Activation of a novel microglial gene encoding a lysosomal membrane protein in response to neuronal apoptosis

Molecular Brain Research 88 (2001) 1–13 www.elsevier.com / locate / bres

Research report

Activation of a novel microglial gene encoding a lysosomal membrane protein in response to neuronal apoptosis Mizuho Origasa a , Shuuitsu Tanaka a , Kazuhiko Suzuki a ,1 , Shigenobu Tone b , Bing Lim c , a, Tatsuro Koike * a

Molecular Neurobiology Laboratory, Division of Biological Science, Graduate School of Science, Hokkaido University, North Ward N10 W8, Sapporo 060 -0810, Japan b Department of Biochemistry, Kawasaki Medical College, Kurashiki 701 -1092, Japan c Hematology /Oncology Division, Harvard Institutes of Medicine, Harvard Medical School, Boston, MA 02115, USA Accepted 19 December 2000

Abstract In an attempt to understand the molecular mechanism of microglial activation in response to neuronal death or degeneration, we have employed cerebellar cell cultures prepared from P7 rats and grown in normal K 1 (5.4 mM) medium. Under this condition, glial cells respond to degeneration and cell death of granule neurons that begins to occur at 4 days in vitro (DIV). Here we describe a novel gene, granule cell death-10 ( gcd-10) that is expressed in microglia and up-regulated in an early period of granule cell death. gcd-10 is homologous to the mouse lysosomal-associated multispanning membrane protein (LAPTm5) with hematopoietic origin. Immunocytochemistry and vital staining with acridine orange revealed that GCD-10 was localized at the perinuclear area of cultured microglia and COS 1 cells infected with a GCD-10-expressing adenoviral vector. In cerebellar cell cultures, however, GCD-10 was markedly up-regulated and widely distributed to the cytoplasm, which paralleled the localization of the ED1 antigen, the lysosomal marker. In vivo, gcd-10 is expressed mainly in the brain and the spleen, and was up-regulated upon nerve injury in retina 7 days after optic nerve transection. These findings suggest that gcd-10 is involved in the dynamics of lysosomal membranes associated with microglial activation both in vitro and in vivo.  2001 Elsevier Science B.V. All rights reserved. Theme: Development and regeneration Topic: Glia and other non-neuronal cells Keywords: Microglia; Cerebellar granule neuron; Apoptosis; Lysosome

1. Introduction During development of the mammalian nervous system, about half of neurons initially generated die around the time of their functional contact with targets [21,23,35]. This programmed cell death (PCD) in neurons accompanies cell shrinkage and chromatin condensation, both of which are hallmarks of apoptosis [23,26]. Neurons deprived of support by trophic factors also undergo apoptosis in vitro. Extensive studies have revealed that this apoptotic *Corresponding author. Tel.: 181-11-706-3446; fax: 181-11-7064448. E-mail address: [email protected] (T. Koike). 1 Present address: Pharmaceutical Research Institute, Kyowa Hakkow Kogyo Co., Ltd., Shizuoka 411-0943, Japan.

pathway involves the transfer of BAX to mitochondria followed by the release of cytochrome c, which activates a cascade of caspases [36]. In spite of this progress, little is known about the manner in which glial cells — and microglia in particular — respond to neuronal apoptosis. PCD in the development of Caenorhabditis elegans is subject to strict genetic regulation [13]: seven genes have been defined as those relating to the engulfment of dead cells. These genes are thought to constitute two separate pathways through which the cells execute the recognition, engulfment, and digestion of the corpses. These pathways may be conserved across evolution. For example, the CED-5 homologue is part of a human CrkII-DOCK180Rac signaling pathway that controls phagocytosis and cell migration [37,46], while CED-7 is a homologue to mammalian ABC transporters [28,54]. Furthermore, overex-

0169-328X / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0169-328X( 01 )00005-5

2

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

pression of the human homologue of ced-6 promotes phagocytosis of the cells undergoing apoptosis [46]. These results suggest that engulfment is also genetically controlled in PCD, occurring as a natural part of mammalian development. However, the details of the cascade in mammals remain largely undefined. Microglia are brain macrophages and play important immune roles in the mammalian brain. Microglia in normal adult brain are in an immunologically resting state exhibiting no phagocytotic activity. When brain develops inflammatory conditions or neurodegenerative diseases, microglia become activated and transformed from the ramified type into the amoeboid type [19]. Thus, it is considered that microglial activation is associated with various neuropathological states [48]. Activated amoeboidal microglia are also frequently observed in various regions of the brain during normal development in which programmed cell death takes place [51]. In an attempt to define the steps involved in glial responses to cell death, we here employed cerebellar cell cultures in which glial cells respond to degeneration and cell death of granule neurons that are grown in vitro in a medium containing physiological K 1 (5.4 mM). Under these conditions, granule neurons undergo the initial differentiation steps, begin to degenerate at |4 days in vitro (DIV), and then die progressively via an apoptotic cascade by 8 DIV [4,17,27,47,49]. We have searched for genes that are up-regulated during cell death by differential hybridization, and found a novel microglial gene, granule cell death-10 ( gcd-10). We show here that gcd-10 is homologous to two mouse cDNAs, LAPTm5 [1] and E3 [45], expressed in hematopoietic cells, and that the GCD10 protein is localized in lysosomes of microglia that respond to neuronal cell death by up-regulation.

2. Materials and methods

(Molecular Probes, Eugene, OR). Under these standard conditions, OX-42-positive microglia and GFAP-positive astrocytes were 5 and 2% of the number of granule neurons, respectively. For a long-time culture (|10 days) of granule neurons, a high concentration of potassium (final 30 mM) was added to the culture medium at 2 DIV as described [49]. To eliminate the contamination of nonneuronal cells, cerebellar cells were incubated in the presence of 10 mM aphidicolin instead of FudR [49]. Microglia were isolated as described previously [49]. Briefly, the cerebral cortices were dissected from neonatal rat pups and were dissociated with 250 U / ml Dispase for 60 min at 378C. The dissociated cells were grown in a mixture of DME / F-12 Ham (Sigma) containing 10% heatinactivated FCS, 50 U / ml penicillin, 50 mg / ml streptomycin until confluency (|8–10 days). Microglia were then collected and replated on 35-mm dishes. They were identified by staining with the microglial marker OX-42 [40] (mouse monoclonal antibody; BMA, Tavistock Square, London, UK) or anti-MRF-1 [49]. More than 95% of all cells were microglia. The experimental procedures conformed to the guidance by the committee of the Research Center of Laboratory Animal, Hokkaido University.

2.2. Extraction and purification of RNA Cells were washed with a Ca 21 , Mg 21 -free phosphatebuffered saline (PBS), pH 7.2, and solubilized with 4 M guanidinium thiocyanate, pH 7.0, containing 25 mM sodium citrate, 0.5% sarkosyl, and 0.1 M 2-mercaptoethanol [8]. The total RNA fraction was extracted from one or two 60-mm dishes (1310 7 cells / dish) for cerebellar cell culture or one dish for microglia culture. The RNA was precipitated by isopropanol. After centrifugation for 15 min, the pellet was washed with 70% ethanol and dried with flowing air.

2.1. Cell culture Cerebellar cell cultures were prepared from the cerebella of P7 rats (Sprague–Dawley) as described previously [47]. The cells were plated on poly-L-lysine (Sigma, St. Louis, MO)-coated 60-mm dishes (1310 7 cells / dish) or 35-mm dishes (0.3310 7 cells / dish) for RNA isolation or Western blots, and cover glasses (diameter 12 mm; Matsunami Glass, Osaka) (1310 4 cell / cm 2 ) for immunocytochemistry. The plated cells were cultured in MEM supplemented with 10% heat-inactivated fetal calf serum (FCS; Trace Biosciences, Sydney, Australia), 50 U / ml penicillin, 50 mg / ml streptomycin, and 20 mM fluorodeoxyuridine (FudR) at 368C in a humidified atmosphere of 5% CO 2 / 95% air (standard culture). Cell viability was assessed by incubating the cultures with calcein acetoxymethyl ester (1.2 mM) and propidium iodide (7.5 mM) for 30 min as described in the manufacturer’s instruction

2.3. Differential hybridization and isolation of a partial cDNA fragment of the gcd-10 clone Differential hybridization was carried out as described previously [49]. In brief, RNA was isolated from cells of the original cerebellar cell cultures both 3 DIV (healthy granule neurons) and 5 DIV (degenerating granule neuron) [47] to construct cDNA libraries of each day. The probe, 32 P-labeled 5 DIV cDNA library generated by reverse transcription with random hexamer primer was hybridized with |50-fold excess of mRNA isolated from 3-DIV cells, and then unhybridized labeled cDNA was purified as a probe with a hydroxyapatite column. These steps for the construction of subtracted cDNA probe were repeated twice. Approximately 5000 plaques of the cDNA library of 5-DIV cells were screened with the subtracted probe. Clones that expressed messages only in the culture con-

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

taining degenerating-granule neurons were subcloned into pUC18 plasmid vectors with the SureClone Ligation Kit. The nucleotide sequence of the partial cDNA clone of gcd-10 (97 bp) was determined using a DNA sequencing system with dye primer cycle sequencing (DSQ-2000L; Shimazu, Kyoto, Japan).

2.4. Cloning of gcd-10 cDNA Based on a partial gcd-10 cDNA fragment, the gcd-10 cDNA clones were isolated by the rapid amplification of cDNA ends (RACE) [16] using the Marathon cDNA Amplification Kit (Clontech, Palo Alto, CA). The RACE cDNA library was constructed from the total RNA of the 5-DIV cerebellar cell culture. The RACE reaction products were subjected to a polyacrylamide gel electrophoresis (7%), and specific products were excised, which was followed by subcloning into pUC18. Nucleotide sequence analysis was performed as described above. Sequence searches of the nucleotides and the predicted amino acids of the GenBank database were performed using fasta software (DDBJ, Mishima, Sizuoka, Japan).

2.5. Northern blot analysis Equal amounts of total RNA (10–20 mg) were loaded per lane onto formaldehyde denaturing gel (1% agarose) and separated by electrophoresis. The RNAs were transferred to a nylon membrane (Hybond-N 1 ; Amersham, Arlington Heights, IL). Hybridization was performed as described previously [47]. The cDNA probes were labeled with [ 32 P]dCTP to a specific activity of 1–1.5310 9 dpm / ng by the random priming method using the Ready-To-Go DNA Labeling Kit (Pharmacia Biotech, Piscataway, NJ). The gcd-10 or G3 PDH probe was prepared from a cloned partial cDNA fragment (786 bp), or a PCR product subcloned into pUC18 (983 bp) [47], respectively. The washed membranes were visualized, and the amount of radioactivity of specific transcripts was measured using a Bio-imaging Analyzer (BAS2000; Fuji Photo Film, Tokyo, Japan).

2.6. Southern blot analysis Genomic DNA was extracted from rat cerebellar cells with the SDS-proteinase K method [44]. The isolated DNA (18 mg) was digested with BamHI or HindIII or XbaI, and then separated on a 0.7% agarose gel. The DNA was transferred to a nylon membrane (Hybond-N 1 ) using 203 SSC (333 mM NaCl, 333 mM sodium citrate). Blotting and hybridization were done as described for Northern blot analysis.

2.7. Adenoviral gene transfer of gcd-10 Recombinant adenovirus was generated as described

3

[31]. In brief, a fragment of gcd-10 S cDNA (from 17 to 816 nucleotides) was integrated into a Swa I-cleavage site of pAdex1CAwt that is in the downstream of CAG promoter [33]. The cassette cosmid bearing an expression unit for gcd-10, pAxCATgcd-10 was co-transfected into 293 cells together with EcoT22I-digested DNA-TPC of Ad5-dlX by the calcium phosphate method using a CellPhect Transfection kit (Pharmacia). Recombinant viruses containing the expression unit for gcd-10 (AxCATgcd10) were isolated and propagated further to assess restriction analysis and expression of gcd-10. For infection to COS 1 cells or purified microglia, the virus was purified with CsCl [24]. After COS 1 cells or microglia were incubated in a medium (1 ml for a 35-mm dish) containing the recombinant adenovirus at 10–50 multiplicities of infection (moi) for 1 h at 378C, a normal quantity of medium was added and the cells were further incubated for 2 days before being subjected to experiments.

2.8. Western blot analysis Anti-human LAPTM5 antibody was generated based on a peptide from the carboxyl terminus, CSKTPEGGPAPPPYSEVSKTP, as described previously [1]. Cultured cerebellar cells and microglia untreated or infected with AxCATgcd10 were homogenized in a RIPA lysis buffer (150 mM NaCl, 1.0% NP-40, 0.5% dexycholic acid, 0.1% SDS, 5 mM EDTA, 50 mM Tris–HCl, pH 8.0, 1 mM sodium orthovanadate, 2 mM NaF, 1 tablet / 50 ml of the protein inhibitor cocktail (Boehringer Mannheim, Mannheim, Germany)). Protein extracts were separated on a 15% SDS–polyacrylamide gel and then transferred to a polyvinylidene difluoride membrane (PVDF) (Millipore, Bedford, MA). Immunostaining was performed with an enhanced chemoluminescence detection kit (ECL; Amersham) according to the manufacturer’s directions, with a 0.37-mg / ml concentration of the primary antiserum.

2.9. Immunocytochemical analysis and staining with acridine orange Cultured cells on glass coverslips were fixed with 4% paraformaldehyde (PFA) / 0.12 M Na 1 -phosphate buffer, pH 7.2. After being treated with 0.2% Triton X-100 for 5 min, the cells were pre-incubated with PBS containing 10% horse serum for 1 h. The cells were then incubated with anti-LAPTM5 (0.37 mg / ml) or ED1 (monoclonal antibody recognizing an antigen in lysosomal membrane, mouse IgG, 1:500 dilution; Serotec, Oxford, UK) [5,10,11] for 1 h at room temperature. Detection of the primary antiserum was performed using biotinylated IgG according to the procedure provided by the manufacturer (Histofine kit; Nichirei, Tokyo, Japan). Staining was made visible by horseradish peroxidase-conjugated streptavidin and aminoethercarbazol (AEC) reaction product (Zymed, South San

4

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

Pharmaceuticals) for 30 min at |258C. The fluorescences were then analyzed with a laser scanning microscope (Fluoview; Olympus, Tokyo, Japan). The localization of lysosomes was visualized by incubating the cells with 5 mg / ml acridine orange at 378C for 10 min. After three washes, they were immediately examined on a fluorescence microscope (Nikon, Tokyo, Japan).

2.10. Axotomy of the optic nerve Fig. 1. Immature granule neurons in cerebellar cell culture. Live and dead cells were stained with 1.2 mM calcein acetoxymethyl ester and 7.5 mM propidium iodide, respectively, as described in the manufacturer’s instruction. Granule neurons at 3 DIV were calcein AM-positive (yellow) with extensive neurites (A). At 5 DIV, however, the neurons were dying and became PI-positive (red) (B). Scale bar, 25 mm.

Francisco, CA). Horseradish peroxidase-conjugated antirabbit or -mouse IgG (1:200 dilution) was also used as a secondary antibody for anti-LAPTM5 or ED1 antibodies, respectively. One of the sister coverslips was stained with Mayer’s hematoxylin solution. Fixed cerebellar cells were also used for double staining with anti-LAPTM5 and OX-42, the microglial marker (mouse IgG, 1:300 dilution; BMA Biomedicals). The cells were incubated with the primary antiserum for 1 h followed by fluorescence-conjugated secondary IgG (FITC-conjugated anti-rabbit IgG or TRITC conjugated anti-mouse IgG, 1:20 dilution; ICN

Under anesthesia, the right optic nerve of a Wistar rat (postnatal 7 weeks) was transected. The rat was sacrificed after 7 days of survival. The optic nerve of the right eye was completely transected intraorbitally in all experimental animals 3 mm away from the optic nerve head sparing the central retinal artery [7], and the left eye was sham operated. Both right and left retinas were isolated and treated with GTC solution for RNA isolation as above.

3. Results

3.1. Isolation and characterization of the gcd-10 cDNA clone Immature granule neurons isolated from 7-day-old rat cerebella were grown under normal conditions with phys-

Fig. 2. Nucleotide and deduced amino acid sequences of gcd-10S cDNA. A TGA stop codon (*) is followed by a 39-untranslated region containing an AATAAA polyadenylation signal (double underlined). The five putative transmembrane domains are underlined.

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

iological K 1 (5.4 mM). These neurons extensively elaborated neurites by 3 DIV (Fig. 1A), began to degenerate at 4 DIV, and then underwent apoptosis at 5 DIV (Fig. 1B). Over 50% of the neurons died at 6 DIV and most of them died by 8 DIV. The apoptotic nature of the cell death of these granule neurons has been well documented [6,43,47,49]. These cerebellar cell cultures contain |5% microglia under standard conditions, which respond to granule cell death by activation [49]. To search for genes that are regulated during this neuronal death, mRNAs from both normal (3 DIV) and dying cells (5 DIV) were isolated and analyzed by differential hybridization. We herein describe the gcd-10 (granule-cell-death-10) gene, which is specifically up-regulated in 4-DIV cells. The partial gcd-10 cDNA fragment was subcloned and sequenced. Information from this partial sequence allowed us to design specific primer sets and perform a RACE reaction to isolate full-length clones of gcd-10. Fig. 2

5

shows that the full-length cDNA is 1309-bp long and contains a 19-bp 59-untranslated region, an entire openreading frame of 786 bp encoding 261 amino acid residues, and a 504-bp 39-untranslated region including a polyadenylation signal. As shown in Fig. 2, five putative transmembrane domains were found, suggesting that the GCD-10 protein is likely to be a membrane protein. Database searches revealed homology with two different genes both derived from mouse hematopoietic cells: the lysosomal-associated multitransmembrane protein mRNA (LAPTM5) [1], and the retinoic acid induced E3 protein mRNA (E3) [45]. As shown in Fig. 3A, these mouse genes are different in size, but their coding regions show a striking sequence homology (99.5%). The function of these genes remains elusive, but LAPTM5 is reported to be a lysosomal membrane protein in hematopoietic cells. Northern blotting using a probe designated in Fig. 3A revealed the presence of two transcripts (1.3 kbp, gcd-10S;

Fig. 3. Comparison of nucleotide and deduced amino acid sequences among gcd-10 homologues. Comparison of cDNA lengths (A) and amino acid sequences (B) of gcd-10 homologues. E3, E3 protein; GCD-10, rat granule cell death-10; h, human; LAPTM5, lysosomal-associated multitransmembrane protein; m, mouse. Dashes denote identical residues. Underlines denote the five putative transmembrane domains.

6

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

2.4 kbp, gcd-10L) (Fig. 4A). Fig. 3A shows that all the homologues have a similar coding region but with total cDNA lengths of different sizes (S or L). It is noted that gcd-10S and L may correspond to mouse LAPTM5 and E3, respectively, in terms of their size. In humans, however, only the longer form of the cDNA has been found [1,45]. Fig. 3B shows the alignment of the amino acid sequences of GCD-10 and its homologues for comparison. The coding region of both mouse LAPTM5 and E3 has a 94.3% similarity to the GCD-10 sequence. The putative five transmembrane domains are well preserved across the species.

3.2. Up-regulation of gcd-10 during apoptosis of cultured granule neurons

Fig. 4. Northern blot analyses showing the up-regulated expression of gcd-10 in response to granule cell death in cerebellar cell culture. Sister cerebellar cells were incubated in normal K 1 (5.4 mM) medium for 3–7 days (A) or in a medium containing high K 1 (30 mM) for more than 5 days followed by a switch from the high K 1 medium to a serum-free, normal K 1 medium (B). At each day or hour indicated at the top of the autoradiograms, the cells were subjected to RNA isolation according to the guanidinium thiocyanate method. Total RNA (20 mg) was separated on a 1% agarose gel and transferred to a nylon membrane. 32 P-labeled gcd-10 or G3 PDH cDNA was used as a hybridization probe. Both gcd-10 and G3 PDH mRNAs were detected on the same membrane. The positions of the rRNA are indicated on the left. The amount of separated rRNA in each lane was almost equal when the transferred membrane was stained with methylene blue (data not shown). The data presented are representative of three separate experiments.

To study the expression pattern of the gcd-10 gene during neuronal death in the cerebellar cell culture, we examined the relative transcript levels as a function of incubation time by Northern blot analysis. In Fig. 4, blotting revealed a gcd-10 mRNA of 1.3 kbp ( gcd-10S) in length, which was consistent with the expected length of the full-length gcd-10S cDNA. Fig. 4 also shows the presence of a band of 2.4 kbp ( gcd-10L). Genomic Southern blot analysis revealed the presence of a main band in the DNA preparations digested by each restriction enzyme indicated (Fig. 5). Since the probe used for the Southern blotting has no digestion sites for the three restriction enzymes, it is likely that the observed sub-bands were produced by digestion of introns. This suggests that gcd-10S and L may be encoded at a single locus. Fig. 4A also shows that in the cerebellar culture, both transcript levels were up-regulated at 4 DIV (1.6-fold increase in the intensity compared to that of 3 DIV), when granule neurons began to degenerate, and had a peak at 6 DIV (3.0-fold increase in the intensity compared to that of 3 DIV), followed by a decrease at 7 DIV, at which time most granule neurons had died and were left on the surface of the dishes. It has been well documented that granule neurons maintained under depolarizing conditions with elevated high K 1 (30 mM) undergo apoptosis when the high K 1 medium and serum are replaced with a normal K 1 medium [12,30,55]. Up-regulation of gcd-10 mRNA was also detected in this ‘low’ K 1 -induced apoptosis of granule neurons (Fig. 4B). The transcripts of gcd-10 were remarkably up-regulated for 0–8 h, and peaked at around 48 h after the medium change. These results indicate that the expression of the gcd-10 gene is indeed up-regulated in response to the apoptosis of granule neurons.

3.3. Immunocytochemical and biochemical analysis of the cell type that synthesizes the GCD-10 protein Cerebellar cultures under our conditions (the standard culture described in Section 2.1) contain |90% granule neurons and |10% non-neuronal cells (Fig. 6A). To

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

7

(AxCATgcd10), the band with 29 kDa became detectable and increased in intensity according to the degree of the moi of the vector used for the infection (Fig. 7Ab–d). Fig. 7B shows that the expression level of GCD-10 was clearly elevated as a function of incubation time: the intensity of the band at 6 DIV increased 2.3-fold compared to that at 3 DIV, consistent with the finding by Northern blot analysis. These results strongly indicate that microglia indeed express the GCD-10 protein.

3.4. Up-regulation of gcd-10 mRNA in microglia occurs independently of their proliferation

Fig. 5. Southern blot analysis of gcd-10. Each lane contains 18 mg of genomic DNA isolated from rat cerebral cells followed by digestion with BamH1 (B), HindIII (H) or XbaI (X). The digested DNAs were separated on 0.7% agarose gel, and transferred to a nylon membrane. The membrane was then hybridized with 32 P-labeled fragments of gcd-10 (shown as Fig. 3A). Arrowheads indicate a main band for each enzyme digest. The data presented are representative of three separate experiments.

identify the cell type that expresses the GCD-10 protein, immunocytochemistry and Western blotting were carried out. Immunocytochemistry using anti-LAPTM5 antibody revealed that GCD-10 was localized at the cytoplasmic areas of microglia, but not in granule neurons in this culture (Fig. 6B). No staining in astrocytes or fibroblasts was observed (data not shown). Double staining using anti-LAPTM5 and OX-42 [40], the latter as a marker for microglia, shows that FITC-labeled GCD-10 positive cells (Fig. 6C) were well overlapped with a population of TRITC-labeled OX-42 positive cells (Fig. 6D). Western blot analysis using anti-LAPTM5 antibodies also supported the idea that microglia expressed the GCD-10 protein (|29 kDa, Fig. 7A), which was consistent with the band in the hematopoietic cells [1]. It should be noted that purified microglia expressed this protein in an extremely limited amount: the band was made visible only when 65-mg proteins were loaded (Fig. 7Aa), whereas loading of 30-mg proteins gave no visible band under the present condition (data not shown). Moreover, when purified microglia were infected with an adenovirus carrying gcd-10

To exclude the possibility that this gcd-10 up-regulation was simply a result of the proliferation of microglia, we examined the gcd-10 mRNA level in the presence of a DNA polymerase inhibitor, aphidicolin [49]. Northern blot analysis revealed that the gcd-10 mRNA was undetectable at 4 DIV of the cerebellar culture, when granule neurons normally begin to degenerate [43], and remained undetectable at 5–7 DIV in the presence of this most potent mitotic inhibitor [30]. However, when purified microglia (1310 5 cells) were added to the 3-DIV cerebellar culture (1310 7 cells) and incubated for 18 h in the presence of aphidicolin, the level of gcd-10 mRNA was markedly up-regulated (2.2-fold increase in the intensity compared to that of microglia alone, Fig. 8). These results indicate that upregulation of gcd-10 in response to neuronal cell death occurred independently of the proliferation of microglia.

3.5. Subcellular localization of the GCD-10 protein in microglia and cerebellar cell culture Immunocytochemistry using anti-LAPTM5 antibody was performed to identify subcellular localization of the GCD-10 protein. Fig. 6F shows that in purified microglia, the GCD-10 positive area was localized at the perinuclear region, consistent with the localization of LAPTM5 in hematopoietic cells. In purified microglia infected with AxCATgcd10, the expression of GCD-10 was extremely enhanced, but the localization was also well confined to the perinuclear area (Fig. 6G). This observation was further supported by the ectopic expression of GCD-10 in AxCATgcd10-infected COS 1 cells, which showed that GCD-10 was also localized at these areas (Fig. 6I), whereas untreated COS 1 cells showed no staining (Fig. 6H). In contrast, the GCD-10 positive area was not limited to this region, but rather was widely distributed in the cytoplasm of microglia in cerebellar cell cultures (Fig. 6C). In spite of this apparent difference, a primary localization of GCD-10 to lysosomes was demonstrated by the coincidence of these areas with regions of acridine orange uptake. In purified microglia, uptake of acridine orange, an acidophilic dye, was highly condensed at the perinuclear areas, giving a red vesicular distribution indicating lower pH in lysosomes (Fig. 6J). This observation

8

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

was consistent with the localization of LAPTM5 positive areas shown in Fig. 6G and I. In contrast, in cerebellar cell cultures, the dye distribution was shifted from the perinuclear to cytoplasmic vesicles, and it appeared as a widely distributed yellow-orange color indicative of higher pH

under these conditions (Fig. 6K). In addition, immunocytochemistry using ED1, a monoclonal antibody which recognizes a single chain glycoprotein that is predominantly expressed on the lysosomal membranes of phagocytes [5,10,11], revealed a similar pattern of positive areas:

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

9

Fig. 7. Western blot analysis of the GCD-10 / LAPTM5 protein expressed in purified microglia and cerebellar cell cultures. Protein fractions extracted from purified microglia or cultured cerebellar cells at each indicated day were separated on a 15% acrylamide gel, and then transferred to PVDF membranes. The membranes were exposed to anti-LAPTM5 antibodies. Chemiluminescent detection of the primary antibody shows the presence of a band with 29 kDa. (A) Purified microglia, 20 mg protein / each lane except for a: (a) untreated (65 mg); (b) infected with wild type adenovirus (30 moi); (c) infected with AxCATgcd10 with 10 moi or (d) with 30 moi. (B) Increased band intensity (29 kDa) as a function of incubation time in cerebellar cell culture (120 mg protein / each lane). These data presented are representative of three separate experiments.

perinuclear in purified microglia (Fig. 6L) or widely distributed into the cytoplasm of microglia in cerebellar cell cultures (Fig. 6M). These results suggest altered dynamics of the lysosomal membranes under these conditions.

3.6. The expression of gcd-10 in vivo and its response to nerve injury We next examined gcd-10 expression in vivo. Fig. 9 shows that both gcd-10S and L mRNA were strongly expressed in the spleen, which result was consistent with the origin of the two homologues. The two gcd-10 transcripts were also weakly detected in the brain (Fig. 9A, exposure for |20 h). When the membrane was exposed to an imaging plate for a longer period (|48 h), both forms of gcd-10 were faintly detected in the liver and testis (Fig. 9B). We also prepared a cerebellar cell culture from mice

and performed a similar Northern blot analysis, with the result that only the transcript of 2.4 kb corresponding to gcd-10L was detected (data not shown). To elucidate whether or not gcd-10 transcripts are up-regulated in response to neuronal apoptosis or degeneration in vivo, we also examined mRNA levels in the retina after optic nerve dissection. Fig. 10 shows that gcd-10 also increased in the retina following optic nerve transection of a postnatal 7 weeks rat. The intensity of the ipsilateral retina increased 5.3-fold compared to that of the control retina (Fig. 10). It is well known that neurons in the retina die by apoptosis in accompaniment with caspase activation, and that microglia are activated by cutting of optic nerve [7,50]. This is consistent with the finding that the expression levels of gcd-10 markedly increased on the retina at 7 days after cutting of the optic nerve. This indicates that gcd-10 was up-regulated in response to neuronal apoptosis or injury occurring after axotomy of the optic nerve in the retina.

Fig. 6. Immunocytochemistry and vital staining with acridine orange of cerebellar cells, purified microglia and COS 1 cells. Cultured cerebellar cells (4 DIV), purified microglia and COS 1 cells were fixed with 4% PFA and then treated with anti-LAPTM5 antibodies, the ED1 monoclonal antibody or the OX-42 monoclonal antibody. The OX-42 was used as a microglial marker. The staining with anti-LAPTM5 antibodies or the ED1 antibody was made visible by the AEC reaction product. (B, F–I, L, M). (A) Cerebellar cells (4 DIV) stained with Mayer’s hematoxylin solution: arrowheads, granule neurons; asterisks, microglia. (B) GCD-10 / LAPTM5-positive cells in cultured cerebellar cells (red): scale bar, 25 mm. (C–E) Identification of GCD-10 / LAPTM5positive cells in cultured cerebellar cells by a confocal laser microscope. (C) FITC-labeled GCD-10 / LAPTM5 positive cells (green). (D) TRITC-labeled OX42 positive cells (red). (E) Double stained cells (yellow). (F–I) Localization of LAPTM5 / GCD-10 in cells left untreated or infected with AxCATgcd10 cells. (F) Untreated microglia. (G) Microglia infected with AxCATgcd10. (H) Untreated COS 1. (I) COS 1 infected with AxCATgcd10. (J–M) Cells were stained with acridine orange for 10 min at 378C and examined on a fluorescence microscope (J, K) or stained with ED1 (L, M), a monoclonal antibody that recognizes an antigen in lysosomal membranes of phagocytes. J and L, purified microglia; K and M, microglia in cerebellar cell culture.

10

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

Fig. 8. Autoradiogram showing the up-regulation of gcd-10 mRNA under the inhibitory condition of cell proliferation. Granule cells (35-mm dishes) were grown in presence of 10 mM aphidicolin, and cortical cells were prepared separately. Purified microglia obtained from cortical culture (1310 5 cells) were replated or added into the 3-DIV granule culture and incubated for an additional 18 h in the presence of aphidicolin. All RNA was extracted from 4-DIV granule cells (G), purified microglia alone incubated for 18 h (M), or the mixed cells (G1M), and separated in an agarose gel, and hybridized with gcd-10 or G3 PDH probes. The data presented are representative of three separate experiments.

4. Discussion By utilizing cerebellar cell cultures under normal condition with physiological K 1 , we have successfully isolated a novel microglial gene, gcd-10, that responds to granule cell death. Moreover, we show that gcd-10 is expressed in the brain in vivo and markedly up-regulated in response to neuronal injury in the retina. By homology search of this sequence, we found several human and mouse homologues that were classifiable into two groups according to their lengths: 1.3 kbp (rat gcd10S, mouse LAPTm5) and 2.4 kbp (rat gcd-10L, mouse E3, human LAPTm5). Each of these gcd-10 transcripts is expressed in a tissue-dependent manner. For example, both gcd-10S and L are expressed in the rat spleen and brain, whereas gcd-10S is predominantly expressed in the rat liver and testis. In contrast to the expression of both forms of gcd-10 in rat cerebellar cell culture, only gcd-10L is expressed in mouse cerebellar culture. This is also the case with humans. All of these homologues, however, appear to encode a single protein with highly conserved amino acid sequences. To date, no homology has been found between GCD-10 / LAPTM5 and lysosomal proteins, which are all highly glycosylated. It is noted that all five transmembrane

Fig. 9. Northern blot analysis showing expression levels of gcd-10 in various tissues of the rat. Total RNAs were isolated from 5-DIV cerebellar cell culture (C, 25 mg), brain (B), liver (L), spleen (S), testis (T) (B, L, S, T: 35 mg), and analyzed by Northern blotting as described in Section 2.5. The same membrane was used for the detection of gcd-10 and G3 PDH mRNAs. (A) Exposure for |20 h: gcd-10 is detected in cerebellar cell cultures, the brain and spleen. (B) Exposure for a longer period (|48 h): gcd-10 is weakly expressed in the liver and the testis. The data presented are representative of three separate experiments.

domains and a proline-rich region in the carboxyl tail are particularly well conserved among the proteins of these three species. This suggests that these highly conserved domains may have a functional significance. The prolinerich region (PXXXPXXXPXPPP) would be of special interest. Several proteins with proline-rich domains have been shown to be involved in interaction with SH3 domain-containing proteins [9,38,53]. A proline-rich motif has also been identified in ion channels [41]. Many of the lysosomal membrane proteins may be transporters or channel proteins, while others, like the tetraspanning melanoma-associated antigen CD63 / ME491, are mobilized to the cell surface upon activation and are thought to play a role in cell adhesion and tumor metastasis [2]. Lysosomes are also involved in the modification of proteins prior to antigen presentation [18,44]. A previous attempt has been made to isolate interacting proteins by screening a yeast cDNA library using the proline-rich sequence of the LAPTM5 protein as bait in a two-hybrid yeast system [14]. Based on the results of this study, the

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

Fig. 10. Northern blot showing elevated gcd-10 mRNA in retina after optic nerve dissection. Total RNA (20 mg) was separated on a 1% agarose gel and transferred to a nylon membrane. 32 P-labeled gcd-10 or G3 PDH cDNA was used as a hybridization probe. Both gcd-10 and G3 PDH mRNAs were detected on the same membrane. (1) Control, (2) 7 days after optic nerve dissection. The positions of the rRNA are indicated on the left. The data presented are representative of two separate experiments.

authors argued that protein ubiquitination might serve as a signal for protein uptake in the lysosomal system. These findings raise the possibility that GCD-10 / LAPTM5 may mediate the degradation of ubiquitinated proteins in lysosomes. On the other hand, E3, another homologue of gcd-10, has been isolated as a retinoic acid (RA) inducible transcript that is regulated directly by the RA receptor alpha during myelopoiesis. RA is known to play a pivotal role in cellular growth, differentiation, homeostasis, and tumor suppression. To date, however, little is known about its role in microglial development and functions. Retinoic acid is reported to alter microglial morphology [19] and stimulate secretory functions of these cells in vitro [32]. Recently, it has been shown that RA is capable of altering the intracellular trafficking of the lysosomal enzymes to be transported to their destinations in a macrophage cell line [25]. However, this effect is not mediated through retinoid receptors. Since E3 contains RARE (retinoic acid receptor element) in its promoter region, one of the DNA binding sites, the gcd-10 transcripts may be regulated by RAR directly, if this is also the case with gcd-10. Preliminary evidence suggests that gcd-10 expression is indeed altered by treatment of cerebellar cell cultures with RA (data not shown), although the physiological relevance of RA remains to be investigated. The results of vital staining with the acidophilic dye,

11

acridine orange, also suggest that GCD-10 is likely to be localized at lysosomes. Lysosomes are well known to be a major proteolytic compartment containing various enzymes. It has been reported that lysosomal enzymes of macrophages, e.g. DNaseII-like acid DNase and cathepsins, are responsible for the degradation of nucleosomes in apoptotic cells [34]. In addition, DNA fragmentation in apoptotic cells caused by CAD (caspase-activated DNase) is sensitive to inhibitors that block the acidification of lysosomes [29]. It has also been shown that in Alzheimer senile plaques, activated microglia with enhanced lysosomal or phagocytotic activity are more strongly stained with lysosomal marker 25F9 than those not showing such activity [52]. We also found that purified microglia were stained with anti-LAPTM5 antibodies much more strongly than untreated cells when neuronal debris was added to their culture medium (data not shown). These observations are further strengthened by the findings in regard to the ED1 antigen, a single chain glycoprotein of 90–100 kDa that is expressed predominantly on the lysosomal membrane of phagocytes [5,10,11]. Previous studies have shown that, because of its increased expression during phagocytic activity, the ED1 antigen can be considered a marker of activated microglia [5,15,39]. Under our condition, ED1 immunoreactivity was extremely low in purified microglia, but was significantly enhanced in microglia in cerebellar cell cultures, suggesting that lysosomal activities are enhanced in these cells in response to neuronal cell death. A primary localization of GCD-10 to lysosomes was further demonstrated by the coincidence of the areas stained with anti-LAPTM5 antibodies with those stained by acridine orange uptake and those stained with ED1. A major shift of the area stained with anti-LAPTM5 and ED1 antibodies, as well as a major shift area stained by acridine orange uptake, took place from the perinuclear region in purified microglia to the cytoplasmic vesicles widely distributed in the cytoplasm of microglia in cerebellar cell cultures. This striking difference could not be attributed simply to intracellular trafficking, since GCD-10 is not glycosylated and thus is not targeted by mannose-6-phosphate. It is thus likely that the lysosomal membranes are under highly mobile conditions in cerebellar cell cultures. In this regard, the color of acridine orange, an acidophilic dye, taken up into the cells is of special interest. Purified microglia incubated with acridine orange demonstrated red punctate fluorescence at the perinuclear area. Acridine orange is a lipophilic weak base, which is protonated and stacked into multilayers and emits a red fluorescence at lower pH. In contrast, microglia in the cerebellar cell culture demonstrated yellow-orange fluorescences indicating an increase in lysosomal pH. While lower pH of lysosomes is necessary for lysosomal enzymes to be stored in the organelles by de-glycosylation, it is likely that the elevation of the pH of acidic organelles is caused by newly synthesized lysosomal enzymes being rerouted from the

12

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13

lysosomal to a secretary pathway [3,20,22,42]. Thus, the increase of the lysosomal pH of microglia in the cerebellar cell culture may be indicative of lysosomal enzymes in the secretary pathway. It remains undetermined, however, whether or not this altered distribution of GCD-10 in cerebellar cell cultures occurred in response to cell death. Some degree of altered distribution was observed even in the early stages of cerebellar cell culture, i.e. at 2 DIV. Thus, it is possible that some soluble factor(s) or cellular interactions alter this distribution in microglia. However, our preparations of cerebellar cell cultures normally contain cell debris to which microglia may respond, such that the lysosomal membranes were likely to be propagated under these conditions. In any case, it is rather safe to say that the propagation of GCD-10 and the up-regulation of gcd-10 are regulated by separate mechanisms. In our cerebellar cell culture, both gcd-10 mRNAs and GCD-10 protein were up-regulated at 4 DIV, when neurons begin to degenerate under normal conditions with physiological K 1 . Major cell death begins to occur at 5 DIV under these conditions [43]. Moreover, during apoptosis of mature granule neurons, gcd-10 mRNAs were also increased within 8 h following a switch from a high K 1 (30 mM) to a combined treatment with low K 1 (5 mM) and serum deprivation. In either case, the increase of GCD-10 occurred prior to neuronal cell death. These results indicate that the up-regulation of gcd-10 in response to apoptosis may be induced at least by an early signal, but not as a result of phagocytosis of dead neurons. Up-regulation of gcd-10 may be a prerequisite for phagocytosis of dead neurons. We show here that gcd-10 is a lysosomal protein of microglia and is up-regulated in response to neuronal death or degeneration both in vitro and in vivo. The precise role of microglial gcd-10 during cell death will require further investigation.

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9] [10]

[11]

[12]

[13] [14] [15]

Acknowledgements [16]

We thank Dr Akira Matsuda, Department of Ophthalmology, Faculty of Medicine, Hokkaido University for his assistance in the optic nerve experiments. This work is supported by a Grant-in-Aid of the Promotion of Research and Education in Hokkaido University, Grant-in-Aid for Scientific Research No. 09480217 from MESCS, Japan, Grant-in-Aid No. 64-14 from the Promotion of Fundamental Studies in Health Sciences in Japan (to T.K.) and by grants RO1-DK54417 and RO1-DK7636-6 from the National Institutes of Health, USA (to B.L.).

References

[17]

[18]

[19]

[20]

[21] [22]

[1] C.N. Adra, S. Zhu, J. Ko, J. Guillemot, A.M. Cuervo, H. Kobayashi, T. Horiuchi, J. Lelias, J.D. Rowley, B. Lim, LAPTM5: a novel

lysosomal-associated multispanning membrane protein preferentially expressed in hematopoietic cells, Genomics 35 (1996) 328–337. D.O. Azorsa, J.A. Hyman, J.F.K. Hildreth, CD63 / Pltgp40: a platelet activation antigen identical to the stage-specific, melanoma-associated antigen ME491, Blood 78 (1991) 280–284. A.M. Badger, J.A. Handler, C.A. Genell, D. Herzyk, E. Gore, R. Polsky, L. Webb, P.J. Bugelski, Atiprimod (SK&F 106615), a novel macrophage targeting agent, enhances alveolar macrophage candidacidal activity and is not immunosuppressive in candida-infected mice, Int. J. Immunopharmacol. 21 (1999) 161–176. R. Balazs, O.S. Jorgensen, N. Hack, N-methyl-D-aspartate promotes the survival of cerebellar granule cells in culture, Neuroscience 27 (1988) 437–451. J. Bauer, T. Sminia, F.G. Wouterlood, C.D. Dijkstra, Phagocytic activity of macrophages and microglial cells during the course of acute and chronic relapsing experimental autoimmune encephalomyelitis, J. Neurosci. Res. 38 (1994) 365–375. S.V. Bhave, L. Ghoda, P.L. Hoffman, Brain-derived neurotrophic factor mediates the anti-apoptotic effect of NMDA in cerebellar granule neurons: signal transduction cascades and site of ethanol action, J. Neurosci. 19 (1999) 3277–3286. P. Chaudhary, F. Ahmed, P. Quebada, S.C. Sharma, Caspase inhibitors block the retinal ganglion cell death following optic nerve transection, Mol. Brain Res. 67 (1999) 36–45. P. Chomczynski, N. Sacchi, Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction, Anal. Biochem. 162 (1987) 156–159. G.B. Cohen, R. Ren, D. Baltimore, Modular binding domains in signal transduction proteins, Cell 80 (1995) 237–248. J.G. Damoiseaux, E.A. Dopp, W. Calame, D. Chao, G.G. MacPherson, C.D. Dijkstra, Rat macrophage lysosomal membrane antigen recognized by monoclonal antibody ED1, Immunology 83 (1994) 140–147. C.D. Dijkstra, E.A. Dopp, P. Joling, G. Kraal, The heterogeneity of mononuclear phagocytes in lymphoid organs: distinct macrophage subpopulations in the rat recognized by monoclonal antibodies ED1, ED2 and ED3, Immunology 54 (1985) 589–599. S.R. D’Mello, C. Galli, T. Ciotti, P. Calissano, Induction of apoptosis in cerebellar granule neurons by low potassium: inhibition of death by insulin-like growth factor I and cAMP, Proc. Natl. Acad. Sci. USA 90 (1993) 10989–10993. R.E. Ellis, J. Yuan, H.R. Horvitz, Mechanisms and functions of cell death, Annu. Rev. Cell Biol. 7 (1991) 663–698. S. Fields, O. Song, A novel genetic system to detect protein–protein interactions, Nature 340 (1989) 245–246. M. Frank, H. Wolburg, Cellular reactions at the lesion site after crushing of the rat optic nerve, Glia 16 (1996) 227–240. M.A. Frohman, Rapid amplification of complementary DNA ends for generation of full-length complementary cDNAs: thermal RACE, Methods Enzymol. 218 (1993) 340–358. V. Gallo, A. Kingsbury, R. Balazs, O.S. Jorgensen, The role of depolarization in the survival and differentiation of cerebellar granule cells in culture, J. Neurosci. 7 (1987) 2203–2213. R.N. German, MHC-dependent antigen processing and peptide presentation: providing ligands for T lymphocyte activation, Cell 76 (1994) 287–299. D. Giulian, T.J. Baker, Characterization of ameboid microglia isolated from developing mammalian brain, J. Neurosci. 6 (1986) 2163–2178. A. Gonzalez-Noriega, J.H. Grubb, V. Talkad, W.S. Sly, Chloroquine inhibits lysosomal enzyme pinocytosis and enhances lysosomal enzyme secretion by impairing receptor recycling, J. Cell Biol. 85 (1980) 839–852. C.E. Henderson, Programmed cell death in the developing nervous system, Neuron 17 (1996) 579–585. S.S. Huang, H.A. Koh, J.S. Huang, Suramin enters and accumulates in low pH intracellular compartments of v-sis-transformed NIH 3T3 cells, FEBS Lett. 416 (1997) 297–301.

M. Origasa et al. / Molecular Brain Research 88 (2001) 1 – 13 [23] M.D. Jacobson, M. Weil, M. Raff, Programmed cell death in animal development, Cell 88 (1997) 347–354. [24] Y. Kanegae, M. Makimura, I. Saito, A simple and efficient method for purification of infectious recombinant adenovirus, Jpn. J. Med. Sci. Biol. 47 (1994) 157–166. [25] J.X. Kang, J. Bell, A. Leaf, R.L. Beard, R.A.S. Chandraratna, Retinoic acid alters the intracellular trafficking of the mannose-6phosphate / insulin-like growth factor II receptor and lysosomal enzymes, Proc. Natl. Acad. Sci. USA 95 (1998) 13687–13691. [26] J.F.R. Kerr, B.V. Harmon, Definition and incidence of apoptosis: an historical perspective, in: L.D. Tomei, F.O. Cope (Eds.), Apoptosis the Molecular Basis of Cell Death, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1991, pp. 5–29. [27] T. Koike, Neuronal survival of cerebellar granule cells in vitro is regulated by levels of intracellular calcium, Soc. Neurosci. Abstr. 17 (1991) 1499. [28] M.-F. Luciani, G. Chimini, The ATP binding cassette transporter, ABC1, is required for the engulfment of corpses generated by apoptotic cell death, EMBO J. 15 (1996) 226–235. [29] D. McIlroy, M. Tanaka, H. Sakahira, H. Fukuyama, M. Suzuki, K. Yamamura, Y. Ohsawa, Y. Uchiyama, S. Nagata, An auxiliary mode of apoptotic DNA fragmentation provided by phagocytes, Genes Dev. 14 (2000) 549–558. [30] T.M. Miller, E.M. Johnson Jr., Metabolic and genetic analyses of apoptosis in potassium / serum-deprived rat cerebellar granule cells, J. Neurosci. 16 (1996) 7487–7495. [31] S. Miyake, M. Makimura, Y. Kanegae, S. Harada, Y. Sato, K. Takamori, C. Takuda, I. Saito, Efficient generation of recombinant adenoviruses using adenovirus DNA-terminal protein complex and a cosmid bearing the full length virus genome, Proc. Natl. Acad. Sci. USA 93 (1996) 1320–1324. [32] K. Nakajima, N. Takemoto, S. Kohsaka, Retinoic acid enhances the secretion of plasminogen from cultured rat microglia, FEBS Lett. 314 (1992) 167–170. [33] H. Niwa, K. Yamamura, J. Miyazaki, Efficient selection for highexpression transfectants with an oval eukaryotic vector, Gene 108 (1991) 193–200. [34] C. Odaka, T. Mizuochi, Role of macrophage lysosomal enzymes in the degradation of nucleosomes of apoptotic cells, J. Immunol. 163 (1999) 5346–5352. [35] R.W. Oppenheim, Cell death during development of the nervous system, Annu. Rev. Neurosci. 14 (1991) 453–501. [36] G.V. Putcha, M. Deshmukh, E.M. Johnson Jr., BAX translocation is a critical event in neuronal apoptosis: regulation by neuroprotectants, BCL-2, and caspases, J. Neurosci. 19 (1999) 7476–7485. [37] P.W. Reddien, H.R. Horvitz, CED-2 / CrkII and CED-10 / Rac control phagocytosis and cell migration in Caenorhabditis elegans, Nat. Cell Biol. 2 (2000) 131–136. [38] R. Ren, B.J. Mayer, P. Cicchetti, D. Baltimore, Identification of a ten-amino acid proline-rich SH3 binding site, Science 259 (1993) 1157–1161. [39] W.A. Rinner, J. Bauer, M. Schmidts, H. Lassmann, W.F. Hickey, Resident microglia and hematogenous macrophages as phagocytes in adoptively transferred experimental autoimmune encephalomyelitis: an investigation using rat radiation bone marrow chimeras, Glia 14 (1995) 257–266.

13

[40] A.P. Robinson, T.M. White, D.W. Mason, Macrophage heterogeneity in the rat as delineated by two monoclonal antibodies MCR OX-41 and MRC OX-42, the latter recognizing complement receptor type 3, Immunology 57 (1986) 239–247. [41] D. Rotin, D. Bar-Sagi, H. O’Brodovich, J. Merilainen, V.P. Lehto, C.M. Canessa, B.C. Rossier, G.P. Downey, An SH3 binding region in the epithelial Na 1 channel (alpha rENaC) mediates its localization at the apical membrane, EMBO J. 13 (1994) 4440–4450. [42] A.Y. Rudensky, M. Maric, S. Eastman, L. Shoemaker, P.C. DeRoss, J.S. Blum, Intracellular assembly and transport of endogenous peptide –MHC class II complexes, Immunity 1 (1994) 585–594. [43] K. Sakai, K. Suzuki, S. Tanaka, T. Koike, Up-regulation of cyclin D1 occurs in apoptosis of immature but not mature cerebellar granule neurons in culture, J. Neurosci. Res. 58 (1999) 396–406. [44] J. Sambrook, E.F. Fritsch, T. Maniatis, Analysis and cloning of eukaryotic genomic DNA, in: Molecular Cloning: A Laboratory Manual, 2nd Edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989, pp. 92–962. [45] L.M. Scott, L. Mueller, S.J. Collins, E3, a hematopoietic-specific transcript directly regulated by the retinoic acid receptor alpha, Blood 88 (1996) 2517–2530. [46] E. Smits, W. Van Criekinge, G. Plaetinck, T. Bogaert, The human homologue of Caenorhabditis elegans CED-6 specifically promotes phagocytosis of apoptotic cells, Curr. Biol. 9 (1999) 1351–1354. [47] K. Suzuki, T. Koike, Brain-derived neurotrophic factor suppresses programmed death of cerebellar granule cells through a post-translational mechanism, Mol. Chem. Neuropathol. 30 (1997) 101–124. [48] W.J. Streit, Microglial response to brain injury: a brief synopsis, Toxicol. Pathol. 28 (2000) 28–30. [49] S. Tanaka, K. Suzuki, M. Watanabe, S. Tone, T. Koike, Upregulation of a new microglial gene, mrf-1, in response to programmed neuronal cell death and degeneration, J. Neurosci. 18 (1998) 6358–6369. [50] S. Thanos, Neurobiology of the regenerating retina and its functional reconnection with the brain by means of peripheral nerve transplants in adult rats, Surv. Ophthalmol. 42 (1997) S5–26. [51] M.B. Upender, J.R. Naegele, Activation of microglia during developmentally regulated cell death in the cerebral cortex, Dev. Neurosci. 21 (1999) 491–505. [52] M.M. Verbeek, I. Otte-Holler, P. Wesseling, W.E. Van Nostrand, C. Sorg, D.J. Ruiter, R.M. de Waal, Differential expression of intercellular adhesion molecule-1 (ICAM-1) in the A beta-containing lesions in brains of patients with dementia of the Alzheimer type, Acta Neuropathol. (Berl.) 90 (1995) 493–503. [53] M.C. Weiler, J.L. Smith, J.N. Masters, CR16, a novel proline-rich protein expressed in rat brain neurons, binds to SH3 domains and is a MAP kinase substrate, J. Mol. Neurosci. 7 (1996) 203–215. [54] Y.C. Wu, H.R. Horvitz, The C. elegans cell corpse engulfment gene ced-7 encodes a protein similar to ABC transporters, Cell 93 (1998) 951–960. [55] G.-M. Yan, B. Ni, M. Weller, K.A. Wood, S.M. Paul, Depolarization or glutamate receptor activation blocks apoptotic cell death of cultured cerebellar granule neurons, Brain Res. 656 (1994) 43–51.