Calpains Released by T Lymphocytes Cleave TLR2 To Control IL-17 Expression Free

Calpains are cytosolic, calcium-activated neutral cysteine proteases (1). Two ubiquitously expressed isoforms, calpain 1 (or μ-calpain) and calpain 2 (or m-calpain), require micromolar and millimolar Ca²⁺ concentrations, respectively, for an activity that is tightly controlled by calpastatin. Calpains play an important proinflammatory role; they are involved in the NF-κB–dependent expression of proinflammatory cytokines and adhesion molecules (2). Calpains are also critical for inflammatory cell adhesion and chemotaxis, as well as inflammatory mediator processing (3). Finally, we demonstrated that calpains are implicated in the cleavage of hsp90, which is required to maintain glucocorticoid receptor in a ligand-binding conformation and, thereby, to settle anti-inflammatory effects of glucocorticoids (4).

Calpains participate in immune responses as well. T lymphocyte activation after TCR/CD3 complex engagement increases calpain expression (5). In turn, calpain activity is involved in T lymphocyte migration (6), secretion of IL-2 and cell surface expression of IL-2R subunit α (CD25) (7), secretion of IFN-γ and Th1 commitment (8), and secretion of IL-17 and Th17 commitment (9).

Although calpains are considered intracellular enzymes, a few studies showed that they are partially externalized. Calpains are secreted by lymphocytes, endothelial cells, and parathyroid cells, among others (5, 10, 11). Calpain secretion is thought to be, in an unconventional way, due to the lack of N-terminal classic secretion signal peptide. Interestingly, a novel hypothesis proposes that unconventional secretion provides a mechanism through which the consequences of a single enzymatic activity differ dramatically, according to intracellular or extracellular localization (12). This notion is strengthened by the observation that, when externalized, calpains seem to promote resolution and tissue repair instead of inflammation/immunity development. For instance, externalized calpains activate anti-inflammatory cytokines (TGF-β) (13) and inactivate proinflammatory proteins (chemerins) (14). In addition, we demonstrated that extracellular calpains participate in both epithelium regeneration in a model of acute kidney injury (15) and angiogenesis in models of kidney inflammation (10). The opposite activities of intra- and extracellular calpains raise questions regarding the mechanisms of calpain secretion. We present evidence that mouse and human lymphocytes secrete calpains through a pathway that involves the ABCA1 transporter. In turn, extracellular calpains primarily affect IL-17 expression through cleavage of the TLR2 extracellular domain.

Spleen cells from C57BL/6J mice were isolated by passing the tissue through a nylon membrane. They were depleted of erythrocytes by a 90-s exposure to ACK lysing buffer (BioWhittaker), washed, and resuspended in DMEM supplemented with 10% FCS, 1% glutamine, 10 mM HEPES, 0.05 mM 2-ME, and penicillin/streptomycin. Lymphocytes were isolated from this preparation using the mouse CD3⁺ T Cell Enrichment Kit (Stem Cell Technologies), according to the manufacturer’s instructions, or the Lympholyte-M Kit (CEDARLANE), and T cells were purified by negative selection and magnetic separation with a CD3⁺ Cell Isolation Kit (Miltenyi Biotec).

For PBMC and polymorphonuclear neutrophil (PMN) isolation, blood was collected from subjects in full agreement with institutional guidelines. After centrifugation (100 × g for 15 min) to remove platelets, EDTA blood samples were diluted (1:1) in RPMI 1640 medium, and PBMCs and PMNs were isolated by density centrifugation on Ficoll-Paque (GE Healthcare).

Isolated mouse spleen lymphocytes or negatively selected T cells (Pan T Cell Isolation Kit II; Miltenyi Biotec) were cultured at 5 × 10⁶ cells/well in a 24-well plate in 500 μl MEM supplemented with 1% FBS, 2 μM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. They were stimulated or not with 1 μg/ml soluble anti-CD3 or plate-bound anti-CD3 (BD Pharmingen) and soluble anti-CD28 (1 μg/ml; eBioscience), 0.5–4 μg/ml μ-calpain (Merck Chemicals), 20 μg/ml calpastatin (Sigma-Aldrich), 10 ng/ml Pam3Cys (Cayla InvivoGen), 0.001–1 ng/ml recombinant mouse IL-2 (BD Pharmingen), 50 ng/ml neutralizing anti–IL-2 (BD Pharmingen), or IgG2a isotype control (eBioscience), alone or in combination. After 1–72 h, the cells were collected, centrifuged (1500 rpm for 3 min), and suspended in 500 μl PBS supplemented with 0.1% FBS.

T cell proliferation was determined with BrdU-labeling solution. The uptake of BrdU was detected using the Cell Proliferation BrdU Kit (Roche Diagnostics).

Intracellular calpain activity was determined in spleen lymphocytes or PBMCs (3 × 10⁶), as previously described (9). For measuring calpain activity in the extracellular milieu, isolated spleen lymphocytes or PBMCs were cultured in 24-well tissue culture dishes in RPMI 1640 medium (3 × 10⁶ cells in 500 μl). After the indicated culture period, cell-conditioned medium was diluted (1:1) in Krebs–Ringer HEPES (KRH) solution (pH 7.4) containing 4 mM CaCl₂, with or without 100 μM calpain inhibitor 1, and incubated for 10 min before the addition of 50 μM calpain substrate N-succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (AMC) (Sigma-Aldrich). After a 90-min incubation period, fluorescence was detected at 360 nm excitation and 460 nm emission, using the FLx800 spectrofluorometer (BioTek Instruments). Calpain activity was determined as the difference between fluorescence measured with and without calpain inhibitor 1.

Spleen lymphocytes (6 × 10⁶/ml) were incubated for 4 h in the presence or absence of μ-calpain (1 μg/ml). Cytokines (IFN-γ, IL-2, IL-4, IL-6, IL-10, IL-12, IL-17, and TNF-α) were measured in the supernatants using the Mouse Inflammatory Cytokines and Chemokines Multi-Analyte ELISArray Kit (SABiosciences). Other ELISAs included calpastatin (Antibodies-Online.com), soluble TLR2 (Clinisciences), IL-17A (QIAGEN), and IL-17F (eBioscience).

Proteins extracted from whole lymphocytes or lymphocyte cell surfaces, using a Cell Surface Protein Isolation Kit (Perbio Science), and from cell supernatants after concentration (10×) on an Amicon filter (Millipore) were resolved with 4–12% Bis-Tris gels (Fisher Bioblock Scientific), under nonreducing conditions and in the presence of protease inhibitors, and transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore). Membranes were blotted with specific primary Abs, including anti-ABCA1 rabbit polyclonal PA1-16789 (1/500; Thermo Fisher Scientific), anti-gp96 rabbit polyclonal 14491 (1/1000; Abcam), anti–IL-6R AF1830 (1/1000; R&D Systems), anti-gp130 rabbit polyclonal (0.5 μg/ml; Abcam), anti-TLR2 rabbit polyclonal 7966-1 (1/1000; Abcam), and anti-TLR1 (1/1000; OriGene), followed by HRP-conjugated goat secondary Ab (Jackson ImmunoResearch). ECL (Fisher Bioblock Scientific) was used for developing. Quantification was performed using ImageJ processing and analysis.

One microgram of recombinant human TLR2 extracellular domain (R&D Systems) was incubated at 37°C in KRH solution (pH 7.4) containing 2 mM CaCl₂, with or without 1 μg/ml μ-calpain. After 30 s, samples were analyzed by SDS-PAGE.

Cells (10⁶/100 μl) were incubated for 1 h at 4°C with the following surface Abs: FITC-conjugated anti-mouse ABCA1 (Bio-Rad Serotec), allophycocyanin-conjugated anti-mouse TLR2 (R&D Systems), violetFluor-conjugated anti-mouse CD3 (Tonbo biosciences), FITC-conjugated anti-mouse CD4 (BD Pharmingen), VioBlue-conjugated anti-mouse TCR γδ (Miltenyi Biotec), PerCP-Cy5.5–conjugated anti-mouse Ly6G (GR1; eBioscience), or isotype-control Abs. Data were collected on a MACSQuant cytofluorometer (Miltenyi Biotec) and analyzed with FlowJo software (TreeStar).

Total cellular RNA was collected from mouse T lymphocytes using a RNA Miniprep Super Kit (Bio Basic), checked for quality by measuring the ratio of optical densities at 260 and 280 nm, reverse transcribed into cDNA with Superscript II (Life Technologies BRL), and amplified by PCR using a LightCycler 480 (Roche Diagnostic) with SYBR Green (Fast Start DNA Master SYBR Green I; Roche Applied Science, Roche Diagnostic) and specific primers for mouse IL-17A, IL-21, and IL-23R and two housekeeping genes (β-actin and Gusb). The two reference genes were used to normalize the quantitative PCR results, using Roche LightCycler 2.0 software (Roche Diagnostic).

To modulate ABCA1 expression, HEK293 cells were transfected with small interfering RNA (siRNA) targeting ABCA1 human gene (SI03025193) or negative-control siRNA (SI03650318; both from QIAGEN). Briefly, HEK293 cells were seeded in a 24-well plate at a density of 40 × 10³ cells/well in 500 μl DMEM with GlutaMAX and 10% FBS. After overnight incubation at 37°C, cells were transfected with 6 pmol siRNA (100 μl) through the use of 3 μl HiPerFect transfection reagent (QIAGEN). Two days after transfection, cell culture medium was changed to 250 μl KRH with 2 mM calcium, and cells were stimulated for 4 h, with or without 10 μg/ml 4F (Proteogenix), for functional assay (i.e., intra- and extracellular calpain activity measurements). In independent wells, cells were lysed in RLT, and their RNA was isolated with an EZ-10 Spin Column Kit (Proteogenix) and reverse transcribed with a First-Strand cDNA Synthesis Kit (Thermo Scientific). Then, cDNA was amplified on a Light Cycler 480 system (Roche) using ABCA1 gene-specific primers (5′-TGCTGCATAGTCTTGGGACTC-3′ and 5′-ACCTCCTGTCGCATGTCACT-3′) or GAPDH gene-specific primers (5′-AGCCACATCGCTCAGACAC-3′ and 5′-GCCCAATACGACCAAATCC-3′) and SYBR Green (Roche).

Samples were mixed in a 1:2 ratio with a freshly prepared α-cyano-4-hydroxycinnamic acid matrix (4 mg/ml in acetonitrile 60% [v/v], H₂O 40% [v/v], ammonium citrate 10 mM, and trifluoroacetic acid 0.1% [v/v]) and allowed to dry at room temperature for crystallization. MALDI-TOF analyses were performed on a tandem time-of-flight mass spectrometer (4800 MALDI TOF/TOFTM Analyzer; Applera Applied Biosystems, Framingham, MA) after calibrating the instrument in external mode with synthetic peptides (mass range 500–3400 Da) in positive reflector mode. Then, a list of precursors with signal/noise > 40 were generated for fragmentation (tandem mass spectrometry analysis) with a 2 kV-TIS200/CID OFF method.

The cohort of 24 adult patients with type 1 diabetes was described previously (16). They were randomly assigned to placebo or IL-2 (0.33, 1, or 3 MIU/d) for a 5-d course and were followed on days 0, 6, and 15. On these days, blood samples were obtained, and ABCA1 gene expression was analyzed in PBMCs using the SurePrint G3 Human GE v2 8×60K Microarray (Agilent Technologies; ID: 039494), according to the manufacturer’s protocol, as reported (17). ABCA1 gene expression is the average expression measured by two probes (A_24_P235429, A_33_P3422897).

The human C-reactive (CRP) gene promoter/signal peptide was chosen for inflammation-inducible calpastatin transgene expression and secretion. Using the strategy previously described to generate transgenic mice with inflammation-inducible overexpression of GM-CSF (18), the Mouse Clinical Institute (Illkirch-Graffenstaden) amplified a construct, using PCR, including the human CRP promoter, a signal peptide sequence, and the entire coding sequence of the mouse calpastatin gene Cast-201, as depicted in Supplemental Fig. 2A. Transgenic mice were generated by microinjection of the construct into fertilized eggs of C57BL/6 mice. Offspring tail DNA was analyzed for the presence of transgene by PCR using the primers shown in Supplemental Fig. 2B. Two lines were obtained by two different founders; CRP/Calpast and CRP/Calpast2 were deposited in the European Mouse Mutant Archive repository with IDs EM:05917 and EM:06130, respectively. All animal procedures were performed in accordance with the European Union Guidelines for the Care and Use of Laboratory Animals (Authorization for the document reference 00515.01).

Age-matched male CRP/Cast C57BL/6 mice and wild-type (WT) littermates were injected i.p. with 20 μg Pam3Cys in 100 μl PBS. After 18 h, mice were euthanized by a lethal dose of pentobarbital, followed by cervical dislocation. Peritoneal exudates were collected by lavage with 1 ml HBSS. Supernatants were collected by centrifugation (1500 rpm for 5 min) and subjected to assay for calpain activity and ELISA for calpastatin (antibodies-online.com) and IL-17 (QIAGEN). Total peritoneal cells and the absolute numbers of neutrophils were counted using a hemocytometer and FACS, respectively.

Transgenic male CRP/Cast C57BL/6 mice and WT littermates were used at 10–14 wk. Mice were anesthetized using ketamine and xylazine prior to all procedures, which were approved by the Animal Care Use Committee of the Paris 13 University (Bobigny, France). Arthritis was induced with native chicken collagen type II (Morwell Diagnostics, Zurich, Switzerland). Briefly, mice were injected s.c. at the base of the tail and on the back with 50 μg collagen type II emulsified in Freund’s adjuvant in each location. A boost was performed on day 21 (19). The mice were monitored for arthritis using a blind procedure. Disease severity was assessed using a macroscopic scoring system: 0, normal; 1, detectable arthritis with erythema; 2, significant swelling and redness; 3, deformity; and 4, ankylosis. In each mouse, the score was determined for three joints of each hind paw (toes, tarsus, and ankle) and two joints of each fore paw (toes and carpus) (20).

For the histological evaluation, animals were killed, and their hind paws were dissected free and processed as previously described (21). A blinded procedure was used to evaluate the severity of inflammation (synovial proliferation and inflammatory cell infiltration) and joint destruction (cartilage, irregularities, and bone erosions) using a four-point scale (0–3, where 0 was normal and 3 was severe). For incidence determinations, histological inflammation or destruction was defined as an inflammation or destruction score ≥ 0.5.

Data are expressed as mean ± SEM. Comparisons between two groups of values were made with the two-tailed unpaired Student t test using Microsoft Excel or the Mann–Whitney rank U test using StatView. Where indicated, dose- and time-response were analyzed using two-way ANOVA. The p values < 0.05 were considered statistically significant.

There are two potential explanations for finding calpains in the extracellular milieu of lymphocytes: passive release due to cell death and active transport through cell membrane. Previous studies excluded a defect in lymphocyte viability, thus suggesting a secretion process (5). For unconventional secretion of proteins that lack an N-terminal signal-peptide sequence, such as calpains (1), different pathways are potentially involved, including, in particular, ATP-binding cassette (ABC) family transporters (22). To address whether lymphocytes secrete calpains through an ABC-dependent process, we initially analyzed the effects of ABC inhibitory drugs on calpain activity in the intra- and extracellular milieu (Fig. 1A, 1B). Calpain release from TCR-activated mouse spleen lymphocytes or human PBMCs was reduced by glyburide, an ABCA1 transporter inhibitor, but not by cyclosporine A, a p-glycoprotein (ABCB1) inhibitor, or by MK571, a multidrug resistance-related protein (ABCC) inhibitor. The extracellular differences between glyburide-treated and untreated cells were reflected in the retention of calpain activity in the intracellular compartment. Conversely, apoA-I and apoA-I mimetic peptide 4F, which increase ABCA1 membrane expression (Supplemental Fig. 1A), displayed dose-dependent stimulatory effects on calpain secretion by TCR-activated mouse spleen lymphocytes, thus limiting their cytosolic activity (Fig. 1C). These compounds exerted no cytotoxic effects, as assessed by annexin V/propidium iodide staining (>98% cell viability). These results suggest a role for ABCA1 in calpain exteriorization, consistent with ABCA1 expression in lymphocytes (23). To conclusively demonstrate the role of ABCA1, we used an siRNA-mediated gene-silencing approach in HEK293 cells. The efficiency of knockdown was 34.9%. Downmodulation of ABCA1 expression led to a 36% decrease in extracellular calpain activity while increasing intracellular calpain activity by 120% (Fig. 1D). As expected under these conditions, 4F was no longer able to enhance calpain exteriorization.

Finally, we identified, using Western blotting, calpain isoforms exteriorized under the control of ABCA1 (Fig. 2A). Calpain 1 and 2 were found within TCR-activated mouse spleen lymphocytes, whereas calpain 4 remained barely detectable. Similar results were found in the extracellular milieu of these cells, with the exception that calpain 4 was significantly expressed. As expected, addition of glyburide to inhibit ABCA1-dependent processes induced the retention of all of the calpain isoforms within lymphocytes. The ability of ABCA1 to promote the shedding of microvesicles (24), as well as the marked release of microvesicles from activated T lymphocytes (25), led us to study whether calpains are secreted within microvesicles. Round membrane-coated vesicles of variable size (100–200 nm in diameter) were identified by electron microscopy in the supernatant of TCR-activated mouse spleen lymphocytes (Fig. 2B). Western blot analysis of these microvesicles demonstrated the presence of calpains, including calpain 1, 2, and 4. Taken together, these results indicate that activated lymphocytes secrete calpains through an ABCA1-dependent process that involves, at least in part, microvesicle shedding.

To gain insights into the effects of secreted calpains, we initially examined whether the addition of exogenous μ-calpain affected the production of cytokines by mouse spleen lymphocytes in vitro (Fig. 3A). The levels of IFN-γ (Th1), IL-4 (Th2), IL-6, IL-12, and TNF-α were not modified. In contrast, addition of μ-calpain decreased the levels of IL-2 (but the differences did not reach statistical significance), IL-10, and, even more markedly, IL-17.

Given the effect of exogenous μ-calpain on IL-17 production, experiments next determined whether secretion of endogenous calpains through an ABCA1-dependent process was limiting IL-17 expression as well. To this aim, naive mouse lymphocytes were stimulated under minimal Th17-polarizing conditions, including TCR activation and IL-6 addition. Chemicals increasing ABCA1 expression or activity, apoA-I and apoA-I mimetic peptide 4F, blunted IL-17 production in a dose-dependent manner (Fig. 3B). As expected, addition of calpastatin to inactivate calpains specifically in the extracellular medium amplified IL-17 production and completely suppressed the inhibitory effect of 4F. Importantly, the inhibitory effect of either drug on IL-17 production persisted for 20 h after its withdrawal (Fig. 3C). Taken together, these data suggest that ABCA1-dependent secretion of calpains controls IL-17 expression.

Inhibition of IL-17 expression by μ-calpain occurred in mouse lymphocytes, such as Th17 cells (Supplemental Fig. 1B); it was significant after 4 h and had its maximum effect at a concentration of 4 μg/ml (Fig. 3D). This could not be due to cell death because mouse lymphocyte viability was slightly increased under these experimental conditions. Comparison of the effects of extracellular μ-calpain on the two main members of IL-17 family revealed that it limited IL-17A expression, whereas it substantially enhanced that of IL-17F (Supplemental Fig. 1C).

Inhibition of IL-17 expression by μ-calpain was not attributable to a direct degradation of this cytokine, because IL-17 immunoreactivity was maintained at 93.9 ± 2.0% of its initial value after exposure to 2 μg/ml μ-calpain (n = 3). To address whether extracellular calpains influence IL-17A expression by regulating gene transcription and/or mRNA stability, we carried out a quantitative RT-PCR assay. The culture of mouse T lymphocytes under minimal Th17-polarizing conditions resulted in a rapid induction of the mRNA coding for RORγt, the master regulator for Th17 differentiation (Fig. 3E). This was almost completely blunted in the presence of extracellular μ-calpain, which also blunted the expression of other genes required to expand the Th17 cell population and stabilize their phenotype, such as IL-21 (Fig. 3E).

Next we sought to examine the molecular mechanisms whereby extracellular calpains control IL-17 expression. We hypothesized that they would cleave membrane receptor and/or ligand involved in the expression of IL-17. To explore this possibility, we compared the profile of membranous proteins in mouse spleen lymphocytes treated or not with μ-calpain. The one-dimensional SDS-PAGE analysis of proteins isolated after cell surface biotinylation revealed that μ-calpain primarily diminished two proteins with apparent molecular masses ∼40 and ∼ 20 kDa (Fig. 4, far left panel). The main ∼20-kDa band was excised from the gel and, after in-gel tryptic digestion, the resulting peptide digests were analyzed by MALDI-TOF/TOF. The obtained mass spectrometry and tandem mass spectrometry data were used for a data search (National Center for Biotechnology Information), leading to the identification of the N terminus (DDEVDVDGTVEEDLGKSR) of the chaperone glucose-regulated protein 94 (also known as gp96) with the highly significant score of 94. To confirm the relationship between the ∼20-kDa band and the gp96 fragment, we next performed Western blot experiments using an Ab directed against the N terminus of gp96 (Fig. 4, near left panel). In mouse lymphocyte membranes, this Ab recognized an intact form of gp96 and N terminus fragments of ∼72, ∼52, ∼40, and ∼20 kDa. The intensity of all of these bands was decreased markedly by cell exposure to μ-calpain; conversely, it was amplified by calpastatin. Importantly, the disappearance of the main membrane-associated form of gp96 was concomitant with its release without further cleavage in extracellular medium (Fig. 4, near right panel). These results strongly support that extracellular calpains limit the association of gp96 with lymphocyte plasma membrane by increasing its detachment rather than its degradation.

Even if gp96 is generally localized in the endoplasmic reticulum, where it is required for cell surface expression of TLRs, it may translocate to the cell surface and gain extracellular access after cell stress (26). When externalized, gp96 or its N terminus fragments eventually serve as endogenous ligands for TLR2 (27). Thus, we considered that extracellular calpains could affect the association of gp96 with lymphocyte plasma membrane by limiting its binding to TLR2. To explore this hypothesis, we assessed the expression of gp96 at the surface of mouse lymphocytes in the presence of neutralizing Ab against TLR2 (Fig. 4, far right panel). As expected, TLR2 inactivation markedly reduced membrane association of gp96 for N terminus fragments ∼52 and ∼40 kDa and moderately reduced its associated for the intact form.

To directly assess the degradation of TLR2 by extracellular calpains, its lymphocyte expression pattern was analyzed by Western blotting. A 1 h exposure of mouse spleen lymphocytes to exogenous μ-calpain effectively limited TLR2 expression at the cell surface while increasing the appearance of a soluble form of TLR2 in the extracellular milieu (Fig. 5A). Membrane TLR2 had an apparent mass of 95 kDa, whereas its soluble form had an apparent mass of 83 kDa, which would correspond to the full TLR2 extracellular domain (28). In contrast, exogenous μ-calpain did not promote the shedding of TLR1, which forms heterodimers with TLR2 (Fig. 5B), or that of TLR4 (data not shown) or the IL-6R complex, including the α-chain and the signal transducer gp130, both of which are essential for Th17 commitment and IL-17 expression (Fig. 5B). We confirmed these findings further using FACS analysis: μ-calpain limited TLR2 expression in mouse lymphocytes (Fig. 5C), dendritic cells (data not shown), and human neutrophils (Fig. 5 D), whereas calpastatin had the opposite effect. Similarly, a 1-h exposure of human PBMCs to exogenous μ-calpain limited TLR2 expression at the cell surface while increasing the appearance of a soluble form of TLR2 in the extracellular milieu (Fig. 5E). Importantly, endogenous calpains exteriorized from these cells after exposure to 4F peptide also promoted TLR2 ectodomain cleavage, because calpastatin addition prevented its release (Fig. 5E).

In this context, we observed an almost complete disappearance of membrane TLR2 and modestly abundant soluble TLR2 after human PBMCs were exposed to extracellular calpains. This observation was consistent with the idea that extracellular calpains cleave the extracellular domain of TLR2 and eventually cause its complete degradation. To test this hypothesis, we performed in vitro experiments in which the recombinant extracellular domain of human TLR2 was exposed to μ-calpain. Analysis by SDS-PAGE revealed cleavage of TLR2 as early as 30 s, leading to the appearance of multiple breakdown products (Fig. 5F). The collective data indicate that exteriorized calpains cleave and degrade membrane TLR2, thus preventing gp96 binding and TLR2 engagement.

Because TLR2 activation is known to enhance IL-17 production by T cells (29), we investigated whether the decrease in IL-17 expression observed in response to extracellular calpains could be explained by TLR2 shedding. Treatment of mouse spleen lymphocytes or human PBMCs with a TLR2-TLR1 ligand, Pam3Cys, induced an increase in IL-17 production that was suppressed by μ-calpain (Fig. 6). Conversely, addition of a neutralizing Ab to TLR2 limited IL-17 production, whose residual level was not affected by μ-calpain or calpastatin (Fig. 6, Supplemental Fig. 1D). Similar results were obtained using lymphocytes isolated from the spleen of Tlr2^−/− mice. Taken together, our data clearly demonstrate that extracellular calpains limit IL-17A expression by cleaving TLR2 and, thereby, prevent lymphocyte response to exogenous (e.g., Pam3Cys) or endogenous (gp96) TLR2 ligands.

Next, we sought to examine under what conditions that this novel pathway that limits IL-17 expression might be operational. IL-2 limits Th17 differentiation and IL-17 expression, but the underlying mechanisms remain only partially understood (30). IL-2 could interfere with IL-6–dependent signaling events. To investigate whether calpain secretion could also play a role, we first analyzed the effect of IL-2 on calpain activity in the intra- and extracellular milieu of mouse spleen lymphocytes. We were intrigued to observe that low doses of IL-2 (up to 0.01 ng/ml) increased calpain secretion, thus decreasing cytosolic activity, whereas higher doses (≥1 ng/ml) had the opposite effect (Fig. 7A). Similar dose-dependent effects of IL-2 were obtained when cell surface expression of ABCA1 was examined at the single cell level by flow cytometry (Fig. 7B).

We next looked for molecular events reflecting IL-2–induced calpain externalization. Low doses of IL-2 decreased the lymphocyte expression of TLR2 (Fig. 7C, upper panels). Addition of calpastatin increased TLR2 expression and prevented the response to IL-2, demonstrating that IL-2 controls TLR2 through the externalization of calpains (Fig. 7C, lower panels). Finally, consistent with previous reports, we found that IL-2 limited IL-17 production in a dose-dependent manner (Fig. 7D). More strikingly, inactivation of secreted calpains by the addition of calpastatin reversed the inhibitory effect of IL-2 but only at low concentrations. Taken together, our data support the principle that low doses of IL-2 limit IL-17 production, at least in part through ABCA1-dependent calpain exteriorization and, thereby, TLR2 cleavage.

To further emphasize the clinical relevance of these findings, we also investigated the response of human lymphocytes to IL-2. Western blot experiments demonstrated that addition of low doses of IL-2 increased ABCA1 amounts in human PBMCs in vitro, whereas higher doses did not (Fig. 7E). Furthermore, low-dose IL-2 treatment of adult patients with type 1 diabetes triggered a dose-dependent increase in ABCA1 gene expression in PBMCs (Fig. 7F).

We next investigated whether TLR2 cleavage and IL-17 suppression by exteriorized calpains could be recapitulated in vivo. To conclusively address this role of extracellular calpains, we generated transgenic mice by pronuclear injection of a construct that expresses the mouse calpastatin gene under the control of the CRP gene promoter/signal peptide (CRP/Cast; Supplemental Fig. 2A). The human CRP promoter was shown to combine the advantages of low basal transgene expression with the potential of a marked increase in response to inflammatory stimuli (18). Because this promoter includes a signal peptide sequence, transgene product would be exteriorized from liver cells and reach tissue extracellular medium via the blood stream.

Using these transgenic mice, we first developed a model of sterile peritonitis by injecting the TLR2-TLR1 ligand Pam3Cys into the peritoneal cavity. Although the basal amount of calpastatin mRNA was similar in the liver of CRP/Cast mice and WT littermates, it increased significantly 6 h after Pam3Cys injection (Supplemental Fig. 2C). Twenty hours after Pam3Cys injection, calpastatin reached significantly higher levels in the peritoneal fluid of CRP/Cast mice compared with WT mice, resulting in a decrease in calpain activity (Fig. 8A, 8B). Consistent with the results observed in vitro, there was a significant increase in IL-17 production in CRP/Cast mice (Fig. 8C), Notably, analysis of IL-17 values versus calpain activity in the peritoneal fluid of individual mice confirmed the strong association of low calpain activity with high IL-17 levels in CRP/Cast mice compared with WT mice (Fig. 8D). This high expression of IL-17, which primarily recruits neutrophils, prompted us to assess possible differences in neutrophil infiltration. We found more infiltration of Ly6G⁺ neutrophils in the peritoneal cavity of CRP/Cast mice compared with WT mice (p = 0.0023, Mann–Whitney U test), whereas the total number of peritoneal cells was comparable. Analysis of neutrophil counts versus IL-17 levels in the peritoneal fluid of individual mice confirmed the strong association of high IL-17 levels with high neutrophil counts in CRP/Cast mice compared with WT mice (Fig. 8E).

Next, to explore whether exteriorized calpains also have the capacity to control IL-17–driven inflammatory diseases, we developed a model of rheumatoid arthritis. This autoimmune disease was chosen because it involves processes targeted by extracellular calpains, which are indeed present in increased amounts in pathological synovial fluid (31). For instance, the concentration of gp96 in synovial fluid is greater in patients with rheumatoid arthritis compared with patients with other forms of inflammatory arthritis, and gp96 promotes synovial inflammation through TLR2 signaling (32). As a consequence, in patients with rheumatoid arthritis, blockade of TLR2 prevents the release of inflammatory mediators by synovial tissue (33). Among these mediators, IL-17 is predominantly involved in the induction and progression of rheumatoid arthritis (34). Therefore, we induced collagen-induced arthritis in C57BL/6 mice, as reported previously (20). As shown in Fig. 9A, the incidence of arthritis in two sets of experiments was markedly increased in CRP/Cast mice compared with WT littermates or age- and sex-matched WT mice, whereas the clinical score was not different (Fig. 9B). Histological studies revealed that articular inflammation was not different in the two groups (Figs. 9C, 10A–D), but articular destruction was significantly enhanced in CRP/Cast mice compared with WT mice (Figs. 9D, 10E). Representative histological features are shown in Fig. 9E and 9F. Anti-type II collagen IgG1 and IgG2a were not different in both groups (data not shown).

This study reveals the mode of calpain secretion by lymphocytes and one of the autocrine/paracrine functions of externalized calpains. Using both pharmacological and genetic approaches, we find that calpain exteriorization requires the activity of the ABCA1 transporter. Lymphocytes express several ABC transporters, including ABCA1, whose gene expression is nearly identical in Th1, Th2, and Th17 (23). Like other ABC proteins, ABCA1 is considered responsible for the secretion of proteins lacking a classical hydrophobic signal peptide (16), such as the anti-inflammatory annexin A1 (35), a downstream mediator of glucocorticoids. It is tempting to speculate that, through its floppase activity, ABCA1 participates in the formation of microvesicles (36), which would export calpains. This hypothesis is strengthened by our observation of calpains in membrane vesicles shed from mouse lymphocytes, as previously shown in vesicles derived from parathyroid cells (11). In the extracellular milieu, these vesicles would be broken down rapidly, releasing their content (37), including calpains. In addition to this mode of secretion, calpains could be released in response to tissue damage, thus functioning as endogenous danger signal damage-associated molecular patterns or alarmins.

Once exteriorized from lymphocytes, calpains appear to primarily limit the expression of IL-17A while increasing that of IL-17F. This specificity is in accordance with previous studies demonstrating a negative feedback effect of IL-17A on IL-17F secretion (38). This influences IL-17–producing Th17 cells, whereas the activity of IFN-γ–producing Th1 cells and IL-4–producing Th2 cells is not affected. Interestingly, recent studies consistently documented that ABCA1 is involved in IL-17 control because plasma IL-17 levels increase in Abca1^−/−Abcg1^−/− bone marrow–transplanted mice through the increased production of IL-23 (39). Conversely, agonist of peroxisome proliferator activated receptor-γ was shown to enhance ABCA1 expression (40) and decrease IL-17 generation (41). It remains to be determined whether ABCA1 transport activity participates in IL-17 inhibition through calpain export under these conditions.

Our study also identifies molecular mechanisms responsible for IL-17 inhibition by extracellular calpains. We found that brief exposure of mouse spleen lymphocytes to μ-calpain primarily causes the shedding of membrane-associated gp96. This chaperone, which is abundant in endoplasmic reticulum, can be present in the environment of nonnecrotic cells, playing the role of a damage-associated molecular pattern (42). Among other functions in immunity, extracellular gp96 was shown to increase the release of IL-17 (42). Surprisingly, even if calpain is responsible for the proteolysis of gp96 within cells (43), the release of gp96 from the membrane of lymphocytes is not due to its cleavage by extracellular calpains; rather, it is the result of the shedding of gp96-binding proteins/receptors. CD91 was initially identified as a gp96 receptor, but its actual contribution is controversial (44). More recent evidence suggests that gp96 binds instead to TLR2 (27). TLR2 is expressed in all T cells, including CD4⁺ (mainly Th17), CD8⁺, and NKT cells, and, at higher levels, γδ T cell populations. Its engagement with TLR1 leads to IL-17 production (45). An important finding of our work is that extracellular calpains are responsible for the cleavage of TLR2. Among cell surface TLRs, only TLR2 was shown to undergo proteolytic cleavage (28). This leads to the release of the entire TLR2 extracellular domain, a soluble form of TLR2 that inhibits the function of membrane TLR2, in part by acting as a decoy receptor (46). Thus far, the protease responsible for this processing had not been clearly identified. We show in this article that extracellular calpains are primarily involved. Interestingly, analyzing potential calpain-specific cleavage sites (http://calpain.org) in the extracellular domain of TLR2 identified, with the best score, amino acids in its C-terminal region (586 and 576 aa in human and mouse TLR2, respectively), consistent with our observation that calpains release the entire TLR2 extracellular domain. Subsequent to the shedding of TLR2 extracellular domain, calpains appear to cause its complete proteolysis. This suggests that exteriorized calpains limit TLR2 engagement by decreasing the availability of cell surface TLR2 more than by inducing the appearance of decoy receptors.

Another striking finding of the current study is that IL-2 limits IL-17 expression, at least in part by amplifying ABCA1 expression and, hence, calpain exteriorization. To our knowledge, the regulatory role of IL-2 on ABCA1 has not been reported previously. Effective IL-2 concentrations in human PBMCs were 0.001–0.01 ng/ml in vitro. Importantly, low-dose IL-2 therapy in humans (16), which increases plasma IL-2 levels to such values (47), also amplified ABCA1 expression in PBMCs in vivo. At these concentrations, IL-2 increased ABCA1 gene expression and ABCA1 protein synthesis. The proximal promoter of mouse Abca1 gene contains NF of activated T cells binding sites cooperating with other sites to enhance the transcription of this gene (48). Because IL-2 activates NF of activated T cells (49), transcriptional regulation of ABCA1 by IL-2 is possible. Clearly, further work is needed to elucidate all of the signaling pathways involved in this regulation. Nonetheless, these results might provide an alternative explanation for the anti-inflammatory/immunosuppressive effects of low-dose IL-2 reported, for example, in patients with autoimmune diseases (16).

Finally, the results of two experimental models substantiate the claim that the role of exteriorized calpains that we described in vitro is relevant in vivo. In the sterile peritonitis model, there was an excess of IL-17 and, hence, of neutrophil infiltration in the peritoneal fluid of CRP/Cast mice compared with WT littermates, indicating a critical role for exteriorized calpains in limiting an acute inflammatory process. In collagen-induced arthritis, a more chronic inflammation model, inhibition of exteriorized calpains aggravated joint destruction more than joint inflammation. Consistent with this observation, IL-17 and upstream mechanisms leading to IL-17 expression (e.g., gp96 exteriorization and TLR2 engagement) were shown to promote bone erosion and cartilage degradation in rheumatoid arthritis, thus contributing to the chronicity of the disease (32, 34).

In summary, our results strengthen the hypothesis that intra- and extracellular calpains have opposite effects on inflammation/immunity processes, confirming the idea that the same proteins have distinct, and even opposing functions, in the intracellular and extracellular milieu (12). The identification of such molecular mechanisms provides opportunities for the design of novel anti-inflammatory/immunosuppressive drugs. They would limit IL-17 expression in IL-17–driven chronic inflammatory diseases, such as rheumatoid diseases (rheumatoid arthritis, psoriatic arthritis, ankylosing spondylitis), multiple sclerosis, vasculitis, and atherosclerosis (34). More generally, they would antagonize TLR2 engagement in all TLR2-mediated inflammatory diseases (50). Finally, these drugs would have beneficial effects in diseases improved by low doses of IL-2.

We thank Phong Pham for performing statistical analyses of the clinical data.

The authors have no financial conflicts of interest.