetone-methanol, blocked with BSA and donkey serum, and stained with polyclonal guinea pig anti-insulin antibody (Dako) followed by DyLight 649-conjugated anti-guinea pig secondary antibody. Insulin staining and CFSE fluorescence was visualized on an inverted fluorescence microscope (Nikon). For quantification of pancreas-infiltrating T cells, the pancreas was removed and digested with collagenase as described previously (22, 23) before preparing single-cell suspensions for flow cytometry.

Figure 7. Clonal expansion defect is the mechanism of T1D protection in CD98-null T cells.

Cellular mechanism of CD98hc in Type I Diabetes. (A) Activation/expansion in draining lymph nodes. At the indicated time points after transfer of purified CFSE-labeled naïve T cells from OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ or OT-I+, Slc3a2^fl/fl control littermates, recipient mice were sacrificed. Pancreatic lymph nodes were isolated and cells counted, stained, and analyzed by flow cytometry. Representative histograms, gated on CFSE+ cells, are shown on the left with bar graphs adjacent; error bars represent s.e.m. from 3–4 mice in each group. * P < 0.01 and ** P < 0.04 by two-tailed student's t-test. On day 2, CFSE+ T cells from pancreatic (pLN) and distal axillary (aLN) lymph nodes were also stained for CD69. Overlaid unfilled peak represents staining from OT-I+, Slc3a2^fl/fl control T cells transferred to a non-OVA expressing control mouse. For day 4, migration of CFSE+ T cells to pancreatic islets is shown by immunohistochemistry. (B) Surivival of CD98-null T cells in vivo. Splenic T cells (2×10⁶)were purified from Slc3a2^fl/fl, dlck-Cre⁺ or Slc3a2^fl/fl control mice, injected i.v. into Ly5.1 congenic mice, and enumerated in lymph nodes or spleen 4 days later by Ly5.2 staining and flow cytometry; n = 4 mice per group. (C) Disease transfer with pre-expanded OT-1 CTL. Naïve T cells were purified from OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ or OT-I+, Slc3a2^fl/fl control littermate spleens and expanded/differentiated by culturing with 1ug/ml SIINFEKL and 100U IL-2 for 2 days, and then IL-2 alone for the remaining 4 days. After depleting any remaing CD98hc+ cells in the OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ sample, T cells (7.5 × 10⁶) were injected i.v. into RIPmOVA recipient mice. Daily blood glucose readings were taken, and overt diabetes was confirmed by two consecutive hyperglycemic (>250 mg/dl) readings; n = 4 mice per group.

Multiple Sclerosis disease model (EAE)

Experimental autoimmune encephalomyelitis (EAE) was induced in Slc3a2^fl/fl, dlck-Cre+ mice or control Slc3a2^fl/fl littermates by immunization with murine MOG(35–55) peptide as described (24) Briefly, adult mice were immunized s.c. in the hind flank with 200 μg MOG(35–55) peptide in CFA on day 0. On days 0 and 3, pertussis toxin (200 ng) was injected i.v. Mice were monitored daily for clinical signs of disease, graded on a 0–5 scale as described (24), and weight loss. After 48+ h at clinical score 3 (both hind limbs paralyzed), mice were euthanized by CO₂ inhalation. The last clinical score of these mice continued to be included in mean calculation of clinical score for the following days. Weight was expressed as percent of weight on day 0.

Statistical Analysis

All error bars represent standard error of the mean. P-values listed in figure legends are from analysis using two-tailed Student's t-test unless otherwise indicated.

Supplemental Material

Supplemental Figures 1 (Fig. S1) shows CD98hc expression and T cell population in Slc3a2^fl/fl, dlck-Cre+ mice. Fig. S2 is analysis of regulatory T cells from Slc3a2^fl/fl, dlck-Cre+ mice. Protection from EAE induction in Slc3a2^fl/fl, dlck-Cre+ mice is shown in Fig. S3. Data on CD98hc-null T cell activation and formation of immune synapses with APC are provided as Fig. S4.

Results

Deletion of mature T cell CD98hc protects from experimental T1D

Slc3a2 encodes mouse CD98 heavy chain (CD98hc, or 4F2 antigen) (12), and its deletion results in embryonic lethality(15). To study the role CD98hc in T cell autoimmunity, we crossed an Slc3a2^fl/fl mouse (17) with the dLck-Cre mouse (16), which expresses Cre recombinase in single-positive (CD4 or CD8) T cells. Thymocytes of dlck-Cre+, Slc3a2^fl/fl offspring were null for CD98hc beginning late in thymic development (Supplemental Fig. 1A). The majority (80–90%) of peripheral T cells in dlck-Cre+, Slc3a2^fl/fl mice did not express surface CD98hc protein (Supplemental Fig. 1A), but were able to migrate and populate peripheral lymphoid tissues in normal numbers. We noted no change in the presence of CD4+ and CD8+ subsets in dlck-Cre+, Slc3a2^fl/fl mice (Supplemental Fig. 1B). In addition, the abundance of “natural” (CD25+Foxp3+) Tregs were normal; Tregs could also be induced to develop from dlck-Cre+, Slc3a2^fl/fl splenocytes ex vivo (Supplemental Fig. 2). These data indicated that the dlck-Cre+, Slc3a2^fl/fl mouse is suitable for studies of the function of mature T cell CD98hc in autoimmunity.

Type I diabetes (T1D) is a T cell-mediated autoimmune disease during which autoreactive T cells infiltrate pancreatic islets of Langerhans and destroy beta cells, eliminating the only significant source of insulin and the ability to regulate blood glucose(2). Since the initiating beta cell antigenic targets in human T1D are still not definitively identified, T cell receptor (TCR) transgenic mouse models are often used to study T1D(9, 25). In some of these models, T cells specific for an exogenous protein are transferred to mice engineered to express this protein on the surface of beta cells, marking them for attack (26, 27). To this end, we crossed dlck-Cre+, Slc3a2^fl/fl mice with the OT-1 T cell receptor transgenic (TCR-Tg) strain(18) in which the majority of T cells are specific for the OVA_257–264(SIINFEKL) peptide. Transfer of purified T cells from OT-1+ Slc3a2^fl/fl control mice into a strain (RIPmOVA) that expresses OVA_139–385 protein on the surface of pancreatic beta cells (27) resulted in rapid onset of autoimmune diabetes (Figure 1A). Strikingly, CD98hc-null T cells from dlck-Cre+, Slc3a2^fl/fl donors did not cause disease (Figure 1A). Massive insulitis was seen in pancreatic sections after transfer of CD98hc+ control T cells, but not after CD98-null OT-1 T cell transfer. Transferred T cells localized to islets (Figure 1B), and numbers of OT-1 dlck-Cre+, Slc3a2^fl/fl T cells were not significantly different from OT-1+ Slc3a2^fl/fl control T cells in the pancreas or pancreatic draining lymph nodes two days after transfer (Figure 1B–C).

Figure 1. CD98-null naïve OT-1 T cells do not transfer autoimmune diabetes.

Loss of CD98hc on T cells protects from experimental diabetes transfer. (A) Diabetes incidence. T cells were purified from spleens of OT-1+, Slc3a2^fl/fl, dlck-Cre+ mice or control OT-1+, Slc3a2^fl/fl littermates. T cells (7.5 × 10⁶) were injected i.v. into mice transgenic for OVA on pancreatic beta cells (RIPmOVA). Blood glucose readings were taken each day to monitor for hyperglycemia. Mice with blood glucose of >250 mg/dl for 2 days in a row were considered diabetic. At the conclusion (day 17 after T cell transfer), pancreata from RIPmOVA mice receiving either CD98-null T cells or control T cells were fixed in formalin, sectioned, and stained with Hematoxalin and Eosin to identify cellular infiltrate. The experiment was repeated with an additional 6 mice per group, yielding similar results. (B) Migration to pancreatic islets. Naïve T cells were purified from spleens of OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ or OT-I+, Slc3a2^fl/fl control littermates, labeled with CFSE, and transferred to recipient RIPmOVA mice. Two days after transfer, mice were sacrificed and pancreata isolated. Frozen pancreatic sections (5μm) were stained with anti-insulin antibodies to identify islets and analyzed by fluorescence microscopy (200× magnfication). These data are representative of two independent experiments. (C) Migration to draining lymph nodes. RIPmOVA recipient mice were sacrificed two days after transfer of CFSE+ T cells as in (B). CFSE+ T cells were enumerated by flow cytometry from single-cell suspensions of pancreatic lymph nodes (n=3–4 mice per group.) The experiment was repeated with an additional 3–4 mice per group with similar results.

OT-1 T cell transfer into RIPmOVA is well accepted as a model of autoimmune T1D (21, 27–29); however, the antigen, an ovalbumin peptide, is not a natural constituent in mammals. To examine the role of T cell CD98hc in an authentic autoimmune model, we assessed the effect of its deletion in experimental allergic encephalomyelitis (EAE) caused by immunization with murine myelin oligodendrocyte glycoprotein (MOG) peptide (24). dlck-Cre+, Slc3a2^fl/fl were dramatically protected from development of EAE in this model (Supplemental Fig. 3). The fact that T cell CD98hc was also required for development of a T-cell mediated autoimmune disease that targets a different tissue, the CNS, suggests that T cell CD98hc is generally required for T cell-dependent autoimmune diseases.

CD98hc is required for rapid antigen-driven T cell proliferation in vitro

To investigate the mechanism whereby CD98hc participates in experimental T1D, we examined the functions of CD98hc-null T cells in vitro. CD98hc was not required for T cell antigen recognition, polarization, formation of immunological synapses with antigen-presenting cells (APC), or upregulation of the early activation markers CD44 and CD69 (Supplemental Fig. 4). To test the role of CD98hc in T cell proliferation, CD98-null T cells were purified from spleens of OT-1+ dlck-Cre+, Slc3a2^fl/fl or control OT-1+ Slc3a2^fl/fl littermate mice and labeled with the proliferation dye carboxyfluorescein succinimidyl ester (CFSE). CD98-null T cells exhibited defective proliferation in contrast to control OT-1 T cells, which proliferated strongly in response to SIINFEKL peptide + APC (Figure 2A). Similar results were seen when stimulating polyclonal T cells from dlck-Cre+, Slc3a2^fl/fl or control OT-1+ Slc3a2^fl/fl littermate mice with immobilized anti-CD3 plus anti-CD28 antibodies in the presence of IL-2 (Figure 2B). The degree of impairment varied in vitro from moderate to a complete lack of proliferation depending on the cell density and could not be bypassed by adding excess (400U/ml) IL-2 or IL-7 cytokines (data not shown). Homeostatic expansion required to populate and maintain peripheral lymphoid numbers was also (~30%) reduced (Fig. 2C). These data demonstrate that CD98hc is involved in T cell proliferation induced by a variety of stimuli.

Figure 2. CD98hc is required for normal T cell proliferation in vitro.

Proliferation of splenic T cells from OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ mice. (A–B) In vitro proliferation analysis. T cells were purified from splenocytes of 8–12 wk-old OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ (unfilled solid line)and littermate OT-I+, Slc3a2^fl/f control mice (filled peak) and labeled with CFSE. CFSE-labeled T cells were cultured with irradiated APC + SIINFEKL peptide (A) for 5 days, or on platebound anti-CD3 plus anti-CD28 antibodies in the presence of IL-2 (B) for 3 days; proliferation (dye dilution) was measured by flow cytometry. Dashed peak in (A) is data from OT-I+, Slc3a2^fl/f T cells cultured with APC in the absence of antigen. Bar graph summarizes the proliferation index; error bars are SEM from n=3 mice per group. (*p <0.04, **p <0.02 by one-tailed t-test). In vitro proliferation experiments were repeated at least >3 times. (C) Homeostatic proliferation in vivo. T cells were purified from spleens of OT-I+, Slc3a2^fl/fl, dlck-Cre+ or OT-I+, Slc3a2fl/fl littermate mice and labeled with CFSE. T cells (1 × 10⁶) were injected i.v. into sublethally-irradiated (600rads) wildtype BL6 mice. After 7 days, recipient mice were sacrificed, and single-cell suspensions from spleen were stained with fluorochrome-conjugated antibodies to identify OT-1 T cells (CD3+TCRVα2+) and analyzed by flow cytometry. Bar graph summarizes proliferation index from n = 3 mice per group. (*P <0.0005).

The integrin-binding domain of CD98 is necessary and sufficient for rapid T cell proliferation

CD98hc has two documented biochemical functions, each dependent on a distinct domain of the polypeptide. The extracellular domain of CD98hc forms a disulfide bond with a CD98 light chain (LAT-1, etc), enabling proper localization of the heterodimer and thus amino acid co-transport (30, 31). Secondly, the transmembrane and cytoplasmic domains of CD98hc are necessary for interactions with integrin beta subunits and enhanced adhesive signaling (32, 33). Reconstitution with chimeric proteins in which portions of CD98hc have been replaced with another type II transmembrane protein, CD69, allows decisive separation of these two biochemical functions of CD98(34, 35) (Fig. 3A). We utilized a “retrogenic” approach to reconstitute CD98-null T cells with CD98/69 chimeras and test which domain (indicating which biochemical function) of CD98 might rescue T cell proliferation. Bone marrow stem cells from dlck-Cre+, Slc3a2^fl/fl mice were infected with a bicistronic CD98hc or chimera-IRES-Thy1.1 and transplanted into lethallyirradiated recipients. T cells then delete endogeneous CD98hc due to thymic expression of dlck-Cre, and those that express chimera are marked by expression of the congenic Thy1.1 marker. To test which domain of CD98hc rescued T cell proliferation in vitro, (mouse)CD98hc-null splenic T cells were purified from bone marrow transplant recipients, CFSE-labeled, and stimulated with immobilized anti-CD3 and anti-CD28 antibodies in the presence of IL-2. T cells reconstituted with a chimera lacking the CD98hc domain necessary for integrin signaling did not proliferate, but those lacking the amino acid transport domain were able to proliferate almost as well as wild type CD98hc-expressing cells (Fig. 3B). These results suggest that CD98-mediated integrin signaling is required for full T cell proliferation and to rescue the impairment observed in dlck-Cre+, Slc3a2^fl/fl T cells.

Figure 3. CD98 domain required for integrin signaling rescues CD98-null T cell proliferation defect.

The Integrin-binding Domain of CD98hc Drives T cell proliferation. (A) Diagram of CD98hc-CD69 chimeras. CD98hc protein is depicted in gray, and CD69 is depicted in black. Each chimera is defined by its cytoplasmic (C), transmembrane (T), and extracellular (E) domain derived from either CD98hc(98) or CD69 (69). CD98hc extracellular domain is necessary and sufficient for amino acid transport, whereas the intracellular and transmembrane domains are required for integrin interaction. (B) Rescue of T cell proliferation. Eight to 10 wk after transplant of BM cells infected with indicated retroviruses, T cells (mCD98hc⁻) were purified from splenocytes of recipient mice. Cells were labeled with CFSE dye. 200,000 CFSE-labeled t cells were cultured per well in a 48-well plate coated with 5 ug/ml each anti CD3 and anti CD28 + 100U/ml hIL-2 for 3 days. Proliferation was measured as the dilution of CFSE fluorescence by flow cytometry. Bar graph summarizes the division index of mCD98-null T cells expressing the various chimeras. Experiment was performed identically 3 independent times for a total of n=3 per group. (*P <0.015, compared with proliferation of 98-98-98-expressing T cells).

CD98hc is essential for antigen-driven clonal expansion in vivo

We next tested whether the proliferation defect in CD98hc-null T cells was present in vivo. OT-1+ dlck-Cre+, Slc3a2^fl/fl T cells exhibited little proliferation/expansion in response to immunization with SIINFEKL peptide, whereas OT-1+ Slc3a2^fl/fl T cells proliferated rapidly (Figure 4A). We extended these studies by testing clonal expansion of CD98-null T cells on a polyclonal TCR genetic background. To do this, we immunized with a combination of antigen(OVA), anti-CD40 antibody, and a TLR3 agonist (poly I:C) to generate an expansion of CD8+ T cells that approaches that seen during an acute viral infection (19). We used fluorochrome-labeled H-2K^b-SIINFEKL tetramer to track clonal expansion of OVA-specific CD8+ T cells. One week after immunization, the percentage of SIINFEKL-specific CD8+ T cells had increased from <1% to >10% in Slc3a2^fl/fl control mice. In contrast, dlck-Cre+, Slc3a2^fl/fl littermates exhibited almost no clonal expansion (Figure 4B). Taken together, these data show that CD98hc is required for antigen-driven T cell proliferation and resulting clonal expansion in vivo.

Figure 4. CD98hc is critical for clonal expansion in vivo.

Clonal expansion of CD98-null T cells. (A) Adoptively transferred TCR-Tg T cells. T cells were purified from splenocytes of 8–12 wk-old OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ or OT-I+, Slc3a2^fl/fl control littermate mice and labeled with CFSE. CFSE-labeled T cells (2 × 10⁶) were “parked” in a wildtype BL6 mouse for 24h, followed by immunization with 50 μg SIINFEKL peptide in CFA, as indicated in diagram. Mice were sacrificed 2 days after immunization and proliferation was measured as the dilution of CFSE fluorescence by flow cytometry, compared with a mouse receiving CFSE-labeled OT-I+, Slc3a2^fl/fl T cells but no immunization. Bar graph summarizes absolute numbers of CD98+ or CD98-null CFSE+ T cells in immunized recipient mice; n= 3 mice per group. (*P <0.025). Experiment was performed twice. (B) Polyclonal TCR genetic background. Slc3a2^fl/fl, dlck-Cre⁺ and Slc3a2^fl/fl control littermate mice were immunized with a combination of whole ovalbumin protein, anti-CD40 antibody, and a TLR7 agonist (Poly I:C) i.p. in PBS. Six days later, mice were sacrificed and splenocytes stained with flourochrome-conjugated anti CD8 antibody and an H-2K^b-SIINFEKL tetramer. Bar graph summarizes absolute numbers of CD8+ tetramer+ T cells in immunized (n= 3 mice per group) compared with an unimmunized control mouse (**P <0.015). The experiment was performed twice with similar results.

CD98hc is not required for intrinsic T cell effector functions

T cell effector functions such as cytotoxicity or provision of cytokine help for high-affinity antibody responses are the final effectors of T cell immunity. We first tested whether CD98hc is required for CD8+ T cell cytotoxic effector function at the population level in vivo. To test CD8+ cytotoxic activity in vivo, SIINFEKL-pulsed (CFSE^lo) or control (CFSE^hi) target splenocytes were injected i.v. into dlck-Cre+, Slc3a2^fl/fl or control Slc3a2^fl/fl littermate mice one week after immunization with the OVA/anti-CD40/poly(I:C) combination. Four hours later, specific killing was assessed by loss of CFSE^lo vs. CFSE^hi target cells from the spleen using flow cytometry. Slc3a2^fl/fl mice efficiently killed SIINFEKL-pulsed targets, whereas dlck-Cre+, Slc3a2^fl/fl littermates did not (Fig. 5A). To test whether this was due to a lack of clonal expansion or a defect in intrinsic cytotoxic ability of CD98hc-null T cells, OT-1 dlck-Cre+, Slc3a2^fl/fl or control Slc3a2^fl/fl splenocytes were cultured with SIINFEKL peptide and IL-2 (100U/ml) under conditions which allowed some proliferation of CD98hc-null T cells. After 6 days, dlck-Cre+, Slc3a2^fl/fl cultures routinely contained 40–50% fewer cells (data not shown) as expected from impaired proliferation, but remained >90% CD98-negative. When CTL numbers were equalized, OT-1+ CTL from dlck-Cre+, Slc3a2^fl/fl mice were able to degranulate (Fig. 5B) and lyse (Fig. 5C) target cells upon culture with SIINFEKL-pulsed target cells. These data indicate that CD98hc is not necessary for CD8+ effector function at the single cell level, and suggest that the lack of CD8+ effector function at the population level in vivo is solely due to a lack of clonal expansion. To definitively test this idea in vivo, CTL were generated from OT-1 dlck-Cre+, Slc3a2^fl/fl mice and adoptively transferred into recipient C57BL/6 mice. One day later, SIINFEKL-pulsed (CFSE^hi) or control unpulsed (CFSE^lo) target splenocytes were injected i.v. and target lysis was measured 6 hours later. Once CD98-null CD8+ T cells numbers were increased to that of wild type T cells, their capacity to kill targets in vivo was similar (Figure 5D). Thus CD98hc is required for CD8+ T cell expansion and not for cellular effector function.

Figure 5. CD98hc is required for CD8+ T cell effector function at the population, but not the single-cell level.

Cytotoxic effector function of CD98hc-null CD8+ T cells (A) Target cell lysis after in vivo expansion. Slc3a2^fl/fl, dlck-Cre⁺ and Slc3a2^fl/fl control littermate mice were immunized with a combination of whole ovalbumin protein, anti-CD40 antibody, and a TLR7 agonist (Poly I:C) i.p. in PBS. Six days later, mice were injected with a mixture of SIINFEKL-pulsed (CFSE^lo) and unpulsed (CFSE^hi) target splenocytes. 4 hours later, mice were sacrificed and splenocytes analyzed by flow cytometry. Specific killing is calculated as the decrease in the percentage of specific targets in immunized mice divided by the percentage in an unimmunized control mouse. (*P <0.001) Experiment was performed twice. (B) In vitro degranulation. Splenocytes of 8–12 wkold OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ or OT-I+, Slc3a2^fl/fl control littermate mice were differentiated to CTL by culture for 3 days with SIINFEKL (1μg/ml) and IL-2, followed by 3 additional days with IL-2 alone. Resulting CTL were harvested and cultured for 2 h at the indicated effector to target ratios (E:T) for 2 hours in the presence of fluorochrome labeled antibody specific for LAMP-1, followed by staining for CD8 and CD98hc and flow cytometric analysis. Histograms below show staining for CD98hc on the effector cells. Experiment was performed twice, both with 3 mice per group. (C) In vitro target lysis. CTL were generated as in (B) and cultured with SIINFEKL-pulsed or unpulsed targets that had been differentially labeled with CFSE. After overnight culture, target cell lysis was detected by flow cytometry and calculated as in (A). Inset histograms show target cells cultured overnight without CTL, or with 3:1 (E:T) ratios of either Cre+ or No Cre CTL. (D) In vivo target lysis after in vitro expansion. CTL were generated from OT-I+, Slc3a2^fl/fl, dlck-Cre⁺ mice as in (B) and injected i.v. (1.5×10⁷) into BL6 mice. One day later, mice were injected with a mixture of SIINFEKL-pulsed (CFSE^hi) and unpulsed (CFSE^lo) target splenocytes, and target cell lysis detected as in (A). Experiment was performed 3 times.

CD4+ T cells play an important “helper” role in guiding both humoral and cell-mediated immunity through secretion of cytokines. They are critical mediators of autoimmune diabetes in the NOD mouse model (22, 23) and are likely involved in the human disease (36) (37). We therefore tested the requirement for CD98hc in CD4+ T cell effector function. One week after challenge with the T-dependent antigen TNP-KLH, dlck-Cre+, Slc3a2^fl/fl exhibited markedly reduced levels of anti-TNP antibody (Fig. 6A). In sharp contrast, both dlck-Cre+, Slc3a2^fl/fl and control Slc3a2^fl/fl littermate mice responded equally to immunization with a T-independent antigen, TNP-LPS (Fig. 6B.) The impairment of T helper function in CD98hc-null T cells demonstrates a defect in CD98hc-null CD4+ function at the population level. To test whether CD98hc-null T cells have an intrinsic defect in effector function at the single cell level, we stimulated splenocytes from dlck-Cre+, Slc3a2^fl/fl or control Slc3a2^fl/fl littermate mice with a mixture of anti-CD3 and CD28 antibodies under Th1- or Th17-polarizing conditions. After 4–5 days to allow for proliferation and differentiation (dlck-Cre+, Slc3a2^fl/fl cells lagged behind in proliferation, data not shown), splenocyte cultures were stimulated with PMA/Ionomycin in the presence of Brefeldin A. Intracellular cytokine staining showed that CD4+ T cells from dlck-Cre+, Slc3a2^fl/fl or control Slc3a2^fl/fl littermate mice both produced IFN-γ or IL-17A at the single-cell level (Fig. 6C). Thus, CD98hc-null CD4+ T cells do not have a defect in intrinsic effector function, but rather the block in clonal expansion leads to deficient helper T cell responses in vivo.

Figure 6. CD98hc is required for CD4+ T cell effector function at the population, but not the single-cell level.

CD4+ effector function of CD98hc-null T cells. (A) T cell-dependent antibody response. Adult (8–12 wk old) Slc3a2^f/fdlck-Cre⁺ or control mice were immunized with 100 μg of the T cell-dependent antigen TNP-KLH in Complete Freund's Adjuvant (CFA). Mice were bled before immunization (pre-immune, PI) and one week after immunization to obtain serum. Concentrations of anti-TNP IgM or anti-TNP IgG were measured by direct ELISA. Error bars represent s.e.m. from 5 mice in each group. * P < 0.025; ** P < 0.0001 Experiment was performed twice. (B) T cell-independent antibody responses. Slc3a2^f/fdlck-Cre⁺ or control mice were immunized with 50 μg of a T cell-independent antigen, TNP-LPS, in PBS. Anti-TNP IgM and anti-TNP IgG in serum were measured by ELISA. Error bars represent s.e.m. 5 mice for each group. (C) CD4+ cytokine secretion at the single-cell level. Splenocytes from Slc3a2^f/fdlck-Cre⁺ or control mice were cultured with IL-2, IL-12, anti-CD3/28, and anti-IL-4, or IL-6, TGF-β, anti-CD3/28, anti-IL-4, and anti-IFN-γ to differentiate into Th1 or Th17 CD4+ subsets, respectively. After 4 days, cultures were treated with PMA/Ionomycin in the presence of Brefeldin A, followed by surface staining for CD4 and intracellular staining for IFN-γ or IL-17A. Experiment was done 3 times (n=3–4 mice per group for each experiment) with no consistent differences noted between Slc3a2^f/fdlck-Cre⁺ or control CD4+ T cells at the single-cell level. Representive dot plots are shown with the % of the population expressing the indicated cytokine and the Mean Fluorescence Intensity (M.F.I.) of cytokine staining depicted.

Clonal expansion is the cellular mechanism whereby CD98hc mediates experimental T1D

Deleting CD98hc in T cells blocks their ability to cause autoimmune diabetes, and CD98-null T cells exhibit defective clonal expansion but intact effector function. We therefore asked whether the clonal expansion defect explained the protection from T1D observed with loss of T cell CD98hc. First we examined the ability of OT-1 dlck-Cre+, Slc3a2^fl/fl T cells to home and to proliferate in the pancreatic lymph nodes of RIPmOVA recipient mice. CD98hc-null OT-1 CFSE-labeled T cells homed efficiently, as evidenced by similar numbers of purified naïve OT-1 dlck-Cre+, Slc3a2^fl/fl or control OT-1 Slc3a2^fl/fl T cells in pancreatic draining lymph nodes two days after transfer (Fig. 1C bar graph and Fig. 7A histogram). However, control OT-1 Slc3a2^fl/fl T cells were beginning to divide at 2 days after transfer, whereas the CD98-null cells were not (Fig. 7A), despite becoming activated in pancreatic lymph nodes. The results of this clonal expansion defect are apparent 4 days after transfer when the abundance of CD98+ control T cells is 5–6× greater than that of CD98hc-null T cells in pancreatic lymph nodes (Fig. 7A). This proliferation block may also result in activation-induced cell death (AICD) of CD98hc-null T cells, since calculated precursor numbers were fewer than those from CD98hc+ control T cells (data not shown). However, CD98-null T cells do not appear to have an intrinsic survival defect in vivo, as they survive comparably to control T cells 4 days after into congenic Ly5.1 recipients (Fig. 7B). It is thus more likely that the lower-than-expected numbers of recovered T cells from pancreatic lymph nodes at day 4 reflects substantial migration to the pancreas after activation, as CFSE+ T cells are numerous in islets on day 4 (Fig. 7A). The poorly-expanded CD98-null T cells likely cannot efficiently replace this effector cell migration to the pancreas and maintain their numbers in the lymph nodes. If failure of clonal expansion is the mechanism of protection from T1D in CD98hc-null T cells, then transfer of pre-expanded CD98hc-null OT-1 T cells should cause disease similar to control OT-1 T cell transfer. To test this idea, we cultured OT-1 dlck-Cre+, Slc3a2^fl/fl or control OT-1 Slc3a2^fl/fl T cells at high density for 6 days in vitro with SIINFEKL peptide + IL-2. At day 6, any remaining CD98hc+ T cells were eliminated from dlck-Cre+, Slc3a2^fl/fl preparations by depletion with anti-CD98hc antibody, and CD98-null OT-1 (7.5 × 10⁶) or control OT-1 CTL (7.5 × 10⁶) were transferred. Pre-expansion of dlck-Cre+, Slc3a2^fl/fl T cells restored the ability to cause T1D (Fig. 7C). Thus, T Cell CD98hc mediates T1D by enabling antigen-driven proliferation and clonal expansion of T cells.

Discussion

Here we have found that T cell CD98hc (4F2 Antigen, Slc3a2) is dispensable for T cell population, activation, and homeostatic proliferation; however, it is required for clonal expansion and consequent development of experimental autoimmune diabetes. Thus, deletion of CD98hc or inhibition of its function provides a tool to examine the functional effects of blocking T cell clonal expansion without impairing antigen recognition, T cell activation, or intrinsic effector function. Previous studies have shown that the capacity of CD98hc to support proliferation of other cell types (33, 35, 38) depends on its capacity to support integrin signaling rather than amino acid transport. Integrin-mediated signals control adhesion/migration, survival, and proliferation of T cells. In the NOD mouse model for type I diabetes, β2 and αL integrins are required for disease (39), and blockade of α4 integrins delay or prevent diabetes (40–42). Integrin ligands are also overexpressed in inflamed pancreatic islets during type I diabetes in the NOD mouse and in humans (43, 44), and blockade of an α4 integrin ligand, mucosal addressin cell adhesion molecule (MAdCAM-1), reduces diabetes incidence in the NOD mouse (45). Integrins have thus garnered attention as possible targets for therapeutic blockade in type I diabetes and other tissue-specific autoimmune diseases. (46). This total blockade approach can be effective (47–49), but can also lead to life threatening infections such as progressive multifocal leukoencephalopathy (50). The present studies suggest that targeting CD98hc-integrin interaction might selectively ablate clonal expansion and preserve other integrin-dependent T cell functions, such as formation of immune synapses and homing to peripheral tissues (e.g. pancreas, brain).

Importantly, the defect in antigen-driven proliferation was quantitative, in contrast to the complete lack of B cell proliferation that follows CD98 deletion (35). The capacity of the CD98hc-null T cells to proliferate in vitro enabled us to show that a selective block in clonal expansion is the mechanism whereby targeting CD98 protects from experimental T1D. CD98-null T cells could migrate both to pancreatic lymph nodes and to pancreatic islets, but did not proliferate normally. When CD98-null OT-1 T cells were pre-expanded in vitro and numbers increased to match controls, their pathogenicity was restored and they were able to cause diabetes similar to control OT-1 T cells. It is also noteworthy that CD98hc-null T cells were able to develop into Foxp3+ suppressor cells in vitro, confirming that full T cell proliferation is not essential for suppressor T cell development and suggesting that blocking CD98 might serve to impair expansion of effectors while preserving suppressor cell development (51–54). In aggregate, our studies identify CD98hc as a potential target in selectively blocking clonal expansion during T1D, and potentially in other T cell-mediated autoimmune diseases.