Mouse Brain Microglia Express Interleukin-15 and Its Multimeric Receptor Complex Functionally Coupled to Janus Kinase Activity*

The cytokine, interleukin (IL)-15, and the T cell growth factor, IL-2, exhibit a similar spectrum of immune effects and share the IL-2 receptor (IL-2R) subunits IL-2Rβ and IL-2Rγ for signaling in hematopoietic cells. Numerous neuroregulatory activities of IL-2 have been suggested, but its expression in the normal central nervous system (CNS) is apparently very low and regionally restricted. We show by RNA and protein detection that IL-15, its specific receptor molecule, IL-15Rα, and the signal-transducing receptor subunits, IL-2Rβ and IL-2Rγ, are constitutively present in various regions of the developing and adult mouse brain. We further demonstrate, also at the single-cell level, that IL-15 and the components for IL-15Rα/IL-2Rβγ receptors are expressed by microglia. Tyrosine phosphorylation data are presented showing that IL-15 signaling in microglia involves Janus kinase 1 activity. At doses of 0.1–10 ng/ml, IL-15 affected functional properties of these cells, such as the production of nitric oxide, and supported their growth in culture, suggestive of a role as an autocrine growth factor. Microglial IL-15 could thus play a pivotal role in the CNS and may participate in certain CNS and neuroendocrine functions previously ascribed to IL-2.

Interleukin (IL) 1 -15 exhibits a spectrum of immune functions that largely overlaps with that of the T cell growth factor, IL-2, since both cytokines share signal-transducing receptor proteins on hematopoietic cells (1)(2)(3)(4). IL-2 activities are not restricted to the immune system. Experimental and clinical findings show that IL-2 can evoke a variety of physiological responses and toxic alterations in the CNS (5). Bearing in mind the functional similarity (although not identity) of IL-2 and IL-15, we recently proposed that a brain-intrinsic IL-15 could account for or relate to some of the neuromodulatory and neu-roendocrine activities that have formerly been attributed to IL-2 (5).
IL-15 stimulates, like IL-2, the growth of activated T cells and promotes cytotoxic activity and B cell differentiation, but it is, in addition, a potent chemoattractant for T lymphocytes (1). IL-15 was also found to displace IL-2 from its binding sites on immune cells, although both molecules do not share sequence homology (1). The striking similarity in their binding behavior and biological activities is structurally underscored by a predictably similar folding topology. This places IL-15 as a new member of the four-␣-helix bundle cytokine family, to which IL-2 belongs (6). It has been further proposed that IL-15 effects are mediated via components of the IL-2R system (1).
The conceptual model of IL-2R in hematopoietic cells depicts three single chain proteins binding IL-2 either alone or as diand trimeric complexes associating with various enzymatic and modulatory components (7). IL-2R␣ binds IL-2 with low affinity (k D ϳ 10 Ϫ8 M). IL-2R␤ and IL-2R␥ build up the heterodimeric receptor of intermediate affinity (k D ϳ 10 Ϫ9 M). Upon trimerization, IL-2R␣␤␥ represents the high affinity IL-2 receptor (k D ϳ 10 Ϫ11 M). Both the IL-2R ␤ and ␥ chains transmit IL-2 signals to the cytosol, while the inducible IL-2R␣ has no known intracellular consequence and is regarded more as a regulator of IL-2 binding (7,8). IL-2R␥ has been designated the "common ␥ chain," ␥ c , since it is a constituent of IL-4, IL-7, and IL-9 receptors also (7,8).
To substantiate the suggested physiological contribution of the IL-15/IL-15R system to CNS functions, we analyzed the expression of IL-15 in the brain together with that of all the components required to compose functional di-and trimeric IL-15R complexes, including IL-2R␤ and IL-2R␥.
In addition, we addressed the question whether microglia could produce and respond to IL-15, since these cells are a major source and target of cytokine activities in the brain (13). Microglia serve roles in tissue maturation and defense by virtue of their cytotoxic and phagocytotic abilities (14,15); probably exhibit supportive and repair functions, since they produce neurotrophic factors (16); and assist in specific immune responses by taking the role of resident antigen-presenting cells (17). Microglia rapidly respond to changes in brain tissue homeostasis, but their gradual activation and executive responses are under tight control (17,18). Although the underlying mechanisms are poorly understood, cytokines are thought to be crucially involved (13).
After demonstrating that microglia harbor the structural elements for IL-2 and IL-15 receptors, the next question was to ask whether IL-15 signaling involves members of the Janus kinase (JAK) family. This new group of non-receptor proteintyrosine kinases plays a prominent role in the execution of cytokine effects in the immune system and is involved in signaling through IL-2R as well (19).
Our findings on a microglial production of the T cell chemoattractant, IL-15, and the widespread expression in the brain of the IL-2/IL-15 receptor system may also be relevant to the understanding of inflammatory processes, post-traumatic episodes, and the deleterious CNS side effects of immunotherapies (20,21).

EXPERIMENTAL PROCEDURES
Animals and Microglial Cultures-Microglial cell cultures were prepared from newborn NMRI mice or Wistar rats obtained from the local animal facility or from Charles River (Sulzfeld, Germany). The tissue culture preparation has been previously described in detail (22). The purity of cell cultures was verified by Griffonia simplicifolia isolectin B4 staining (22). For sampling of brain tissues, animals were anesthetized with sodium pentobarbital (Nembutal TM , Sanofi, Hannover, Germany; 50 mg/kg, body weight) and transcardially perfused with 20 mM Na 2 HPO 4 /NaH 2 PO 4 buffer, pH 7.4, containing 140 mM NaCl (PBS) to obtain virtually blood-free tissue homogenates. Although removal of circulating blood cells cannot completely rule out contamination of brain tissue with hematopoietic material, it largely reduced the risk of collecting extraneural RNA and protein.
Cytokines-Human recombinant IL-15 and IL-2 were obtained from R & D Systems (Wiesbaden, Germany) and Boehringer Mannheim (Mannheim, Germany). Both human proteins are known to be active in the rodent system.
Electrophysiology-Patch clamp recordings from cultured microglial cells were made in the whole cell voltage clamp configuration as described previously (22).
Immunoprecipitation and Immunodetection-Immunoprecipitations were performed on microglial lysates (5-6 ϫ 10 6 cells, with and without prior stimulation by 150 ng/ml IL-15 for 30 min) and brain tissue homogenates (100 mg, wet weight, in 1 ml) obtained from PBS-perfused P6, P20, and (adult) P63 mice. Cells were lysed for 15 min on ice in 20 mM Tris/HCl buffer, pH 7.3, 140 mM NaCl, 0.5% Triton X-100, 1 tablet/50 ml Complete TM protease inhibitors (Boehringer Mannheim), and 2 mM sodium orthovanadate. Brain tissues were homogenized and extracted with the same buffer. Extracts were cleared by centrifugation and sonicated. Proteins of interest were immunoprecipitated for 12 h at 4°C with 1-2 g/ml concentrations of the respective antibodies (biotinylated rat anti-mouse IL-15, Pharmingen, Hamburg, Germany; goat anti-mouse IL-15R␣, C-19, and N-19, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; rabbit-anti-mouse JAK1, Santa Cruz Biotechnology, Inc. or Pharmingen). Immune complexes were removed by 2 h of incubation with 100 l of a 10% protein A-Sepharose CL-4B suspension (Pharmacia, Freiburg, Germany), equilibrated with 20 mM Tris/HCl buffer, pH 7.3, 140 mM NaCl, 0.1% Triton X-100, 5 mM NaN 3 . The Sepharose was collected by centrifugation, suspended in the same buffer, loaded on 1 M sucrose in buffer, centrifuged again, and resuspended in buffer. The latter washing step was repeated, and the pellet was used for blot analysis.
Assays for Microglial Survival, Growth, and Nitric Oxide (NO) Production-Microglial survival was assayed in 96-well plates with 40,000 cells/well using a lactate dehydrogenase (LDH) kit (Boehringer Mannheim). LDH activity in the supernatants was measured following 24 or 48 h of incubation in the absence (controls) or presence of varying concentrations of IL-15. Similarly, the MTT test involving MTT-3 reagent (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide, Sigma) was carried out in 96-well plates following a standard protocol (25). The assay is based on the enzymatic cleavage of MTT tetrazolium salt to formazan by the succinate-tetrazolium reductase system of the respiratory chain of intact mitochondria and NAD(P)H-dependent enzymes of the endoplasmic reticulum. The effects of IL-2 and IL-15 (1 and 10 ng/ml) on microglial NO production were studied in cultures with and without stimulation by lipopolysaccharide (LPS, Sigma, 50 ng/ml, last 24 h). NO was determined in culture supernatants after 76 h of cytokine exposure by the Griess reaction for the accumulation of nitrite (NO 2 Ϫ ), a breakdown product of NO, using sodium nitrite as a standard. 100 l of each supernatant were mixed with 100 l of Griess reagent and incubated at room temperature for 10 min, and the optical density was determined in a microplate reader (SLT) at a 540-nm wavelength. Protein content was determined by means of a bicinchoninic acid-based assay (Pierce).
Ca 2ϩ Imaging-Imaging of intracellular Ca 2ϩ levels in microglia was performed as described in detail elsewhere (26).
Assays for Motility and Chemotaxis-Microglial motility and chemotaxis were studied by time lapse videomicroscopy and computerassisted motility analysis as described previously (22), using IL-15 and IL-2 at 0.1-10 ng/ml. Briefly, cells on coverslips were placed in a 37°C temperature-controlled chamber on the stage of an inverted microscope (Zeiss IM100, Zeiss) and perfused with prewarmed standard salt solution. Single images were recorded with a frame grabber card (Movieblaster, CPS, Hamburg, Germany) every 12 s at a resolution of 200 ϫ 320 pixels and 128 gray values using a ϫ 40 phase contrast objective and a CCD video camera (Variocam, Phase, Bremen, Germany). The image sequences were analyzed and converted to traces of relative motility by a customer-designed program (22). Chemotaxis was studied as described using a 48-well microchemotaxis chamber (22).
Statistics-Statistical evaluation was carried out with the STATIS-TICA computer program (StatSoft, Tulsa, OK) and using the paired t test, p Ͻ 0.05 being considered significant. Data are given as mean Ϯ S.E.

IL-15 and IL-15R␣ mRNAs Are Expressed in Mouse Brain
and Cultured Microglia-RT-PCR analysis revealed the expression of IL-15 mRNA in mouse brain and cultured microglial cells (Fig. 1A). The corresponding amplimer with an expected size of 422 bp was present in tissue samples obtained from buffer-perfused (virtually blood-free) mouse brain as it was seen in preparations of positive control tissue, the spleen. IL-15-encoding transcripts could also be demonstrated in extracts of cultured microglia.
Similarly, RT-PCR was applied to investigate the expression of IL-15R␣ in extracts of both perfused mouse brain and mouse microglial culture. As shown in Fig. 1B, amplification of IL-15R␣ mRNA-derived cDNA fragments results in a product of the size (386 bp) predicted by sequence and as generated from control spleen RNA material. The identity of amplimers obtained with primers for IL-15 and IL-15R␣ was verified by DNA sequencing.
Expression of IL-15 and IL-15R␣ Can Be Demonstrated at the Single-cell Level-Single-cell RT-PCR analyses were performed on RNA material harvested from electrophysiologically identified microglial cells to support the in vitro data and to rule out any contamination by other cell types (Fig. 2, A-F). Microglial cells were identified by their whole cell currents in addition to inspection of their characteristic morphology. During patch clamp recording, a series of de-and hyperpolarizing voltage steps from the holding potential V h (Ϫ70 mV) was applied to activate voltage-dependent currents. All of the cells used for single-cell analysis exhibited the characteristic current pattern for cultured, unstimulated microglial cells, i.e. expression of inwardly rectifying K ϩ currents elicited at hyperpolarizing potentials and lack of outward K ϩ currents. Fig. 2, C and F, illustrate the current pattern of the microglial cells: hyperpolarizing voltage steps triggered the currents, while with depolarization, no currents above leakage were observed (27).
Cytoplasmic RNA material was collected through the patch pipette by applying negative pressure. Immediately after harvesting, the cytosolic RNA was reverse-transcribed into cDNA, followed by seminested PCR with primer pairs spanning sev-eral exons (Fig. 2, A and D) to amplify cDNA transcripts encoding either IL-15 or its specific receptor subunit, IL-15R␣. The results of the single cell RT-PCR analysis of IL-15 transcripts are given in Fig. 2B. Of 13 cells included in this study, 11 showed the predicted amplimer (85%). As similarly illustrated in Fig. 2E, the set of experiments performed with IL-15R␣-specific primer pairs allowed for the demonstration of the respective transcripts in 63% of the cells identified as microglia (19 of 30 microglial cells).

IL-15 and IL-15R␣ Proteins Are Present in Mouse Brain and Microglial
Culture-Having demonstrated the presence of messages encoding for IL-15 and IL-15R␣, the next step was to show the expression of the respective proteins. Immunoprecipitations were exploited in several sets of preparations (n Ն 3) to isolate IL-15 and IL-15R␣ proteins from mouse brain and microglia. Fig. 3A shows that antibodies against mouse IL-15 precipitated a protein from brain homogenates of young mice of P6, P20, and P63 (adult mice). The protein migrated in SDS-PAGE with an apparent molecular mass of 17-18 kDa, the expected size range of IL-15. It was also found in extracts of microglial cultures (Fig. 3A).
In a similar experiment, anti-IL-15R␣ antibodies allowed for the purification and subsequent detection of the receptor protein from both perfused brain (Fig. 3B) and microglial culture (Fig. 3C), showing an apparent molecular mass of ϳ65 kDa.
Phosphorylation of Tyr residues is an established signaling event following IL-2R/IL-15R activation in lymphocytes. In line with those experiments performed with microglia, we also studied whether phosphorylation of IL-15R␣ would occur when these cells were treated with IL-15. By use of anti-Tyr(P) antibodies, phosphorylation of Tyr residues in immunoprecipitated microglial IL-15R␣ protein was detected. There was no apparent difference in the intensity between material obtained from IL-15-treated cells and control (untreated) cultures (Fig.  3D), indicating that IL-15R␣ is constitutively phosphorylated without further phosphorylation depending on IL-15 signaling. As a target of phosphorylation, Tyr 227 (also Tyr 227 in the human sequence) can be proposed, since the cytosolic moieties of mouse and human IL-15R␣ contain only one Tyr, positioned proximal to the transmembrane region (10,12). Anti-Tyr(P) staining also revealed another phosphorylated, yet unidentified, protein band that coprecipitated with IL-15R␣ and migrated with an apparent molecular mass slightly higher than that of the IgG heavy chain of the precipitating antibody (Ն50 kDa, Fig. 3D).
Mouse Brain and Microglia Express the Components for Trimeric IL-2R/IL-15R Complexes-Functional IL-15 binding on immune cells requires not only IL-15R␣ but also the presence of IL-2R␤ and IL-2R␥. Therefore, the presence of transcripts encoding these subunits was investigated. Fig. 1B reveals that   FIG. 1. RT-PCR analysis of transcripts for IL-15 and the components of the IL-2R/IL-15R system. A, RT-PCR analyses for IL-15 were performed with extracts of mouse microglial cell cultures (MG), perfused (virtually blood-free) mouse brain (Br), and mouse spleen tissue samples (Sp). B, RT-PCR analyses were also performed with extracts of mouse microglial cells, perfused (virtually blood-free) mouse brain, and spleen tissue samples to reveal the presence of mRNAs for IL-15R␣, IL-2R␣, IL-2R␤, and IL-2R␥. The positions of the predicted PCR products are indicated by arrowheads. Molecular size markers (M) are shown with their sizes given in bp.
the corresponding RT-PCR products for IL-2R␤ in microglia and brain extracts were 345 bp as seen with the corresponding spleen standard. IL-2R␥ mRNA-derived PCR products of microglial and mouse brain samples showed the predicted size of 371 bp and were identical to the amplimer produced with spleen tissue material. To obtain a complete pattern of IL-15 and IL-2 receptor subunits, we also studied the expression of IL-2R␣ in microglia in vitro and in brain homogenate. The predicted 301-bp product was obtained from all of the RNA extracts, i.e. material isolated from the spleen, brain, and microglia. The identity of all cDNA fragments specific for the IL-2R subunits was verified by DNA sequencing. The expression of IL-15 and IL-15R␣ protein was detected in extracts of cultured microglia as well as homogenates of (perfused) P6, P20, and adult (P63) mouse brain by immunoprecipitation and subsequent immunoblotting. A, IL-15 protein isolated from microglial culture (MG) and brain of various ages (P6, P20, and P63 (adult)). B, IL-15R␣ protein as precipitated from P6, P20, and P63 mouse brain. C, IL-15R␣ protein detected in microglial culture (MG), either treated (IL-15) or not (control, CTL) with 150 ng/ml IL-15 for 30 min. D, removal of the detecting antibodies from the IL-15R␣ blot and reprobing with antibody against Tyr(P) revealed Tyr phosphorylation of IL-15R␣ as well as a yet unidentified protein slightly bigger by apparent molecular mass than the IgG heavy chain. Treatment of microglia with IL-15 (compare CTL with IL-15) did not alter the phosphorylation intensity significantly. Note that the capturing antibody's 50-kDa IgG heavy chain and 25-kDa light chain were also stained by the immunodetection procedures when the same antibody was used for detection (B and C). Molecular mass markers (kDa) are shown in C.
Taken together, our PCR analyses revealed the presence of transcripts in cultured mouse microglia and normal brain tissue of all of the molecular components forming multimeric IL-15 and IL-2 receptors as described previously for the peripheral immune system.

IL-15 and IL-2R/IL-15R are widely expressed throughout the CNS-
The expression pattern of IL-15 and IL-15R␣ was studied in various regions of perfused mouse brain at different ages, P6, P20, and adulthood (P63). The distribution pattern of transcripts was analyzed for frontal and parietal cortex, hippocampus, medulla oblongata, and cerebellum, demonstrating a widespread expression of IL-15 and IL-15R␣ (Fig. 4A). IL-15R␣ transcripts were found in all tissues at all time points as investigated in two identical but independent sets of experiments. In contrast, IL-15 message in the hippocampal formation appeared only with P20, since it was not detectable in the corresponding region of P6 brain. Similarly, IL-15 mRNA of the medulla oblongata was clearly found only at later developmental stages in the tissue samples obtained from adult mouse brain.
To obtain a complete survey of the IL-2R/IL-15R system, we investigated the regional and ontogenetic expression pattern of the known members of the IL-2R (Fig. 4B). IL-2R␣, IL-2R␤, and IL-2R␥ transcripts were detected in all anatomical subdivisions as investigated at varying developmental stages. These data reveal a widespread presence of the messages encoding for the known subunits constituting di-and trimeric IL-2R and IL-15R.
IL-15 Induces Phosphorylation of Microglial JAK1-As shown in Fig. 5, microglia synthesize the Janus type proteintyrosine kinase, JAK1, with JAK1 phosphorylation occurring in these cells upon stimulation with IL-15.
RT-PCR analyses revealed JAK1-like transcripts in both microglial culture and brain extracts, with the corresponding amplimers occurring at a size (414 bp) identical to the splenic control (Fig. 5, left lanes). Furthermore, immunoprecipitation of JAK1 protein was carried out (in two experiments) with cell lysates from IL-15-treated and control cultures. Using a specific anti-Tyr(P) antibody, pronounced phosphorylation of JAK1 was selectively seen in blots of the experimental sample (Fig. 5, right lanes). Subsequent removal of the detecting antibody complex ("stripping") and reprobing of the blot with anti-JAK1 showed the presence of similar protein amounts in both extracts, the control and the IL-15-treated microglia (Fig. 5,  middle lanes). As a comparison, it should be noted that demonstration of JAK phosphorylation during IL-2 and IL-15 signaling in lymphocytes often involves larger cell numbers and higher concentrations of the cytokines (10) than those used in the present investigation. Consequently, the successful demonstration of JAK1 activation in microglia indicates either relatively high expression of this protein-tyrosine kinase or a sensitive response to IL- 15   IL-15 Supports Microglial Survival-IL-15 supported the survival of microglia in culture at doses of 0.1-10 ng/ml as determined (in repeated assays) by means of a conventional cytotoxicity detection kit based on LDH activity (Fig. 6A). In this assay, the extracellular presence of the cytoplasmic LDH is taken as an indicator of plasma membrane damage. When the cells were given fresh medium containing IL-15, we measured significantly less LDH activity in the supernatants after 24 h of incubation compared with cells lacking IL-15 supplement. Decreases in LDH activity of 60 -80% were seen over a dose range determined for various IL-15 effects on immune cells (3,4). 10 ng/ml of IL-15 were less effective than lower doses, resembling the bell-shaped dose-response relationship known from many cytokine activities. However, when LDH was determined in microglial supernatants after 48 h, IL-15 decreased LDH activity only in the range of 15-50%, compared with the untreated control. A loss of IL-15 bioactivity over the incubation time is unlikely, since IL-15 is a stable protein (R & D Systems). Alternatively, microglial cells can synthesize IL-15 themselves, which over time could sufficiently reaccumulate in the freshly replaced culture supernatant, rendering the cells less dependent on exogenous IL-15 supply after 48 compared with 24 h.
Similarly, about a 40% increase in cell numbers over control was determined in microglial cultures incubated with 1 ng/ml IL-15, using the MTT assay (Fig. 6B). Complementary to the LDH test, this assay measures the activity of living cells to reduce MTT reagent. The data from both experiments suggest that IL-15 supports microglial survival in culture.
IL-15 Attenuates Microglial NO Production-NO is a release product of activated microglia and is thought to play a major role in microglia-mediated cytotoxicity. The modulatory effects of IL-15 and IL-2 on microglial NO release are compared in Fig.  7. As expected, basal NO production of resting (unstimulated) microglia was low, while treatment with bacterial lipopolysaccharide (LPS, 50 ng/ml for 24 h) induced a marked (about 20-fold) increase in NO release. IL-15 at a concentration of 10 ng/ml led to a 30% reduction of LPS-induced NO production, as shown in Fig. 7, and to 25 and 43% reduction in two additional sets of experiments. On the contrary, IL-2 at 10 ng/ml augmented LPS-stimulated NO secretion by about 65%. The data indicate that both cytokines may exert contrasting modulatory influences on microglial NO production.
IL-15 Does Not Affect Microglial Motility-Migration and chemotaxic movement are characteristic microglial responses to activating stimuli (18). Since IL-15 is known to be a chemoattractant for T cells (29), we tested the hypothesis that IL-15 affects microglial motility. We studied random, nondirected microglial motility by computer-assisted time lapse video microscopy (three experiments) and active, directed microglial translocation along a cytokine concentration gradient in a microchemotaxis assay (two experiments) (22). Neither IL-15 nor IL-2 was able to influence significantly any type of motile activity (data not shown), while another known chemotaxic factor, complement factor C5a, had a clear dose-dependent effect, as previously reported (22). DISCUSSION This study demonstrates that mouse brain tissues as well as microglia in vitro express IL-15, the recently identified IL-2like cytokine, together with its specific receptor subunit, IL-15R␣. Mouse brain and microglial cells were also found to harbor the signal-transducing subunits, IL-2R␤ and IL-2R␥, indispensable for assembling functional di-and trimeric IL-2 and IL-15 receptor complexes. In addition, microglial IL-15R is shown to be linked to JAK1, indicating the presence of the novel Janus kinase messenger system in these cells.
The IL-15/IL-15R System Can Account for IL-2 Effects in the CNS-We recently proposed that IL-15 may be found relevant to CNS functions, most notably those that have been suggested for IL-2 (5). The extensively studied cytokine IL-2 was shown to induce pronounced effects in neurons and glia of various brain regions. However, based on current evidence, there is weak anatomical match between IL-2 responsiveness/binding molecules and the expression of endogenous IL-2 in the brain, with IL-2 being hardly detectable by means of conventional immunostaining or in situ hybridization techniques (5,30). Very low brain tissue concentrations of IL-2 could relate to the sometimes very local action and extreme potency of IL-2 (31). On the other hand, lymphocytic IL-2 was shown to cross the bloodbrain barrier and may have access to IL-2R on brain cells upon massive peripheral induction (5).
IL-15, distinct by sequence but IL-2-like by activity, could well account for at least some IL-2-like effects in the CNS. We demonstrate that IL-15 and IL-15R␣ transcripts are both expressed in the mouse brain throughout ontogeny. Sequencing of the cDNA products revealed identity to the immune systemderived counterparts. In addition, microglia are shown (also at the single-cell level) to serve as a cellular source for both mRNAs. These data and a recent report on IL-15 mRNA in fetal human cell cultures (32) together reveal IL-15 expression in CNS cells at the transcriptional level. Nevertheless, at least two mechanisms additionally control IL-15 production at posttranscriptional level (1). First, multiple AUG start codons in the long 5Ј-untranslated region of human and mouse IL-15 mRNA effectively interfere with translation. Second, IL-15 contains an unusually long signal peptide with its experimental manipulation dramatically increasing IL-15 levels. It is worth emphasizing that our data also reveal IL-15 and IL-15R␣ protein synthesis in vivo and in vitro, since we could extract these molecules from mouse brain tissue and microglial culture.
Mouse Brain and Microglia Express IL-2R/IL-15R Complexes-The principle of subunit sharing is a feature of many cytokine receptors (6), but in functional IL-15 receptors, it extends to the incorporation of both IL-2R signaling proteins, IL-2R␤ and IL-2R␥ (1). PCR data, Northern blot, and in situ hybridization for IL-2R␤ expression have recently been reported for mouse and human brain, whereas IL-2R␥ has not yet been detected in normal brain cells or tissues, although this particular subunit is also an integral part of IL-4, IL-7, and IL-9 receptors (5).
We show that the normal CNS and, at a cellular level, microglia have the potential to form dimeric IL-2R␤␥ as well as both trimeric IL-2R␣␤␥ and IL-15R␣/IL-2R␤␥ complexes. This does not imply parallel IL-2 and IL-15 signaling per se. The relative expression levels of the proteins in a given cell may dictate trimerization favoring either IL-2 or IL-15 signaling (10,11). IL-2R␣ was recently shown to be enhanced in microglial cells upon stimulation with LPS (33). This could suggest that under resting conditions (without stimulation) microglia may carry constitutive IL-15R (trimers), since we detected significant amounts of IL-15R␣ transcripts and protein in untreated cultures. The ratio of both ␣ subunits could itself be subject to variation during development or changes in CNS activity or homeostasis as it is similarly discussed for monocytes, T, and NK cells (7).
JAK Kinases and Signaling through Microglial IL-2R/IL-15R-The present work shows that JAK1, the Janus-type protein-tyrosine kinase associated with IL-2R␤ in immune cells, is expressed in mouse CNS and microglia. Both JAK1 transcripts and protein were detected in extracts obtained from these cells in vitro. Stimulation of microglial cultures with IL-15 resulted in Tyr phosphorylation of JAK1, indicating that the microglial IL-15R is able to transmit a signal to the cytosol and that JAK1 has catalytic activity in this process.
Members of the JAK family of protein-tyrosine kinases were recently shown to play a predominant role in cytokine receptor signaling and to physically and functionally interact with the IL-2R ␤ and ␥ chains (19,34). IL-2R/IL-15R stimulation in immune cells results in the phosphorylation of JAK1 and JAK3 and the subunits ␤ and ␥. This creates binding sites for signal transducer and activator of transcription (STAT) proteins, such as STAT 3 and STAT 5, that upon phosphorylation and homo-or heterodimerization, translocate to the nucleus, where they act as activators of gene transcription (35,36). We also obtained preliminary evidence for IL-2R␥ protein expression in mouse brain and microglia as well as JAK3 message and protein in mouse brain (data not shown). Since JAK3 is believed to be present primarily in hematopoietic cells (37), microglia are the most likely candidate in the CNS. Further recruitment of cytosolic enzymes and adapters, such as p85 of the phosphoinositide 3-kinase, Shc, Grb2/Sos, Ras, and Raf, by IL-2R␤ and IL-2R␥ are thought to link the JAK/STAT pathway to extracellular signal-regulated kinase/mitogen-activated protein kinase (MAPK/ERK) routes of signaling, which enables several ways of transcriptional activation (7, 38 -41).
Together, the present findings reveal a microglial IL-2R/IL-15R complex that is coupled to JAK-mediated signaling. This novel messenger system may prove to be of similarly basic importance for CNS cells as is now rapidly becoming apparent for immune cells.
Physiological Functions and Clinical Implications of the Microglial IL-15/IL-15R System-The functional in vitro data, in conjunction with the demonstration of IL-15 synthesis and receptor expression, are suggestive of IL-15 being an auto/ paracrine factor regulating microglial survival and function. Moreover, neural cells and brain structures found to be responsive to IL-2 will conceivably also respond to IL-15. While few regions, such as the hippocampal formation or the median eminence/arcuate nucleus, are enriched with both IL-2R and IL-2-like molecules (5), other areas may rely to a greater degree on IL-15.
Microglia are known for their fast response to changes in tissue homeostasis (18,42). Activation of microglia following infection or trauma or during pathological alterations involves changes in their morphology, enhanced motility, expression of cell surface antigens, and increased phagocytotic and secretory activities (14,15). Although IL-15 did not affect microglial motility in the present experiments, it could assist in the recruitment of lymphoid cells into brain parenchyma. IL-15 is not only a growth factor but a chemoattractant for T lymphocytes, the latter feature not being shared by IL-2 (29,43). Microglial IL-15 may thus serve as a signal for T cell extravasation. In return, activated T cells, once in the brain, will release cytokines with established effects on macrophages, such as IL-3, granulocyte-macrophage colony-stimulating factor, or IL-2, shown to drive proliferation of activated microglia (33). Reciprocal supply and shared use of cytokines by invading immune and resident brain cells, including microglia, astrocytes, and endothelial cells, may crucially participate in the initiation and progression of CNS inflammation (13,44).
NO is a neuromodulatory agent and is, for example, involved in IL-2-mediated release of hypothalamic corticotropin-releasing factor (5,45). Released by microglia upon activation, NO is also considered a cytotoxic factor playing crucial roles in both microglia-mediated tissue defense and damage (46). We observed that IL-15 and IL-2 modulate microglial NO production, indicative of their involvement in the regulation of this microglial activity. Increasing concentrations of IL-15 attenuated LPS-induced NO production, while IL-2 had the opposite effect; thus, the sharing of receptor subunits does not imply identical physiological consequences (1,11). Indeed, functional differences between IL-2 and IL-15 are known from immune cells. They may result from the actual availability of ␣ receptor proteins, a yet unknown signaling contribution of the longer IL-15R␣ cytoplasmic domain or the different ligand-receptor interactions, since the IL-15⅐IL-2R␤␥ complex is less stable than the IL-2⅐IL-2R␤␥ trimer (9).
Finally, expression of IL-2R/IL-15R by neural cells may re-late to CNS and neuroendocrine side effects associated with either the enhancement or suppression of IL-2R␤␥ signaling during immunotherapies (20,21). Conclusion-The rodent CNS expresses an IL-15/IL-15R␣ system together with the IL-2R subunits ␤ and ␥ that are shared for signaling by both IL-15 and IL-2. IL-15 and IL-15R␣/ IL-2R␤␥ are widely distributed across brain tissues. Endogenous IL-15 may thus relate to several CNS effects that have formerly been described for IL-2. Microglia are proposed to be an ubiquitous carrier of the IL-15/IL-15R system throughout the brain and to respond to IL-15, also in an autocrine manner. Microglial IL-15 could play a pivotal role in CNS inflammation by assisting the extravasation of T lymphocytes. The demonstration of IL-15-inducible JAK1 phosphorylation in microglia indicates the presence of functional JAK/STAT signaling pathways in these brain cells. Future studies on IL-2/IL-15 functions in the CNS will unravel how their receptor signaling is integrated in neural cell physiology.