Stomatin-like Protein 2 Deficiency in T Cells Is Associated with Altered Mitochondrial Respiration and Defective CD4⁺ T Cell Responses

Stomatin-like protein 2 (SLP-2) is an abundant protein in the proteome of detergent-insoluble microdomains of mitochondrial membranes (1). It is a member of the highly conserved stomatin family, which is composed of stomatin, SLP-1, SLP-2, and SLP-3 (2–7) and as such is part of the SPFH superfamily that also includes prohibitins (PHB), flotillins, and the bacterial proteins HflC and HflK (8, 9). The function of these proteins has been linked to the regulation of assembly and function of multichain receptors (6, 10–13). In line with this proposed function, we have recently shown that SLP-2 binds cardiolipin (CL) and promotes the formation of specialized CL-enriched microdomains that could facilitate the assembly of protein complexes within mitochondrial membranes (10).

SLP-2 can be detected in many tissues including lymph nodes, thymus, and tonsils. In addition, it is expressed at low levels in human PBMCs including T cells, B cells, and monocytes, and its expression is upregulated upon activation in vitro (1). Consistent with this distribution, it has been shown that SLP-2 plays a role in T cell activation (1): upregulation of SLP-2 correlates with enhanced activation and functional responses, whereas downregulation of SLP-2 expression is associated with decreased duration of TCR signaling as well as a decrease in T cell activation.

To study the mechanism of action of SLP-2 in vivo, we initially generated an SLP-2 knockout mouse strain. However, this strain was embryonic lethal at the preimplantation stage (C.D. Lemke and J. Madrenas, unpublished observations). To circumvent this limitation, we generated a T cell-specific SLP-2 knockout mouse model. Using this model, we report in this study that SLP-2–deficient T cells show abnormal CL compartmentalization in mitochondrial membranes, decreased levels of the NADH dehydrogenase (ubiquinone) iron-sulfur protein 3, NADH dehydrogenase (ubiquinone) 1β subcomplex subunit 8, and NADH dehydrogenase (ubiquinone) 1α subcomplex subunit 9 of complex I of the electron transport chain and decreased activity of this complex, as well as decreased activity of complex II plus III of this chain. SLP-2–deficient T cells also show increased uncoupled respiration (respiration that is not associated with ATP production) and greater reliance on glycolysis, which may partially compensate for the altered mitochondrial respiration. Together, these defects are associated with a selective posttranscriptional reduction in IL-2 production and decreased responses to stimulation through the TCR in vivo and ex vivo. We propose that SLP-2 contributes to T cell activation primarily by regulating the compartmentalization of CL in mitochondrial membranes and ensuring proper assembly and function of the respiratory chain components in these membranes. This, in turn, ensures optimal metabolic response during T cell activation.

A full-length genomic fragment containing the mouse SLP-2 gene was cloned within the lox sites of the 3LoxP3NwCD vector. Upon sequence confirmation, the plasmid was electroporated into C57BL/6 embryonic stem cells, and clones were selected for neomycin resistance. The resistant clones were screened by Southern blot and PCR analysis to confirm homologous recombination of the SLP-2–floxed DNA sequence. Clones containing the SLP-2–floxed sequence were injected into B6/Try blastocysts and implanted into a pseudopregnant mouse. The offspring were selected by chimeric coat color and backcrossed with C57BL/6 mice to produce SLP-2^lox/wt mice in the C57BL/6 background. SLP-2–floxed mice were crossed with C57BL/6 mice transgenic for Cre recombinase under the control of the CD4 promoter, purchased from Taconic Farms (Hudson, NY) (14), to generate a T cell-specific knockout of SLP-2 strain (SLP-2 T-K/O). SLP-2 floxed mice lacking Cre were used as control mice. Breeding colonies were derived from the same original SLP-2–floxed breeders and kept in parallel. C3H/HeJ mice were obtained from The Jackson Laboratory. Mice were maintained in the animal facility at the University of Western Ontario with approval from the Animal Use Subcommittee in accordance with the Canadian Council on Animal Care Guidelines.

T cells were isolated by magnetic separation using the MACS Pan T cell isolation kit (Miltenyi Biotec, Auburn, CA). To confirm deletion at the genetic level, total genomic DNA was extracted using the Qiagen DNeasy blood and tissue kit (Qiagen, Mississauga, ON), and full-length SLP-2 PCR was performed using the following primers: forward primer 5′-ACTTCCACCCTTCAGTCCAGGTCG-3′ and reverse primer 5′-ACTTGGATTCTGTGAAAGCAGACAC-3′. Samples were separated by DNA electrophoresis in a 1% agarose gel and visualized by ethidium bromide staining. To confirm deletion at the protein level, thymocytes and T cells were pelleted in PBS containing sodium orthovanadate (400 μM) and EDTA (400 μM) and resuspended in lysis buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris [pH 7.6], 5 mM EDTA, 1 mM sodium orthovanadate, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 25 μM p-nitrophenyl-p’ guanidinobenzoate) at 4°C for 30 min. Lysates were then cleared of debris by centrifugation (10,700 × g, 4°C, 10 min), mixed with sample buffer containing 2-ME, and boiled. Cell lysates were resolved by standard SDS-PAGE and immunoblotted with rabbit polyclonal anti–SLP-2 antisera (Proteintech Group, Chicago, IL) and mouse monoclonal anti-GAPDH Ab (Chemicon International, Temecula, CA). Detection of proteins after Western blotting was performed using the BM Chemiluminescence Substrate POD system (Roche, Laval, QC, Canada).

Intact mitochondria were isolated from primed T cells using the Qproteome Mitochondria Isolation Kit (Qiagen). Detergent-insoluble fractions were isolated by incubation on ice in 25 mM HEPES buffer with 0.1% Triton X-100, as previously described (15). Fractions were analyzed by Western blotting or TLC.

Soluble and insoluble mitochondrial fractions were washed in 2 ml methanol/water (1:1, by volume), and 25-μl aliquots were taken for protein determination (16). Subsequently, 0.5 ml water and 2 ml chloroform were added to initiate phase separation. Samples were centrifuged at 2000 rpm for 10 min, and the upper phase was aspirated. Two milliliters theoretical upper phase (methanol/0.9% NaCl/chloroform 48:47:3, by volume) was added, and centrifugation was repeated for 5 min. The organic phase was removed, dried under nitrogen gas, and resuspended in chloroform/methanol (2:1, by volume). A 50-μl aliquot was spotted onto thin-layer plates for separation of CL from other phospholipids as described (17). Phospholipids were removed from the plate and phospholipid phosphorus content of CL determined as described (18).

NADH dehydrogenase activity was measured from whole-cell lysates using the complex I enzyme activity microplate assay kit (Mitosciences). The activity of complex II plus III (succinate cytochrome c reductase), measured as reduction of cytochrome c after addition of succinate, was determined as described (19).

ATP levels in whole-cell lysates were measured using an ATP Determination Kit from Molecular Probes (Eugene, OR). Briefly, 10 μl ATP standards (from 1.95 to 1000 nM) or lysates were added in triplicate to 90 μl standard reaction solution in a 96-well Chromalux Luminescent Assay microplate (Dynex Technologies, Chantilly, VA). Sample luminescence was measured using an MLX Microtiter Plate Luminometer (Dynex Technologies). Sample ATP concentrations were normalized to cell equivalents.

Thymocytes, splenocytes, and lymph node cells were isolated from control and SLP-2 T-K/O mice and stained with the following Abs: PE-conjugated or allophycocyanin-conjugated rat anti-CD4 Ab, FITC-conjugated or PerCP-Cy5.5–conjugated rat anti-CD8 Ab, allophycocyanin-conjugated Armenian hamster anti-CD3 Ab, PE-conjugated rat anti-CD25 Ab, FITC-conjugated rat anti-CD44 Ab, PE-Cy7–conjugated mouse anti-Tbet, eFluor 660-conjugated rat anti-GATA3, PE-conjugated rat anti-retinoic acid-related orphan receptor γt (RORγt), or with isotype control Abs (all from BD Biosciences, San Jose, CA).

Flow cytometry was performed with a BD FACSCalibur flow cytometer and CellQuest computer software (BD Biosciences) and analyzed with FlowJo flow cytometry analysis software (Tree Star, Ashland, OR).

Splenocytes isolated from control and SLP-2 T-K/O mice were stimulated with staphylococcal enterotoxin E (SEE) superantigen (Toxin Technologies, Sarasota, FL) or 0.01–1 μg/ml anti-CD3 Ab (Cedarlane Laboratories, Burlington, ON, Canada) and 0.2 μg/ml anti-CD28 Ab (eBioscience, San Diego, CA) in 96-well round-bottom or flat-bottom plates for 24 h at 37°C (20). After 24 h stimulation, supernatants were collected for IL-2, IL-17, IL-4, and IFN-γ measurement by ELISA (21). Cultures were then supplemented with [³H]thymidine and incubated for 24 h to measure cell proliferation.

For Th polarization experiments, CD4⁺ T cells were purified with the MACS CD4⁺ T cell isolation kit. Cells were incubated in vitro with 1 μg/ml plate-bound anti-CD3 and 1 μg/ml anti-CD28 under polarizing conditions for 4 d: Th1, 3.5 ng/ml IL-12, 2 μg/ml anti-IL-4; Th2, 10 ng/ml IL-4, 2.5 μg/ml anti–IFN-γ; and Th17, 3 ng/ml TGF-β, 20 ng/ml IL-6, 10 ng/ml anti-CD23, 2 μg/ml anti–IL-4, and 2.5 μg/ml anti–IFN-γ (22, 23). Cells were washed and restimulated with 1 μg/ml plate-bound anti-CD3 for 24 h, and supernatants were collected for IFN-γ, IL-4, and IL-17 measurement, whereas cells were stained for transcription factor expression.

T cells from control and SLP-2 T-K/O mice were stimulated with plate-bound anti-CD3 and anti-CD28 (5 and 2 μg/ml, respectively) for 48 h. Following incubation, T cells were suspended in Seahorse XF assay media supplemented with 25 mM glucose and 2 mM l-glutamine and seeded onto XF24 V7 24-well cell-culture microplates (Seahorse Bioscience, North Billerica, MA) coated with 50 μg/ml poly-d-lysine at 500,000 cells/well. T cells were placed at 37°C in a CO₂-free atmosphere for 1 h prior to analysis. Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR, a measure of lactic acid formed during glycolysis) were measured prior to and following the sequential addition of 10 μM oligomycin, 15 μM carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone, 100 nM rotenone, and 1 μM antimycin A using the Seahorse XF-24 analyzer (Seahorse Bioscience). Total OCR refers to the steady-state cellular OCR measured prior to the addition of mitochondrial inhibitors. The oligomycin-sensitive respiration is a measure of coupled respiration (i.e., coupled to ATP synthase activity), whereas the oligomycin-insensitive respiration is a measure of the uncoupled respiration. Mitochondrial respiration was inhibited with complex I and III inhibitors rotenone and antimycin A, respectively, the remaining OCR being attributed to nonmitochondrial sources. Mitochondrial OCR was calculated as the difference between total OCR and nonmitochondrial OCR.

Mice were immunized s.c. with 100 μg OVA in emulsion with Freund’s complete adjuvant (Difco; BD Biosciences). Eight to 10 d later, mice were sacrificed, lymph nodes collected, and single-cell suspensions prepared for in vitro stimulation with OVA or pigeon cytochrome c (PCC). Culture supernatants were collected after 48 h for cytokine assays and cultures supplemented with [³H]thymidine for proliferation assay 24 h later. APCs were treated with 50 μg/ml mitomycin C for 30 min at 37°C followed by washing and used at a 1:4 ratio with CD4⁺ T cells. For measurement of activated T cell populations upon secondary stimulation, mice were immunized with 50 μg OVA in emulsion with CFA. Ten days after the primary immunization, mice were boosted with 50 μg OVA in emulsion with CFA, and 7 d later, lymph nodes were isolated and analyzed for CD4-, CD8-, CD44-, and CD25-expressing populations.

Splenocytes were stimulated with 1 μg/ml anti-CD3 Ab and 0.2 μg/ml anti-CD28 Ab in flat-bottom plates for 24–36 h at 37°C in the presence of brefeldin A (eBioscience). Cells were surface stained for CD4-allophycocyanin and CD8-PerCP-Cy5.5 followed by intracellular staining for IL-2–PE and IFN-γ–FITC using Invitrogen fix and permeabilization reagents (Life Technologies, Grand Island, NY).

Splenocytes were stimulated with 1 μg/ml anti-CD3 Ab and 0.2 μg/ml anti-CD28 Ab in flat-bottom plates for 6 h at 37°C. Total RNA was isolated from stimulated cells or an equal number of unstimulated cells using the Qiagen RNeasy kit, and RNA was quantified using a NanoDrop (Thermo Scientific, Wilmington, DE). cDNA was synthesized from 500 ng RNA using the iScript cDNA synthesis kit (Bio-Rad, Mississauga, ON, Canada), and transcript levels were quantified on the Bio-Rad CFX96 real-time PCR detection system using iTaq SYBR Green Supermix with ROX (Bio-Rad). IL-2 primer sequences, forward 5′-CCTGAGCAGGATGGAGAATTACA-3′ and reverse 5′-TCCAGAACATGCCGCAGAG-3′ (22), and mouse GAPDH real-time primers were purchased from SABiosciences (Qiagen). Relative levels of IL-2 transcript were determined by comparing the threshold cycle detection values of IL-2 in wild-type (WT) and SLP-2 T-K/O cells in stimulated and unstimulated cells and were normalized to GAPDH levels by the ΔΔthreshold cycle method (24).

Single-cell suspensions were made from control and SLP-2 T-K/O mouse lymph nodes 8 d after OVA immunization. Cells were stimulated with SEE, anti-CD3 Ab and anti-CD28 Ab or OVA for up to 30 min. Cells were lysed and lysates used for serial immunoblotting with rabbit anti-active ERK-1 and ERK-2 E10 (Cell Signaling Technology) Ab and total ERK 1/2 antisera (StressGen Biotechnologies).

Abdominal heterotopic transplants were performed as described (25). After transplant, graft viability was assessed daily by abdominal palpation of the transplant heartbeat. Graft rejection was determined by lack of heartbeat.

Data analysis was performed with GraphPad Prism software (GraphPad Software, La Jolla, CA). Data are presented as mean ± SEM. Two-tailed Student t test, paired t test, or ANOVA with Bonferroni post hoc testing was used to determine p values. The p values <0.05 were considered significant.

To study the effect of SLP-2 in vivo, we first tried to generate conventional SLP-2 knockout mice, but no viable embryos were obtained after >300 term pregnancies (C.D. Lemke and J. Madrenas, unpublished observations), suggesting that SLP-2 deficiency is embryonic lethal at the preimplantation stage. To circumvent this problem, conditional SLP-2 knockout mice were generated in which the whole SLP-2 gene was replaced by a lox-flanked genomic SLP-2 gene. The resulting SLP-2^lox mouse was bred together with a transgenic mouse containing the Cre-recombinase gene under the control of the CD4 promoter (14) to generate mice in which T cells lacked SLP-2 expression. These resulting knockout mice are called SLP-2 T-K/O throughout this study.

SLP-2 knockout in the T cell lineage was confirmed at the genomic level by PCR analysis and at the protein level by Western blot analysis. Using genomic PCR with primers flanking the entire SLP-2 sequence to give a 4-kb fragment for full-length SLP-2 and a 500-bp fragment for SLP-2 deletion, we detected the full-length SLP-2 fragment in CD4⁺ thymocytes, naive and primed T cells, and splenic non-T cells from control mice. In contrast, we failed to detect the full-length SLP-2 in SLP-2 T-K/O CD4⁺ thymocytes and naive and blast T cells. In these mice, only the truncated product, indicative of genomic deletion of SLP-2, was observed (Fig. 1A). In the non-T cell population from spleen, we still detected the truncated SLP-2 fragment in addition to the full-length SLP-2 fragment, likely due to contamination of the prep by T cells (up to 2% of total cell numbers in the prep).

The deletion of genomic SLP-2 translated into significantly lower SLP-2 protein expression in SLP-2 T-K/O CD4⁺ thymocytes and complete loss of SLP-2 expression in peripheral T cells but not in spleen CD4⁻CD8⁻ cells (Fig. 1B), confirming the restriction of SLP-2 deletion to conventional T cells. Moreover, and as reported with primary human T cells (1), these experiments confirmed upregulation of SLP-2 protein levels upon activation of WT T cells (see blast versus naive in Fig. 1B).

Proteins of the SPFH superfamily have been linked to the organization of membrane microdomains (26–31). Based on this assumption, and because SLP-2 is mostly a mitochondrial protein localized in the intermembrane space (32, 33), we examined if SLP-2 deficiency would alter mitochondrial membrane compartmentalization. We first looked at the effect of SLP-2 deficiency on PHB because these proteins interact with each other and contribute to the formation of detergent-insoluble microdomains in human mitochondria (15, 34). Triton X-100–insoluble domains from T cell mitochondria from control and SLP-2 T-K/O mice were prepared and blotted for PHB1 and SLP-2. Although total cellular levels of PHB1 were not affected in SLP-2–deficient T cells (see below, Fig. 2E), the level of PHB1 in the mitochondrial detergent-insoluble microdomain fraction was reduced, although the levels in the detergent-soluble fraction were similar between control and SLP-2 T-K/O cells (Fig. 2A). A loss of PHB from the insoluble domain would predict a gain in PHB in the soluble fraction. However, this assay was performed in such a way that the entire insoluble fraction was analyzed on the Western blot, whereas only a portion of the soluble fraction was analyzed. Thus, the change detected in the insoluble fraction is not detected as an increase in the soluble fraction due to the much larger pool of soluble PHB compared with insoluble. As the amount of PHB1 in detergent-soluble fractions of control and SLP-2 T-K/O cells was similar, the loss of PHB1 from the insoluble fraction suggested an abnormal compartmentalization in the SLP-2 T-K/O mitochondrial membrane.

We have recently shown that SLP-2 not only interacts with PHB but also binds CL, the signature phospholipid linked to detergent-insoluble microdomains of mitochondrial membranes (10). Therefore, the defect in mitochondrial membrane compartmentalization of PHB1 associated with SLP-2 deficiency prompted us to test if the content or distribution of CL was also impaired. Total CL content in T cells from control and SLP-2 T-K/O mice as detected with nonyl-acridine orange staining was not significantly different (Fig. 2B). However, the CL content in mitochondrial detergent-insoluble microdomains isolated from primed SLP-2 T-K/O T cells was significantly (p < 0.05) lower than that of control samples (Fig. 2C). Altogether, these results imply that SLP-2 deficiency in T cells is associated with impaired mitochondrial membrane compartmentalization of CL.

Proper compartmentalization of mitochondrial membranes into CL-enriched microdomains is required for optimal assembly and function of the components of the respiratory chain (35, 36). Therefore, we examined the effects of SLP-2 deficiency in assembly and function of the complexes of the respiratory chain. First, we analyzed the levels of mitochondrial respiratory chain components in SLP-2–deleted T cells using a mixture of Abs against different subunits of complexes I to V of the respiratory chain. SLP-2 T-K/O cells had lower levels of the NDUFB8 subunit of complex I and slight decrease of the 30-kDa subunit of complex II compared with control cells (Fig. 2D). We corroborated this finding for two other subunits of the 45 described subunits of complex I, including the NDUFS3 and NDUFA9 subunits (Fig. 2E). Note that the total cellular levels of PHB1 were not affected by SLP-2 deficiency. SLP-2 deficiency did not grossly affect the levels of complexes III to V, including ATP synthase, which were similar in control and SLP-2 T-K/O T cells.

The observed decrease in levels of complex I subunits translated into a significant (p < 0.01) decrease in the functional activity of complex I NADH dehydrogenase in SLP-2 T-K/O T cells (Fig. 2F). In addition, an assay of the combined activity of succinate dehydrogenase and cytochrome c reductase (complex II plus III activity) also showed a significant decrease in SLP-2–deficient T cells (p < 0.05, Fig. 2G).

Despite the decrease in complex I and complex II plus III activities, neither naive nor activated SLP-2–deficient T cells showed a decrease in total cellular ATP levels (Fig. 2H). To further characterize this finding, we assessed mitochondrial and nonmitochondrial respiration as well as glycolysis in these cells using Seahorse XF technology (Seahorse Bioscience) (Fig. 3). We found that SLP-2–deficient T cells had similar levels of total and mitochondrial OCR as WT T cells (Fig. 3A, 3B). However, SLP-2–deficient T cells exhibited significantly (p < 0.05) higher nonmitochondrial OCR (Fig. 3C). More importantly, these cells also had a significant (p < 0.05) increase in mitochondrial uncoupled respiration (Fig. 3D, 3E) (i.e., respiration that is not coupled with mitochondrial ATP production). The uncoupled respiration in SLP-2 T-K/O T cells represented 24% of mitochondrial respiration versus 17% in WT T cells, whereas the coupled respiration was substantially reduced in SLP-2 T-K/O mice (76%) compared with WT mice (83%). The increase in uncoupled respiration was associated with elevated glycolysis, as we found significantly increased ECAR, a measure of glycolysis, in 71% of SLP-2 T-K/O mice (p < 0.05) (Fig. 3F). Together, these results indicated that SLP-2–deficient T cells have altered respiration with a compensatory greater reliance on glycolysis.

Next, we examined the effect of SLP-2 deletion and the altered mitochondrial respiration on T cell function. First, we looked at the cellularity of thymus, spleen, and lymph nodes. Organ sizes were similar between control and SLP-2 T-K/O mice. The total thymocyte and splenocyte cell counts (Fig. 4A) as well as the number of double-negative, double-positive, CD4⁺ single-positive, and CD8⁺ single-positive thymocytes and splenic cells (Fig. 4B, 4C) were similar in unmanipulated SLP-2 T-K/O and control mice, implying a grossly normal T cell development in these mice. Although the total cell count in lymph nodes was similar in control and SLP-2 T-K/O mice (Fig. 4A), there was a slight but consistent decrease in CD4⁺ T cells in the lymph nodes of SLP-2 T-K/O mice (p = 0.09, Fig. 4D).

Next, we tested the functional responses of SLP-2–deficient T cells. Naive splenocytes from control and SLP-2 T-K/O mice were stimulated ex vivo with increasing concentrations of SEE superantigen. We found that SLP-2 T-K/O mice produced significantly less IL-2 upon superantigen stimulation compared with control mice (p < 0.05) (Fig. 5A). The decrease in IL-2 production, however, did not translate into significantly less T cell proliferation (Fig. 5B).

We have previously reported that SLP-2 levels increase upon priming (1), implying that the role of this protein is greater upon T cell priming. To test the secondary response of SLP-2–deficient T cells, control and SLP-2 T-K/O splenocytes were stimulated in vitro with PMA and ionomycin for 72 h followed by 48 h resting before being stimulated with SEE superantigen. Again, primed SLP-2 T-K/O cells produced significantly less IL-2 compared with control cells (p < 0.001, Fig. 5C). We observed a trend toward less proliferation in SLP-2 T-K/O cells, but it did not reach statistical significance (Fig. 5D). The significant decrease in IL-2 responses seen in SLP-2 T-K/O T cells was also observed when naive and in vitro-primed cells were stimulated with anti-CD3 and anti-CD28 Abs (data not shown). Together, these results demonstrate that SLP-2–deficient T cells have decreased activation leading to decreased IL-2 responses.

To determine if SLP-2–deficient T cells showed impaired T cell differentiation, control and SLP-2 T-K/O splenocytes were stimulated with SEE under nonpolarizing conditions, and Th1, Th2, or Th17 cytokine production was measured. Under these conditions, no IL-4 was detected, but SLP-2 T-K/O cells produced similar levels of IFN-γ and IL-17 compared with control cells (Fig. 6A, 6B). Likewise, primed cells also showed similar IFN-γ and IL-17 production in response to SEE stimulation (Fig. 6C, 6D). To specifically examine CD4⁺ T cell differentiation, isolated CD4⁺ T cells were stimulated in vitro under Th1-, Th2-, or Th17-polarizing conditions and analyzed for cytokine production and transcription factor expression. SLP-2 T-K/O cells produced similar levels of IFN-γ, IL-4, and IL-17 compared with control cells under the appropriate differentiation conditions (Fig. 6E–G). As well, the expression of Tbet, GATA3, and RORγt transcription factors was similar between control and SLP-2 T-K/O mice (Fig. 6H–J). Together, these results demonstrate that, once activated, SLP-2–deficient CD4⁺ T cells can undergo normal T cell differentiation.

To assess how the impairment in T cell responses observed ex vivo translated into T cell responses in vivo, we first assessed the capacity of SLP-2–deficient T cells to respond to immunization. Mice were immunized with OVA and ex vivo recall responses tested 10 d after immunization. The response to PCC was used as a control. In these experiments, SLP-2–deficient cells had significantly decreased responses (p < 0.05) measured by both IL-2 production and proliferation (Fig. 7A, 7B).

As an additional model for in vivo T cell activation, we examined the response to vascularized allografts in SLP-2 T-K/O mice compared with that in SLP-2 competent mice. Heterotopic heart transplants from C3H mice (H-2^k) into C57BL/6 (H-2^b) WT and SLP-2 T-K/O mice were performed and graft survival monitored. In these experiments, we observed a slight but consistent and significant (p < 0.05) delay in the rejection of allografts in SLP-2 T-K/O mice (Fig. 7C). Together, these results corroborated that SLP-2 deficiency in T cells is associated with impaired responses to TCR-dependent stimulation in vivo.

The lymph nodes from control and SLP-2 T-K/O mice had comparable numbers of CD19⁺ B cells and CD3⁺ T cells both before and after immunization (data not shown). Although there was no difference in the total number of T cells, SLP-2 T-K/O mice had a consistent and significant decrease in the percentage of CD4⁺ T cells but not in the percentage of CD8⁺ T cells after immunization (Table I). The decrease in the percentage of CD4⁺ T cells was primarily due to decreased proportion of activated T cells, as demonstrated by the decrease in CD4⁺CD25⁺ and CD4⁺CD44^hi T cell populations. The percentage and function of regulatory CD4⁺CD25⁺Foxp3⁺ T cells were comparable between control and SLP-2 T-K/O mice (Table I and data not shown). Furthermore, the decrease in activated CD4⁺ T cells was not due to an increased susceptibility to activation-induced cell death (data not shown).

To further examine the impairment of secondary CD4⁺ T cell responses, mice were immunized s.c. with OVA/CFA and 10 d later boosted with OVA/CFA. Seven days after receiving the boost, lymph nodes were isolated and analyzed for CD4, CD8, CD25, and CD44 expression. As shown in Table II, SLP-2 T-K/O mice had a significant decrease in the percentage of CD4⁺ T cells (p < 0.05) as well as the percentage of activated subsets CD4⁺CD25⁺ and CD4⁺CD44^hi (p < 0.01 and p < 0.05, respectively). Furthermore, a comparison of changes in these two subsets between primary and secondary responses in SLP-2 T-K/O mice demonstrated a more intense defect in secondary responses (with decreases of 47.2 and 34% in CD4⁺CD25⁺ and CD4⁺CD44^hi T cells, respectively) than in primary responses (with decreases of 16.7 and 5.5% for the same subsets; Table I). Furthermore, although the total number of the CD4⁺ cells were similar between control and SLP-2 T-K/O mice (2.4 ± 0.45 × 10⁶ compared with 2.2 ± 0.24 × 10⁶), there was a higher number of CD8⁺ T cells in SLP-2 T-K/O mice (2.4 ± 0.48 × 10⁶ compared with 3.2 ± 0.35 × 10⁶), further indicating an impairment in CD4⁺ T cell responses in SLP-2 T-K/O mice.

To further investigate the defect in IL-2 responses in SLP-2 T-K/O mice, we examined IL-2 production at the single-cell level with intracellular staining and flow cytometry. Naive splenocytes were isolated from WT and SLP-2 T-K/O mice and stimulated for 24 h in the presence of anti-CD3 and anti-CD28 Abs. In agreement with the IL-2 data presented above, significantly fewer CD4⁺ and CD8⁺ SLP-2 T-K/O cells produced IL-2 in response to stimulation compared with WT T cells (p < 0.001 and p < 0.01, respectively, Fig. 8A–F). In addition, the geometric mean of the IL-2 fluorescence staining was slightly but significantly lower in the SLP-2 T-K/O CD4⁺IL-2⁺ and CD8⁺IL-2⁺ cell population compared with WT populations (p < 0.001 and p < 0.01, respectively, Fig. 8G, 8H). Together, these data indicate that fewer SLP-2–deficient T cells became activated to express IL-2, and those T cells that were activated produced less IL-2 compared with WT T cells. Surprisingly, when we investigated the transcript levels of IL-2 in stimulated T cells, we found no significant difference in IL-2 transcript levels between SLP-2 T-K/O and WT cells (Fig. 8I), which may point to a possible role for SLP-2 in the processing of IL-2 mRNA.

Although SLP-2 regulates human T cell signaling in cell lines (1), we found that mouse SLP-2–deficient T cells had comparable levels of ERK-1/-2 phosphorylation (i.e., activation) in response to both SEE stimulation and anti-CD3/anti-CD28 stimulation in vitro during early time points and up to 15 min (data not shown and Fig. 8J). This result suggests that the defect in IL-2 production is not due to a drastic impairment of TCR-dependent ERK activation in SLP-2–deficient T cells. An assessment of signaling through other pathways is needed to fully address the effect of SLP-2 deficiency on TCR signaling.

Of interest, the defect in IL-2 production in SLP-2–deficient T cells seemed to be limited to this cytokine. In addition to normal IFN-γ and IL-17 secretion from stimulated SLP-2 T-K/O cells, we found that a similar proportion of WT and SLP-2 T-K/O cells became IFN-γ–expressing T cells in response to stimulation (Fig. 8K, 8L). Furthermore, quantification of fluorescent signal from IFN-γ⁺ cells indicated a similar expression level in both WT and SLP-2 T-K/O cells (Fig. 8M, 8N).

We report in this study the characterization of SLP-2 deficiency in T cells in vivo using a conditional knockout mouse strain with the Cre-Lox system and Cre recombinase under the control of the CD4 promoter. In this mouse strain, deletion of the SLP-2 gene occurred during the double-positive stage of thymocyte development. The SLP-2 T-K/O mice were viable and bred well, giving sufficient numbers for biochemical and functional studies of the effects of SLP-2 deficiency in T lymphocytes.

We found that SLP-2 deficiency in T cells is associated with abnormal CL compartmentalization in mitochondrial membranes. SLP-2 is tightly associated with the mitochondrial inner membrane (32, 33), and other members of the SPFH family have been linked to the organization of various membrane microdomains (26–31). Therefore, it is plausible to suggest that SLP-2 promotes the compartmentalization of the mitochondrial inner membrane into CL-enriched microdomains. This proposed role is consistent with the effects of SLP-2 deficiency on mitochondrial function because formation of CL microdomains is required for optimal assembly of the electron transport complexes in the mitochondrial inner membrane (37). Thus, in the absence of SLP-2, mitochondrial inner membrane compartmentalization is altered, and this interferes with the optimal assembly of the respiratory chain, ultimately resulting in altered respiration.

Although we have previously shown that SLP-2 overexpression leads to an increase in CL biosynthesis (10), we did not find a decrease in total CL levels in the SLP-2 T-K/O mice. This may reflect a difference in the response of cells to SLP-2 overexpression versus depletion: whereas increased SLP-2 expression activates CL biosynthesis, basal CL biosynthesis proceeds as normal in the absence of SLP-2. In this way, SLP-2 overexpression may use additional pathways to increase CL levels.

How does SLP-2 contribute to the formation of CL-enriched microdomains? We have recently shown that SLP-2 binds CL and interacts with the mitochondrial resident proteins PHB1 and 2 (10, 32). Thus, SLP-2 may link PHB with CL-enriched microdomains. This model supports the proposed involvement of PHB in the compartmentalization of the mitochondrial inner membrane (38). Also, by facilitating mitochondrial membrane organization, SLP-2 enhances mitochondrial function by optimizing respiration and increasing resistance to apoptosis as shown under conditions of SLP-2 upregulation in vitro (10). In addition, this model helps to explain the observations that CL is required for proper function of the respiratory chain complexes (39) and that CL-enriched microdomains play a role in the induction of apoptosis (15, 34, 40).

Deficiency in SLP-2 is associated with altered mitochondrial respiration. This is illustrated by defective activities of complexes I and II plus III of the electron transport chain and by increased respiration uncoupled to ATP production. The physiological relevance of a reduction of complex I activity is likely greater than what is measured because the assay commonly used to determine the activity of complex I only detects partial reaction of complex I (41–43). Defective complex I activity is considered the most common defect of the respiratory chain in humans (44) and causes a diverse set of diseases, including Leigh syndrome and renal tubular acidosis. Complex I represents the major entry point into the respiratory chain through the oxidation of NADH to NAD⁺ with the transfer of two electrons into the respiratory chain (45). In the absence of SLP-2, there is inefficient assembly of complex I subunits, confirming a previous suggestion that downregulation of SLP-2 reduced the level of components of the respiratory chain (32). CL-enriched microdomains may be especially required for optimal assembly of complex I because of the large, multichain nature of this complex, which includes at least 45 subunits (45).

The defect in mitochondrial membrane compartmentalization of CL in the absence of SLP-2 not only affects the function of complex I but also of complex II plus III and likely of other protein supercomplexes in the mitochondrial membrane as suggested by the data presented in this study and by preliminary proteome characterization of detergent insoluble microdomains (S.D. Dunn and J. Madrenas, unpublished observations).

In T cells, the impairment in mitochondrial membrane compartmentalization and mitochondrial respiration associated with SLP-2 deficiency did not lead to a complete block in T cell responses. Our data indicate that this may be due to a compensatory greater reliance in glycolysis in SLP-2–deficient T cells that may mask the defect in respiratory chain activity. Furthermore, the partial phenotype seen in our mouse model is in line with what is observed in mitochondrial diseases of humans in which the penetrance and phenotypic manifestations of the disease within the same pedigree is variable ranging from almost no phenotype to death (46). A paucity of immune manifestations is also observed in humans with complex I deficiency. Although these patients have a reduction of complex I activity in peripheral blood leukocytes, they do not show overt immune deficiency (47, 48), likely due to a compensatory reliance on glycolysis for ATP production (49). However, these effects are not negligible as illustrated by the recent observation that complex I plays an important role in the activation of T cells (50, 51) and also by the decrease in CD4⁺ T cell expansion under conditions of low NAD⁺ levels from decreased complex I activity (47, 52–54).

A surprising finding in SLP-2–deficient T cells is the selective posttranscriptional defect in IL-2 production in response to TCR stimulation, documented by the decrease in both intracellular and secreted levels of IL-2 protein with abundant IL-2 transcripts. A similar phenotype has been reported in anergic T cells in which IL-2 transcript levels were similar in T cells stimulated under both activating and anergic conditions, but IL-2 protein was only detected in activated T cells (55, 56). This posttranscriptional block in cytokine production affected other cytokines, such as IFN-γ, IL-4, and IL-17 and was dependent on AU-rich regions in the 3′ untranslated region of cytokine mRNA (55–57). In contrast to anergic conditions, the posttranscriptional block in SLP-2–deficient T cells appears to be limited to IL-2 expression, as IFN-γ and IL-17 were expressed normally. The greater reliance on glycolysis found in SLP-2–deficient T cells may explain why SLP-2 T-K/O cells produce normal levels of IFN-γ, because work from other groups has shown that IFN-γ production is more sensitive to glucose levels than IL-2 production (58, 59). The precise nature of the SLP-2–dependent metabolic checkpoint of IL-2 mRNA processing is currently unknown.

Why the metabolic demands upon activation would be higher in CD4⁺ T cells than in CD8⁺ T cells as implied by the selective effect of SLP-2 deficiency on CD4⁺ T cell expansion remains to be determined. Recent work has started to unravel the role of metabolic pathways in the differentiation of various T cell subsets (51, 60, 61). Memory CD8 T cells seem to rely on fatty acid oxidation, wherein mice with a block in fatty acid metabolism showed significantly decreased memory CD8 T cells, and upregulation of fatty acid oxidation with metformin, an AMPK activator, increased the generation of CD8⁺ T cells (61). Furthermore, rapamycin treatment increased CD8 memory generation, likely through the upregulation of fatty acid oxidation (60, 61). Regulatory T cells also require high levels of fatty acid oxidation (51). Indeed, precursor cells can be driven into the regulatory T cell lineage by the addition of excess fatty acid, metformin, or rapamycin. In contrast, effector CD4⁺ T cells (Th1, Th2, and Th17 cells) require high levels of glycolysis (49). The consistently predominant defect in the response of CD4⁺ T cells observed in SLP-2 T-K/O mice, despite compensatory glycolysis, points to additional metabolic checkpoints that require further investigation.

In summary, we show that the loss of SLP-2 in T cells alters CL compartmentalization in mitochondrial membranes and the proper assembly and function of the respiratory chain. These defects are associated with altered mitochondrial respiration that is increasingly uncoupled from ATP production and may be partially compensated by a greater reliance on glycolysis. As a result, SLP-2–deficient T cells have defective in vitro and in vivo effector responses, mostly within the CD4⁺ compartment. Based on these findings, we conclude that SLP-2 is an important regulator of mitochondrial membrane compartmentalization. Availability of the SLP-2 T-K/O mouse strain may thus be helpful to dissect fundamental aspects of metabolic regulation of T cell responses.

We thank Drs. Dameng Lian and Hao Wang for microsurgical expertise and Eric H. Ball, Fred Possmayer, Heidi McBride, Andrew D. Wells, and Wayne W. Hancock and the members of the Madrenas laboratory for helpful comments and discussion.

The authors have no financial conflicts of interest.