INTRODUCTION

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 (1-4). 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 neuroendocrine 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 di- and trimeric complexes associating with various enzymatic and modulatory components (7). IL-2R binds IL-2 with low affinity (k_D ~ 10⁸ M). IL-2R and IL-2R build up the heterodimeric receptor of intermediate affinity (k_D ~ 10⁹ M). Upon trimerization, IL-2R represents the high affinity IL-2 receptor (k_D ~ 10¹¹ 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).

IL-15 recruits both the IL-2R  and  chains for its signaling but does not bind to IL-2R (1, 9). Instead, IL-15 interacts specifically with its own  subunit, IL-15R, a 60-kDa protein that is structurally related to IL-2R (10-12). This new receptor subunit binds IL-15 with high affinity (k_D ~ 10¹¹ M) and assists in the stable formation of an IL-15·IL-2R complex (9).

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 protein-tyrosine 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

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 Na₂HPO₄/NaH₂PO₄ 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.

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.

Patch clamp recordings from cultured microglial cells were made in the whole cell voltage clamp configuration as described previously (22).

Amplification of mRNA transcripts by RT-PCR was performed with total RNA obtained from virtually blood-free (perfused) brain regions of postnatal day 6 (P6), day 20 (P20), and day 63 (P63) mice (30-100 mg, wet weight) or cultured microglia (1-10 × 10⁶ cells) using the Trizol method (Life Technologies, Inc., Eggenstein, Germany) (23). After normalizing for the sample amount, cDNAs were prepared using Superscript RT^TM (Life Technologies, Inc.). cDNAs were amplified in a Thermocycler 9600 (Applied Biosystems, Weiterstadt, Germany) by hot start PCR (5 min, 94 °C) in 35-40 cycles (30 s, 94 °C denaturation; 30 s, 55-60 °C annealing; 30 s plus 1 s/cycle, 72 °C elongation) and 10 min at 72 °C for elongation. Subsequently, the tubes were cooled to 4 °C, and the products were analyzed by agarose gel electrophoresis. Primers for IL-15, IL-15R, IL-2R, IL-2R, IL-2R, and JAK1 were derived from published sequences delivered to the EMBL data base: mouse IL-15 (EMBL accession number U14332), sense: mmil15-492s (IL15-1s), 5-GAA TAC ATC CAT CTC GTG CTA CT-3 and antisense: mmil15-651as (IL15-2as), 5-GCT TTC AAT TTT CTC CAG GTC-3, mmil15-913as (IL15-1as), 5-TTT GCA AAA ACT CTG TGA AGG-3; mouse IL-15R (EMBL accession number U22339), sense: mmil15ra-121s (IL15R-1s), 5-TGC TGC TGC TGC TGT TGC TA-3, mmil15ra-138s (IL15R-2s), 5-CTA CTG TTG CTC CGC TGA G-3 and antisense: mmil15ra-506as (IL15R-1as), 5-TGT CTC TGT GGT CAT TGC GGT AT-3; mouse IL-2R (EMBL accession number K02891), sense: mmil2ra-105s, 5-TTG CTG ATG TTG GGG TTT CTC-3 and antisense: mmil2ra-405as, 5-TGT CTG TTG TGG TTT GTT GCT CT-3; mouse IL-2R (EMBL accession number M28052), sense: mmil2rb-410s, 5-GTG GAC CTC CTT GAC ATA-3 and antisense: mmil2rb-754as, 5-GTT TCG TTG AGC TTT GAC CCT CA-3; mouse IL-2R (EMBL accession number D13565), sense: mmil2rg-750s, 5-CTG GGG GAG TCA TAC TGT AGA GG-3 and antisense: mmil2rg-1120as, 5-AGG CTT CCG GCT TCA GAG AAT-3; mouse JAK1 (EMBL accession number S63728), sense: mmjak1-301s, 5-CCA TGG CGT TCT GTG CTA AAA TG-3 and antisense: mmjak1-714as, 5-GGA GTG GGG TTG CTT CTG GAA-3. Sizes of amplimers were 422 or 160 bp (for IL-15), 386 or 369 bp (for IL-15R), 301 bp (for IL-2R), 345 bp (for IL-2R), 371 bp (for IL-2R), and 414 bp (for JAK1). All primer pairs were selected to cross exon borders distinguishing genomic contaminations. Each PCR analysis was run 2-5 times, starting from independent preparations of RNA material. As negative controls, selected hematopoietic cell lines were used. Fainter bands occurring in some of the experiments were probably due to mispriming artifacts. Some of them were investigated by PCR sequencing revealing an origin from unrelated messengers. For single-cell RT-PCR, microglial cells were electrophysiologically identified by means of patch clamp recording (24). The cytoplasm of individual cells was harvested through the patch pipette containing 130 mM KCl, 5 mM EGTA, 0.5 mM CaCl₂, 3 mM MgCl₂, 3 mM ATP, 10 mM HEPES, pH 7.2, by applying negative pressure. Subsequently, cDNA was synthesized from the obtained cytoplasmic RNA by reverse transcription. The fragments of interest were amplified by seminested PCR approaches. Amplimer identity was confirmed by DNA sequencing carried out in either the sense or antisense direction using the Taq DyeDeoxy Terminator Cycle sequencing kit on an automated ABI DNA sequencer (model 373A, Applied Biosystems, Weiterstadt, Germany). For negative control purposes, solution was expelled from pipettes without patch-clamp recording and treated in the very same way as the experimental samples.

Immunoprecipitations were performed on microglial lysates (5-6 × 10⁶ 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 µg/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 µl 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 NaN₃. 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.

Immunodetections were performed following SDS-PAGE (8-15% gels, using sample buffer containing 2-mercaptoethanol) and transfer to nitrocellulose. After blocking with bovine serum albumin and appropriate normal, nonimmune serum (5% each, in PBS, 0.05% Tween 20, 0.1% bovine serum albumin), the antigen was visualized using the precipitating antibodies against IL-15, IL-15R, and JAK1 or mouse anti-Tyr(P) antibody (Sigma, Deisenhofen, Germany), followed by treatments with horseradish peroxidase-conjugated secondary antibodies or ExtrAvidin^TM (Sigma) and SuperSignal^TM 228 ULTRA ECL detection reagent (Pierce). For reprobing, blots were stripped in 100 mM glycine/HCl buffer, pH 2.5, or 62 mM Tris/HCl-buffer, pH 6.8, containing 2% SDS and 100 mM 2-mercaptoethanol, again bovine serum albumin- and serum-blocked and stained with the second immunodetection cascade.

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 (NO₂), a breakdown product of NO, using sodium nitrite as a standard. 100 µl of each supernatant were mixed with 100 µl 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).

Imaging of intracellular Ca²⁺ levels in microglia was performed as described in detail elsewhere (26).

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 × 320 pixels and 128 gray values using a × 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).

Statistical evaluation was carried out with the STATISTICA computer program (StatSoft, Tulsa, OK) and using the paired t test, p < 0.05 being considered significant. Data are given as mean ± S.E.

RESULTS

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.

[View Larger Version of this Image (28K GIF file)]

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 (70 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⁺ currents elicited at hyperpolarizing potentials and lack of outward K⁺ 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).

Detection of transcripts encoding IL-15 and IL-15R in microglia by single-cell RT-PCR.

[View Larger Version of this Image (32K GIF file)]

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 several 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  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).

Demonstration of IL-15 and IL-15R protein by immunoblotting.

[View Larger Version of this Image (42K GIF file)]

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 ~65 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, Tyr²²⁷ (also Tyr²²⁷ 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 (50 kDa, Fig. 3D).

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 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.

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.

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.

Widespread expression of IL-15, IL-15R, and IL-2R transcripts in the mouse brain.

[View Larger Version of this Image (44K GIF file)]

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.

As shown in Fig. 5, microglia synthesize the Janus type protein-tyrosine kinase, JAK1, with JAK1 phosphorylation occurring in these cells upon stimulation with IL-15.

[View Larger Version of this Image (52K GIF file)]

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 or both.

Cytokines may induce both protein phosphorylation and/or rises in intracellular Ca²⁺ ([Ca²⁺]_i) (28). Changes in [Ca²⁺]_i are unlikely to play a role in IL-2R signaling. The longer intracellular C terminus of IL-15R, however, when compared with IL-2R (12), may allow for novel contributions to IL-15R signaling. We monitored [Ca²⁺]_i levels in individual cultured microglial cells while exposing them for varying periods of time (1-10 min) to different concentrations of IL-15 and IL-2 (1 and 10 ng/ml). In none of the cases were significant changes in [Ca²⁺]_i observed (four experiments, data not shown).

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.

[View Larger Version of this Image (32K GIF file)]

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.

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.

[View Larger Version of this Image (26K GIF file)]

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-2-like 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.

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 blood-brain 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 system-derived 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 post-transcriptional level (1). First, multiple AUG start codons in the long 5-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.

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).

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.

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-15·IL-2R complex is less stable than the IL-2·IL-2R trimer (9).

Finally, expression of IL-2R/IL-15R by neural cells may relate to CNS and neuroendocrine side effects associated with either the enhancement or suppression of IL-2R signaling during immunotherapies (20, 21).

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.