Natural Soluble Interleukin-15R (cid:1) Is Generated by Cleavage That Involves the Tumor Necrosis Factor- (cid:1) -converting Enzyme (TACE/ADAM17)*

This study shows that the high affinity (cid:1) -chain of the interleukin (IL)-15 receptor exists not only in mem-brane-anchored but also in soluble form. Soluble IL-15R (cid:1) (sIL-15R (cid:1) ) can be detected in mouse sera and cell-conditioned media by enzyme-linked immunosorbent assay and by immunoprecipitation and Western blotting. This protein has a molecular mass of about 30 kDa because of the presence of a single N -glycosylation site, which is reduced to 26 kDa after N- glycosidase treatment. Transmembrane IL-15R (cid:1) is constitutively con-verted into its soluble form by proteolytic cleavage that involves tumor necrosis factor- (cid:1) -converting enzyme (TACE), and this process is further enhanced by phorbol 12-myristate 13-acetate (PMA) stimulation. The hydroxamate GW280264X, which is capable of block-ing TACE and the closely related disintegrin-like metalloproteinase

Many cytokine receptors are naturally expressed in both membrane-linked and soluble forms (1)(2)(3)(4). A number of soluble proteins corresponding to the extracellular portions of transmembrane receptors and adhesion molecules have been identified in biological fluids (1,2). The progress in this field balanced the concept of free ligands and bound receptors by demonstrating a widespread existence of soluble receptors and membrane-anchored ligands. Soluble receptors commonly consist of the extracellular portions or ectodomains of membranebound forms and thereby retain the ability to bind ligand. They typically function as natural antagonists, carrier molecules, or chaperones to protect their ligands from proteolytic degradation and increase the half-life and in some cases act as biological agonists (1)(2)(3). Two major mechanisms account for the physiological or pathological generation of soluble receptors. These include alternative mRNA splicing that gives rise to a secreted polypeptide lacking a transmembrane and/or other regions or proteolytic cleavage of membrane-bound receptor proteins from the cell surface by proteases, a process also known as receptor shedding (1)(2)(3).
binding of IL-15 (K d ϳ 10 Ϫ11 M) (23). Both IL-15R␣ and IL-15 are expressed by a variety of tissues and cell types, including monocytes/macrophages, keratinocytes, fibroblasts, nerve, muscle, and epithelial cells (25)(26)(27)(28)(29). Several different isoforms of human and murine IL-15R␣ as a result of alternative splicing of the IL-15R␣ gene were recently described (23,30,31). Recombinant soluble IL-15R␣ prevents collagen-mediated arthritis (32), inhibits short-term carrageenan-induced inflammation (33), and enhances cardiac allograft survival (34). Although recombinant soluble IL-15R␣ has been used for several in vivo studies (32)(33)(34), no data concerning the existence of natural sIL-15R␣ were reported thus far. In this study, we demonstrated the presence of natural soluble IL-15R␣ (sIL-15R␣) in mouse serum. Furthermore, murine fibroblasts constitutively release sIL-15R␣ into the culture medium, and this process is further stimulated by PMA. We provide several lines of evidence that constitutive and PMA-inducible IL-15R␣ cleavage involves the activity of TACE by using specific MP inhibitors, by overexpression of the enzyme, and by using TACEdeficient fibroblasts.
Mice-Female mice of the inbred strains Balb/c, CH3, and C57BL/6 were obtained from Charles River Laboratories (Sulzfeld, Germany). IL-15R␣Ϫ/Ϫ mice were bred in Research Center Borstel under specific pathogen-free conditions. Mice were bled from the tail vein, and sera were analyzed by ELISA and Western blotting.
For transient transfection, cells were seeded at 5 ϫ 10 5 /well in 6-well plates. COS-7 cells were transfected using the GenePORTER 2 transfection kit (Gene Therapy Systems, San Diego, CA), and MEFs were transfected using LipofectAMINE 2000 (Invitrogen). Transfection efficiency was confirmed by Western blotting using specific antibodies. Conditioned medium was collected 48 h after transfection, concentrated 10-fold using 10-kDa cutoff filtration units (Vivaspin; Vivascience, Hannover, Germany), and analyzed by immunoprecipitation and Western blotting for the presence of sIL-15R␣.
ELISA for sIL-15R␣-A 96-well plate (Greiner, Hamburg, Germany) was coated overnight at 4°C with 1 g/ml of IL-15-IgG 2b FP, which served as a capture protein for sIL-15R␣. Wells were blocked with 3% BSA in PBS for 2 h. Samples (50 l/well) were added to the plate and incubated overnight. Serial dilutions of murine recombinant sIL-15R␣ were used for standardization. Bound sIL-15R␣ was detected using biotinylated anti-mouse IL-15R␣ antibodies followed by incubation with streptavidin-peroxidase. Chromogenic substrate (R&D Systems) was used for visualization, and reaction was stopped after 20 min of incubation by addition of 1 N H 2 SO 4 . Optical density was determined at 450 nm using ELISA reader (Dynatech, Denkendorf, Germany). The detection limit of ELISA was 20 pg/ml of recombinant sIL-15R␣. The specificity of newly developed sIL-15R␣ ELISA was convincingly demonstrated by the complete absence of detectable sIL-15R␣ using IL-2-IgG 2b FP as a coating reagent.
Immunoprecipitation and Western Blotting-Nonidet P-40 (0.5% final concentration) and mixture of protease inhibitors were added to supernatants and immunoprecipitation with anti-IL-15R␣ antibodies was performed for 2 h at 4°C. Immunocomplexes were captured on protein G-agarose. To analyze glycosylation, after immunoprecipitation, the samples were treated with 250 mU of N-glycosidase F (Roche) for 3 h at 37°C according to the manufacturer's instructions. Samples were resuspended in SDS-PAGE sample buffer (62.5 mM Tris-HCL, pH 8.0, 1% glycerol, 2% SDS, 5% ␤-mercaptoethanol, and 0.01% bromphenol blue), boiled for 5 min, and analyzed on 10% SDS-PAGE. The resolved proteins were transferred onto nitrocellulose (Bio-Rad). Blots were blocked for 1 h in phosphate-buffered saline containing 0.05% Tween 20 (phosphate-buffered saline-T) and 3% BSA (Sigma). After incubations with primary and secondary antibodies and washing with phosphate-buffered saline/Tween 20, visualization of specific proteins was carried out by an enhanced chemiluminescence (ECL) method using ECL Western blotting detection reagents (Amersham Biosciences) according to the manufacturer's instructions. Cellular extracts were prepared and analyzed by Western blotting as described elsewhere (31).
CTLL Assay-Serial 2-fold dilutions of the supernatant were added to microtiter plates containing 1 ϫ 10 4 CTLL-16 cells in a final volume of 200 l. The cells were incubated for 20 h at 37°C and pulse-labeled for another 4 h with 0.5 Ci of [ 3 H]thymidine (5 Ci/mmol). Thymidine incorporation was quantified by liquid scintillation counting (PerkinElmer Life and Analytical Sciences). Standards of IL-15, IL-2, and sIL-15R␣ were included in each experiment.
Flow Cytometric Analysis-Adherent cells were harvested from culture plates using acutase (PAA Laboratories, Coelbe, Germany). IL-15R␣ expression was evaluated by incubation of cells with IL-15-IgG 2b FP as described previously (31), and analyzed by flow cytometry using FACScalibur (BD Biosciences) and CELLQuest software. Negative controls consisted of isotype-matched, nonspecific antibodies (BD Phar-Mingen). The fluorescence signal of the labeled cells was calculated as median fluorescence intensity of the cell population.
Data Analysis-All experiments were performed in at least three independent assays that yielded highly comparable results. Data are summarized as mean Ϯ S.D. Statistical analysis of the results was performed by Student's t test for unpaired samples. A p value of Ͻ 0.05 was considered as statistically significant.

Murine Fibroblasts Release Soluble IL-15R␣-Murine
L929 fibrosarcoma cells abundantly express IL-15R␣ mRNA and protein (27). Membrane-bound IL-15R␣ could be detected on the cell surface of these fibroblasts by flow cytometry and confocal microscopy (data not shown). To investigate whether these cells release sIL-15R␣ into the culture medium, conditioned media were collected from L929 fibroblasts after 3-5

W I T H D R A W N W I T H D R A W N
F e b r u a r y 9 , 2 0 1 1 F e b r u a r y 9 , 2 0 1 1 days of culture and analyzed by CTLL assay for the ability to block specifically IL-15-mediated proliferation of CTLL cells.
In parallel, the stimulation with IL-2 was assessed, which served here as a control. In fact, IL-15-but not IL-2-induced proliferation of CTLL cells was strongly inhibited by the addition of the conditioned medium from L929 fibroblasts (Fig. 1A).
Reportedly, a variety of proteins may be released from the cell surface after stimulation with PMA, a potent activator of protein kinase C, by triggering a metalloproteinase-dependent ectodomain sheddase machinery (37,38). This pathway is defined as inducible shedding and seems to be highly conserved among multiple cell types. Thus, we investigated whether PMA treatment could enhance the release of sIL-15R␣ from murine fibroblasts. To this end, L929 cells were treated with PMA for different time intervals, or left untreated. Interestingly, these experiments revealed that PMA had the ability to up-regulate the amount of membrane IL-15R␣ within the first hour of stimulation as demonstrated by IL-15-IgG 2b FP staining of the membrane-bound IL-15R␣ and flow cytometry analysis (Fig.  1B). However, the increase in membrane-linked IL-15R␣ was transitory and dropped substantially within the second hour of PMA treatment. After 2 h, L929 fibroblasts clearly had less membrane-bound IL-15R␣, resulting in 52% reduction of median fluorescence intensity compared with the untreated control (Fig. 1B). Given that down-modulation of IL-15R␣ from the cell surface may result from its release into the culture medium, cell-free supernatants from PMA-stimulated and untreated cells were assayed for the presence of the soluble ␣ chain by ELISA. Indeed, PMA treatment induced the release of the soluble form to the culture medium from L929 cells, increasing the amount of sIL-15R␣ 3-fold (Fig. 1C). These results were further corroborated by Western blotting analysis of the culture medium. Conditioned media from these cells were harvested after 2 h of culture, and sIL-15R␣ was immunoprecipitated using specific antibodies. As illustrated in Fig. 1D, immunoprecipitation and Western blotting experiments confirmed that sIL-15R␣ is present in the cell supernatants as a single band protein with molecular mass of about 30 kDa. Re-appearance of the membrane-bound IL-15R␣ was detected after 24 h of PMA removal (data not shown).
It has been reported that certain ligands could induce release of respective receptors to the culture medium. Such liganddriven release of soluble form of the cognate receptor has been demonstrated for IL-4 and TNF␣ (10, 39). Thus, the ability of IL-15 to induce sIL-15R␣ release was tested. Parallel with FIG. 1. Murine fibroblasts release soluble IL-15R␣. A, suppression of IL-15-mediated CTLL proliferation by conditioned media obtained from L929 fibroblasts. Cells were cultured for 3-5 days and conditioned media were concentrated 10-fold. The ability of concentrated media to block IL-15-induced CTLL proliferation was tested. IL-2-mediated proliferation was used as a positive control and proliferation of cells without cytokines as a negative control. *, p Ͻ 0.05 versus IL-2induced proliferation. B, effect of PMA on the surface expression of IL-15R␣ in L929 cells. Cells were treated with 200 ng/ml PMA for 1 or 2 h, and expression of IL-15R␣ was detected by fluorescence-activated cell sorting analysis. Untreated cells were used as a control. C, analysis of IL-15R␣ shedding in L929 cell line induced by different factors. Cells were incubated in the presence of PMA (200 ng/ ml), TNF␣ (10 ng/ml), IL-15 (50 ng/ml), or LPS (10 g/ml) for 2 h, and conditioned media were concentrated 10-fold followed by quantitation of sIL-15R␣ by ELISA. Recombinant sIL-15R␣ was used as a standard. **, p Ͻ 0.01 versus control samples. D, detection of sIL-15R␣ in conditioned media from untreated or PMAtreated L929 cells by Western blotting after immunoprecipitation with specific anti-IL-15R␣ antibodies. E, effect of TNF␣ on the expression of IL-15R␣ in L929 cells. Cells were treated with 10 ng/ml of TNF␣ for 1 or 2 h, and expression of IL-15R␣ was detected by fluorescenceactivated cell sorting analysis. Untreated cells were used as a control. F, glycosylation pattern of sIL-15R␣. sIL-15R␣ was immunoprecipitated from L929-conditioned medium and treated with N-glycosidase or left untreated, as described under "Materials and Methods." After treatment, protein lysates were analyzed by Western blotting using anti-IL-15R␣ antibodies.
IL-15, we also tested for the ability to induce release of sIL-15R␣ several other agents, including TNF␣ and LPS. In L929 cells, TNF␣ can reportedly induce apoptosis, necrosis, or even both at once (27,40), whereas LPS treatment triggers expression of chemokines by these cells. 2 L929 cells were treated with IL-15, TNF␣, or LPS for different time intervals, and shedding of the IL-15R␣ chain was evaluated by FACS and ELISA. It is interesting that only TNF␣ was able to induce shedding of IL-15R␣ in L929 cells to the levels comparable with PMA treatment, leading to the 2.5-fold increase of sIL-15R␣ concentration in the culture medium (Fig. 1C), whereas IL-15 and LPS had no effect. Furthermore, TNF␣ treatment mimicked PMA action, also leading to the increase of the membrane-bound IL-15R␣ within 1 h of stimulation followed by down-regulation of the protein within next hour, as assessed by flow cytometry. The time-course changes in the expression of membrane-linked IL-15R␣ are shown in Fig. 1E. Taken together, these experiments demonstrate that murine L929 fibroblasts release sIL-15R␣ into the culture medium under natural conditions, whereas stimulation with PMA or TNF␣ enhances this process.
The Soluble IL-15R␣ Is N-glycosylated-The molecular mass of the natural sIL-15R␣ is about 30 kDa, which is approximately 5 kDa higher than the reported molecular mass of the recombinant sIL-15R␣ produced in bacteria (32). The difference in the molecular mass might result from the presence of one potential N-glycosylation site in the extracellular part of IL-15R␣ (24). To analyze a glycosylation pattern of sIL-15R␣, IL-15R␣ was precipitated from L929-conditioned medium us-ing specific antibodies. The immunoprecipitates were treated with N-glycosidase, and the resulting products were analyzed by Western blotting. As shown in Fig. 1F, treatment of sIL-15R␣ with N-glycosidase leads to the shift of the specific band to lower molecular mass position (about 26 kDa), which corresponds to non-glycosylated recombinant sIL-15R␣ (32).
Metalloproteinase Inhibitors That Discriminate between TACE and ADAM10 Differentially Block Constitutive and Inducible Shedding of IL-15Ra-Disintegrin-like metalloproteinases (ADAMs) have been implicated in the constitutive as well as the PMA-inducible ectodomain shedding of a number of cell surface-expressed molecules (5,38). Both types of shedding are blocked by broad-spectrum metalloproteinase inhibitors, such as batimastat (22). To test whether the shedding of IL-15R␣ is mediated by metalloproteinases, cells were first incubated with this inhibitor. These experiments demonstrated the ability of batimastat to suppress constitutive and PMA-induced IL-15R␣ shedding, as assessed by ELISA ( Fig. 2A) and Western blotting (Fig. 2B).
Next, we sought to determine the identity of a specific metalloproteinase responsible for the shedding of IL-15R␣. Given the important role of TACE and ADAM10 in ectodomain shedding and soluble receptor generation, we investigated the involvement of these two proteases. A panel of hydroxamatebased metalloproteinase inhibitors was recently used to analyze the involvement of the ADAM family members in fractalkine shedding (22). Among these inhibitors, two compounds were found to differ in their ability to abrogate TACE and ADAM10 action. GW280264X inhibitor blocked both enzymes, whereas GI254023X compound had potent ADAM10 inhibitory 2 E. Bulanova and S. Bulfone-Paus, unpublished observations. activity with minimal influence on TACE. Thus, the effect of these compounds on constitutive and inducible shedding of IL-15R␣ in L929 cells was analyzed. To this end, cells were treated with 10 M of either inhibitor for 1 h followed by PMA stimulation for another 2 h, and concentration of sIL-15R␣ in the cell supernatants was assessed by ELISA (Fig. 2C). The presence of sIL-15R␣ was also evaluated by Western blotting analysis after immunoprecipitation from the cell supernatants (Fig. 2D). These experiments showed that GW280264X compound clearly inhibited constitutive and PMA-induced IL-15R␣ shedding, whereas GI254023X had no effect. These chemical agents also had similar effect upon TNF␣-induced release of sIL-15R␣ (data not shown). It is noteworthy that the inhibition of the IL-15R␣ shedding by the respective metalloproteinase inhibitors (the broad-spectrum inhibitor batimastat and the inhibitor of TACE and ADAM10 GW280264X) was associated with an increase of IL-15R␣ on the cell surface or in the cell lysates, whereas all used inhibitors did not affect the total protein content in the cell lysates and cell viability within 48 h, compared with cycloheximide (data not shown). These experiments provide evidence that those chemical compounds do not act through nonspecific inhibition of protein synthesis. The selective action of the inhibitors upon respective metalloproteinase(s) is in agreement with previous studies (22,41). Taken together, these results show that TACE plays a major role in the constitutive and inducible IL-15R␣ cleavage, whereas ADAM10 seems to be dispensable for this process.

Overexpression of TACE Enhances Constitutive and Inducible IL-15R␣ cleavage-
To assess the role of TACE and ADAM10 in the IL-15R␣ shedding in more detail, we overexpressed IL-15R␣ alone or together with TACE or ADAM10 in COS-7 cells. To this end, COS-7 cells were transiently transfected with a vector coding for IL-15R␣ in combination with TACE or ADAM10. Co-transfection of IL-15R␣ with an empty vector was used as a control. The shedding of IL-15R␣ was analyzed 48 h after transfection by ELISA and Western blotting. Overexpression of IL-15R␣ alone resulted in an increased release of sIL-15R␣ into the culture medium, which was further up-regulated by PMA (Fig. 3A). These experiments indicate that endogenous MP activity is present in COS-7 cells. Co-transfection of TACE and IL-15R␣ significantly enhanced both constitutive and PMA-inducible IL-15R␣ cleavage (2.1and 1.7-fold increase, respectively, versus overexpression of IL-15R␣ alone). These results were further confirmed by the detection of sIL-15R␣ in the culture medium by Western blotting after immunoprecipitation with specific antibodies (Fig.  3B). The equal expression of transfected proteins was evaluated by probing the membrane with respective antibodies (Fig.  3C). ADAM10 is detectable as a dual band with 90-and 65-kDa molecular mass corresponding to the precursor and mature protein. Mature TACE is detectable as a band with molecular mass of 80 kDa, whereas IL-15R␣ is a 65-kDa protein. Thus, the constitutive and PMA-induced IL-15R␣ cleavage upon overexpression in COS-7 cells is mediated by TACE. As a next step, we evaluated the difference in the release of sIL-15R␣ between parental and derivative L929 cell lines (referred to as L929R), which is resistant to the cytotoxic action of TNF␣ (data not shown). Interestingly, that the spontaneous and PMA-induced release of sIL-15R␣ from L929R cells was about 1.7-2.2-fold higher compared with parental L929 fibroblasts (Fig. 3D). L929R cells constitutively express higher amounts of TACE at mRNA and protein levels (Fig. 3, E and F). Noteworthily, both cell lines express equivalent amounts of IL-15R␣ protein (Fig. 3F). Finally, we wanted to verify that the observed ability of TNF␣ to increase the production of sIL-15R␣ is not caused by its direct cytotoxic effect upon target cells. Importantly, TNF␣ was able to induce the increased release of sIL-15R␣ both in TNF␣-sensitive and -resistant cell lines (L929 and L929R, respectively), whereas the concentration of sIL-15R␣ was ϳ2.5-fold higher in the culture medium from L929R cells (Fig. 3D). Thus, TNF␣ preserves the ability to stimulate sIL-15R␣ production in the fibroblasts that are not sensitive to its cytotoxic action. Taken together, these results provide further support to the idea that a stable increase in the expression of TACE is sufficient to shed more IL-15R␣ from the cell surface.

Cleavage of IL-15R␣ Is Reduced in Murine Fibroblasts Lacking TACE but Remains Intact in ADAM10-deficient Cells-
More evidence for the role of TACE and ADAM10 in the cleavage of IL-15R␣ was obtained by using mouse embryonic fibroblasts that were generated from mouse embryos with a targeted deletion of TACE or ADAM10 as described previously (7,36). The absence of TACE or ADAM10 in respective cells was confirmed by Western blotting using specific anti-ADAM10 or anti-TACE antibodies (Fig. 4A, top and bottom). However, the endogenous expression level of IL-15R␣ was rather low in ADAM10Ϫ/Ϫ, TACEϪ/Ϫ, and respective wild type MEFs as detected by reverse transcription-PCR and Western blotting (data not shown). Therefore, MEFs from these mice were transiently transfected with a vector coding for IL-15R␣ and analyzed for the constitutive or PMA-induced shedding of IL-15R␣ into the culture medium 48 h after transfection. Both constitutive and PMA-induced release of sIL-15R␣ was dramatically lower in TACEϪ/Ϫ MEFs compared with their normal WT counterparts (Fig. 4B). However, a small amount of sIL-15R␣ was still detectable in the supernatants from these cells. As expected, the reconstitution of TACE into TACEϪ/Ϫ MEF (TACEϪ/ϪRe) completely restored sIL-15R␣ generation, in either the presence or the absence of PMA. No significant changes in sIL-15R␣ production were observed in ADAM10Ϫ/Ϫ versus ADAM10ϩ/ϩ MEFs. The expression levels of IL-15R␣ are shown as controls (Fig. 4A, middle). These results further support an important role for TACE in the constitutive and inducible release of sIL-15R␣.
Detection of sIL-15R␣ in Mouse Serum-To identify sIL-15R␣ in the body fluids of mice, sera from at least five animals per strain were first tested by ELISA. As shown in Fig. 4C, Balb/c, C57BL/6, and CH3 mice have different levels of circulating sIL-15R␣ varying from 18 ng/ml in Balb/c to 40 ng/ml in C57BL/6 mice. The specificity of the newly developed sIL-15R␣-ELISA was convincingly demonstrated by the complete absence of the detectable protein in the sera of IL-15R␣Ϫ/Ϫ mice (Fig. 4C). These results were further corroborated by Western blotting analysis, which showed the presence of a band representing a protein with a molecular mass of ϳ30 kDa corresponding to the N-glycosylated form of sIL-15R␣ after immunoprecipitation from serum using specific antibodies (Fig. 4D).

DISCUSSION
In this study, we show the existence of a natural soluble form of the high affinity IL-15 receptor ␣ chain in the mouse serum and cell-conditioned medium. Furthermore, our data indicate that proteolytic cleavage represents a major mechanism for sIL-15R␣ generation in mice by providing several lines of evidence that unambiguously demonstrate an important role for TACE in the IL-15R␣ shedding. Murine fibrosarcoma cells constitutively release detectable amounts of sIL-15R␣ into the culture medium, and the generation of sIL-15R␣ is further enhanced by PMA stimulation. Both a broad-spectrum metalloproteinase inhibitor batimastat and an inhibitor of TACE and ADAM10 (GW280264X) inhibited the spontaneous and inducible shedding of IL-15R␣, whereas a selective inhibitor of ADAM10 (GI254023X) did not affect sIL-15R␣ production. Moreover, the release of sIL-15R␣ by fibroblasts derived from mice with a homozygous disruption of the gene encoding TACE was dramatically reduced compared with shedding by corresponding wild-type cells. The reconstitution of active TACE expression in TACE-deficient cells restored both spontaneous and PMA-inducible IL-15R␣ shedding.
IL-15R␣ is thus a novel substrate for TACE, although our results do not exclude a possibility that TACE is not the only MP involved in the proteolytic cleavage of IL-15R␣. The fact that some sIL-15R␣ is still detectable in the supernatants of TACEϪ/Ϫ cells is a very strong argument for the involvement of other proteases. Current studies in our laboratory are focused on the identification of putative additional members of the ADAM family, which may contribute to the constitutive and/or inducible shedding of this receptor. Notwithstanding, how TACE recognizes its diverse targets and features in ADAM substrates that allow their proteolysis remains obscure (4). It has been suggested that structural and kinetic characteristics rather than minimal consensus shedding sequences and identity of amino acids consisting them regulate access of the protease to the cleavage site (42,43), although the Ala-Val preference for TACE has recently been reported (44). The most consistent feature indicative of cleavage site among ADAM substrates is that they usually reside in a stalk region between the membrane and an initial globular extracellular subdomain (45). The identification of amino acids comprising a potential cleavage site for TACE in the extracellular region of IL-15R␣ will provide helpful insights to our understanding of this process.
The proteolytic cleavage and release of transmembrane cell surface proteins or ectodomain shedding has emerged as an important post-translational mechanism to regulate the function of cell surface proteins (2,7). Many structurally and functionally diverse proteins, which are initially synthesized as membrane-anchored moieties, can be shed from the cell surface in a biologically active form, including adhesion molecules, cytokines, growth factors, and their receptors (1,2). Ectodomain cleavage and shedding of N-glycosylated IL-15R␣ might have a particular physiological and/or pathological relevance for the biology of IL-15, and supposedly affect the function of this pleiotropic cytokine in several aspects. IL-15 plays important roles in both innate and adaptive immunity, and is associated with a wide range of immunopathological reactions, including rheumatoid arthritis and allograft rejection (32,34). The bioavailability of sIL-15R␣ might contribute to complex regulatory mechanisms of IL-15 secretion, which is tightly controlled at multiple levels (46). Regulation of the availability of free IL-15 may be an important protective mechanism against excessive IL-15 activity. The presence of sIL-15R␣ in biological fluids could affect function of IL-15 in the first place by competing for the cytokine with the cellular receptors and inhibiting its action. Given that IL-15 has many features of an attractive therapeutic target in inflammatory synovitis (47)

W I T H D R A W N W I T H D R A W N
F e b r u a r y 9 , 2 0 1 1 F e b r u a r y 9 , 2 0 1 1 15R␣ might also affect the bioavailability of IL-15 by prolonging the cytokine half-life and slowing the release of the ligand from a complex between the soluble receptor and IL-15 to provide physiological concentrations of the cytokine in tissues. Furthermore, sIL-15R␣ may expand IL-15 action from autocrine or juxtacrine to paracrine or endocrine modes, resulting in local or systemic effects of the cytokine, and influence the nature and/or duration of the signaling event. High or low levels of sIL-15R␣ may correlate with distinct pathological conditions and possibly serve as a prognostic marker for certain types of diseases, such as allergy and inflammation. Moreover, membrane-bound IL-15 present on monocytic THP-1 cell line and human monocytes participates in reverse signaling, inducing activation of extracellular signal-regulated kinase 1/2 and p38, and production of IL-8 (48). The ability of membranebound IL-15 to mediate reverse signaling is also supported by data from our laboratory (57 Alternative splicing of specific transcripts that gives rise to a polypeptide lacking transmembrane and/or other regions represents another mechanism, which accounts for the appearance of secreted forms of diverse membrane-linked proteins. For example, sIL-5R and sIL-7R are generated by alternative splicing (50,51), but TNF receptors are cleaved from the cell surface (10), whereas for sIL-6R, sIL-4R, and growth hormone receptor, both mechanisms were shown (3,52,53). In addition, a machinery of soluble receptor generation may undergo evolutionary divergence in different species. A classic example of such phenomenon is generation of growth hormone receptor by proteolytic cleavage in rabbit and human and by alternative splicing in mouse and rat (42,53,54). Thus, cellular mechanisms accounting for the release of sIL-15R␣ in mouse and human may be different. Moreover, a release of full-length p55 TNFR in exosome-like particles has recently been demonstrated as a novel mechanism for soluble cytokine receptor generation (55). Although our results support the idea that sIL-15R␣ in mouse is produced mainly by proteolysis that involves TACE, it is not clear yet whether an alternative splicing mechanism and/or a release in exosome-like particles may also contribute to the generation of sIL-15R␣, and additional studies are required to address this question.
Despite the fact that certain ligands have been shown to enhance the release of their soluble receptors (10, 37), IL-15 treatment clearly had no effect on sIL-15R␣ production by murine fibroblasts. Instead, TNF␣ was able to trigger the release of sIL-15R␣ from these cells. Noteworthily, the proteolytic cleavage of soluble TNF receptors I and II in human polymorphonuclear leukocytes reportedly relies upon TACE and is enhanced by TNF␣ treatment (10). Thus, an increase in the production of sIL-15R␣ may represent a concomitant event that results from an enhanced proteolytic activity of TACE induced by TNF␣ stimulation. The involvement of TACE is further supported by the ability of GW280264X to block TNF␣induced sIL-15R␣ release into the culture medium, whereas a FIG. 4. Analysis of sIL-15R␣ release in MEFs lacking TACE or ADAM10 and detection of sIL-15R␣ in mouse sera. A, deficiency of TACE or ADAM10 in respective MEFs, and equal expression of IL-15R␣ after transfection was confirmed by Western blotting using specific antibodies. B, IL-15R␣ shedding in TACEϪ/Ϫ, TACEϩ/ϩ, ADAM10Ϫ/Ϫ, ADAM10ϩ/ϩ, and TACEϪ/Ϫ reconstituted with TACE (TACEϪ/ϪRe) MEF cell lines. Cells were transfected with IL-15R␣ expression vector and stimulated 48 h after transfection with PMA for 2 h or left untreated. Thereafter, conditioned media were analyzed for the presence of sIL-15R␣ by ELISA. C and D, sera from C57BL/6, CH3, Balb/c, and IL-15R␣Ϫ/Ϫ mice were tested for the presence of sIL-15R␣ by ELISA (C) and Western blotting (D). Bars represent the mean of at least five mice per strain (**, p Ͻ 0.01). selective ADAM10 inhibitor GI254023X was ineffective (data not shown). Notwithstanding the above explanation, the mechanism underlying the ability of TNF␣ to enhance sIL-15R␣ production is not yet clear and deserves a systemic exploration in follow-up experiments. Furthermore, PMA-induced up-regulation of membrane-anchored IL-15R␣ within the first hour of the treatment is interesting and is not seen with IL-6R and fractalkine (11,23). Given that IL-15R␣ could be detected as a protein associated with the endoplasmic reticulum, the Golgi, and perinuclear space by confocal microscopy and Western blotting (31), and the ability of this receptor to recycle between endosomes and the cell surface (56), the observed increase in the amount of membrane-bound IL-15R␣ within the first hour of treatment with PMA or TNF␣ is probably a result of mobilization of the receptor from intracellular pools.
In summary, our data provide evidence that, like many proteins with a single transmembrane domain, the high affinity ␣-chain of the murine IL-15 receptor exists not only in membrane-anchored but also in soluble form, and implicate TACE as a metalloproteinase responsible for the shedding of this receptor. Given that an excessive or insufficient production of cytokines is often the main cause of the pathology in many types of diseases and the emerging importance of soluble cytokine receptors as novel potential immunotherapeutic agents, a dissection of the role played by sIL-15R␣ in a complex biology of IL-15 under physiological or pathological conditions will facilitate the development of new tools for therapeutic intervention.