Sustained Expression of Osteopontin Is Closely Associated with Calcium Deposits in the Rat Hippocampus After Transient Forebrain Ischemia

Abstract

The present study was designed to evaluate the extent and topography of osteopontin (OPN) protein expression in the rat hippocampus 4 to 12 weeks following transient forebrain ischemia, and to compare OPN expression patterns with those of calcium deposits and astroglial and microglial reactions. Two patterns of OPN staining were recognized by light microscopy: 1) a diffuse pattern of tiny granular deposits throughout the CA1 region at 4 weeks after ischemia and 2) non-diffuse ovoid to round deposits, which formed conglomerates in the CA1 pyramidal cell layer over the chronic interval of 8 to 12 weeks. Immunogold-silver electron microscopy and electron probe microanalysis demonstrated that OPN deposits were indeed diverse types of calcium deposits, which were clearly delineated by profuse silver grains indicative of OPN expression. Intracellular OPN deposits were frequently observed within reactive astrocytes and neurons 4 weeks after ischemia but rarely at later times. By contrast, extracellular OPN deposits progressively increased in size and appeared to be gradually phagocytized by microglia or brain macrophages and some astrocytes over 8 to 12 weeks. These data indicate an interaction between OPN and calcium in the hippocampus in the chronic period after ischemia, suggesting that OPN binding to calcium deposits may be involved in scavenging mechanisms.

Osteopontin (OPN) is an adhesive glycoprotein containing the peptide sequence Arg-Gly-Asp (RGD). OPN has been proposed to act as a chemoattractant that recruits microglia, macrophages, and astrocytes in response to ischemic injury and has also been identified as a potent neuroprotectant (Ellison et al. 1998, 1999; Wang et al. 1998; Lee et al. 1999; Meller et al. 2005; Schroeter et al. 2006; Choi et al. 2007; Yan et al. 2009; Chen et al. 2011; van Velthoven et al. 2011). In addition, several lines of evidence suggest that OPN acts as a phagocytosis-inducing opsonin (McKee and Nanci 1996; Pedraza et al. 2008; Schack et al. 2009; Shin et al. 2011). Thus, it appears that OPN action in nervous tissue, especially in pathogenic processes in the ischemic brain, is much more extensive than was initially supposed.

Calcification within the central nervous system is a common finding associated with several neuropathological disorders, including brain ischemia (Gutierrez-Diaz et al. 1985; Dux et al. 1987; Bonnekoh et al. 1992; Kato et al. 1995; Ohta et al. 1996; Puka-Sundvall et al. 2000; Ramonet et al. 2002, 2006; Oliveira et al. 2003; Bueters et al. 2008). Several lines of evidence suggest that OPN plays a major role in the regulation of ectopic calcification and pathological mineralization of the vasculature (Steitz et al. 2002; Fujita et al. 2003; Makiishi-Shimobayashi et al. 2004; Pampena et al. 2004; Giachelli 2005; Scatena et al. 2007). In addition, a recent study by Maetzler et al. (2010) demonstrated that the lack of OPN induces co-occurring progressive neurodegeneration and microcalcification after an excitotoxic insult, suggesting that OPN is involved in brain calcification. In support of this, we found that calcium precipitation may provide a matrix for the binding of the OPN protein to cell debris or degenerating neurites induced by ischemic injury (Shin et al. 2012). Taken together, these data suggest that OPN participates in the regulation of ectopic calcification in the ischemic brain. In this regard, it is interesting to note that in addition to being expressed in activated microglia/macrophages within the first few days after transient forebrain ischemia, OPN protein accumulated in the CA1 hippocampal region over a 4-week period following the insult (Choi et al. 2007). Thus, long-lasting and progressive OPN accumulation in the ischemic hippocampus may be associated with calcium deposits, although the precise nature of these OPN deposits is still unknown.

The present study was designed to characterize OPN accumulation elicited by cerebral ischemia. Given the possible relationship between OPN and calcium deposits, we investigated, over a 12-week survival period, the extent and topography of OPN protein expression following transient forebrain ischemia and compared this with patterns of calcium deposition using alizarin red Staining, a sensitive and specific method for detection of calcium deposits (Nitsch and Scotti 1992; Mori et al. 2000). In addition, we focused on the relationship between OPN expression and glial reactions using confocal and immunogold-silver electron microscopy (EM), as persistent expression of OPN in the ischemic hippocampus has been ascribed to reactive astrocytes (Choi et al. 2007), and glial cells may participate in the formation of calcium deposits following excitotoxin lesioning (Oliveira et al. 2003).

Materials and Methods

Induction of Transient Forebrain Ischemia

All surgical interventions and presurgical and postsurgical animal care were provided in accordance with the Laboratory Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals, and the Guidelines and Policies for Rodent Survival Surgery provided by the IACUC (Institutional Animal Care and Use Committee) in the School of Medicine, The Catholic University of Korea.

Adult male Sprague-Dawley rats (250–300 g) were used in this study. Transient forebrain ischemia was induced by the four-vessel occlusion and reperfusion method described by Pulsinelli and Brierley (1979) with minor modifications (Lee et al. 1999). Briefly, the vertebral arteries were electrocauterized and cut to stop circulation in these vessels. After 24 hr, both common carotid arteries were occluded for 10 min with miniature aneurysmal clips. Only those animals showing completely flat electroencephalograms after vascular occlusion were classified as ischemic and used in the study. Body temperatures (measured rectally) were maintained at 37.5 ± 0.3C with a heating lamp during and after ischemia. Sham-operated rats, with cauterized vertebral arteries and ligatures placed around the carotid arteries, were used as controls. No animal convulsed or died following reperfusion or sham operation.

Animals were allowed to live for 4, 8, and 12 weeks after reperfusion. At each time point following reperfusion, animals (n=11/time points for the ischemic group; n=5/time points for the sham operation group) were deeply anesthetized with 16.9% urethane (10 ml/kg) and killed by transcardial perfusion with a fixative containing 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Additional groups of rats were used for immunogold-silver EM at 4 and 8 weeks after ischemia (n=5/time points).

Immunohistochemistry and Double Labeling

Free-floating sections (25 µm thick) were processed for double-immunofluorescence histochemistry. The sections were incubated at 4C overnight with a mix of monoclonal mouse anti-rat OPN (American Research Products, Belmont, MA; dilution 1:100) and one of the following primary rabbit polyclonal antibodies: ionized calcium-binding adaptor molecule 1 (Iba1; Wako Pure Chemical Industries, Ltd., Osaka, Japan; dilution 1:500), glial fibrillary acidic protein (GFAP; Chemicon, Temecula, CA; dilution 1:1500), or biotin-conjugated mouse monoclonal antibody to neuronal nuclear antigen (NeuN; Chemicon; dilution 1:100). The sections were then incubated with a mixture of Cy3-conjugated anti-mouse antibody (Jackson ImmunoResearch, West Grove, PA; dilution 1:1500), fluorescein isothiocyanate (FITC)–conjugated anti-mouse antibody (Jackson; dilution 1:50), FITC-conjugated anti-rabbit antibody (Jackson; dilution 1:50), or FITC-conjugated streptavidin (Jackson; dilution 1:50) for 1 hr. The specificity of immunoreactivity was confirmed by the absence of the immunohistochemical reaction in sections from which primary or secondary antibodies were omitted or from which the primary antibody was substituted with purified mouse IgGĸ Isotype control (BD Biosciences; Franklin Lakes, NJ). Counterstaining of cell nuclei was carried out with DAPI (4′,6-diamidino-2′-phenylindole; Roche, Basel, Switzerland; dilution 1:1000) for 10 min. Slides were viewed with a confocal microscope (LSM 510 Meta; Carl Zeiss Co. Ltd., Jena, Germany). Images were converted to the TIFF format, and contrast levels were adjusted using Adobe Photoshop version 7.0 (Adobe; San Jose, CA).

Histochemistry and Immunohistochemistry

To simultaneously detect calcium and the OPN protein, we employed alizarin red S staining and immunostaining of OPN. Free-floating sections were first incubated at 4C overnight with monoclonal mouse anti-rat OPN (American Research Products; dilution 1:100). Antibody staining was visualized with FITC-conjugated anti-mouse antibody (Jackson ImmunoResearch; dilution 1:50). Sections were then mounted on gelatinized glass and immersed in 2% (w/v) alizarin red S (Sigma; St. Louis, MO) for 30 sec followed by a rinse in distilled water. Counterstaining of cell nuclei was carried out with DAPI for 10 min. Control sections were prepared as described above.

Immunoelectron Microscopy and Electron Probe Microanalysis

We used a preembedding immunogold-silver technique to detect the OPN protein. Floating vibratome sections (50 µm thick) were blocked with PSG solution (0.05% saponin/0.2% gelatin/0.01 M phosphate-buffered saline) containing 1% bovine serum albumin and then incubated with a mouse monoclonal anti-rat OPN antibody overnight at 4C. The sections were then incubated with an anti-mouse secondary antibody conjugated with 1.4-nm gold particles (Nanoprobes, Stony Brook, NY; dilution 1:100) for 2 hr, and silver enhancement was performed with the HQ silver enhancement kit (Nanoprobes) for 3 min. After fixation, dehydration, and embedding in Epon 812, areas of interest were excised and glued onto resin blocks. Ultrathin sections (70–90 nm thick) were lightly stained with uranyl acetate and observed with an electron microscope (JEM-1010; JEOL, Tokyo, Japan).

For electron probe microanalysis to analyze the spatial distribution of selected elements within OPN-labeled structures, thin sections (80–100 nm) of selected regions were placed on copper grids and carbon coated (approximately 20 nm thickness). The chemical composition of silver-enhanced immunogold-labeled profiles was analyzed using field emission–transmission electron microscopy (FE-TEM; JEM-2100F, JEOL) equipped with EDAX (for elemental detection by X-ray analysis), installed at the Korea Basic Science Institute. The operative voltage was 200 kV.

Results

Spatiotemporal Relationship between OPN Expression and Calcium Deposits in the Ischemic CA1 Hippocampus

Our previous report provided a comprehensive description of the spatiotemporal distribution of OPN in rats at 1 day to 4 weeks following a 10-min ischemic insult (Choi et al. 2007). To investigate the relationship between OPN and calcium deposition in the hippocampus over the chronic interval of 4 to 12 weeks following ischemic injury, we performed double labeling with OPN and alizarin red S. No specific staining for alizarin red S or OPN was detected in the hippocampus of sham-operated rats at any time point (Suppl. Fig. S1). Four weeks after ischemia, tiny granular deposits showing OPN immunoreactivity were diffusely distributed throughout the hippocampal CA1 region (Fig. 1A), as observed previously (Choi et al. 2007). In the same section, amorphous to granule-like alizarin red staining in the CA1 region corresponded well with that of OPN immunostaining (Figs. 1B, C). Over the chronic interval of 8 (Fig. 1D) and 12 weeks (Fig. 1G) after ischemia, distinct patterns of OPN staining were observed; non-diffuse, ovoid to round deposits of OPN immunostaining with variable size appeared to become enlarged and occupied the CA1 pyramidal cell layer over time. Interestingly, at higher magnification, it was clear that OPN immunostaining was much more intense at the periphery than in the core region of these deposits (Fig. 1J). As revealed by alizarin red staining of the corresponding profiles, the distribution pattern of OPN protein closely matched that of calcium deposits (Fig. 1D–L). All immunoreactivity specific to OPN was eliminated when the primary antibody was omitted or substituted by purified mouse IgGĸ Isotype control (Suppl. Fig. S2), confirming the specificity of this antibody.

Figure 1.

Spatiotemporal relationships of osteopontin (OPN) and alizarin red staining in the hippocampal CA1 region at 4 (A–C), 8 (D–F), and 12 weeks (G–L) after transient forebrain ischemia. (A–C) Tiny granular OPN deposits were diffusely distributed over the hippocampal CA1 region, which colocalized with amorphous to granule-like alizarin red Staining. pcl, the pyramidal cell layer; sr, stratum radiatum. (D−L) Over the chronic interval of 8 (1D) and 12 weeks (1G) after ischemia, non-diffuse, ovoid to round deposits of OPN immunostaining with variable size appeared to become enlarged and occupied the CA1 pyramidal cell layer, and their distribution closely matched that of alizarin red Staining. (J−L) Higher magnification views of the boxed areas in G to I, respectively. Note that the surfaces of OPN deposits were more intensely labeled than their cores. Cell nuclei appeared blue after DAPI staining. Scale bars A-I = 200µm; J-L = 50µm.

Spatiotemporal Correlation among OPN, Astroglial, and Microglial Responses in the Ischemic CA1 Hippocampus

To define the relationship between OPN expression and astroglial and microglial responses in the ischemic hippocampus, we performed double labeling with OPN and two glial-specific markers: GFAP and Iba1. As described above, no specific staining for OPN was detected in the hippocampal CA1 region of sham-operated rats, where GFAP-positive astrocytes showed thin glial processes and Iba1-positive microglia exhibited ramified morphology (Suppl. Fig. S3). At 4 weeks after ischemia (Fig. 2A–C), immunostaining for OPN and GFAP was observed throughout the CA1 hippocampus, and it exhibited an overlapping regional distribution. At higher magnification, tiny, diffuse granules of OPN immunostaining were distributed cytoplasmically in astrocytes that appeared hypertrophic (Fig. 2D–F). In addition, dot-like deposits of OPN immunoreactivity appeared to be located outside the astrocytes, probably in the extracellular matrix. At this time point, double labeling for OPN and Iba1 showed that some amoeboid brain macrophages contained dot-like labeled profiles but not tiny OPN granules, as seen in reactive astrocytes (Fig. 2G–L).

Figure 2.

Relationship between osteopontin (OPN) protein expression and astroglial (A−F) and microglial (G−L) responses in the hippocampal CA1 region 4 weeks after transient forebrain ischemia. (A−F) Immunostaining for OPN and glial fibrillary acidic protein (GFAP) was observed throughout the CA1 hippocampus and revealed an overlapping regional distribution. (D−F) Higher magnification views of the CA1 region. Note that tiny OPN granules were observed within the cytoplasm of astrocytes with a hypertrophic appearance (asterisks in D−F). (G−L) Double labeling for OPN and Iba1 showed that some amoeboid brain macrophages contained dot-like labeled profiles. (J−L) Higher magnification views of the CA1 region. Note that tiny OPN granules (asterisks in J−L), which appeared to be localized in astrocytes, were not observed within brain macrophages. Cell nuclei appeared blue after DAPI staining. Scale bars A-C and G-I = 200µm; D-F and J-L = 20µm.

At 8 weeks after ischemia, non-diffuse, ovoid to round deposits of variable sizes and shapes were noted, which appeared to be preferentially localized in the extracellular matrix (Fig. 3A, D). Some astrocytes showed tiny granular deposits of immunostaining, but most astrocytes were unlabeled or weakly labeled (Fig. 3A–C, Suppl. Fig. S4A). At this time point, double labeling for OPN and Iba1 showed that brain macrophages were frequently located in close proximity to OPN deposits and sometimes contained such deposits within the cytoplasm (Fig. 3D–F, Suppl. Fig. S4B).

Figure 3.

Relationship between osteopontin (OPN) protein and astroglial (A−C, G−I) and microglial (D−F, J−L) responses in the hippocampal CA1 region 8 (A−F) and 12 weeks (G−L) after transient forebrain ischemia. (A−C) Double labeling for OPN and glial fibrillary acidic protein (GFAP) showed that although some astrocytes showed tiny granular deposits of immunostaining (asterisks in A−C), most astrocytes were unlabeled or weakly labeled. Note that non-diffuse OPN-labeled deposits of variable sizes and shapes appeared to be located in the extracellular matrix. (D−F) Double labeling for OPN and Iba1 showed that brain macrophages were frequently located in close proximity to OPN deposits and sometimes contained these deposits within the cytoplasm (arrowheads in D−F). (G−I) Double labeling for OPN and GFAP showed that hypertrophic astroglial cytoplasm and processes were in close proximity to conglomerates of OPN-labeled deposits and sometimes encompassed them. Note that the cores of these deposits were devoid of specific labeling for OPN. (J−L) Double labeling for OPN and Iba1 showed that most clusters of OPN deposits were localized within cell bodies of brain macrophages. Cell nuclei appeared blue after DAPI staining. Scale bars A-L = 20µm.

At 12 weeks after ischemia, OPN-labeled deposits with variable size were observed, the centers of which were devoid of specific labeling (Fig. 3G, J). Hypertrophic astroglial cytoplasm and processes were closely associated with these conglomerates of labeled deposits but did not show tiny granular immunostaining (Fig. 3G–I). Double labeling for OPN and Iba1 showed that most conglomerated OPN deposits were localized within cell bodies of brain macrophages (Fig. 3J–L).

Electron Microscopic Analysis of OPN-Immunoreactive Profiles in the Ischemic CA1 Hippocampus

Localization of OPN in the ischemic hippocampal CA1 region was further analyzed by immunogold-silver EM. In animals subjected to ischemic insult followed by 4 weeks of reperfusion, profuse accumulation of silver-enhanced immunogold particles was observed along the surface of irregularly shaped, electron-dense structures, which were located both intracellularly and, frequently, extracellularly. Extracellular calcium deposits, which appeared to be more electron dense at the periphery, were associated with remnants of cellular debris (Fig. 4A) and occasionally contained degenerated mitochondria, which were small and highly electron dense (Fig. 4B). Intracellular calcium deposits were observed within degenerating neurons containing several lipid inclusions (Fig. 4C, D) and also within reactive astrocytes, which were characterized by relatively sparse organelles and the presence of bundles of fibrils or filaments (Fig. 4E, F). The latter cells contained multiple labeled calcium deposits, which were closely associated with glial filaments. Some of these intracellular deposits appeared to be less electron dense compared with extracellular deposits and showed evenly distributed silver grains (insets in Fig. 4D, E). Labeled deposits were sometimes observed within cells with ultrastructural characteristics of phagocytic macrophages (data not shown).

Figure 4.

Ultrastructural localization of osteopontin (OPN) by the immunogold-silver method in the CA1 region at 4 weeks following ischemia. (A, B) Silver grains indicative of OPN accumulated along the surface of electron-dense calcified deposits, which were irregular in shape and size and were located extracellularly surrounded by a large amount of cellular debris. Note that these deposits showed increased electron density at the periphery and sometimes contained degenerated mitochondria (arrowheads in B). (C, D) OPN-labeled deposits were observed within the soma (arrowheads in C) and dendrites (arrowheads in D) of degenerating neurons containing several lipid inclusions (arrows in C, D). nu, nucleus. (Inset in D) Higher magnification view of the boxed area in D. Note that silver grains were evenly localized within the intracellular deposit. (E, F) Some astrocytes contained lipid inclusions (arrow in E) and multiple OPN-labeled deposits, some of which were closely associated with glial filaments (f in F). (Inset in E) Higher magnification view of the boxed area in E. Note that silver grains were rather diffusely localized in the intracellular deposit. Scale bars A, C-E = 4µm; B and F = 1µm; D and E Insets = 0.5µm.

At 8 weeks after ischemia, calcium deposits appeared to be more electron dense and homogeneous in profile compared with those at 4 weeks (compare Fig. 4 with Fig. 5) and to be aggregated and fused together. Calcium deposits delineated by profuse accumulations of silver grains were localized within degenerating neurons, some of which were degenerated to an extent that precluded identification of the cell type (Fig. 5A, B). Silver-enhanced grains were also closely associated with the surface of calcium deposits located in the extracellular space (Fig. 5C, D) or internalized by brain macrophages showing large dark nuclei with dense and highly clumped heterochromatin along the nuclear membrane (Fig. 5D–F). The latter cell occasionally contained large clusters of calcium deposits with profuse silver grains (Fig. 5E) and showed prominent features of phagocytic activity, such as lipophagosomes and lipid droplets (Fig. 5F).

Figure 5.

Ultrastructural localization of osteopontin (OPN) by the immunogold-silver method in the CA1 region at 8 weeks following ischemia. (A, B) Silver grains indicative of OPN accumulated along the surface of electron-dense calcified deposits, which were located intracellularly in dead neurons with few recognizable organelles. Note that these calcium deposits appeared to be fused to each other and much more electron dense than those at 4 weeks following ischemia. (Insets in A and B) Higher magnification views of the boxed areas in A and B, respectively. (C) The aggregated, extracellular calcium deposits (inset shows lower magnification) were delineated by profuse accumulations of silver grains. (D) A brain macrophage containing multiple calcium deposits with conspicuous enrichments of silver grains. Note that the silver grains also accumulated along the surface of aggregated extracellular calcium deposits (E in D). nu, nucleus. (E) A brain macrophage containing large clusters of calcium deposits with profuse silver grains. (F) A brain macrophage containing labeled deposits showed prominent features of phagocytic activity, such as lipophagosomes and lipid droplets (arrows in F). (Inset in F) Higher magnification view of the boxed area in F. Note silver grains in the engulfed deposits (arrowheads). Scale bars A, B, and C inset = 4µm; D and F = 2µm; B, C, E, and A and F Insets = 1µm.

At this same time point, astrocytes showing extensive cell body hypertrophy and cytoplasmic processes with dense bundles of intermediate filaments were frequently observed in close proximity to OPN-labeled deposits (Fig. 6A, B, D) and sometimes completely surrounded them (Fig. 6C, E). In addition, some astrocytes containing labeled deposits also contained phagolysosomes, suggesting phagocytosis by astrocytes (Fig. 6F, G).

Figure 6.

Ultrastructural localization of osteopontin (OPN) in association with reactive astrocytes (A−G) and elemental maps of OPN-labeled deposit (H−K) in the CA1 region at 8 weeks following ischemia. (A, D) Astrocytes showing extensive hypertrophy were frequently observed in close proximity to OPN-labeled calcium deposits. (B, C, E) Higher magnification views of the boxed areas in A and D, respectively. Note that astroglial processes were closely associated with OPN-labeled deposits (B) or completely encompassed them (C, E). Also note dense bundles of parallel-running glial filaments in astrocytes (f in B, C, E). nu, nucleus. (F, G) A reactive astrocyte containing an OPN-labeled deposit (asterisk in G) with a phagolysosome (arrowhead in G), providing evidence of phagocytosis by an astrocyte. A phagocytic brain macrophage (M) is observed in the vicinity. (G) Higher magnification view of the boxed area in F. f denotes bundles of glial filaments in astrocytic cytoplasm. (H−K) Transmission electron microscopy (TEM) and scanning TEM (STEM) images and corresponding compositional images of an identical ultrathin section. Electron probe microanalysis indicates the presence and the spatial distribution of silver (Ag, red) and calcium (Ca, green) within the labeled profile. Note that the prominent signals for calcium were observed within the labeled profile and that silver signals corresponded with silver grains for OPN. Scale bars A, D, F = 4µm; B, C, and G = 1µm; E, H-K = 0.5µm.

The spatial distribution of calcium in the labeled profiles was analyzed by electron probe microanalysis from the same ultrathin section stained with the immunogold-silver technique (Fig. 6H–K). On elemental maps of the labeled profile at 8 weeks after ischemia, calcium signals were much more abundant and evenly distributed within the profile (Fig. 6K). As shown in a TEM image of the corresponding profile, silver signals corresponded with silver grains for OPN (compare Fig. 6H with 6J). These findings suggest that the OPN protein was localized along the surface of electron-dense deposits, which were indeed calcium deposits.

Discussion

Our findings demonstrate sustained and progressive expression of OPN protein in regions most affected by ischemic injury, the CA1 hippocampal region, in rats analyzed between 4 and 12 weeks after a 10-min period of global forebrain ischemia. We found that the distribution of OPN in these affected regions was very closely related to calcium precipitation and glial reactions based on the following observations. First, simultaneous alizarin red staining and OPN immunohistochemistry showed that the spatial and temporal expression patterns of OPN closely resembled those of calcium deposits. Second, OPN-labeled deposits were closely associated with reactive astrocytes and microglia/brain macrophages. Third, the combination of immunogold-silver EM and electron probe microanalysis showed that silver-enhanced grains (indicating OPN) were clearly associated with the surface of electron-dense precipitates, typical of calcium deposits.

Two distinct forms of OPN-labeled deposits were observed by light microscopy: 1) tiny, diffuse granular deposits distributed throughout the CA1 region at 4 weeks after ischemia and 2) larger, ovoid to round deposits, which appeared to become enlarged in the CA1 pyramidal cell layer over the chronic interval of 8 to 12 weeks. Immunogold-silver EM demonstrated that these OPN deposits were indeed diverse types of calcium deposits that were located both intracellularly and extracellularly. It is noteworthy that although the silver grains indicative of OPN accumulated along the surface of extracellular calcium deposits, the grains were evenly distributed within some intracellular deposits. Given that progressive calcium deposition appeared to occur within OPN-labeled deposits, thereby rendering them more electron dense and homogeneous over time (compare Fig. 4 with Fig. 5), it is likely that mature calcium deposits became filled with nearly evenly dense calcium precipitates and that this may have led to accumulation of OPN along the periphery of the calcium deposits.

Intracellular calcium deposits labeled by OPN were frequently observed within astrocytes and neurons at 4 weeks. In particular, astrocytes contained varying amounts of lipid droplets and multiple calcium deposits, most of which were closely associated with glial filaments. At later time points (8–12 weeks), however, when progressive calcium deposition occurred in affected areas, only a few astrocytes showed granular OPN labeling indicative of multiple intracellular calcium deposits. A possible explanation for these results is that the multiple foci of intracellular calcification in astrocytes led to cellular damage and the complete calcification of whole cells. In support of this idea, previous studies have described astroglial calcification in areas of neurodegeneration (Agnew et al. 1979; Mahy et al. 1999).

Given that calcium precipitation in the ischemic brain is associated with impaired calcium homeostasis and neurodegeneration (Gutierrez-Diaz et al. 1985; Dux et al. 1987; Bonnekoh et al. 1992; Shirotani et al. 1994; Ohta et al. 1996; Watanabe et al. 1998; Puka-Sundvall et al. 2000; Bueters et al. 2008) and that neuronal death was restricted mainly to the hippocampal CA1 region after a four-vessel occlusion insult of the same period (Pulsinelli et al. 1982), progressive calcium deposition in the ischemic hippocampal CA1 regions may be due to long-lasting degenerative processes occurring in these regions. Several studies have documented progressive neuronal degeneration in the hippocampus following brief periods of ischemia or trauma (Smith et al. 1997; Bendel et al. 2005; Bueters et al. 2008; Kiryk et al. 2011). Notably, Bueters et al. (2008) found that an accumulation of large calcium deposits occurred in the late neurodegeneration phase, affecting newly generated CA1 neurons, but not at early time points (2–14 days after ischemia) despite extensive neuronal cell death. The results suggest that these calcification processes may be associated with neuroinflammatory and astroglial reactions. In addition, astrocytes participate in the accumulation and structural expansion of calcium deposits in excitotoxic lesions (Oliveira et al. 2003), and heterogeneity of the astroglial population is a contributing factor to variations in regional vulnerability to neurodegeneration and calcium deposition (Rodríguez et al. 2004). In this context, it is of interest that astroglial OPN expression in the hippocampus is first induced at 10 days after ischemia (Choi et al. 2007), further increased at 4 weeks, and maintained at elevated levels for at least 8 weeks. These observations, together with our new findings, suggest that astrocytes express OPN in the early phase of calcium deposition in the ischemic hippocampus and thus astroglial OPN is centrally involved in calcium precipitation in the ischemic brain. The mechanisms underlying this process are poorly understood at present.

We recently suggested that OPN is involved in phagocytosis of fragmented cell debris by macrophages in response to cerebral ischemia (Shin et al. 2011) and that calcium precipitation in cell debris facilitates local binding of OPN, leading to OPN-mediated phagocytosis (Shin et al. 2012). This proposal was supported by the present finding that most OPN deposits were encompassed by brain macrophages at 12 weeks, the latest time point examined (Fig. 3J–L). At 12 weeks, hypertrophic astrocytes were also frequently observed in close proximity to OPN-labeled calcium deposits, some of which appeared to be surrounded and engulfed by astrocytes. Several studies have documented phagocytic activity of astrocytes after various brain injuries (Watabe et al. 1989; al-Ali and al-Hussain 1996; Sobaniec-Lotowska 2003; Mázló et al. 2004; Ito et al. 2007). Our observations suggest that surviving astrocytes may participate in phagocytosis of calcium deposits, although brain macrophages appear to be the cells primarily responsible for their removal and that OPN may be involved in the scavenging mechanism. However, several lines of evidence suggest that OPN regulates ectopic mineralization in response to injury in many soft tissues, including brain tissue subjected to excitotoxic insult (Boskey et al. 1993; Steitz et al. 2002; Makiishi-Shimobayashi et al. 2004; Giachelli 2005; Speer et al. 2005; Maetzler et al. 2010). Steitz et al. (2002) suggested in an elegant study that OPN initially physically inhibits crystal growth and simultaneously provides a recognition site and/or concentration gradient for macrophages, thereby leading to acidification and dissolution of residual bioapatite. Thus, the possibility that binding of OPN to calcium deposits inhibits the calcification process cannot be excluded, and further studies are needed to identify the roles of OPN in calcification in the ischemic brain.

In conclusion, our data indicate that OPN is closely associated with calcium precipitation in the ischemic CA1 hippocampus. The OPN protein accumulated profusely along the surface of diverse types of calcium deposits, ranging from small intracellular deposits to large extracellular conglomerates. The labeled calcium deposits appeared to be gradually phagocytized by microglia/brain macrophages and some astrocytes over 8 to 12 weeks after ischemia. Thus, our data suggest that OPN binding to calcium deposits is involved in the mechanisms by which such deposits are scavenged.