Involvement of Cathepsin E in Exogenous Antigen Processing in Primary Cultured Murine Microglia*

We have attempted to elucidate an involvement of cathepsin E (CE) in major histocompatibility complex class II-mediated antigen presentation by microglia. In primary cultured murine microglia, CE was localized mainly in early endosomes and its expression level was markedly increased upon stimulation with interferon-γ. Pepstatin A, a specific inhibitor of aspartic proteases, significantly inhibited interleukin-2 production from an OVA-(266–281)-specific T helper cell hybridomas upon stimulation with native OVA presented by interferon-γ-treated microglia. However, pepstatin A failed to inhibit the presentation of OVA-(266–281) peptide. The possible involvement of CE in the processing of native OVA into antigenic peptide was further substantiated by that digested fragments of native OVA by CE could be recognized by OVA-specific Th cells. Cathepsin D also degraded native OVA into antigenic peptide, whereas microglia prepared from cathepsin D-deficient mice retained an ability for antigen presentation. On the other hand, the requirement for cysteine proteases such as cathepsins S and B in the processing of invariant chain (Ii) was confirmed by immunoblot analyses in the presence of their specific inhibitors. In conclusion, CE is required for the generation of an antigenic epitope from OVA but not for the processing of Ii in microglia.

We have attempted to elucidate an involvement of cathepsin E (CE) in major histocompatibility complex class II-mediated antigen presentation by microglia. In primary cultured murine microglia, CE was localized mainly in early endosomes and its expression level was markedly increased upon stimulation with interferon-␥. Pepstatin A, a specific inhibitor of aspartic proteases, significantly inhibited interleukin-2 production from an OVA-(266 -281)-specific T helper cell hybridomas upon stimulation with native OVA presented by interferon-␥treated microglia. However, pepstatin A failed to inhibit the presentation of OVA-(266 -281) peptide. The possible involvement of CE in the processing of native OVA into antigenic peptide was further substantiated by that digested fragments of native OVA by CE could be recognized by OVA-specific Th cells. Cathepsin D also degraded native OVA into antigenic peptide, whereas microglia prepared from cathepsin D-deficient mice retained an ability for antigen presentation. On the other hand, the requirement for cysteine proteases such as cathepsins S and B in the processing of invariant chain (Ii) was confirmed by immunoblot analyses in the presence of their specific inhibitors. In conclusion, CE is required for the generation of an antigenic epitope from OVA but not for the processing of Ii in microglia.
Cathepsin E (CE, 1 EC 3.4.23.34) is an intracellular aspartic protease of the pepsin family, which is highly homologous to the lysosomal aspartic protease cathepsin D (CD, EC 3.4.23.5). In contrast to CD, CE has a limited distribution in tissues and cell types such as lymphoid tissues, gastrointestinal tracts, blood cells, and microglia (1)(2)(3)(4). Furthermore, the intracellular localization and molecular form of CE in microglia contrast sharply with those in other tissues and cell types. CE is mainly localized in the endosomal structures possibly as the mature form in microglia (4), whereas this enzyme was found to be localized in various cellular compartments such as the plasma membrane, endoplasmic reticulum, and Golgi apparatus as the enzymatically inactive proform in other tissues and cell types (4 -6). The localization of CE in the endosomal system has been also demonstrated in antigen-presenting B cells lymphoblasts (7). Thus it is conceivable that the mature form of CE is closely associated with endosomal localization of this enzyme because the proform of CE appears to be converted into mature form only after entering in intracellular acidic compartments. Microglia are known to interact with invaded CD4 ϩ T helper (Th) cells in the central nervous system (8 -10). Through this interaction, microglia may contribute to tissue damage and repair during autoimmune disease, viral infections, and chronic inflammatory disease (see reviews, see Refs. [11][12][13]. The major histocompatibility complex (MHC) class II-mediated antigen presentation requires the participation of endosomal/lysosomal proteases in endocytic route to degrade exogenous antigens and invariant chain (Ii) associated with MHC class II (see reviews, Refs. 14 and 15). Despite the strategic localization of CE in microglia, no information is available about the involvement of CE in MHC class II-mediated antigen presentation by microglia.
There is increasing evidence that two classes of intracellular proteases, aspartic and cysteine proteases, are involved in proteolytic steps required for MHC class II-mediated antigen presentation. On the basis of studies utilizing specific inhibitors or cathepsin-deficient mice, cathepsin L (CL) and cathepsin S (CS), lysosomal cysteine proteases have been recently demonstrated to be responsible for the terminal degradation of Ii to class II-associated Ii peptide (CLIP) during maturation of MHC class II molecules in the thymus and the peripheral lymphoid organs, respectively (16 -22). Furthermore, cathepsin F is also implicated in CLIP generation in peripheral macrophages (23). On the other hand, aspartic protease inhibitors such as pepstatin A have been shown to prevent antigen processing and Ii degradation (24,25). In B cells, expression level of CE was up-regulated upon cellular activation and specific inhibitor of CE blocked the presentation of ovalbumin (OVA) in these cells (7,26). Furthermore, CD has been reported to generate antigenic peptides from OVA or hen egg lysozyme that could be presented to T cells (27)(28)(29). More recently, however, the experiments conducted with splenocytes and macrophages prepared from CD-deficient (CDϪ/Ϫ) mice have concluded that CD is dispensable for degradation of Ii and processing of a number of exogenous and endogenous antigens (17,30). These observations may suggest that CE plays a pivotal role in the MHC class II-mediated antigen presentation as an aspartic protease. To elucidate a specific involvement of CE and CD in the MHC class II-mediated antigen presentation by microglia, we conducted experiments utilizing specific inhibitors for cathepsins, purified enzymes and microglia prepared from CDϪ/Ϫ mice.
Animals-C57BL/6 mice were purchased from Seac Yoshitomi (Fukuoka, Japan). Heterozygous (ϩ/Ϫ) mice (36) were transferred to the Institute of Experimental Animal Sciences, Kyushu University Faculty of Dental Sciences, and kept in conventional facilities. Selection of CDϪ/Ϫ mice from littermates obtained by heterozygous coupling was performed according to the method as previously reported (37,38).
Cell Culture-Microglia were isolated from a primary culture of mouse brain as described previously (4,39). Briefly, the whole brain from 3-day-old C57BL/6 mice was minced and treated with papain (90 units/ml) and DNase (2000 units/ml) at 37°C for 30 min. The mechanically dissociated cells were seeded into plastic flasks at a density of 10 7 pre-300 cm 2 in Eagle's minimum essential medium, 0.3% NaHCO 3 , 2 mM glutamin, 0.2% glucose, 5 g/ml insulin, and 10% fetal calf serum (FCS, Mi medium), and maintained at 37°C a 10% CO 2 , 90% air atmosphere. Subsequent medium replacement was carried out every 3 days. After 10 -14 days in culture, floating cells and weakly attached cells on the mixed primary cultured cell layer were isolated by gentle shaking of the flask for 10 min. The resulting cell suspension was transferred to Petri dish (Falcon 1001, Lincolon Park, NJ) and allowed to adhere at 37°C. Unattached cells were removed after 30 min, microglia were isolated as strongly adhering cells. About 90% of these attached cells were positive for F4/80, makers for macrophage/microglial cell types. In some experiments, microglia prepared from CDϪ/Ϫ, wildtype and hetero-type littermate mice were also used.
Peritoneal macrophages were obtained from 10-week-old male C57BL/6 mice injected peritoneally with 4.05% thioglycollate (Nissui) (2 ml/mouse) 3 days before harvesting cells. Thioglycollate-elicited macrophages were isolated by peritoneal lavage with DMEM followed by adherence to Petri dishes in the presence of 10% FCS in DMEM. Adherent macrophages were washed vigorously with phosphate-buffered saline (PBS) and scraped off the dish. Primary cultures were used for experiments immediately after their isolation.
Splenocytes and thymocytes were obtained from 10-week-old male C57BL/6 mice. The spleen and thymus were excised and cells dispersed into tissue culture dishes containing Hank's balanced salt solution. Cell suspensions were filtrated through cell strainer (Falcon 2350). Eryth-rocytes were lysed by incubation with 1.66% NH 4 Cl for 5 min at room temperature. The splenocyte and thymocyte were washed and resuspended in RPMI 1640 medium supplemented with 10% heat-inactivated FCS, 100 units/ml penicillin, and 100 g/ml streptomycin.
Bone marrow cells (BMCs) were obtained from 10-week-old male C57BL/6 mice femur and tibia. Using a 25-gauge needle, ice-cold PBS was forced through the bone cavity. Cells were drawn several times through the syringe to break up clumps. Cell suspensions were filtrated through cell strainer. Erythrocytes were lysed by incubation with 1.66% NH 4 Cl for 5 min at room temperature. The bone marrow cells were washed and resuspended in RPMI 1640 medium supplemented with 10% heat-inactivated FCS, 50 M 2-mercaptoethanol, 100 units/ml penicillin, and 100 g/ml streptomycin.
Generation of OVA-(266 -281)-specific I-A b Restricted Th Cell Hybridomas-OVA-specific Th cell hybridomas were prepared from C57BL/6 mice which were immunized with protein emulsified in complete Freund's adjuvant (100 g, subcutaneously). After 10 days, draining lymph nodes cells were cultured (4 ϫ 10 6 cells/ml) with OVA (50 g) for 7 days. Viable cells were collected by centrifugation over Ficoll-Hypaque (Amersham Biosciences InC., Piscataway, NJ) and cultured (10 6 cells/ml) for another 4 days with irradiated syngenic splenocytes as APC (7 ϫ 10 6 cells/ml) in the presence of OVA (10 g). Th cell blasts were collected by centrifugation over Ficoll-Hypaque and used for fusion. Th cell blasts were fused with BW5147, a NFAT-LacZ ϩ , ab-TCR-AKR thymoma line (a kind gift from Dr. Kenneth Rock, Harvard Medical School, Boston, MA) using polyethylene glycol (PEG 1500, Roche Molecular Biochemicals). Briefly, Th cell blasts and BW5147 fusion partner cells were mixed with the same ration (10 6 -10 7 cells) in a 50-ml conical centrifuge tube, and washed once in pre-warmed serum-free medium. The supernatant was aspirated off and the pellet loosened by gentle tapping. Then, 1 ml of 50% PEG was slowly added over the course of 1 min and 45 s. PEG was then diluted out with 10 ml of warm serum-free medium and the tube placed in a 37°C water bath. After 8 min, cells were centrifuged and resuspended in complete medium to a density of 2-3 ϫ 10 5 /ml, and 0.1 ml of this suspension was added to yield a final concentration of 1.36 mg/ml hypoxanthine (Sigma), 1.76 mg/ml aminopterin (Sigma), and 388 mg/ml thymidine (Sigma). The hypoxanthine/aminopterin/thymidine-resistant Th cell hybrids were generally observed within 2 weeks of culture. Individual clones were expanded and initially screened for expression of CD3. CD3 ϩ hybridomas were then screened for OVA and OVA-(266 -281) peptide with I-A b specificity in the interleukin (IL)-2 production assay described below. The hybridoma clone was established by subcloning a hybridoma clone which showed both OVA and OVA-(266 -281) peptide specificity. Stock OVA-(266 -281)-specific Th cell hybridomas were cultured in Iscove's modified Dulbecco's medium containing 10% heat-inactivated FCS, 100 units/ml penicillin, 100 g/ml streptomycin, and 50 M 2-mercaptoethanol.
Antigen Presentation Assay-Microglia or macrophages were seeded in 96-well flat-bottom tissue culture plates at a density 3 ϫ 10 4 per well. After 18 h, the culture medium was replaced with fresh medium without or with IFN-␥ (100 units/ml), and cells were incubated at 37°C for 24 h. Immediately before the addition of OVA-specific Th cell hybridomas, microglia, or macrophages were gently washed three times with medium to completely remove IFN-␥. Th cells were added at a density 5 ϫ 10 4 in Iscove's modified Dulbecco's medium, 10% FCS to wells containing the microglia in presence of native OVA (1 M) or OVA-(266 -281) peptide (0.3 M). For analysis of Th cell-derived IL-2, supernatants from triplicate cultures were harvested after 24 h. IL-2 was quantified by using mouse IL-2 enzyme-linked immunosorbent assay (ELISA) kit (BioSource International Inc.).
In Vitro Digestion of OVA and Antigen Presentation Assay-OVA was dissolved in 100 mM citrate buffer, pH 5.5, at a final concentration of 10 mg/ml. CD and CE were added at a final concentration of 200 units/ml, and control sample contained 1 M pepstatin A. The sample sterilized by passing through 0.2-m filters, and incubated overnight at 37°C. For CB digestions, OVA was dissolved at 10 mg/ml in 250 mM citrate buffer at pH 5.5 containing 1 mM EDTA and 2 mM dithiotheitol. The 20 g/ml CB was added, and control sample contained 1 M leupeptin. Microglia cells at a density 1 ϫ 10 5 per well were fixed for 5 min at room temperature in 0.5% paraformaldehyde before assay. Fixed microglia cells were incubated with 1 ϫ 10 5 per well of OVA-specific Th cell hybridomas in the presence of the digested OVA (final concentration 1 mg/ml) for 24 h at 37°C. The digests were neutralized and diluted in fresh culture medium.
Electrophoresis and Immunoblotting-Detailed electrophoresis and immunoblotting procedures were described previously (41,42). Briefly, cells in dishes were washed with phosphate-buffered saline and mechanically removed. Cells were pelted by centrifugation and resuspended in phosphate-buffered saline containing 0.1% Triton X-100. The soluble cell extracts were prepared by ultrasonication followed by centrifugation at 105,000 ϫ g for 30 min. The extract were subjected to SDS-polyacrylamide gel electrophoresis under reducing conditions after the heat treatment in the solubilizing buffer at 100°C for 5 min. For immunoblotting the proteins on gels were electrophoretically transferred to nitrocellulose membranes and then immunostained with In-1 (anti-mouse invariant chain antibody) or anti-CE antibody. As a control, the primary antibody was replaced by preimmune rat IgG or rabbit IgG.
Immunocytochemistry-Detailed immunohistochemical procedures have been described elsewhere (4). Briefly, microglia plated on chamber slide (Lab-Tek) were fixed with paraformaldehyde for 30 min at room temperature. After washed in PBS, cells on the slides were treated with PBS containing 3% normal goat serum for 3 h at room temperature and then stained with the following combinations of antibodies for 1 h at 37°C: CE and Lamp-2, CE and TfR, CD and Lamp-2, CD and TfR. After washing, the cells were incubated with a mixture of Alexa Fluor TM 488 anti-rat IgG and Alexa Fluor TM 594 anti-rabbit IgG (Molecular Probes Inc.). After several washes with PBS, samples were mounted in the anti-fading medium Vecta shield (Vector Laboratory) and examined with a confocal laser scanning microscope MRC-1000 (Bio-Rad, Hertz, United Kingdom).
Statistical Analysis-Data are expressed as mean Ϯ S.D. The significance of differences between groups was determined with two-way analysis of variance (ANOVA), followed by Scheffe's post-hoc test for multiple comparison when F ratios reached significance.

Distribution and Expression of CE in Various Antigen
Presenting Cells-The distribution and the molecular form of CE was examined in cell extracts of various antigen presenting cells (APCs), thymocytes, BMCs, and hepatocytes using immunoblot experiments under the reducing condition. Microglial cell extracts produced an intense protein band with an apparent mass of 42-kDa corresponding to the mature form. Similar intense 42-kDa band was also observed in cell extracts from peripheral macrophages (Fig. 1A). Under the same condition, however, extracts from splenocytes as well as the BMCs and thymocytes showed a protein band with an apparent molecular mass of 46-kDa (Fig. 1A). A brief acid treatment (at pH 3.5 and 37°C for 30 min) rapidly converted the 46-kDa polypeptide to the 42-kDa form suggesting that the 46-kDa polypeptide corresponds to proform of CE. On the other hand, there was no detectable band corresponding to CE in cell extracts derived from hepatocytes. In contrast, CD has been shown to be present in all cell types examined as a 42-kDa enzyme corresponding to the mature form (data not shown). Fig. 1B shows mRNA expression of cathepsins in primary cultured murine microglia before and after treatment of IFN-␥. Although messages amplified by RT-PCR makes quantitation difficult, the expression level of CE amplified in microglia was always smaller than those of CD and cysteine proteases including CB, CL, and CS amplified under the same condition. At 24 h after treatment of IFN-␥, the expression level of CE was increased by ϳ2-fold to reach the similar level of other cathepsins. On the other hand, the expression levels of other cathepsins remained constant even after treatment of IFN-␥.

Intracellular Localization of CE and CD in Microglia-
The intracellular localization of CE and CD in murine microglia was examined by indirect immunofluoresence double-staining with endosomal/lysosomal markers. Indirect immunofluoresence staining was performed in microglia using affinity purified polyclonal antibodies specific for CD and CE, affinity purified monoclonal antibodies specific for mouse Lamp-2 (a marker for lysosomes) and mouse TfR (a marker for early endosomes). As shown in Fig. 2, immunoreactivity of CE was mainly observed in the early endosomes identified by the presence of TfR. Double staining with anti-Lamp-2 antibody also showed that CE was partially localized in lysosomes. By contrast, immunoreactivity of CD corresponded well with that of Lamp-2 but not with that of TfR indicating the lysosomal localization of CD. These results indicate that the localization of CE is much more suitable for processing of exogenous antigens than that of CD because exogenous antigens are known to be first processed in the early endosome after endocytosis.
Antigen Presenting Abilities of the Primary Cultured Microglia-Next we examined the antigen presentation abilities of primary cultured murine microglia. When microglia treated with IFN-␥ (100 units/ml, for 24 h) were cultured with OVAspecific Th cell hybridomas, Th cell hybridomas strongly responded to both native OVA and OVA-(266 -281) peptide secreting IL-2 in the culture medium (Fig. 3). The concentration of native OVA and OVA-(266 -281) peptide which induced the maximal IL-2 secretion from Th cell hybridomas was determined to be 1 and 0.3 M, respectively. It was also noted that the level of IL-2 was dependent on the density of microglia and reached the plateau level at 10 5 cells/ml. In contrast, non- treated microglia failed to activate Th cell hybridomas even after treatment with the optimal concentration of native OVA or OVA-(266 -281) peptide. These results indicate that microglia treated with IFN-␥, but not untreated cells, show strong antigen presenting ability for both native OVA and OVA-(266 -281) peptide.
Effects of Protease Inhibitors on Antigen Presentation by Microglia-To assess roles of cathepsins in antigen presentation by microglia, IFN-␥-activated microglia pretreated with various inhibitors for 24 h were cultured with OVA-specific Th cell hybridomas. Upon stimulation with native OVA or OVA-(266 -281) peptide, the amounts of IL-2 released from OVA-specific Th cell hybridomas into the culture medium were normalized by the control value. As shown in Fig. 4, pepstatin A (10 M), a specific inhibitor of aspartic proteases, significantly inhibited the IL-2 production from OVA-specific Th cells upon stimulation with native OVA presented by IFN-␥-activated microglia. However, OVA-specific Th cells produced IL-2 upon stimulation with OVA-(266 -281) peptide presented by IFN-␥-activated microglia in the presence of pepstatin A. These observations were interpreted as that aspartic proteases, CE and CD, were involved in the degradation of OVA into antigenic peptide but not in the degradation of Ii chain in microglia.
CLIK-060, a specific inhibitor of CS, and CA074Me, a specific inhibitor of CB, significantly inhibited the IL-2 production from OVA-specific Th cell hybridomas upon stimulation with both native OVA and OVA-(266 -281) peptide (Fig. 4). Leupeptin and E64-d, broad spectrum cysteine protease inhibitors, also almost completely suppressed the production of IL-2 upon stim-ulation with both native OVA and OVA-(266 -281) peptide (data not shown). However, CLIK-148, a specific inhibitor of CL, had no significant effect on the IL-2 production upon stimulation with either native OVA or OVA-(266 -281) peptide. These results indicate that CS and CB are involved in the degradation of Ii chain or in both proteolytic processes for antigen and Ii chain in microglia.
We also examined effects of protease inhibitors on activation of OVA-specific Th cell hybridomas upon stimulation with phorbol myristate acetate (10 ng/ml) and ionomycin (1 M). Protease inhibitors used in the present study did not affect the activity of Th cell hybridomas quantitated by measuring activity of ␤-galactosidase with the concentration up to 10 M (data not shown). Although Ascaris pepsin inhibitor, a specific inhibitor of CE, with the concentration of 10 M inhibited the antigen presentation of both native OVA and OVA-(266 -281) peptide, we omitted this observation from the present results because Ascaris pepsin inhibitor with the concentration of 10 M significantly reduced ␤-galactosidase activity.
Similar effects of those protease inhibitors were also obtained when IFN-␥-activated peripheral macrophages were used as APCs. Pepstatin A (10 M) significantly inhibited the IL-2 production from OVA-specific Th cell hybridomas upon stimulation with native OVA but not with OVA-(266 -281) peptide (Fig. 5). CLIK-060 significantly inhibited the IL-2 production from OVA-specific Th cell hybridomas upon stimulation with native OVA and OVA-(266 -281) peptide. On the other hand, CLIK-148 had no significant effect on the IL-2 production upon stimulation with either native OVA or OVA-(266 -281) peptide.
Effects of Protease Inhibitors on the Degradation of Ii Chain in Microglia-To further elucidate the role of aspartic and cysteine proteases on proteolytic processing of Ii chain in microglia, effects of protease inhibitors on degradation of Ii chain were examined by 15% SDS-PAGE under the reduced condition and immunoblotting by using the Ii chain cytoplasmic tailspecific monoclonal antibody In-1. IFN-␥-activated microglia were cultured for 24 h in the absence and presence of protease inhibitors. In cell extracts from non-treated microglia, two bands corresponding to p31 and p41 isoforms of intact Ii chain were detected (Fig. 6). The 10-kDa proteolytic fragment corresponding to Ii-p10, an NH 2-terminal fragment of Ii chain, was slightly detectable. On the other hand, there was a marked accumulation of Ii-p10 in leupeptin-treated microglia (Fig. 6). The accumulation of Ii-p10 was also detected in microglia treated with CLIK-060 or CA074Me. On the other hand, CLIK-148 showed no detectable effect on the accumulation of Ii-p10 (data not shown). These results indicate that CB and CS play a pivotal role in the degradation of Ii chain in microglia.
In contrast, accumulation of Ii-p10 was not detected in microglia treated with pepstatin A (Fig. 6). To further assess the requirement of aspartic proteases in the initial stages of Ii chain processing, microglia were treated with leupeptin in the presence of pepstatin A. Inhibition of total aspartic protease activity by pepstatin A did not prevent the accumulation of Ii-p10 in leupeptin-treated microglia indicating that aspartic proteases were not required for even in the initial stages of Ii chain processing.
Requirement of Proteolytic Processes by Aspartic Proteases to Generate Antigen Peptide from Native OVA-To directly assess the requirement of aspartic proteases, CE and CD, for generation of antigenic peptide from native OVA, native OVA was digested with, respectively, purified CE or CD in vitro for 24 h at the relatively mild acidic condition similar to the early endosome (i.e. pH 5.5). The digested fragments were added to the co-culture system consisting of fixed microglia and OVA- specific Th cell hybridomas. In this system, OVA-specific Th cell hybridomas can be activated to produce IL-2 only in the presence of antigenic peptide, because microglia were lightly fixed with 0.5% paraformaldehyde to prevent further processing of native OVA through endocytic uptake (Table I). Upon stimulation with OVA digested by purified CE, OVA-specific Th cell hybridomas could produce a significant amount of IL-2 (Table I). A similar result was also obtained upon stimulation with OVA digested by purified CD. When OVA was digested by purified CE or CD in the presence of pepstatin A, digested fragments failed to produce a significant amount of IL-2 (data not shown). On the other hand, there was no significant pro- duction of IL-2 upon stimulation with OVA digested by purified CB. These results indicate that CE and CD are able to digest native OVA into antigenic peptide.
Generation of Antigenic Epitope in Microglia Prepared from CDϪ/Ϫ Mice-Finally we conducted experiments utilizing microglia prepared from CDϪ/Ϫ mice to better characterize the role of aspartic proteases in the generation of OVA antigenic epitope. Primary cultured microglia prepared from CDϪ/Ϫ, wild-type, and hetero-type littermate mice were stimulated with IFN-␥ for 24 h. After co-culture with OVA-specific Th cell hybridomas in the presence of native OVA for 24 h, IL-2 released in the culture medium was measured. Although the efficiency for presenting OVA antigenic peptide to Th cell hybridomas in microglia prepared from CDϪ/Ϫ mice was slightly reduced as compared with those from their wild-and heterotype littermates, the difference did not reach the statistical significance (Fig. 7). These results strongly suggest that CD is dispensable for presenting OVA-derived peptide to OVA-specific Th cell hybridomas by microglia. DISCUSSION Our observations here show that CE exists in murine microglia/macrophages as the mature enzyme, whereas this enzyme exists in other peripheral APCs such as splenocytes and BMCs as enzymatically inactive proform. The difference in molecular form of CE among APCs is considered to reflect the difference in its intracellular localization, because CE was found to be converted into the mature form in an autocatalytic manner only after entering acidic compartments such as endosomes (4). Furthermore, the expression level of CE mRNA in microglia was found to be markedly increased upon stimulation with IFN-␥ which is well known to induce MHC class II expression as well as co-stimulatory molecule which are required for Th cell activation (43,44). These unique properties of CE in microglia/macrophages prompted us to further investigate the role of CE in the MHC class II-mediated antigen presentation by microglia. The involvement of CE in antigen presentation of OVA by microglia was first assessed by using pepstatin A, a specific inhibitor of aspartic proteases. Pepstatin A inhibited the processing of OVA, but not inhibited the presentation of OVA antigenic peptide. The possible involvement of pepstatin A-sensitive aspartic proteases in the processing of antigenic peptide of native OVA was further substantiated by the present observations that digested fragments of native OVA by CE or CD could be recognized by OVA-specific Th cell hybridomas. In the present study, we could obtain no evidence indicating the involvement of aspartic proteases in the processing of Ii chain. However, there are several discrepant reports showing that aspartic protease inhibitors effectively prevented the processing of Ii chain (24,25). One reason for this discrepancy may be that aspartic proteases are involved in antigen presentation by microglia in a different way from that by peripheral APCs such as B cells. Additional experiments are necessary to elucidate this deduction.
Our observations here further showed that microglia isolated from CDϪ/Ϫ mice retained an ability for antigen presentation. This is consistent with recent studies utilizing peripheral APCs prepared from CDϪ/Ϫ mice indicating that CD is not involved in degradation of Ii chain or generation of antigenic peptides (17,30). The present finding that pepstatin A inhibited the antigen presentation of OVA by microglia suggest that CE is essentially involved in processing of OVA because CD and CE are only known for pepstatin A-sensitive endosomal/lysosomal aspartic proteases in mammalian cells. Recent evidence has also shown that CE is closely linked with the proteolytic processing of tetanus toxin (45). There is increasing evidence that agrees with a role for CE in antigen processing. CE is expressed in APCs such as Langerhans cells and B cells (26,46). Furthermore, it has been reported that CE is responsible for the processing of OVA in A20 cells (7). Although CE in B cells is found in the endoplasmic reticulum, Golgi, and endosome-like compartment (26), the distinct subcellular localization and molecular form is unclear. We have previously reported that CE was found to be localized in the early endosome as well as endoplasmic reticulum and Golgi of primary cultured rat microglia (4). Thus it is likely to consider that the expression of CE in activated microglia is mainly associated with processing of endocytic antigen into the endosomal compartments. The requirement of CE for the processing of exogenous antigens in microglia is to be further elucidated in future studies utilizing CEϪ/Ϫ mice that have been under generation.
Although cysteine proteases are considered to be critical for the terminal step of Ii breakdown, their exact role varies among FIG. 6. Effects of protease inhibitors on the Ii-chain processing in primary cultured murine microglia. Immunoblot analysis of the cell extract of microglia by the Ii chain cytoplasmic tail-specific monoclonal antibody In-1. The cell extracts from microglia treated with protease inhibitors were analyzed by SDS-PAGE under reduced condition following immunoblotting. Leu, leupeptin; Pep, pepstatin A.  7. Ability of antigen presentation of OVA by primary cultured murine microglia defected of CD. OVA-specific Th cell hybridomas were cultured together with IFN-␥-treated microglia prepared from CDϪ/Ϫ mice as well as their wild (ϩ/ϩ)-and hetero (ϩ/Ϫ)type littermates in the presence of native OVA. After 24 h, supernatant from triplicate cultures were harvested and Th cell-derived IL-2 was measured by ELISA. Each column and bar represents the mean Ϯ S.D. of three experiments, respectively. different APCs (14,18,22,23). In B cells and dendritic cells, CS exclusively mediates the degradation of Ii to CLIP. For this purpose, thymic cortical cells and macrophages use CL and cathepsin F, respectively. Microglia have been suggested to use CS or CL for the degradation of Ii to CLIP, because microglia express CS and CL, but not cathepsin F (47,48). To address this suggestion, we conducted experiments utilizing specific inhibitors for cysteine proteases and found that the degradation of Ii to CLIP in microglia was mediated by both CS and CB but not CL. Similar results were also obtained when peripheral macrophages were used as APCs. Although there is increasing evidence that microglia is distinct from peripheral macrophages in various aspects, microglia were found to closely resemble peripheral macrophages as APCs.