Astrocytes release peptide and nonpeptide transmitters that influence neuronal development, function, and plasticity. However, the molecular components of the astroglial secretory pathways in vivo are largely unknown. Here, we analyze in astrocytes the production, expression regulation, trafficking, and release of secretogranin III (SgIII), a member of the multifunctional granin family. We show that astroglial cells in culture synthesize and release a nonprocessed form of SgIII. In vivo studies show that many neuronal populations produce and transport SgIII. In particular, the highest SgIII expression in the cerebral cortex in vivo is present in astroglial cells. Both SgIII protein and mRNA are abundantly detected in cortical astrocytes and in Bergmann glial cells. Moreover, the levels of SgIII mRNA and protein in reactive astrocytes, induced by perforating injury increase dramatically. These results implicate SgIII in the astrocyte secretory pathway in vivo and show that its expression is finely regulated during glial activation. The robust expression of SgIII in astrocytes and its regulation in the injured brain suggest both intracellular and extracellular roles for this glial granin in the physiology and repair/damage of neuronal circuits.
Exocytotic vesicular secretion of signaling molecules contributes to glia–neuron communication (Araque et al. 2001; Volterra and Meldolesi 2005). In particular, astrocyte-secreted glutamate, adenosine triphosphate (ATP) and D-serine regulate neuronal physiology by modulating synaptic transmission and network activity (Pascual et al. 2005; Panatier et al. 2006; Perea and Araque 2007). In addition to nonpeptide transmitters, peptides and proteins released from astrocytes, such as growth factors, cytokines, and extracellular components, control the development, plasticity, and pathology of neuronal circuits (Allan and Rothwell 2001; Christopherson et al. 2005). Although astrocyte vesicular secretion is crucial for neuronal function, secretory pathways remain poorly characterized. It has now been shown that astroglial cells contain competent compartments and molecular machinery for a regulated secretory pathway (for review, see Montana et al. 2006). Thus, astrocytes can release gliotransmitters, such as glutamate, ATP, and neuropeptide Y, from synaptic-like microvesicles, lysosomes, and secretory granules in a Ca2+-dependent manner (Bezzi et al. 2004; Zhang et al. 2007; Ramamoorthy and Whim 2008). However, the molecular components of the transmitter secretory pathway in glia in vivo are largely unknown (Fiacco et al. 2009).
A biochemical hallmark of neuroendocrine secretory granules is the storage and release of uniquely acidic proteins known as granins, which are considered markers of the regulated secretory pathway (Taupenot et al. 2003; Helle 2004). The granin family comprises the “classical” members: chromogranin A (CgA), chromogranin B (CgB), and secretogranin II (SgII); and the 5 less known secretogranin III (SgIII, 1B1075 gene product), secretogranin IV (HISL-19 antigen), secretogranin V (neuroendocrine secretory protein 7B2), secretogranin VI (NESP55), and secretogranin VII (the nerve growth factor inducible protein VGF; Taupenot et al. 2003; Helle 2004; Montero-Hadjadje et al. 2008). Intracellularly, granins have a role in the sorting and aggregation of secretory products in the trans-Golgi network and in the subsequent biogenesis of secretory granules (Taupenot et al. 2003; Helle 2004). Moreover, proteolytic processing of these proteins, as for prohormones, gives rise to bioactive peptides (e.g., catestatin and secretoneurin) with autocrine, paracrine, and endocrine functions (Simon et al. 1988; Shyu et al. 2008).
In endocrine cells, SgIII is a key sorting receptor for peptide hormones (for review, see Takeuchi and Hosaka 2008). Moreover, SgIII has recently been shown to be expressed and released by nonneuroendocrine cell types that display a regulated secretory pathway, such as platelets and mast cells (Coppinger et al. 2004; Prasad et al. 2008). Here, we studied the little known SgIII in rodent astrocytes. We show that a nonprocessed form of SgIII is produced and released by cortical astrocytes in primary cultures. In brain sections, SgIII transcripts and protein were found in neuronal populations and abundantly in cortical astrocytes. Moreover, SgIII expression was specifically upregulated in reactive astrocytes after perforating brain injury. These results show that SgIII is a reliable component of the astrocyte secretory pathway and suggest important roles for glial SgIII in the glia–neuron communication.
Materials and Methods
Antibodies and Reagents
Polyclonal antibodies against SgIII were purchased from Sigma–Aldrich (Deisenhofen, Germany) and Proteintech Group Inc. (Chicago, IL). SgII antibodies were kindly provided by Dr R. Fischer-Colbrie. Monoclonal antibodies against synaptosome-associated protein of 25 kDa (SNAP-25), glial fibrillary acidic protein (GFAP), neuronal nuclei (NeuN), carboxypeptidase E (CPE), FLAG, β-tubulin, and vimentin were obtained from Sternberg Monoclonals Inc. (Baltimore, MD), Chemicon (Temecula, CA), BD Transduction Laboratories (San Jose, CA), Sigma–Aldrich, and Developmental Studies Hybridoma Bank (Department of Biological Sciences, The University of Iowa, Iowa city, IA). Polyclonal antibody to Olig2 was from Chemicon. FLAG-tagged rat SgIII (1–471) cDNA was a gift of Dr M. Hosaka. Brefeldin A was obtained by Sigma–Aldrich and 8Br-cyclic adenosine 3′,5′-monophosphate (cAMP) was from Biolog Life Science Institute (Bremen, Germany). Most chemicals and cell culture reagents were from Sigma–Aldrich and GIBCO (Invitrogen, Paisley, UK), respectively.
OF1 mice were provided by Charles River Laboratories, Inc. (Lyon, France), and Sprague–Dawley rats were obtained from the Laboratory Animal Service of the University of Barcelona. Animals were anesthetized with ketamine (Ketolar, Parke-Davis, Barcelona, Spain) and xylazine (Rompun, Bayer Healthcare, Kiel, Germany). Stab wound lesions were made as described previously (Aguado et al. 2002). Under anesthesia, adult animals were fixed to a stereotactical frame, and the cerebral cortex was then stabbed with a #11 scalpel blade. The wounds were 3-mm long and 2-mm deep at 1.5 mm from and parallel to midline and ran 0.5 mm away from the Bregma over the right parietal cortex. To provoke secretory protein accumulation in neural bodies, colchicine was administrated intracerebroventricular (i.c.v.) (Aguado et al. 1999). Seventy-five microgram of colchicine in 10 μL of phosphate-buffered saline (PBS) was stereotaxically injected in the lateral ventricle of adult rats. Sham-operated controls only received 10 μL of PBS. Animals were kept under controlled temperature (22 ± 2 °C), humidity (40–60%), and light (12-h cycles) and treated in accordance with the European Community Council Directive (86/609/EEC) on animal welfare. Every effort was made to minimize animal suffering. In all in vivo experimental conditions, at least 4 animals were used per time-point.
The COS kidney epithelial cell line was grown in Dulbecco's Modified Eagle's Medium supplemented with 10% fetal bovine serum with penicillin/streptomycin at 37 °C in a 5% CO2 atmosphere. COS cell line was transfected with FLAG-SgIII using Lipofectamin 2000, and peptide expression was assayed 48 h later.
Astrocyte cultures were prepared from the cerebral cortex of 2-day-old mice as described previously (Paco et al. 2009). Briefly, the cerebral cortex was isolated, and the meninges were carefully dissected away. Cortical tissues were then minced and incubated in 0.25% trypsin and 0.01% DNase (Roche Diagnostics, Mannheim, Germany). Dissociated cells were seeded in flasks and grown in high-glucose Dulbecco's Modified Eagle's Medium, and F-12 (1:1) containing 10% fetal bovine serum, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid and penicillin/streptomycin at 37 °C in a 5% CO2 incubator. At confluence (days 10–12), the flasks were shaken over night, and the cells were rinsed, detached, and subcultured at a density of 50 000 cells/cm2 on poly-L-lysine-coated plastic culture dishes and coverslips. Under these conditions, cell cultures were essentially formed by astrocytes (>92% GFAP+ cells) and virtually devoid of neurons (<0.5% Tuj-1+ cells). A small percentage of microglia and immature oligodendrocytes (<8% CD11b+ and NG2+ cells) was also detected (Paco et al. 2009). Long-term treatment with 8Br-cAMP did not change the cellular composition (Paco et al. 2009). Neuronal cultures were prepared from the cerebral cortex of E16 mouse embryos. After trypsin and DNase treatment, dissociated cells were seeded on poly-L-lysine-coated culture plates. Neurons were grown at a density of 50 000 cells/cm2 in Neurobasal medium containing B27 and N2 supplements, glutamine, and penicillin/streptomycin at 37 °C in a 5% CO2 atmosphere for 10 days. During the first 4 days in culture, the media were also supplemented with 20 μg/mL 5-fluoro-2′-deoxyuridine and 50 μg/mL uridine to inhibit the mitotic activity of glial cells. More than 99% of the cells were Tuj-1+ neurons and less than 1% of the cells were GFAP+ astrocytes (Paco et al. 2009). In all in vitro conditions, at least 3 independent experiments in duplicate were performed.
Reverse Transcriptase-Polymerase Chain Reaction
To perform polymerase chain reaction (PCR) experiments, total RNA was purified from the forebrain and from cultured astrocytes of mice using Trizol Reagent (Invitrogen) following the manufacturer's instructions. All RNA obtained was reverse transcribed using pdN6 (random hexamers) and AMV reverse transcriptase (Amersham Biosciences, GE Healthcare, Buckinghamshire, UK). Then, cDNAs were then amplified with specific primer pairs (5′–3′) for SgIII, forward CAATTCAAGCTTTCCCCAAA and reverse TTTCACTCGGCTTGCTTTCT (NM_009130); GFAP, forward TGCTAGCTACATCGAGAAGG and reverse TCCTCTGTCTCTTGCATGTT (NM_001131020.1); SNAP-25, forward CCTCCACTCTTGCTACCTGC and reverse CTCCTCTGCATCTCCTCCAG (NM_011428.3); and actin, forward ATATCGCTGCGCTGGTCGTC and reverse AGGATGGCGTGAGGGAGAGC (NM_007393.3). PCR products were analyzed on 1–2% agarose gels, and images were captured with a scanner.
Cultured cells and tissues were homogenized in ice-cold lysis buffer containing 50 mM Tris–HCl pH 7.4, 150 mM NaCl, 5 mM MgCl2, 1 mM ethyleneglycol-bis(2-aminoethylether)-N,N,N',N'-tetra acetic acid, 1% Triton X-100, and protease inhibitor cocktail (Roche Diagnostics). Samples of postnuclear lysates were electrophoresed in 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE; BioRad Laboratories, Hercules, CA) and then transferred to nitrocellulose membranes (Whatman Schleicher & Schuell, Keene, NH). The membranes were blocked in a solution containing 5% nonfat milk powder in tris-buffered saline tween-20 (140 mM NaCl, 10 mM Tris–HCl, pH 7.4, and 0.1% Tween 20) for 1 h at room temperature and then incubated with primary antibodies in blocking buffer for 2 h at room temperature. After several washes in blocking solution, the membranes were incubated for 1 h with horseradish peroxidase-conjugated secondary antibodies (VECTOR, Burlingame, CA). Bound antibodies were visualized with enhanced chemiluminescence reagents (GE Healthcare). Blot images were captured with a scanner.
Astrocytes and neurons were seeded on poly-L-lysine-coated tissue culture dishes, to minimize cell loss. Before the release experiment, cells were serum starved. Brefeldin A 5 μg/mL was incubated with astrocytes for 12 h and with neurons for 4 h. After release, conditioned medium was collected, and lysis buffer was added to the cells (see above). Cell media were centrifuged at 600 g for 5 min to remove dislodged cells, and proteins were precipitated with 5% trichloroacetic acid. Secretory proteins were analyzed in both media and cells by SDS–PAGE and immunoblotting.
Cells grown on glass coverslips were washed in ice-cold PBS and fixed with 4% paraformaldehyde in PBS for 15 min. Animals were perfused transcardially under deep anesthesia with 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4. The brains were removed from the skulls, postfixed for 4 h in the same fixative solution, and cryoprotected overnight at 4 °C by immersion in a 30% sucrose solution in 0.1 M PB. Forty-micrometer thick frozen sections were obtained with a cryostat and collected in PBS. Sections processed for the peroxidase method were soaked for 30 min in PBS containing 10% methanol and 3% H2O2 and subsequently washed in PBS. To suppress nonspecific binding, cell cultures and brain sections were incubated in 10% serum–PBS containing 0.1% Triton X-100, 0.2% glycine, and 0.2% gelatin for 1 h at room temperature. Incubations with the primary antibodies were carried out overnight at 4 °C in PBS containing 1% fetal calf serum, 0.1% Triton X-100, and 0.2% gelatin. Some histological sections were processed using the avidin–biotin–peroxidase method (Vectastain ABC kit, VECTOR). The peroxidase complex was visualized by incubating the sections with 0.05% diaminobenzidine and 0.01% H2O2 in PBS. Sections were mounted, dehydrated, and coverslipped in Eukitt. Cell cultures and some brain sections were processed for immunofluorescence using secondary fluorochrome-conjugated antibodies (Alexa Fluor 488 and Alexa Fluor 568, Molecular Probes, Eugene, OR), and cell nuclei were stained with 4′,6-diamidino-2-phenylindole (Molecular Probes). Cell-containing coverslips and histological sections were mounted with Mowiol. The specificity of the immunostaining was tested by omitting the primary antibodies or by replacing them with an equivalent concentration of nonspecific IgG. No immunostaining was observed in these conditions. Bright field and fluorescent images were obtained with the Olympus fluorescent BX-61 and Leica TCS SPE scanning confocal microscopes.
Double immunolabeling using primary antibodies raised in the same species (SgIII and Olig-2) was performed by a sequential avidin–biotin–peroxidase method (Levey et al. 1986). The first immunostaining was performed as above. The second immunolabeling reaction was developed in a medium containing 0.01% benzidine dihydrochloride, 0.025% sodium nitroprusside (Merck, Darmstadt, Germany), and 0.005% H2O2 in PBS, pH 6. Double immunolabeling using primary antibodies raised in different species was performed by incubation with different fluorescent-conjugated secondary antibodies. Colocalization data are expressed as mean ± standard error of the mean of the colocalization percentage for at least 7 brain sections from 3 animals.
In Situ Hybridization
Nonradioactive hybridizations were performed on free-floating forebrain sections according to the method of Aguado et al. (2003). Briefly, 4% paraformaldehyde-perfused brains were cryoprotected in 30% sucrose, frozen, and coronally sectioned at 40 mm. Antisense, sense, and nonrelated probes were obtained by in vitro transcription of the rat SgIII and the mouse neuron-specific cotransporter KCC2 cDNAs in the presence of digoxigenin-UTP (Aguado et al. 2003; Han et al. 2008). Free-floating coronal sections were hybridized with RNA probes. After hybridization, sections were incubated with sheep polyclonal antidigoxigenin Fab fragments conjugated to alkaline phosphatase (Roche Diagnostics) and developed with a BCPI/NTB substrate. To reveal the strong SgIII mRNA expression in glial cells compared with neurons, hybridization reaction was limited. In some hybridized sections, SgIII, GFAP, and NeuN were subsequently detected by immunofluorescence. No signal was detected in control hybridizations performed with sense riboprobe, and a neuronal-restricted expression pattern was obtained with the KCC2 riboprobe.
A Nonprocessed Form of SgIII Is Produced and Released by Cultured Astrocytes
To identify SgIII, we used 2 different polyclonal antibodies, HPA006880 and 10954-1-AP. We confirmed the specificity of the antibodies detecting FLAG-tagged SgIII expressed in COS cells by western blotting (Fig. 1). HPA006880 and FLAG antibodies recognized the same bands corresponding to SgIII in both cell lysates and media of transfected cultures (Fig. 1). No signal was found when the SgIII antibodies were incubated with mock-transfected COS cells (Fig. 1). Immunofluorescence staining revealed that transfected SgIII was sorted to cytoplasmic compartments (data not shown). Similar results were obtained with 10954-1-AP antibodies (data not shown).
We examined the expression of SgIII by astrocytes analyzing tissues and primary cell cultures. As previously reported (Ottiger et al. 1990), anterior pituitary was a major source of SgIII, whereas brain tissues showed moderate quantities of the granin (Fig. 2A). In endocrine tissues (Fig. 2A), both polyclonal antibodies against SgIII recognized the precursor, mature, and cleaved proteins (ranging from 70 to 20 kDa; Holthuis et al. 1996; Han et al. 2008). Precursor and mature SgIII (∼70 and 55 kDa bands) were also found in brain and cultured neurons, whereas a higher electrophoretic mobility band (∼75–80 kDa; Han et al. 2008) was occasionally detected in cultured neurons (Fig. 2A). SgIII was also observed in lysates of cultured astrocytes but at lower levels. Remarkably, the detected SgIII forms in glial cells corresponded with the largest size forms (Fig. 2A). Electrophoretic mobility of SgIII in cultured astrocytes was the same, irrespective of whether it was detected using either HPA006880 or 10954-1-AP antibodies (data not shown). We also compared SgIII content in cell cultures with another astroglial granin, SgII (Fischer-Colbrie et al. 1993). SgII was hardly detected and only revealed by increasing the amount of total protein loaded on gel (Fig. 2A). Finally, expression of SgIII by cultured astrocytes was further confirmed by reverse transcriptase-polymerase chain reaction (Fig. 2B). These data demonstrate that SgIII is expressed by cultured astrocytes.
Next, we analyzed SgIII trafficking and release from cultured astrocytes. SgIII was abundant in the culture medium of astrocytes (as early as 30 min after release) but scarce in cell lysates (Fig. 3A). A double band corresponding to nonprocessed precursor forms was generally observed in astrocyte culture media and cell lysates (Fig. 3A). In agreement with the secretory pathway involved, SgIII released from astrocytes was markedly reduced after administration of Brefeldin A, a drug that blocks secretion by inhibiting transport from endoplasmatic reticulum to the Golgi (Fig. 3A). In contrast, Brefeldin A treatment dramatically increased the SgIII content in cell lysates (Fig. 3A). Immunofluorescence labeling showed the retention of de novo synthetized SgIII in secretory compartments of astrocytes during Brefeldin A treatment (Fig. 3B). Moreover, the granin accumulation in GFAP-positive cells after Brefeldin A treatment further confirmed the SgIII expression by astroglial cells. In contrast to astrocytes, basal secretion of SgIII was extremely low in neurons (Fig. 3A). Moreover, Brefeldin A incubation in neuronal cultures increased the content of the precursor SgIII form, indicating that SgIII retention in early secretory compartments prevents its proteolytic processing in neurons (Fig. 3A). Because it has been proposed that SgIII exerts a role in peptide sorting thought its interaction with CPE and astrocytes do synthesize CPE (Vilijn et al. 1989; Hosaka et al. 2005), we compared release of SgIII with CPE from astroglial cells. Brefeldin A-sensitive secretion of glial CPE was very similar to that observed for SgIII, displaying high rates of basal release (Fig. 3A). We conclude that SgIII is produced by cultured astrocytes as a nonprocessed form which is poorly stored and mostly secreted into the media concurrently with CPE.
It has been shown that granin synthesis can respond differently in different cell types to a variety of signaling pathways, including cAMP (Taupenot et al. 2003). To examine whether the content of SgIII in cultured astrocytes is regulated differentially to other glial granins, cell cultures were treated with the permeable cAMP analog 8Br-cAMP (1 mM) for 8 days, and levels of SgIII and SgII were analyzed. Long-term cAMP elevation in astrocytes upregulated SgII in both culture media and cell lysates (Fig. 4). This regulation is consistent with the presence of a functional cAMP response element in the SgII promoter (Taupenot et al. 2003). In contrast, released SgIII was reduced by 8Br-cAMP administration (Fig. 4). No apparent changes were detected in cell lysates (Fig. 4). These results show that SgIII and SgII are differentially regulated in astrocyte by the cAMP signaling pathway.
Cortical Astrocytes Robustly Express SgIII In Vivo
To analyze expression of SgIII by astrocytes in vivo, we combined immunohistochemistry with nonradioactive in situ hybridization on tissue sections. The strongest SgIII immunostaining was detected in endocrine cell subsets of the anterior pituitary and neurosecretory areas of the brain (Fig. 5A,B). Endocrine pituitary cells showed granular staining throughout the cytoplasm (Fig. 5A), whereas SgIII signal in hypothalamic areas was mainly associated with neuronal fibers (Fig. 5B). The immunostaining for SgIII in neuroendocrine areas was consistently correlated with an mRNA hybridization signal (Fig. 5B,C). In addition to neuroendocrine nuclei, varicose axons, punctate structures, and neuronal cell bodies were immunolabeled for SgIII in most forebrain areas (data not shown). Although many neuronal somata displayed SgIII mRNA hybridization signal, SgIII immunolabeling was weak in neuronal cell bodies (Fig. 5D,E). We examined whether synthesized neuronal SgIII is transported along axons by arresting the axonal transport with colchicine. Although the microtubule disruption colchicine may alter gene expression, protein accumulation in the neuronal bodies and the initial axonal segment by i.c.v. administration is considered the consequence of a transport arresting (Cortés et al. 1990; Aguado et al. 1999). In colchicine-treated brains, SgIII immunolabeling was found accumulated in neuronal bodies in different areas, as shown in the cerebral cortex (Fig. 5E,F). Specifically, strong SgIII immunostaining was detected in dilated proximal axons of large-projection neurons (Fig. 5F). Accordingly with the high levels of SgIII mRNA in neuroendocrine cells, the colchicine-induced SgIII increase in neuronal somata was outstandingly prominent in hypothalamic neurons (Fig. 5G,H). These data show that SgIII is expressed and transported along axons in forebrain neurons.
In addition to neurons, a prominent SgIII immunolabeling was observed in cell bodies and thin processes of nonneuronal star-shaped cells in the forebrain (Fig. 6A,B). SgIII stained astroglial-like cells were consistently observed throughout different layers of the cerebral cortex (i.e., neocortex and hippocampus), mainly in the upper layers (Fig. 6A,B). Other areas of the forebrain and mid- and hindbrain also displayed non neuronal Sg-III positive cells (Supplementary Figure S1). Some star-shaped cells showing a faint SgIII labeling were detected in myelinated tracts, such as corpus callosum, fimbria, and the optic nerve (Fig. 6C). Strong labeling for SgIII was detected in Bergmann glial cell bodies located in the Purkinje cell layer of the cerebellum and their radial processes to the pial surface (Fig. 6D). Double labeling with SgIII and the marker Olig2 indicated that oligodendrocytes did not contain this granin (Fig. 6E). In contrast, we found a widespread colocalization of SgIII with the astrocyte marker GFAP (Fig. 7). For instance, about 96% of SgIII-labeled nonneuronal cells in the piriform cortex (96.71 ± 0.96%) and hippocampus (97.12 ± 0.95%) were GFAP+ (Fig. 7). Characteristically, other GFAP structures, such as cell bodies in the white matter and astrocytic end-feet surrounding blood vessels and the glia limitans displayed weak or absent SgIII immunostaining (Fig. 7). High magnification images performed by confocal microscopy showed that SgIII labeling in astrocytes was typically found as scattered punctate staining in cell bodies and processes (Fig. 7). Parallel double immunofluorescence staining was performed with antibodies against SgII. However, no signal for SgII was detected in GFAP+ structures (data not shown). These results show that SgIII is a distinctive granin of astrocytes in situ.
Next, expression of SgIII by astrocytes was further investigated by analyzing SgIII transcripts in brain sections. Examination of forebrain showed that the strongest SgIII hybridization signal corresponded to nonneuronal cells (Fig. 8). The SgIII mRNA expression pattern in nonneuronal cells was similar to that observed for SgIII protein. Thus, glial cells displaying SgIII mRNA transcripts were mainly detected in the cerebral cortex and to a lesser extend in other forebrain regions (Fig. 8A–D). Nonneuronal cells expressing high SgIII mRNA levels were distributed throughout the different layers of cortical areas, that is, neocortex and hippocampus (Fig. 8A–D). Hybridization signal was detected as a thin punctate in the processes and somata of glial cells (Fig. 6A,D). Astrocytic identity of glial SgIII expressing cells was demonstrated by double labeling with the marker GFAP in each region analyzed (Fig. 8D,E). Moreover, location of SgIII mRNA and protein in the same glial cells further substantiates the SgIII synthesis in astrocytes (Fig. 8F,H). Taken together, these data show that SgIII is robustly expressed by astrocytes in vivo.
Injury-Induced Reactive Astrocytes Upregulate SgIII Gene Expression
The functional role of granins in glial cells is unknown. To investigate potential roles of SgIII in astrocyte in vivo, we analyzed regulation of SgIII expression in the injured brain. In response to CNS insults, astrocytes become reactive and exert important roles in repair and damage of neuronal circuits (Sofroniew 2005). We examined responses of SgIII protein and mRNA expression in a model of traumatic brain injury induced by stab wounds (Mathewson and Berry 1985).
SgIII protein was examined in sections from brains obtained 2, 5, and 9 days postinjury. We observed an increase of SgIII immunoreactivity in the vicinity of the wound site in the injured hemisphere (Fig. 9A). This enhancement was noticed after 2 days and peaked at 9 days postinjury (data not shown). Upregulation of SgIII was restricted to hypertrophied astroglial-shaped cells (Fig. 9C,E). Glial SgIII labeling was strongest around the lesion and declined in intensity in proportion to the distance from the wound (Fig. 9A). The highest levels of wound-induced SgIII expression were located throughout the hippocampus, as well as in the white matter tracts, corpus callosum, and fornix (Fig. 9D,E). SgIII also increased in astroglia in the injured neocortex but to a lesser extent (Fig. 9B,C). Patterns of SgIII staining in the noninjured hemisphere resembled those observed in the sham-injured brain (Fig. 9B,D). However, in some cases, glial cells exhibiting upregulated levels of SgIII were observed in the hippocampus and white matter contralateral to the lesion (data not shown). A 95.8 ± 0.72% of the cells showing stabbing-induced upregulation of SgIII were also costained with GFAP (Fig. 9H–J). Astrocytes showing SgIII overexpression in the vicinity of the wound exhibited the main characteristics of reactive cells, such as enhanced content of GFAP, hypertrophy, and thickened processes (Fig. 9H). Reactive distinctiveness of astrocytes was more obvious in the hippocampus and white matter tract than in the neocortex (data not shown). To further confirm the activated state of SgIII+ glial cells, we performed double labeling with the reactive astrocyte marker vimentin. Astrocytes exhibiting upregulated levels of SgIII in the vicinity of the wound site were strongly immunostained for vimentin (Fig. 9G). In contrast, vimentin immunofluorescence was absent in normal SgIII+ astrocytes in the contralateral hemisphere (Fig. 9F). For comparative purposes, we examined immunostaining for SgII in damaged sections. In the hippocampus and neocortex, SgII staining was associated with varicose fibers and neuronal cell bodies but consistently not with GFAP+ astrocytes (Fig. 9K–M). We conclude that perforating brain injury differentially upregulates SgIII content in activated astrocytes.
To assess whether brain stabbing regulates SgIII expression at the mRNA level, we analyzed SgIII transcripts in brains obtained 2 and 4 days postinjury. Although upregulation changes in astrocyte SgIII transcripts were detected in the ipsilateral hemisphere 2 days after brain stabbing, a dramatic increase of SgIII mRNA hybridization signal was observed 4 days postinjury (Fig. 10A). Changes of glial SgIII mRNA levels in wounded brains mirrored those found for the protein. Thus, we detected an SgIII mRNA overexpression in glial cells around the insult (Fig. 10A). Glial cells showing the most intense SgIII hybridization signal were distributed throughout the hippocampus, corpus callosum, and the neocortex (Fig. 10A–C). Increased levels of SgIII transcripts in glial cells were maximal in the vicinity of the wound and decreased with the distance from the lesion (Fig. 10A). Conversely, SgIII mRNA expression was slightly decreased in neuronal cell bodies adjacent to the lesion (Fig. 10B,C). As expected from the parallel expression changes of SgIII protein and mRNA in glial cells of the damaged hemisphere, upregulated SgIII transcripts and protein were colocalized in 98.92 ± 0.39% of glial cells. Furthermore, GFAP immunostaining of SgIII-hybridized sections from wounded brains revealed that almost all (98.91 ± 0.33%) of the glial cells exhibiting SgIII upregulation were astrocytes (Fig. 10D–F). Taken together, these results show that brain injury markedly induces SgIII gene expression in reactive astrocytes.
In the present study, we show the expression, transcriptional regulation, trafficking, and release of the secretory pathway component SgIII in astroglial cells. Originally, SgIII was identified in CNS neurons and pituitary endocrine cells (Ottiger et al. 1990). More recent reports have analyzed the SgIII expression and function in different secretory cells, but the study of this granin in the brain has been poorly reported (Holthuis et al. 1996; Hosaka et al. 2002; Coppinger et al. 2004; Han et al. 2008; Prasad et al. 2008). We show here that SgIII is synthesized and transported along axons in many neuronal populations, especially in neuroendocrine nuclei and inner layers of the cerebral cortex. In addition to neurons, we reveal that high levels of SgIII mRNA and protein are expressed by astrocytes. These results are consistent with the in situ hybridization data provided by the Allen Brain Atlas indicating that SgIII mRNA is targeted in neurons and glial cells (Lein et al. 2007). By double-labeling analysis, we show that nonneuronal SgIII expression is restricted to astroglial cell populations and that the intensity of such expression varies. Thus, SgIII expression was high in cortical astrocytes and Bergmann glia but low in astrocytes located in white matter tracts. Typical components of the neurotransmitter secretory pathway, such as vesicular transporters and sorting/aggregation and trafficking/exocytic proteins have been shown in astroglial cells (Montana et al. 2006). However, most of these studies were performed on cultured or freshly isolated cells, whereas expression of the secretory pathway machinery in glial cells in situ is largely unknown. Our results implicate SgIII in the astroglial secretory pathway in vivo.
Regulation of the classical granins has been extensively analyzed in the brain under different conditions (Winkler and Fischer-Colbrie 1992; Fischer-Colbrie et al. 1995). CgA, CgB, and SgII/secretoneurin are distinctively regulated in neuronal subpopulations after brain insults but not in nonneuronal cells (Martí et al. 2001; Pirker et al. 2001). To our knowledge, the present study is the first to provide evidence of transcriptional control of a granin in glial cells in vivo. Both SgIII mRNA and protein levels were dramatically increased in reactive astrocytes after traumatic brain injury. Moreover, our in vitro and in vivo observations on SgIII and SgII indicate that granin expression in glial cells is differentially controlled. Coordinated regulation of SgIII and the hormone precursor proopiomelanocortin has been shown in Xenopus intermediate pituitary during color adaptation (Holthuis and Martens 1996). Based on this parallel transcriptional control, an active role of SgIII in the production and release of secretory peptides was initially suggested (Holthuis and Martens 1996). Reactive astrocytes, induced by different brain insults, overexpress a variety of secreted peptides (Ridet et al. 1997). For instance, expression of peptide hormones, growth factors, and cytokines, such as adrenomedullin, endothelin, normal growth factor, IL-6, and tumor necrosis factor α are upregulated in reactive astrocytes (Nie and Olsson 1996; Goss et al. 1998; Acarin et al. 2000; Jahnke et al. 2001). Therefore, we hypothesize an important functional involvement of SgIII in peptide trafficking in activated astrocytes in vivo.
We show here that the SgIII found in and released from astrocytes is larger than that detected in endocrine cells and neurons. A similar noncleaved form of SgIII has recently been identified in PC12 cells and adrenal glands (Han et al. 2008). Because the SgIII prohormone convertases PC1/3 and PC2 have not been detected in astrocytes (Winsky-Sommerer et al. 2000), glial cells may contain and release the unprocessed SgIII. In agreement, the prohormone convertase substrates SgII and proenkephalin were primarily detected as uncleaved forms in astrocytes (Batter et al. 1991; Fischer-Colbrie et al. 1993). Whether the noncleaved form of SgIII exerts a distinctive functional role or is processed extracellularly in vivo to peptides remains to be elucidated.
SgIII and its interacting protein CPE are essential sorting receptors for peptide hormones (Takeuchi and Hosaka 2008). Through its high cholesterol–binding domain, SgIII sorts adrenomedullin and proopiomelanocortin to the secretory granules (Hosaka et al. 2004; Han et al. 2008). Because astrocytes are equipped with molecular components of the secretory granules and can release peptide hormones in a Ca2+-dependent manner (Vilijn et al. 1989; Batter et al. 1991; Takahashi et al. 2000; Krzan et al. 2003; Ramamoorthy and Whim 2008; Paco et al. 2009), a function for SgIII in the regulated secretory pathway of glial cells is expected. However, the low cellular content of SgIII, together with its abundance in cell media, indicates that synthesized SgIII is poorly stored and promptly released from cultured astrocytes. Several mechanisms may explain the Ca2+-dependent release of peptides, accompanied by low SgIII retention in astrocytes. First, it has been demonstrated that a small quantity of SgIII is enough to sort a large number of peptides to secretory granules in PC12 cells (Han et al. 2008). CgA is considered a key aggregating factor involved in the biogenesis of secretory granules, which interacts with SgIII (Kim et al. 2001; Hosaka et al. 2002; Prasad et al. 2008). Thus, low expression of CgA in astrocytes could reflect the limited storage/aggregation of SgIII in glial cells (Majdoubi et al. 1996; Woulfe et al. 1999). Moreover, astrocytes display higher basal secretion and lower stimulus-triggered release of granule peptides than “professional” secretory cells (Fischer-Colbrie et al. 1993; Calegari et al. 1999; Paco et al. 2009). Finally, because astrocytes grown in culture do not accurately reflect their attributes in vivo, storage of peptides in astroglial cells may be higher in vivo than in culture. The present results and a recent microarray analysis show that SgIII mRNA expression and protein content in astrocytes are higher in vivo than in vitro (Cahoy et al. 2008; this study).
In addition to their intracellular functions, intact granins and their proteolytic-derived peptides exert important roles in cell-to-cell signaling, including homeostatic processes, inflammatory reactions, and the innate immunity (Helle 2004). For instance, CgA and the SgII fragment secretoneurin are potent microglia activators and chemotactic and angiogenic factors, respectively (Taupenot et al. 1996; Fischer-Colbrie et al. 2005). Our data clearly show that SgIII gene expression is upregulated in activated astroglia after perforating brain damage. We propose an extracellular role for glia-released SgIII in damage/repair of neuronal circuits after brain injury.
In summary, this study demonstrates that cortical astrocytes synthesize and release the secretory pathway component SgIII. Moreover, the robust expression of SgIII in astrocytes in vivo and its regulation in the damaged brain suggest important intracellular and/or extracellular roles for this granin in the glia–neuron communication.
Spanish Ministry of Education and Science Grants (BFU-2004-01154 and BFU2007-67889 to F.A.) and Carlos III Health Institute (PI070917 to E.P.).
We are grateful to Drs M. Hosaka (Gunma University) and R. Fisher-Colbrie (Innsbruck University) for FLAG-SgIII construct and SgII antibody, respectively; Dr E. Soriano (University of Barcelona) for helpful discussions; and Robin Rycroft for editorial assistance. Antibodies to β-tubulin were obtained from the Developmental Studies Hybridoma Bank at the University of Iowa. Conflict of Interest: None declared.