Regulation of the Heparan Sulfate Proteoglycan, Perlecan, by Injury and Interleukin-1␣

*Esther Garcı´a de Ye´benes, *Angela Ho, *Tanuja Damani, *†Howard Fillit, and *†Mariann Blum

*Fishberg Research Center for Neurobiology and †Henry L. Schwartz Department of Geriatrics and Adult Development, Mount Sinai School of Medicine, New York, New York, U.S.A.

Abstract: Perlecan is a specific proteoglycan that binds al., 1991). Perlecan is an HSPG formed by a core protein to amyloid precursor protein and ␤-amyloid peptide, is of 400–470 kDa in the human and 396 kDa in the present within amyloid deposits, and has been implicated mouse, attached to three chains of the glycosaminogly- in plaque formation. Because plaque formation is asso- can heparan sulfate. The role of HSPGs, specifically ciated with local inflammation, we hypothesized that the perlecan, in Alzheimer’s disease (AD) plaque formation mechanisms involved in brain inflammatory responses could influence perlecan biosynthesis. To test this hy- is being extensively investigated. Its presence has been pothesis, we first studied perlecan regulation in mice after demonstrated in several types of amyloidosis (Snow et inflammation induced by a brain stab wound. Perlecan al., 1990b). Furthermore, perlecan has been localized mRNA and immunoreactivity were both increased 3 days immunohistochemically not only in neuritic but also in after injury. Interleukin-1␣ (IL-1␣) is a cytokine induced diffuse plaques (Snow et al., 1988, 1990a,b, 1994b). Its after injury and plays an important role in inflammation. early presence in AD pathological alterations suggests As such, IL-1␣ may be one of the factors participating in that perlecan could play an active role in plaque forma- regulating perlecan synthesis. We thus studied perlecan tion. It has been shown that perlecan binds with high regulation by IL-1␣, in vivo. Regulation of perlecan mRNA affinity to ␤-amyloid peptide (A␤) and amyloid precur- by this cytokine was area-specific, showing up-regulation sor protein (APP) (Narindrasorasak et al., 1991; Bueé et in hippocampus, whereas in striatum, perlecan mRNA was unchanged. To support this differential regulation of al., 1993a,b; Castillo et al., 1997) and to other molecules perlecan mRNA by IL-1␣, basic fibroblast growth factor implicated in AD such as apolipoprotein E (Mahley et (bFGF), a growth factor also present in plaques, was al., 1979). It has been demonstrated as well that heparin, studied in parallel. bFGF mRNA did not show any re- a highly sulfated glycosaminoglycan related to perlecan gional difference, being up-regulated in both hippocam- heparan sulfate glycosaminoglycans, promotes ␤-secre- pus and striatum in vivo. In vitro, both astrocyte and tase cleavage of APP (Leveugle et al., 1997). Moreover, microglia were immunoreactive for perlecan. Moreover, in rodents, fibrillar amyloid deposition after the infusion perlecan mRNA was increased in hippocampal glial cul- of A␤ is increased and more persistent if A␤ is coinfused tures after IL-1␣ but not in striatal glia. These results with HSPGs (Snow et al., 1994a). Finally, in vitro mi- show an increase in perlecan biosynthesis after injury and croglia phagocytosis of A␤ aggregates is inhibited by the suggest a specific regulation of perlecan mRNA by IL-1␣, which depends on brain area. Such regulation may have presence of heparan sulfate glycosaminoglycans (Shaffer important implications in the understanding of regional et al., 1995). However, it is unknown what is leading to brain variations in amyloid plaque formation. Key Words: the accumulation of perlecan in AD. Perlecan—Injury—Interleukin-1—Basic fibroblast growth Although, at present, factors triggering A␤ deposit factor—Alzheimer’s disease—Inflammation. formation are unknown, inflammatory mechanisms have J. Neurochem. 73, 812–820 (1999). been implicated in AD and in altered metabolism of A␤ (Rogers et al., 1996; Tocco et al., 1997). Furthermore,

Heparan sulfate proteoglycans (HSPGs) are typical Received December 21, 1998; revised manuscript received March 24, 1999; accepted March 25, 1999. constituents of basement membranes together with col- Address correspondence and reprint requests to Dr. M. Blum at lagen IV, laminin, and entactin/nidogen. Classically, the Fishberg Research Center for Neurobiology, Mount Sinai School of functions attributed to the HSPGs have been (a) molec- Medicine, One Gustave Levy Place, Box 1065, New York, NY 10029, ular impermeability of the basement membranes through U.S.A. Abbreviations used: A␤, ␤-amyloid peptide; AD, Alzheimer’s dis- anionic interchange, (b) sequestration of growth factors ease; APP, amyloid precursor protein; bFGF, basic ﬁbroblast growth that bind to heparin, and (c) basement membrane degra- factor; GFAP, glial ﬁbrillary acidic protein; HSPG, heparan sulfate dation during the process of tumor metastasis (Noonan et proteoglycan; IL-1, interleukin-1; PBS, phosphate-buffered saline.

812 REGULATION OF PERLECAN BY INJURY AND IL-1␣ 813 there are several clinical and epidemiological reports AD pathology that it exhibits. We analyzed, as well, indicating beneficial effects of antiinflammatory drugs in perlecan regulation in the striatum, a typical brain area AD (for review, see McGeer and McGeer, 1995). Re- where fibrillar deposits of A␤ are not found (Wisniewski gardless of what is the initial event leading to AD pa- et al., 1989; Gearing et al., 1993). Moreover, as a posi- thology, the CNS responds with the establishment of a tive control, we studied the response of bFGF mRNA to local inflammatory reaction with microglial activation IL-1␣ treatment in vivo in both brain areas. We decided and production of cytokines such as interleukin-1 (IL-1) to use bFGF as the positive control because of its local (Griffin et al., 1989), which, in turn, may stimulate the elevation after cortical injury (Logan et al., 1992), its production of neurotrophic factors or extracellular matrix requirement of HSPGs to bind and to activate its receptor components by astrocytes (Giulian and Lachman, 1985). (Yayon et al., 1991; Ornitz et al., 1992), and, finally, its The interaction between these factors in AD may con- increased deposition in association with plaques and tribute to the establishment of a “vicious cycle” (Fillit tangles in AD brain (Go´mez-Pinilla et al., 1990; Stopa et and Leveugle, 1995; Merrill and Benveniste, 1996) in al., 1990). which one event precipitates and aggravates the next. Because glial cells are locally increased after injury During an inflammation reaction in the brain a series and they are responsible for the synthesis of IL-1␣,we of cellular changes and release of cytokines occur. In decided to analyze whether glial cells are able as well to response to trauma, for example, glial cells change their produce perlecan protein. With this goal, we character- morphology, increase their number, and establish new ized hippocampal glial cultures by performing double- interactions with tissue components located in the in- labeled immunocytochemistry to perlecan, with immu- jured area. The goal of these reactive glial changes is the nocytochemical markers for astrocytes and microglia. removal of cellular debris and edema from the wound Finally, to investigate if the regional specificity in the site, reestablishment of the blood–brain barrier, forma- regulation of perlecan was conserved in vitro, we ana- tion of an astrocytic scar, and induction of neuronal lyzed perlecan mRNA regulation by IL-1␣ in both hip- regeneration. In addition, glial cells participate in the pocampal and striatal glial cells in culture. regulation of extracellular matrix components and inflammatory proteins (Reier and Houle, 1988). Initially, microglial cell changes occur in response to neuronal MATERIALS AND METHODS injury, showing cell proliferation, migration to the injury site, and phenotypic and functional changes, as, for ex- Animals and in vivo procedures ample, appearance of phagocytic activity and expression Three-month-old male C57BL/6 mice (Harlan, Indianapolis, of major histocompatibility complex (MHC) class II IN, U.S.A.) were used. Animals were anesthetized with 287.5 mg/kg 2,2,2-tribromoethanol (Avertin; stock of 12.5 mg/ml; antigens. As well, after injury, there is liberation of Aldrich Chemicals Co., Milwaukee, WI, U.S.A.), injected in- cytokines and growth factors such as basic fibroblast traperitoneally. Animals were placed in a stereotaxic frame growth factor (bFGF) (Logan et al., 1992) and local (David Kopf Instruments, Tujunga, CA, U.S.A.), and were used increases in levels of A␤ (Otsuka et al., 1991; Cotman et either to perform brain stab wound or for intraventricular al., 1996) and extracellular matrix component proteins injections. (Laywell et al., 1992; Brodkey et al., 1995). Among the Stab wound. A1-␮l Hamilton syringe was placed in the right inflammatory cytokines locally produced after brain in- side of the brain through the cortex and the hippocampus, jury is IL-1 (Giulian et al., 1986; Nieto-Sampedro and perpendicular to the pial surface of the cortex, and was left in Berman, 1987). It has been demonstrated that IL-1 is one place for 5 min. Coordinates were 2.2 mm caudal to bregma, of the inflammatory cytokines that modulate the reactive 2–3 mm lateral to the midline suture, and 2–3 mm ventral to the astrogliosis to mechanical injury (Balasingam et al., surface of the brain. Animals were killed 3 days after the stab wound, and brains were processed for immunocytochemistry 1994; Rostworowski et al., 1997) and that IL-1 is also and in situ hybridization. responsible for induction of intercellular adhesion mol- Intraventricular injections. IL-1␣ (100 units; specific activ- ecule-1 (Shibayama et al., 1996). Little is known about ity, Ͼ1 ϫ 107 units/mg; Genzyme, Cambridge, MA, U.S.A.) or the regulation of perlecan after brain injury. However, a vehicle [phosphate-buffered saline (PBS), pH 7.4] in a final recent report shows an increase in perlecan immunore- volume of 0.5 ␮l was injected into the right lateral ventricle activity in hippocampus, following intracerebroventricu- using a 1-␮l Hamilton syringe. This dose was chosen based on lar kainate injections (Shee et al., 1998). results obtained after the performance of a dose–response curve In light of all these previous studies, we further inves- for IL-1␣ (data not shown). Coordinates were 1.5 mm rostral to tigated perlecan regulation after brain injury. We ana- bregma, 1 mm lateral to the midline suture, and 2 mm to the lyzed perlecan mRNA and immunoreactive material fol- surface of the brain. The needle was kept in place for 5 min lowing a brain needle stab wound. Because an increased after the injection and withdrawn slowly to prevent backflow of the solution injected along the needle track. Animals were production of IL-1 after brain injury has been reported, it killed 6, 8, and/or 24 h after the injection (n ϭ 3–5 for each seems likely that this cytokine may regulate perlecan group). Brains were quickly removed, and the hippocampus synthesis after stab wound injury. To test this hypothesis, and/or the striatum were dissected, immediately frozen in dry we investigated IL-1␣ regulation of mouse perlecan ice, and stored at Ϫ80°C. Cytoplasmic total RNA was isolated, mRNA expression in vivo. The hippocampus was se- and a nuclease protection assay (Blum, 1989) was performed to lected as the region of study based on the representative quantify perlecan mRNA.

J. Neurochem., Vol. 73, No. 2, 1999 814 E. GARCI´ADEYE´ BENES ET AL.

Quantitative solution hybridization nuclease Laboratories, Burlingame, CA, U.S.A.) for2hatroom tem- protection assay perature; sections were then washed in PBS and coverslipped in Unlabeled sense and high-specific-activity (ϳ1 ϫ 109 cpm/ Permafluor (Lipshaw, Pittsburgh). ␮g) 32P-labeled antisense RNAs were transcribed according to Hippocampal glial cultures. Petri dishes 35 mm in diameter the manufacturer’s recommendation. A standard curve with were fixed with 5% acetic acid/95% alcohol for 30 min at room increasing amounts [0–10 ␮l of 100 fg/␮l(ϩ)-strand] of sense temperature, washed in PBS, and then incubated in blocking RNA was used for quantification. The standard and known buffer (0.3% Triton X-100 and 5% rabbit or goat serum in PBS) amounts of cytoplasmic RNA were hybridized with ϳ200 pg of for 1 h. Double-labeled immunofluorescence was performed, antisense 32P-labeled RNA probe. Samples were heat-dena- combining antibodies to glial fibrillary acidic protein (GFAP) tured at 85°C for 5 min and hybridized overnight at 45°C. After or Mac-1 with perlecan antibody. Cells were incubated over- hybridization, samples were treated with a nuclease, phenol/ night with primary antibodies for perlecan: (a) rat monoclonal chloroform-extracted, precipitated, resuspended in 1ϫ TE anti-HSPG core protein (5 ␮g/ml; Upstate Biotechnology) or buffer, and electrophoresed on a nondenaturing 5% acrylamide (b) anti-perlecan antibody made in the goat [1:100; a generous gel. Gels were dried and quantitated by phosphorimage analy- gift of Dr. R. Kisilevsky, Kingston, Ontario, Canada (for more sis. Results were determined by linear regression analysis from detailed information, see Narindrasorasak et al., 1991)], anti- the standard curve and presented as attomoles of mRNA per GFAP (mouse monoclonal anti-GFAP; 1:50; Boehringer microgram of total RNA. Mannheim, Indianapolis), or anti-microglia (rat monoclonal anti-Mac-1; 1:50; Boehringer Mannheim), in blocking buffer, Isolation of cDNA clones at 4°C. Cultures were washed with PBS and incubated in the Mouse perlecan cDNA clone was generously provided by corresponding combination of secondary antibodies for2hat Dr. J. R. Hassell (Noonan et al., 1991). A 322-bp EcoRI/PstI room temperature. GFAP and microglia were visualized using fragment from BPG7 was subcloned into vector Bluescript II anti-mouse IgG-fluorescein (1:200; Vector Laboratories) and SKϩ/Ϫ. A 414-bp fragment corresponding to nucleotides 525– anti-rat IgG-fluorescein (1:200; Vector Laboratories), respec- 1,004 was also subcloned into vector Bluescript/SKϩ from the tively. Perlecan was visualized by incubation with either bio- bFGF cDNA clone, generously provided by Dr. S. Shimasaki tinylated anti-goat IgG or anti-rat IgG-fluorescein (1:200; Vec- (Shimasaki et al., 1988). tor Laboratories). After PBS washes, cultures were incubated with streptavidin conjugated to rhodamine (1:500; Molecular Primary hippocampal and striatal glial cultures and Probes, Eugene, OR, U.S.A.) for1hatroom temperature, in vitro treatment washed in PBS, and coverslipped in Permafluor (Lipshaw). Postnatal day 0–3 C57BL/6 mice were decapitated, the hippocampus was dissected, and meninges were removed. The In situ hybridization tissue was trypsinized and treated with DNase for 15 min at Animals were killed by decapitation, and brains were 37°C. This was followed by a series of washes and centrifu- quickly removed, embedded in OCT, and frozen in ice-cold gations in minimum essential medium/Ham’s F-12 supple- acetone. Tissues were stored at Ϫ80°C until used. Sections 15 mented with 10% fetal bovine serum. The resulting homoge- ␮m thick were fixed in 4% paraformaldehyde for 10 min. nate (cell suspension) was filtered through a sterile nylon Sections were incubated with prehybridization buffer [50% screen (pore size, 37 ␮m). Cells were plated at 1 ϫ 106 cells formamide, 0.6 M NaCl, 10 mM Tris-HCl (pH 7.5), 0.02% per 60-mm-diameter or ϳ300,000 cells per 35-mm-diameter Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albu- dish. Cultures were incubated at 37°C in an atmosphere of 8% min, 1 mM EDTA (pH 8), 0.5 mg/ml salmon sperm DNA, 500

CO2 and 92% air. ␮g/ml yeast total RNA, and 50 ␮g/ml yeast transfer RNA] for Treatment. Cells at 20–23 days in vitro were placed into 3 h at room temperature, followed by an overnight incubation 0.5% serum for 24–48 h and then were treated with IL-1␣ (500 with hybridization buffer [50% formamide, 0.6 M NaCl, 10 units/ml) or vehicle (water) for 0, 6, 8, and 24 h (n ϭ 3–5 for mM Tris-HCl (pH 7.5), 0.02% Ficoll, 0.02% polyvinylpyrroli- each group). RNA was isolated and quantified by a nuclease done, 0.02% bovine serum albumin, 1 mM EDTA (pH 8), 0.1 protection assay for perlecan mRNA. mg/ml salmon sperm DNA, 500 ␮g/ml yeast total RNA, 50 ␮g/ml yeast transfer RNA, and 10% dextran sulfate] containing Immunocytochemistry 10 ng/ml heat-denatured 35S-labeled perlecan riboprobe (spe- Tissue. Animals were killed, and brains were quickly re- cific activity, ϳ1 ϫ 109 cpm/␮g) at 50°C in a humidified moved and fresh-frozen with OCT (Tissue Tek, Torrance, CA, chamber. After hybridization, sections were washed with dif- U.S.A.) embedding medium immersed in a dry ice–chilled ferent concentrations of saline–sodium citrate (SSC) and then acetone bath. Coronal sections were cut 15 ␮m thick in a digested with RNase [0.1 mg/ml in 0.5 M NaCl, 10 mM Tris cryostat and mounted in Superfrost Plus slides (Fisher Scien- (pH 7.5), and 1 mM EDTA] for 30 min at room temperature. tific, Pittsburgh, PA, U.S.A.). Sections were fixed with 5% The sections were exposed to Hyperfilm-␤max for ϳ2 weeks, acetic acid/95% alcohol for 15 min at 4°C and washed with before being coated with liquid photographic emulsion (Kodak PBS. Tissue was then incubated in blocking buffer (0.3% NTB-2) for ϳ1 month. Triton X-100 and 5% goat serum in PBS) for 1 h, followed by an overnight incubation with primary anti-perlecan antibody Statistical analysis (rat monoclonal antibody) that reacts specifically with HSPG Significant differences between IL-1␣-treated animals and core protein (perlecan) and does not present cross-reactivity to control groups were analyzed using Student’s t test or ANOVA other basement membrane proteins (laminin, collagen IV, and followed by Fisher’s protected least significant difference post ectactin/nidogen) or fibronectin (species cross-reactivity, hu- hoc analysis. The level of significance was set at p Ͻ 0.05, p man, bovine, and mouse; 5 ␮g/ml; Upstate Biotechnology, Ͻ 0.01, or p Ͻ 0.001. Lake Placid, NY, U.S.A.). The following day, after PBS All animal experiments were conducted in accord with washes, sections were incubated in blocking buffer containing Guidelines for the Care and Use of Experimental Animals, the secondary antibody anti-rat IgG-fluorescein (1:200; Vector using protocols approved by the Institutional Animal Care and

J. Neurochem., Vol. 73, No. 2, 1999 REGULATION OF PERLECAN BY INJURY AND IL-1␣ 815

FIG. 1. Perlecan mRNA and protein after brain needle stab wound. A: X-ray film autoradiograph shows the hybridization signal in mouse brain that had received a needle stab wound 3 days before the animal was killed. Brains were collected and processed for in situ hybridization, and then sections were exposed to an x-ray film for 2 weeks. The hybridization signal is localized throughout the needle track (arrow). B: Consecutive brain coronal section that was hybridized with perlecan sense riboprobe as the control. No hybridization signal is detected surrounding the needle track (arrow). C: Dark-field photomicrograph of mouse brain section at the needle track site shows the local increase in perlecan mRNA hybridization signal, revealed by the accumulation of silver grains (arrow), in contrast to the nonspecific signal depicted at the periphery. Tissue was exposed to liquid photomicrographic emulsion for 1 month after the hybridization. D: Perlecan immunoreactivity in the mouse brain after injury depicts a perivascular (arrow) and an extracellular (arrowhead) pattern. E: Brain section from the same animal, treated with no primary antibody. F: Immunostaining for perlecan in the area contralateral to the stab wound site shows a typical vascular pattern of perlecan immunoreactivity. Bar ϭ 1 mm for A and B and 50 ␮m for C–F.

Use Committee at Mount Sinai School of Medicine (95- associated with additional extracellular immunoreactive 300NB). material (Fig. 1D, arrowhead). Low levels of nonspecific background binding in the absence of primary antibody RESULTS are shown in Fig. 1E. Altogether, we were able to detect an increase in both Regulation of perlecan mRNA and protein in mouse mRNA levels and immunoreactivity of perlecan after brain after stab wound injury stab injury to the brain. We were not able to determine We used a needle stab wound to the mouse brain to the cellular origin of the perlecan immunoreactive mate- evaluate perlecan mRNA expression levels and immu- rial in vivo owing to technical difficulties matching the noreactivity after injury. Three days after the stab injury, optimal fixations for the antibodies required for double- animals were killed, and brains were processed for in situ labeling for perlecan and glial or neural cells in tissue hybridization or immunocytochemistry. Perlecan mRNA sections. expression was increased along the needle track in both cortex and hippocampus (Fig. 1A), with a net increase in Regulation of perlecan mRNA by IL-1␣ in vivo the hybridization signal compared with the background IL-1 is a cytokine released by reactive glia in response and to the sense-strand negative control (Fig. 1B). At the to brain injury and may be involved in injury-induced microscopic level (Fig. 1C), an increase in the amount of up-regulation of perlecan expression. Therefore, to test silver grains was observed, corresponding to the eleva- this idea we injected IL-1␣ into the lateral ventricle and tion in level of perlecan mRNA above the background measured its effect on perlecan expression. In addition, levels at the site of the stab wound. However, it was not because dense-core senile plaques are present in the possible to determine the cellular source of the hybrid- hippocampus but rarely found in the striatum in AD, both ization signal. brain areas were collected after the intraventricular in- Perlecan immunoreactivity showed a typical labeling jection of IL-1␣ to determine whether there is a brain of blood vessels in the contralateral side of the lesion region-specific regulation of perlecan. Animals were (Fig. 1F). On the injured side of the brain, perlecan killed at 6, 8, and 24 h after IL-1␣ injection, and the right exhibited an increase in the immunoreactivity pattern in hippocampus and striatum were dissected and analyzed. vessels near the stab wound injury site (Fig. 1D, arrow), Perlecan mRNA was quantified by nuclease protection

J. Neurochem., Vol. 73, No. 2, 1999 816 E. GARCI´ADEYE´ BENES ET AL.

FIG. 2. Quantiﬁcation of hippocampal and striatal perlecan and bFGF mRNA expression after intraventricular injection of IL-1␣. IL-1␣ was injected into the right lateral ventricle of 3-month-old mice at a dos- age of 100 units per animal in a ﬁnal volume of 0.5 ␮l. Control animals received PBS injection. Three time points were examined (6, 8, and 24 h after the injection). The value corresponding to zero-time in- dicates the mRNA levels in noninjected animals. Data are mean Ϯ SEM (bars) values (n ϭ 5), expressed as attomoles of perlecan (A and C) or bFGF (B and D) mRNA with respect to the total amount of RNA measured. *p Ͻ 0.05, **p Ͻ 0.01, ***p Ͻ 0.001, as compared with PBS-injected groups by Stu- dent’s t or ANOVA test.

assay. In the hippocampus, perlecan synthesis showed a Characterization of perlecan expression and significant increase at 6 (almost threefold; p Ͻ 0.01) and regulation by IL-1␣ in mixed primary glial cell 8 h (twofold; p Ͻ 0.01) after IL-1␣ injection (Fig. 2A). cultures from hippocampus and striatum At 24 h, perlecan mRNA returned to control values. It is When the brain is injured, glial cells become activated interesting that in the striatum, no increase in perlecan at the site surrounding the lesion. To initiate the removal mRNA level was observed at any of the times analyzed of cellular debris and the repair of the damaged tissue, (Fig. 2C), suggesting a possible link between the brain cytokines and extracellular matrix proteins are released area-specific regulation of perlecan mRNA by IL-1␣ from these activated cells. To determine whether perle- shown here and the selective distribution of the plaques can may be involved in this glial response to injury, we found in AD brains. turned to an in vitro system first to assess whether perlecan is expressed in astrocytes and/or microglia and Regulation of bFGF mRNA in vivo second to determine whether its expression could be Because it is well documented that bFGF mRNA directly regulated by IL-1␣. Thus, we set up mixed hippocampal primary glial cultures and performed dou- levels increase after IL-1␣ treatment, we quantified ble-labeled immunofluorescence using antibodies against bFGF mRNA in the same animals used to study per- perlecan core protein, a mouse-specific antibody to lecan mRNA regulation to serve as a positive control Mac-1 antigen to label microglia and macrophages, and for the region-specific regulation of perlecan mRNA. an antibody to GFAP, an astrocyte marker. The pattern In the hippocampus, bFGF mRNA levels increased at of perlecan immunoreactivity in vitro showed both a 6 (almost twofold; p Ͻ 0.05) and 8 h (3.7-fold; p cytoplasmic (Fig. 3A and C) and a fibril-like (Fig. 3E) Ͻ 0.001) after IL-1␣ injection (Fig. 2B). After 24 h, morphology, often coincidental with cell-to-cell apposi- bFGF mRNA levels did not show any statistically tions. Small dotted extracellular deposits of perlecan significant difference. In the striatum, bFGF mRNA immunoreactivity were also found (Fig. 3C, arrow). Co- showed the same pattern of response following IL-1␣ localization of perlecan immunoreactivity (Fig. 3A and injection as compared with the hippocampus; how- C) in Mac-1 immunoreactive microglial cells (Fig. 3B) ever, it was found to sustain a significant up-regulation and GFAP immunoreactive astrocytes (Fig. 3D) indi- after 24 h (p Ͻ 0.001; Fig. 2D). In conclusion, these cates that both cell types can produce perlecan. data show that intraventricular injection of IL-1␣ is Next, we determined whether perlecan expression in able to up-regulate both perlecan and bFGF mRNAs in glia is regulated by IL-1␣. It has been demonstrated that the hippocampus, whereas in the striatum IL-1␣ only glia isolated from different brain regions have functional was able to up-regulate bFGF expression but not per- differences (Le Roux and Reh, 1994). Therefore, to lecan. determine whether the selective brain region regulation

J. Neurochem., Vol. 73, No. 2, 1999 REGULATION OF PERLECAN BY INJURY AND IL-1␣ 817

changes at any time point with respect to the control group (Fig. 4B). In conclusion, there was a robust and consistent up- regulation by IL-1␣ in hippocampal perlecan mRNA in vitro, whereas no changes were observed in striatal perlecan mRNA levels. This suggests that the differential response depending on the brain area studied detected by in vivo examination is conserved by isolated glial cells from corresponding brain areas. To determine whether a parallel increase in the levels of cellular perlecan protein could be detected in hippocampal glia in response to IL-1␣ treatment, immuno- cytochemical staining for perlecan, microglia, and astrocytes was performed after 24 h of treatment. No increase in perlecan immunoreactivity was observed in either microglia or astrocytes (data not shown). It is possible that the protein increase takes place sooner or later than the time point examined or that the perlecan protein is released into the media and thus becomes undetectable by immunochemical means.

DISCUSSION Our results showed an up-regulation of perlecan mRNA and protein in the mouse CNS after a needle stab wound. We also showed the regulation of perlecan mRNA in vivo by the inﬂammatory cytokine IL-1␣ to be brain region-speciﬁc, with increased perlecan mRNA in hippocampus but no changes in the striatum. In contrast, we found that bFGF mRNA did not show differential regulation according to brain area after IL-1␣ treatment in vivo, being up-regulated in both hippocampus and striatum. Finally, we demonstrated the ability of IL-1␣ to

FIG. 3. Perlecan immunoreactivity in hippocampal glial cell cultures. Glial cells at 23–25 days in vitro were double-labeled for perlecan, Mac-1, or GFAP protein. A and B: Colocalization of perlecan immunoreactivity (A) with Mac-1 immunoreactive microglial cells (B). C and D: Colocalization of perlecan immunoreactivity (C) with GFAP immunoreactive astroglial cells (D). Mi- croglial cells showed an ameboid-like morphology, without pro- cesses. Extracellular deposits of perlecan were also observed (see C, arrow). E: Fiber-like morphology of perlecan immunoreactivity, coincident with cell–cell appositions, was observed in the glial cell cultures. Bar ϭ 10 ␮m for A and B and 20 ␮m for C–E. of perlecan by IL-1␣ observed in vivo could be due to such differences, we challenged mixed glial cultures from hippocampus and striatum with IL-1␣. Hippocam- FIG. 4. Quantitative analysis of perlecan mRNA expression after pal glial cells were treated with IL-1␣ for 6, 8, and 24 h. IL-1␣ treatment in primary hippocampal and striatal glial cell After 6 h, we found a signiﬁcant induction (eightfold; p cultures. Hippocampal (A) and striatal (B) glial cell cultures at Ͻ 0.001) of perlecan mRNA, and we detected a maximal 21–23 days in vitro were treated with IL-1␣. RNA was collected increase at 8 h (10-fold; p Ͻ 0.001). After 24 h, perlecan at 0, 6, 8, and 24 h after treatment. t ϭ 0 was used as the control in both experiments. Data are mean Ϯ SEM (bars) values, ex- mRNA levels remained signiﬁcantly increased compared pressed as attomoles of perlecan mRNA with respect to the total with control values (p Ͻ 0.001; Fig. 4A). It is interesting amount of RNA measured. ***p Ͻ 0.001, compared with the that in primary striatal glial cultures, we did not detect control values, by ANOVA test.

J. Neurochem., Vol. 73, No. 2, 1999 818 E. GARCIÁDEYE´ BENES ET AL. stimulate perlecan mRNA in vitro in hippocampal but tion of IL-1 by microglial cells could, in turn, interact not in striatal glial cells. Furthermore, we identified by with other components of senile plaques such as perle- double-labeled immunocytochemistry the origin of per- can, helping to perpetuate and aggravate the formation of lecan immunoreactive material to both astroglial and plaques. microglial cells in vitro. Snow et al. (1994a) have demonstrated that infusion of To investigate how perlecan, an HSPG, responds to perlecan in the rat brain is able to precipitate plaque brain injury, we performed a needle stab wound to the formation. Together with this finding is the additional brain. Perlecan immunoreactivity in mouse brain showed observation that perlecan binds to A␤ and initiates fibril a vascular distribution in control tissue, as well as the formation (Narindrasorasak et al., 1991; Bueé et al., appearance of extracellular deposits and increase in the 1993a,b; Castillo et al., 1997). It is tempting to hypoth- diameter of blood vessels following the injury. This esize that perlecan participates in producing senile pattern of staining has been described before in different plaques after its stimulation by IL-1 in an inflammatory types of tissue by using different perlecan antibodies milieu. An abnormal increase of perlecan would be a (Perlmutter et al., 1990). Perlecan mRNA was increased consequence of inflammatory events implicated in A␤ after injury, as assessed by in situ hybridization. Al- deposition as well as an important cause of them. Our though, for technical reasons, we were not able to assess results show an up-regulation of perlecan by IL-1␣ in the directly the cell types responsible for this induction, we hippocampus, one of the brain regions that exhibits typ- speculate based on our in vitro findings that the induction ical AD pathology. Moreover, the same type of results of perlecan occurs in both reactive astrocytes and micro- have been obtained in cortex (data not shown), another glia. In summary, we showed that perlecan is up-regu- area presenting senile plaques. In contrast, IL-1␣ does lated after local brain injury, thus indicating that it also not regulate perlecan in the striatum, suggesting that participates in the brain’s reaction to traumatic injury. area-specific regulation of perlecan by IL-1␣ could be a The pathological findings in AD are in many respects crucial factor participating in the pattern of distribution comparable to those after brain injury. In fact, there is of senile plaques in AD. It is important that in vitro clinical evidence for the presence of neurofibrillary tan- regulation of perlecan by IL-1␣ in primary glial cell gles and diffuse plaques in humans, provoked by repet- cultures from hippocampus and striatum supports our in itive brain trauma, as happens in dementia pugilistica vivo results indicating that this differential regulation is (Roberts et al., 1990). Likewise, epidemiological studies observed in isolated glial cells. Further investigation will have shown that a previous history of head trauma rep- be necessary to determine the exact mechanism and the resents a risk factor for the development of AD (Mor- cell or cells responsible in vivo. timer et al., 1991). It has been reported that a rapid Perlecan and bFGF are two molecules closely associ- accumulation of APP and A␤ occurs not only in contu- ated owing to their interaction at the bFGF receptor level. sion-type head injury, but also in the needle stab model In fact, bFGF binds with high affinity to heparan sulfate- (Otsuka et al., 1991). What is currently not understood is related molecules, and the fibroblast growth factor family what are the key steps that lead to a pathological cascade is known as “heparin binding growth factors.” This close with neuronal degeneration as exhibited in AD instead of relationship, both spatially and functionally, is very im- a “normal” scar formation with healing. What is the portant for bFGF stability and receptor binding (Klags- common factor leading to AD pathology regardless of brun and Baird, 1991; Aviezer et al., 1994). To present a the initial triggering factor? As is discussed by Cotman et positive control for our perlecan mRNA data in vivo, we al. (1996), a classical brain response to injury is normally analyzed in parallel the regulation of bFGF by IL-1␣. limited in time, whereas in AD there is the additional bFGF mRNA was increased by IL-1␣ in both hippocam- factor of the transformation of A␤ molecule to a ␤-sheet pus and striatum. Although IL-1 regulation of bFGF has conformation and its deposition into fibrils. These mol- already been reported (Rivera et al., 1994; Ho and Blum, ecules have the ability to initiate and maintain an inflam- 1997), this is the first time that perlecan and bFGF matory reaction that is initially acute but becomes regulation by IL-1␣ is shown in parallel. In contrast to chronic and lasts indefinitely. Thus, in AD, a local in- perlecan, we present here that bFGF is not differentially flammatory response is constantly present, with a grow- regulated by IL-1␣ depending on the brain area, indicating list of inflammatory-associated markers (Rogers et ing that although there is a close relationship between al., 1996) that are believed to play an active role in the these two molecules, they can be independently regu- AD pathogenesis. Interactions between A␤ and some of lated by the same cytokine in the same brain area. Given these inflammatory proteins have already been studied that there are other HSPGs expressed in the CNS besides (Cribbs et al., 1997), as well as interactions between the perlecan, possibly in different brain regions, different amyloid component and microglial cells, with the dem- HSPGs are used to modulate bFGF receptor binding, onstration that specific A␤ domains are able to stimulate which therefore could explain the lack of coregulation of neuronal killing by microglia (Giulian et al., 1996). these two molecules in some brain areas. Alternatively, There is evidence of cytokine production by microglia in this lack of coregulation may serve to modulate bFGF AD brain (Meda et al., 1995; Mrak et al., 1995), includ- action in a brain region-selective pattern. ing IL-1, which could play a determinant role in senile Perlecan is a widely distributed HSPG, found in both plaque formation. In this scenario, the increased produc- normal and AD brains (Couchman and Ljubimov, 1989;

J. Neurochem., Vol. 73, No. 2, 1999 REGULATION OF PERLECAN BY INJURY AND IL-1␣ 819

Perlmutter et al., 1990; Murdoch et al., 1994; Snow et al., Brodkey J. A., Laywell E. D., O’Brien T. F., Faissner A., Stefansson 1994a,b). Nevertheless, the specific cellular source of K., Dorries H. U., Schachner M., and Steindler D. A. (1995) Focal brain injury and upregulation of a developmentally regulated perlecan in the mammalian brain has been little investi- extracellular matrix protein. J. Neurosurg. 82, 106–112. gated. Recent studies have shown, by immunofluores- Bueé L., Ding W., Anderson J. P., Narindrasorasak S., Kisilevsky R., cence and/or western blotting, the presence of perlecan Boyle N. J., Robakis N. K., Delacourte A., Greenberg B., and protein in microglia/macrophages both in vitro and in Fillit H. M. (1993a) Binding of vascular heparan sulfate proteo- vivo in rat and human tissue (Su et al., 1992; Miller et al., glycan to Alzheimer’s disease amyloid precursor protein is medi- ated in part by the N-terminal region of A4 peptide. Brain Res. 1997). We now demonstrate, in agreement with those 627, 199–204. previous studies, that mouse microglial cells are immu- Bueé L., Ding W., Delacourte A., and Fillit H. (1993b) Binding of noreactive for perlecan as revealed by double-labeled secreted human neuroblastoma proteoglycans to the Alzheimer’s immunofluorescence in vitro. Moreover, astrocytes ap- amyloid A4 peptide. Brain Res. 601, 154–163. pear to express perlecan in vitro. These data support the Castillo G. M., Ngo C., Cummings J., Wight T. N., and Snow A. D. (1997) Perlecan binds to the ␤-amyloid proteins (A␤) of Alzhei- hypothesis that inflammatory cells and their products mer’s disease, accelerates A␤ fibril formation, and maintains A␤ such as IL-1␣ could play a key function in senile plaque fibril stability. J. Neurochem. 69, 2452–2465. genesis. Possibly, the stimulation of microglial cells by Cotman C. W., Tenner A. J., and Cummings B. J. (1996) Beta-amyloid A␤, followed by increased production of IL-1␣ and converts an acute phase injury response to chronic injury re- finally activation of perlecan synthesis, could stimulate sponses. Neurobiol. Aging 17, 723–731. Couchman J. R. and Ljubimov A. V. (1989) Mammalian tissue distri- A␤ fibril formation in an autocrine manner. Supporting bution of a large heparan sulfate proteoglycan detected by mono- this idea is the presence of IL-1␣ receptors in both clonal antibodies. Matrix 9, 311–321. microglia and astrocytes (for review, see Otero and Mer- Cribbs D. H., Velazquez P., Soreghan B., Gable C. G., and Tenner A. J. rill, 1994). However, this is not the only possible se- (1997) Complement activation by cross-linked truncated and chi- quence in the pathogeny of plaque formation. For in- meric full-length ␤-amyloid. Neuroreport 8, 3457–3462. Fillit H. M. and Leveugle B. (1995) Disorders of the extracellular stance, the presence of perlecan in diffuse plaques, which matrix and the pathogenesis of senile dementia of the Alzheimer’s have not yet presented an inflammatory reaction and thus type. Lab. Invest. 72, 249–253. do not present activated microglial cells (Snow et al., Gearing M., Wilson R. W., Unger E. R., Shelton E. R., Chan H. W., 1990a,b), as well as the induction of perlecan gene Masters C. L., Beyreuther K., and Mirra S. S. (1993) Amyloid expression before amyloid formation in experimental precursor protein (APP) in the striatum in Alzheimer’s disease: an immunohistochemical study. J. Neuropathol. Exp. Neurol. 52, murine AA amyloidogenesis (Ailles et al., 1993), could 22–30. indicate that other factors are also implicated in perlecan Giulian D. and Lachman L. B. (1985) Interleukin 1 stimulation of deposition. For example, toxic molecules released lo- astroglial proliferation after brain injury. Science 228, 497–499. cally and other types of cells, such as neurons or endo- Giulian D., Young D. G., and Lachman L. B. (1986) Interleukin 1 of thelial cells, could also be responsible for perlecan syn- the central nervous system. Production by amoeboid microglia. J. Exp. Med. 164, 594–604. thesis. Giulian D., Haverkamp L. J., Yu J. H., Karshin W., Tom D., Li J., In summary, the data presented here have implications Kirkpatrick J., Kuo Y. M., and Rohe A. E. (1996) Specific regarding the pathogenesis of senile plaques, as well as the domains of ␤-amyloid from Alzheimer plaque elicit neuron killing involvement of the immune system and perlecan in AD. in human microglia. J. Neurosci. 16, 6021–6037. Moreover, these studies support the use of antiinflammatory Go´mez-Pinilla F., Cummings B. J., and Cotman C. W. (1990) Induction of basic fibroblast growth factor in Alzheimer’s disease pathology. drugs, drugs manipulating perlecan metabolism, or antago- Neuroreport 1, 211–214. nism of perlecan binding to A␤ via the administration of Griffin W. S. T., Stanley L. C., Ling C., White L., McLeod V., Perrot low-molecular-weight heparin (Leveugle et al., 1998) in the L. J., White C. L., and Araoz C. (1989) Brain interleukin 1 and aim to halt the, until now, inexorable evolution of AD. S-100 immunoreactivity are elevated in Down syndrome and Alzheimer disease. Proc. Natl. Acad. Sci. USA 86, 7611–7615. Ho A. and Blum M. (1997) Regulation of astroglial-derived dopami- Acknowledgment: We are thankful to Dr. R. Kisilevsky for nergic neurotrophic factors by interleukin-1␤ in the striatum of kindly providing a perlecan antibody. This work was supported young and middle-aged mice. Exp. Neurol. 148, 348–359. by grant NS-35092 from the National Institutes of Health. Klagsbrun M. and Baird A. (1991) A dual receptor system is required for basic fibroblast growth factor activity. Cell 67, 229–231. REFERENCES Laywell E. D., Dorries U., Bartsch U., Faissner A., Schachner M., and Steindler D. A. (1992) Enhanced expression of the developmen- Ailles L., Kisilevsky R., and Young I. D. (1993) Induction of perlecan tally regulated extracellular matrix molecule tenascin following gene expression precedes amyloid formation during experimental adult brain injury. Proc. Natl. Acad. Sci. USA 89, 2634–2638. murine AA amyloidogenesis. Lab. Invest. 69, 443–448. Le Roux D. P. and Reh A. T. (1994) Regional differences in glial- Aviezer D., Levy E., Safran M., Svahn C., Eckhart B., Schmidt A., derived factors that promote dendritic outgrowth from mouse David G., Vlodavsky I., and Yayon A. (1994) Differential struc- cortical neurons in vitro. J. Neurosci. 14, 4639–4655. tural requirements of heparin and heparin sulfate proteoglycans Leveugle B., Ding W., Durkin J. T., Mistretta S., Eisle J., Matic M., that promote binding of basic fibroblast growth factor to its Siman R., Greenberg B. D., and Fillit H. M. (1997) Heparin receptor. J. Biol. Chem. 269, 114–121. promotes ␤-secretase cleavage of the Alzheimer’s amyloid pre- Balasingam V., Tejada-Berges T., Wright E., Bouckova R., and Yong cursor protein. Neurochem. Int. 30, 543–548. V. W. (1994) Reactive astrogliosis in the neonatal mouse brain Leveugle B., Ding W., Laurence F., Dehouck M. P., Scanameo A., and its modulation by cytokines. J. Neurosci. 14, 846–856. Cecchelli R., and Fillit H. (1998) Heparin oligosaccharides that Blum M. (1989) Regulation of neuroendocrine peptide gene expres- pass the blood–brain barrier inhibit ␤-amyloid precursor protein sion, in Methods in Enzymology, Vol. 168 (Conn P. M., ed), pp. secretion and heparin binding to ␤-amyloid peptide. J. Neuro- 618–633. Academic Press, San Diego. chem. 70, 736–744.

J. Neurochem., Vol. 73, No. 2, 1999 820 E. GARCI´ADEYE´ BENES ET AL.

Logan A., Frautschy S. A., Gonzalez A. M. and Baird A. (1992) A time Rivera S., Gold S. J., and Gall C. M. (1994) Interleukin 1␤ increases course for the focal elevation of synthesis of basic fibroblast basic fibroblast growth factor mRNA expression in adult rat brain growth factor and one of its high-affinity receptors (flg) following and organotypic hippocampal cultures. Mol. Brain Res. 27, 12–26. a localized cortical brain injury. J. Neurosci. 12, 3828–3837. Roberts G. W., Allsop D., and Brunton C. (1990) The occult aftermath Mahley R. W., Weisgraber K. H., and Innerarity T. L. (1979) Interac- of boxing. J. Neurol. Neurosurg. Psychiatry 53, 373–378. tion of plasma lipoproteins containing apolipoproteins B and E Rogers J., Webster S., Lue L. F., Brachova L., Civin W. H., Emmerling with heparin and cell surface receptors. Biochim. Biophys. Acta M., Shivers B., Walker D., and McGeer P. (1996) Inflammation 575, 81–89. and Alzheimer’s disease pathogenesis. Neurobiol. Aging 17, 681– McGeer P. L. and McGeer E. G. (1995) The inflammatory response 686. system of brain: implications for therapy of Alzheimer and other Rostworowski M., Balasingam V., Chabot S., Owens T., and Yong neurodegenerative diseases. Brain Res. Brain Res. Rev. 21, 195– V. W. (1997) Astrogliosis in the neonatal and adult murine brain 218. post-trauma: elevation of inflammatory cytokines and the lack of Meda L., Cassatella M. A., Szendrei G. I., Otvos L. Jr., Baron P., requirement for endogenous interferon-␥. J. Neurosci. 17, 3664– Villalba M., Ferrari D., and Rossi F. (1995) Activation of micro- 3674. glial cells by ␤-amyloid protein and interferon. Nature 374, 647– Shaffer L. M., Dority M. D., Gupta-Bansal R., Frederickson R. C., 650. Younkin S. G., and Brunden K. R. (1995) Amyloid beta protein Merrill J. E. and Benveniste E. N. (1996) Cytokines in inflammatory (Abeta) removal by neuroglial cells in culture. Neurobiol. Aging brain lesions: helpful and harmful. Trends Neurosci. 19, 331–338. 16, 737–745. Miller J. D., Cummings J., Maresh G. A., Walker D. G., Castillo G. M., Shee W. L., Ong W. Y., and Lim T. M. (1998) Distribution of perlecan Ngo C., Kimata K., Kinsella M. G., Wight T. N., and Snow A. D. in mouse hippocampus following intracerebroventricular kainate (1997) Localization of perlecan (or a perlecan-related macromol- injections. Brain Res. 799, 292–300. ecule) to isolated microglia in vitro and to microglia/macrophages Shibayama M., Kuchiwaki H., Inao S., Yoshida K., and Ito M. (1996) following infusion of beta-amyloid protein into rodent hippocam- Intercellular adhesion molecule-1 expression on glia following pus. Glia 21, 228–243. brain injury: participation of interleukin-1␤. J. Neurotrauma 13, Mortimer J. A., van Duijn C. M., Chandra V., Fratiglioni L., Graves 801–808. A. B., Heyman A., Jorm A. F., Kokmen E., Kondo K., and Rocca Shimasaki S., Emoto N., Koba A., Mercado M., Shibata F., Cooksey W. A. (1991) Head trauma as a risk factor for Alzheimer’s K., Baird A., and Ling N. (1988) Complementary DNA cloning disease: a collaborative re-analysis of case-control studies. and sequencing of rat ovarian basic fibroblast growth factor and EURODEM Risk Factors Research Group. Int. J. Epidemiol. 20 tissue distribution study of its mRNA. Biochem. Biophys. Res. (Suppl.), S28–S35. Commun. 157, 256–263. Mrak R. E., Sheng J. G., and Giffin W. S. T. (1995) Glial cytokines in Snow A. D., Mar H., Nochlin D., Kimata K., Kato M., Suzuki S., Hassell J., and Wight T. N. (1988) The presence of heparan sulfate Alzheimer’s disease: review and pathogenic implications. Hum. proteoglycans in the neuritic plaques and congophilic angiopathy Pathol. 26, 816–823. of Alzheimer’s disease. Am. J. Pathol. 133, 456–463. Murdoch A. D., Liu B., Schwarting R., Tuan R. S., and Iozzo R. V. Snow A. D., Mar H., Nochlin D., Sekiguchi R. T., Kimata K., Koike (1994) Widespread expression of perlecan proteoglycan in base- Y., and Wight T. N. (1990a) Early accumulation of heparan ment membranes and extracellular matrices of human tissues as sulfate in neurons and in the beta-amyloid protein containing detected by a novel monoclonal antibody against domain III and lesions of Alzheimer’s disease and Down’s syndrome. Am. J. by in situ hybridization. J. Histochem. Cytochem. 42, 239–249. Pathol. 137, 1253–1270. Narindrasorasak S., Lowery D., Whitt P. G., Poorman R. A., Greenberg Snow A. D., Wight T. N., Nochlin D., Koike Y., Kitama K., DeArmond B., and Kisilevsky R. (1991) High affinity interactions between S. J., and Prusiner S. B. (1990b) Immunolocalization of heparan the Alzheimer’s ␤-amyloid precursor proteins and the basement sulfate proteoglycans to the prion protein amyloid plaques of membrane form of heparan sulfate proteoglycan. J. Biol. Chem. Gerstmann–Straussler syndrome, Creutzfeldt–Jakob disease and 266, 12878–12883. scrapie. Lab. Invest. 63, 601–611. Nieto-Sampedro M. and Berman M. A. (1987) Interleukin-1 like ac- Snow A. D., Sekiguchi R., Nochlin D., Fraser P., Kimata K., Mizutani tivity in rat brain: sources, targets, and effect of injury. J. Neuro- A., Arai M., Schreier W. A., and Morgan D. G. (1994a)An sci. 17, 214–219. important role of heparan sulfate proteoglycan (perlecan) in a Noonan D. M., Fulle A., Valente P., Cai S., Horigan E., Sasaki M., model system for the deposition and persistence of fibrillar Yamada Y., and Hassell J. R. (1991) The complete sequence of ␤-amyloid in rat brain. Neuron 12, 219–234. perlecan, a basement membrane heparan sulfate proteoglycan, Snow A. D., Sekiguchi R., Nochlin D., Kalaria R. N., and Kimata K. reveals extensive similarity with laminin A chain, low density (1994b) Heparan sulfate proteoglycan in diffuse plaques of hip- lipoprotein-receptor, and the neural cell adhesion molecule. pocampus but not in cerebellum of Alzheimer’s disease brain. J. Biol. Chem. 266, 22939–22947. Am. J. Pathol. 144, 337–347. Ornitz D., Yayon A., Flanagan J. G., Svahn C. M., Levi E., and Leder Stopa G. E., Gonzalez A. M., Chorsky R., Corona J. R., Alvarez J., Bird P. (1992) Heparin is required for cell-free binding of basic fibro- D. E., and Baird A. (1990) Basic fibroblast growth factor in blast growth factor to a soluble receptor and for mitogenesis in Alzheimer’s disease. Biochem. Biophys. Res. Commun. 171, 690– whole cells. Mol. Cell. Biol. 12, 240–247. 696. Otero G. C. and Merrill J. E. (1994) Cytokine receptors on glial cells. Su J. H., Cummings B. J., and Cotman C. W. (1992) Localization of Glia 11, 117–128. heparan sulfate glycosaminoglycan and proteoglycan core protein in Otsuka N., Tomonaga M., and Ikeda K. (1991) Rapid appearance of aged brain and Alzheimer’s disease. Neuroscience 51, 801–813. ␤-amyloid precursor protein immunoreactivity in damaged axons Tocco G., Freire-Moar J., Schreiber S. S., Sakhi S. H., Aise P. S., and and reactive glial cells in rat brain following needle stab injury. Pasinetti G. M. (1997) Maturational regulation and regional in- Brain Res. 568, 335–338. duction of cyclooxygenase-2 in rat brain: implications for Alzhei- Perlmutter L. S., Chui H. C., Saperia D., and Athanikar J. (1990) mer’s disease. Exp. Neurol. 144, 339–349. Microangiopathy and the colocalization of heparan sulfate proteo- Wisniewski H. M., Bancher C., Barcikowska M., Wen G. Y., and glycan with amyloid in senile plaques of Alzheimer’s disease. Currie J. (1989) Spectrum of morphological appearance of amy- Brain Res. 508, 13–19. loid deposits in Alzheimer’s disease. Acta Neuropathol. (Berl.) 78, Reier P. J. and Houle J. (1988) The glial scar: its bearing on axonal 337–347. elongation and transplantation approaches to CNS repair, in Ad- Yayon A., Klagsbrun M., Esko J. D., Leder P., and Ornitz D. M. (1991) vances in Neurology, Vol. 47: Functional Recovery in Neurolog- Cell surface, heparin-like molecules are required for binding of ical Disease (Waxman S. G, ed), pp. 87–138. Raven Press, New basic fibroblast growth factor to its high affinity receptor. Cell 64, York. 841–848.

J. Neurochem., Vol. 73, No. 2, 1999