- Email: [email protected]
Expression of fructose-1,6-bisphosphatase mRNA isoforms in normal and basal forebrain cholinergic lesioned rat brain
Int. J. Devl Neuroscience 19 (2001) 279– 285 www.elsevier.nl/locate/ijdevneu
Expression of fructose-1,6-bisphosphatase mRNA isoforms in normal and basal forebrain cholinergic lesioned rat brain Thomas Lo¨ffler a, Samiya Al-Robaiy b, Marina Bigl b, Klaus Eschrich b, Reinhard Schliebs a,* a
Paul Flechsig Institute for Brain Research, Department of Neurochemistry, Jahnallee 59, D-04109 Leipzig, Germany b Institute of Biochemistry, Uni6ersity of Leipzig, Liebigstraße 16, D-04103 Leipzig, Germany
Abstract Fructose-1,6-bisphosphatase is one of the key enzymes in the gluconeogenic pathway predominantly occurring in liver, kidney and muscle. In the brain, fructose-1,6-bisphosphatase has been suggested to be an astrocyte-specific enzyme but the functional importance of glyconeogenesis in the brain is still unclear. To further elucidate the cellular source of fructose-1,6-bisphosphatase in the brain, non-radioactive in situ hybridizations were performed using digoxigenin-labeled RNA probes based on the sequence of recently cloned rat liver and muscle fructose-1,6-bisphosphatase cDNAs. In situ hybridization using a riboprobe for the liver isoform revealed a location of the hybridization signal mainly in neurons, while rat muscle fructose-1,6-bisphosphatase mRNA was detected in both neurons and astrocytes in the hippocampal formation and in layer I of the cerebral cortex. RT-PCR using RNA preparations of rat astrocytes, neurons, and adult whole brain demonstrated a localization of liver fructose-1,6-bisphosphatase mRNA isoform in neurons but not in astrocytes. The muscle fructose-1,6-bisphosphatase mRNA isoform could be detected by RT-PCR in total rat brain, astrocytic, and neuronal mRNA preparations. The isoforms of fructose-1,6-bisphosphatase mRNA seemingly demonstrate a distinct cellular expression pattern in rat brain suggesting a role of glyconeogenesis in both neurons and glial cells. © 2001 ISDN. Published by Elsevier Science Ltd. All rights reserved. Keywords: Gluconeogenesis; Glial fibrillary acidic protein; Rat brain; Riboprobes; In situ hybridization
1. Introduction Fructose-1,6-bisphosphatase (FBPase) is one of the key enzymes in the gluconeogenic pathway catalyzing the hydrolysis of fructose-1,6-bisphosphate into fructose-6-phosphate and inorganic phosphate. In mammals two isoforms of FBPase occur which are encoded by different genes and are denoted as liver and muscle FBPase according to the organs of preferential expression (Mizunuma and Tashima, 1986; Tillmann and Eschrich, 1998; Al-Robaiy and Eschrich, 1999). Gluconeogenesis has been described to occur in various organs including the liver, the kidney and the muscles. However, the brain was initially considered as an organ in which glucose synthesis pathways do not occur
* Corresponding author. Tel.: + 49-341-9725700; fax: +49-3412114492. E-mail address: [email protected] (R. Schliebs).
(Scrutton and Utter, 1968). Later, a number of studies convincingly demonstrated the presence of FBPase in the brain (Majumder and Eisenberg, 1977; Hevor and Gayet, 1978; Liu and Fonnum, 1988), but the functional importance of the gluco- or glyconeogenic pathway in the brain is still a matter of debate (Hevor, 1994). From cell culture studies and immunocytochemistry in brain slices it has been suggested that the FBPase in the brain is an astrocyte-specific enzyme playing a role in the metabolic interaction between neurons and glial cells (Hevor et al., 1986; Schmoll et al., 1995; Verge and Hevor, 1995). Analysis of steadystate mRNA levels by RT-PCR and Northern blotting revealed the presence of one or two FBPase mRNA isoforms in a number of rat tissues including brain (Al-Robaiy and Eschrich, 1999). Non-radioactive in situ hybridization histochemistry in combination with immunocytochemistry for astroglial marker, should allow to further elucidate the
0736-5748/01/$20 © 2001 ISDN. Published by Elsevier Science Ltd. All rights reserved. PII: S0736-5748(01)00011-9
280
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
cellular source of the expression of FBPase mRNA isoforms in rat brain. To demonstrate any functional role of gluco- or glyconeogenesis in the brain, experimental paradigms are required that may affect brain glucose metabolism. Previously, we have shown that basal forebrain cholinergic lesion produced by a single intracerebroventricular injection of the cholinergic immunotoxin 192IgG-saporin, leads to reactive astrogliosis within the lesion site, mostly pronounced 15 days following toxin injection (Lemke et al., 1999). As carbohydrate anabolism may significantly be different between various types of astrocytes (Verge et al., 1996), investigations of the FBPase mRNA expression around or within the cholinergic lesion site should allow to further elucidate functional roles of gluconeogenesis in the brain. The present report demonstrates the expression of the muscle FBPase mRNA in rat brain by both cortical neurons and hippocampal resting astrocytes, while reactive astrocytes induced in the basal forebrain of cholinergic lesioned rats do not express muscle FBPase mRNA. 2. Experimental procedures
2.1. Treatment of animals To selectively degenerate basal forebrain cholinergic neurons, adult male Wistar rats received under anaesthesia an intracerebroventricular injection of 4 mg of the cholinergic immunotoxin 192IgG-saporin (Chemicon International, Temecula, CA, USA) in 7 ml of phosphate buffered saline (PBS, pH 7.4) at a rate of 1.0 ml/min using a Hamilton syringe, as already described previously in more detail (Roßner et al., 1995). Control rats received injections of the vehicle solution only. Animals were analyzed 14 days post surgery. The animal experiments performed in this study have been approved by the Independent Ethical Committee for Animal Experiments of the Regierungspra¨ sidium Leipzig, licence no. TVV 8/97
2.2. Tissue preparation For in situ hybridization, adult Wistar rats were transcardially perfused under deep anaesthesia with saline, followed by 200 ml of fixative (4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4) using a peristaltic pump. Thereafter, the brains were removed from the skull and postfixed in the same fixative overnight at 4°C. After cryoprotection, they were cut on a microtome at 30 mm thickness. For RNA preparations, adult Wistar rats were sacrificed, the brain was rapidly dissected out of the skull and frozen in liquid nitrogen. The tissue was stored at − 80°C until RNA preparation.
Astrocytes were cultured from cerebral cortex of newborn rats with the method described by Hertz et al. (1989), while neuron cultures have been obtained from 15- to 17-day-old rat embryos with the procedure of Hansson and Ro¨ nnba¨ ck (1989). The cells were harvested after 3 weeks using 1× trypsin-EDTA solution (Gibco BRL, Grand Island, USA).
2.3. In situ hybridization of FBPase mRNA Cloned rat muscle and liver FBPase cDNAs were used as templates to generate digoxigenin-labeled cRNA probes by in vitro transcription. The plasmids are pCR-Script SK(+) (Stratagene, La Jolla, CA, USA) containing either an insert of 343 bp including the region 834– 1177 of rat muscle FBPase cDNA or a 422 bp DNA fragment including the region 744–1166 of rat liver FBPase cDNA (when position + 1 is the A of the start codon). Sections were acetylated and pre-hybridized for 6 h at room temperature. For hybridization, sections were covered with the hybridization solution containing 50% formamide, 5× Denhardt’s solution, 5× SSC (150 mM NaCl, 15 mM sodium citrate, pH 7.0), 250 mg/ml yeast tRNA, 500 mg/ml herring sperm DNA and the labeled cRNA probe, hybridized overnight in a humid chamber at 37°C for about 16 h. Following hybridization, the slides were subjected to a high-stringency washing in 0.2× SSC at 72°C for 2 h, followed by rinses at room temperature. The detection of digoxigenin-labeled hybridization products was performed using an anti-digoxigenin antibody conjugated to alkaline phosphatase. As substrate, 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium salt were used. FBPase mRNA positive cells were labeled dark blue. Hybridization with digoxigenin-labeled sense cRNA was used as control. Following in situ hybridization sections were subjected to immunocytochemistry for glial fibrillary acidic protein (GFAP) to visualize astrocytes. Sections were incubated with a rabbit antiserum raised against GFAP from cow brain (Dako, 1:2000) for 14–16 h at 4°C. GFAP positive cells were visualized with the streptavidin/biotin-technique described elsewhere (Bru¨ ckner et al., 1992). Slides were coverslipped, sealed, and analyzed on a phase-contrast microscope (Zeiss).
2.4. RNA preparation and synthesis of the first strand cDNA Total RNA was isolated from 100 mg of rat brain tissue according to Chomczynski and Sacchi (1987) using Trizol (Gibco BRL, Grand Island, USA). Total RNA from astrocytes and neurons was prepared from the cell pellet obtained from one plate using the method described above.
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
2.5. RT-PCR One microgram of total RNA was reverse transcribed with the 1st Strand cDNA Synthesis Kit For RT-PCR (AMV, Roche Diagnostics, Mannheim, Germany) in a final volume of 20 ml using random primer p(dN)6. For the PCR 2 ml of the obtained cDNAs was added to 48 ml of the reaction mixture containing 1× buffer, 1.5 mM MgCl2, 20 pmol of each primer, 0.2 mM of each deoxyribonucleoside triphosphate and 2.5 U Taq polymerase (TaKaRa Biochemicals, Shiga, Japan). Primers used for the amplification are listed in Table 1. For amplification 35 cycles of 94°C for 55 s, 53°C for 45 s and 72°C for 70 s, preceded by 3 min denaturation at 94°C and followed by 3 min extension at 72°C, were used. The PCR products were analyzed on a 2% agarose gel.
3. Results The presence and anatomical distribution of muscle and liver FBPase mRNA isoforms in the rat brain was evaluated by in situ hybridization using digoxigenin-labeled riboprobes. High expression of muscle FBPase mRNA was detected in the hippocampal formation, in upper layers of the cerebral cortex and in the piriform cortex, while a less intense labeling was observed in subcortical and thalamic brain regions. The liver FBPase mRNA demonstrated a similar regional distribution pattern (data not shown). In order to identify those cell populations expressing FBPase mRNA isoforms, in situ hybridization in combination with immunocytochemistry for GFAP to label astrocytes was performed. Representative examples of in situ hybridization to label muscle FBPase mRNA in sections of normal rat brain are shown in Fig. 1. Hybridization with the antisense probe for muscle FBPase mRNA revealed very intense staining in the Table 1 Primer pairs used for the amplification of rat liver (A) and muscle (B) specific FBPase cDNA fragments from the reverse transcribed total RNA of rat brain, cultured astrocytes, and neurons Name
A prRlFo1 prRlRe1 B
prRMFDPFo prRMFDPRe
Sequence
Product length (bp)
CAC CTG CCT GCA CCT TTA GC CAG TTG ACG CCA CAA TTC A
608
GAAGTTTCACTTGCCAC AATG CCTCATCCCCTGTCACA TTC
200
281
hippocampal formation, cerebral cortex and basal forebrain nuclei, whereas hybridization with the sense probe in adjacent sections does not produce any hybridization signals. The pyramidal cells, hilar interneurons, and granular cells of the dentate gyrus display very strong hybridization signals. An analysis of these hippocampal sections at higher magnification discloses that muscle FBPase mRNA is expressed in both neurons and astrocytes. In order to examine this in more detail, in situ hybridization for muscle FBPase mRNA was combined with immunocytochemistry for GFAP, a specific marker for astrocytes, revealing a location of the hybridization signal in both neurons and astrocytes (Fig. 1C, E). A neuronal location of muscle FBPase mRNA was observed in the hippocampus, cerebral cortex, and basal forebrain nuclei, while labeled astrocytes were found in the hippocampal formation and in layer I of the cerebral cortex only (Fig. 1A, C). Liver FBPase mRNA was exclusively expressed in neurons of rat cerebral cortex and hippocampal formation but not in astrocytes, as representatively displayed in Fig. 2. In order to further determine whether the FBPase mRNA isoforms are distinctly expressed by glial and/or neuronal cells, RT-PCR was performed in total RNA isolated from astrocytic and neuronal primary cell cultures as well from the total brain of adult rats using primer pairs specific for muscle and liver FBPase, respectively. As shown in Fig. 3 the RT-PCR experiments demonstrated the localization of liver FBPase mRNA in neurons but not in astrocytes. On the other hand, muscle FBPase mRNA could be detected by RT-PCR in total rat brain, astrocytes, and neurons. In order to reveal whether reactive astrocytes demonstrate altered FBPase expression, a specific cholinergic lesion by intracerebroventricular administration of the cholinergic immunotoxin 192IgG-saporin, was placed in rat basal forebrain, an experimental approach known to produce reactive astrocytes within the lesion site mostly pronounced 15 days after toxin application (Lemke et al., 1999). In situ hybridization to label muscle FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in medial septal sections from rats 15 days after cholinergic basal forebrain lesion by intracerebroventricular 192IgG-saporin demonstrated the presence of reactive astrocytes in the lesion site but they did not express muscle FBPase mRNA (Fig. 4).
4. Discussion The presence of FBPase in the brain has been known for a long time (for historical review, see Hevor, 1994). Cell culture and immunocytochemical studies provided evidence that FBPase represents an astrocyte-specific
282
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
Fig. 1. Representative examples of in situ hybridization to label muscle FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in sections of normal rat brain containing the cerebral cortex (A), medial septum (C), and CA1 subfield of hippocampal formation (E) using digoxygenin-labeled riboprobes. Hybridization with digoxigenin-labeled sense riboprobes in adjacent sections (B, D, F) did not result in any hybridization signals. Magnification: 70 × .
enzyme (Hevor et al., 1986; Schmoll et al., 1995; Verge and Hevor, 1995), indicating neuronal glial interactions with respect to energy metabolism. Recently, the existence of three different FBPase transcripts that might correspond to three genes has been suggested (Cloix et al., 1997). However, the brain-specific fragment was shown to represent an artifact (Stein et al., 2001). Investigations revealing the cellular source of FBPase mRNA expression in the in vivo brain are still lacking. Based on the sequence of recently cloned rat liver and muscle FBPase cDNAs (Al-Robaiy and Eschrich, 1999), we have synthesized digoxigenin-labeled riboprobes that were used to perform in situ hybridization histochemistry in rat brain sections to achieve a cellular
resolution of the FBPase mRNA hybridization signal. Moreover, the hybridization protocol can easily be combined with immunocytochemistry to stain for cellular markers such as the astrocyte-specific glial fibrillary acidic protein (GFAP). The combination of both techniques has demonstrated that a major part of the hippocampal astrocytes as well as astrocytes localized in layer I of the cerebral cortex express muscle FBPase mRNA. We further observed a predominant neuronal expression of the muscle FBPase in most brain regions in the normal, adult rat with higher expression in the cerebral cortex, hippocampus, and basal forebrain nuclei but to a less extent in thalamic and subcortical regions. The expression pattern of the liver FBPase was
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
283
Fig. 2. Representative examples of in situ hybridization to label liver FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in sections of normal rat brain containing the parietal cortex (A), medial septum (B), and CA3 subfield of hippocampal formation (C) using digoxigenin-labeled riboprobes. Hybridization with the digoxigenin-labeled sense riboprobe in adjacent cortical section (D) did not result in any hybridization signals.
similar to that of the muscle FBPase, but contrasting to muscle FBPase, there was no expression of liver FBPase by astrocytes. The findings by in situ hybridization were confirmed by RT-PCR with RNA preparations from total brain and astrocytic as well as neuronal cell populations. Liver FBPase mRNA was detected in RNA preparations of the total brain and in neurons. Muscle FBPase was detectable in RNA preparations from the total brain, astrocytic and neuronal cells. Carbohydrate anabolic difference between various types of astrocytes as observed in cultured astrocytes (Verge et al., 1996), may explain the regionally distinct astrocytic expression pattern of the muscle FBPase detected by in situ hybridization. Our observations that the expression of muscle FBPase is restricted to astrocytes localized in the hippocampal formation and layer I of the cerebral cortex suggest a functional heterogeneity of brain astrocytes with respect to gluco- or glyconeogenesis. Pathological situations in the brain such as neurodegenerative events may certainly affect the cellular energy metabolismas recently described in a cholinergic lesion paradigm (Mehlhorn et al., 1998). A single intracerebroventricular injection of the cholinergic immunotoxin 192IgG-saporin results in selective and specific loss of basal forebrain cholinergic cells in rat basal forebrain that is accompanied by a massive mi-
cro-and astrogliosis (Schliebs et al., 1996). In situ hybridization in brain sections from cholinergic lesioned rats did not reveal any muscle FBPase RNA in reactive astrocytes present in the lesion site. This finding may suggest that in reactive astrocytes the glycolytic cascade predominates presumably due to the high energy demand. While the astrocytic expression pattern of muscle FBPase in rat brain detected in the present report is
Fig. 3. Demonstration of FBPase mRNA isoforms by RT-PCR in rat astrocytes, neurons and brain. RT-PCR was performed using total RNA of rat astrocytes (lane 1), neurons (lane 2) and brain (lane 3) with primers specific for rat liver FBPase cDNA (A) and rat muscle FBPase cDNA (B). PCR controls (lane 4) were performed with the same primers but without cDNA addition. M represents the 123 bp DNA marker.
284
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
Fig. 4. Representative examples of in situ hybridization to label muscle FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in coronal sections through the medial septum from untreated control (A, C) and experimental rats 15 days after cholinergic basal forebrain lesion by intracerebroventricular 192IgG-saporin (B, D) at different magnifications (A, B: 220 × ; C, D: 140× ). Hybridization signals are indicated by blue precipitates, while GFAP-immunoreactive astrocytes are labeled by brown staining (arrows). Note the presence of reactive astrocytes in basal forebrain sections from cholinergic lesioned rats (B) that do not express muscle FBPase mRNA.
partly consistent with the immunocytochemical staining pattern for the FBPase enzyme using a poloclonal antiserum that did not differentiate between isoforms (Schmoll et al., 1995), the neuronal expression pattern observed here for both FBPase mRNA isoforms is in contradiction to the immunocytochemical observations by Schmoll et al. (1995) who did not find FBPase-immunoreactive neurons in rat brain. It could be speculated that muscle FBPase mRNA is present at low concentrations in both neurons and astrocytes in normal rat brain, but the corresponding gene product is preferentially produced in astrocytes that are gluco- or glyconeogenetically active. The presence of FBPase mRNA in neurons may exhibit a reservoir to rapidly induce FBPase enzyme production under extreme metabolic situations such as ischaemia or lack of energy substrates. However, regardless of any speculations, our data clearly demonstrate the expression of FBPase mRNA isoforms in the normal adult brain and a localization in both neurons and astrocytes.
Acknowledgements This work was supported by grants from the Interdisziplina¨ res Zentrum fu¨ r Klinische Forschung (IZKF) at the University of Leipzig, 01KS9504, project C6. We
thank Dr. M. Bru¨ ckner, Paul-Flechsig-Institute for Brain Research, University of Leipzig, for culturing of astrocytes and neurons. T.L. gratefully acknowledges the receipt of a stipend from the Graduiertenkolleg fu¨ r Neurowissenschaften ‘INTERCELL’ at the Medical Faculty of the University of Leipzig.
References Al-Robaiy, S., Eschrich, K., 1999. Rat muscle fructose-1,6-bisphosphatase: cloning of the cDNA, expression of the recombinant enzyme, and expression analysis in different tissues. Biol. Chem. 380, 1079 – 1085. Bru¨ ckner, G., Schober, W., Ha¨ rtig, W., Osterman-Latif, C., Webster, H.H., Dykes, R.W., Rasmusson, D.D., Biesold, D., 1992. The basal forebrain cholinergic system in the raccoon. J. Chem. Neuroanat. 5, 441 – 452. Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by guanidinium thiocyanat-phenol-chloroform extraction. Anal. Biochem. 162, 156 – 159. Cloix, J.-F., Beaulieu, E., Hevor, T., 1997. Various fructose-1,6-bisphosphatase mRNAs in mouse brain, liver, kidney and heart. Neuroreport 8, 617 – 622. Hansson, E., Ro¨ nnba¨ ck, L., 1989. Primary cultures of astroglia and neurons from different brain regions. In: A Dissection and Tissue Culture Manual of the Nervous System. Alan R. Liss, New York, pp. 92– 104. Hertz, L., Juurlink, B.H.J., Hertz, E., Fosmark, H., Schousboe, A., 1989. Preparation of primary cultures of mouse (rat) astrocytes.
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285 In: A Dissection and Tissue Culture Manual of the Nervous System. Alan R. Liss, New York, pp. 105 –108. Hevor, T.K., 1994. Some aspects of carbohydrate metabolism in the brain. Biochimie 76, 111 –120. Hevor, T.K., Gayet, J., 1978. Fructose-1,6-diphosphatase and phosphofructokinase activities in the brain of mice submitted to methionine sulfoxime. Brain Res. 150, 210 –215. Hevor, T.K., Delorme, P., Beauvillain, J.C., 1986. Glycogen synthesis and immunocytochemical study of fructose-1,6-bisphosphatase in methionine sulfoxime epileptogenic rodent brain. J. Cerebr. Blood Flow Metab. 6, 292 –297. Lemke, R., Hartlage-Ru¨ bsamen, M., Schliebs, R., 1999. Differential injury-dependent glial expression of interleukins-1a,b and interleukin-6 in rat brain. Glia 27, 75 –87. Liu, F., Fonnum, H.J., 1988. Purification and characterization of fructose-1,6-bisphosphatase from bovine brain. Arch. Biochem. Biophys. 260, 609 – 615. Majumder, A.L., Eisenberg, F., 1977. Unequivocal demonstration of fructose-1,6-bisphosphatase in mammalian brain. Proc. Natl. Acad. Sci. USA 74, 3222 –3225. Mehlhorn, G., Lo¨ ffler, T., Apelt, J., Roßner, S., Urabe, T., Hattori, H., Nagamatsu, S., Bigl, V., Schliebs, R., 1998. Glucose metabolism in cholinoceptive cortical rat brain regions after basal forebrain cholinergic lesion. Int. J. Devl. Neurosci. 16, 675 – 690. Mizunuma, H., Tashima, Y., 1986. Characterization of rat muscle fructose-1,6-bisphosphatase. J. Biochem. (Tokyo) 99, 1781 – 1788.
285
Roßner, S., Ha¨ rtig, W., Schliebs, R., Bru¨ ckner, G., Brauer, K., Perez-Polo, J.R., Wiley, R.G., Bigl, V., 1995. 192IgG-saporin immunotoxin-induced loss of cholinergic cells differentially activates microglia in rat basal forebrain nuclei. J. Neurosci. Res. 41, 335 – 346. Schliebs, R., Roßner, S., Bigl, V., 1996. Immunolesion by 192IgG-saporin of rat basal forebrain cholinergic system — a useful tool to produce cortical cholinergic dysfunction. Progr. Brain Res. 109, 253 – 264. Schmoll, D., Cesar, M., Fuhrmann, E., Hamprecht, B., 1995. Colocalization of fructose-1,6-bisphosphatase and glial fibrillary acidic protein in rat brain. Brain Res. 677, 341 – 344. Scrutton, M.C., Utter, M.F., 1968. The regulation of glycolysis and gluconeogenesis in animal tissues. Annu. Rev. Biochem. 37, 249 – 302. Stein, S., Liehr, Th., Eschrich, K., 2001. Characterization of the mouse fructose-1,6-bisphosphatase gene. Gene (in press). Tillmann, H., Eschrich, K., 1998. Isolation and characterization of an allelic cDNA for human muscle fructose-1,6-bisphosphatase. Gene 212, 295 – 304. Verge, V., Hevor, T.K., 1995. Regulation of fructose-1,6-bisphosphatase activity in primary cultured astrocytes. Neurochem. Res. 20, 1049 – 1056. Verge, V., Legrand, A., Hevor, T., 1996. Isolation of carbohydrate metabolic clones from cultured astrocytes. Glia 18, 244 – 254.
. .
Expression of fructose-1,6-bisphosphatase mRNA isoforms in normal and basal forebrain cholinergic lesioned rat brain Thomas Lo¨ffler a, Samiya Al-Robaiy b, Marina Bigl b, Klaus Eschrich b, Reinhard Schliebs a,* a
Paul Flechsig Institute for Brain Research, Department of Neurochemistry, Jahnallee 59, D-04109 Leipzig, Germany b Institute of Biochemistry, Uni6ersity of Leipzig, Liebigstraße 16, D-04103 Leipzig, Germany
Abstract Fructose-1,6-bisphosphatase is one of the key enzymes in the gluconeogenic pathway predominantly occurring in liver, kidney and muscle. In the brain, fructose-1,6-bisphosphatase has been suggested to be an astrocyte-specific enzyme but the functional importance of glyconeogenesis in the brain is still unclear. To further elucidate the cellular source of fructose-1,6-bisphosphatase in the brain, non-radioactive in situ hybridizations were performed using digoxigenin-labeled RNA probes based on the sequence of recently cloned rat liver and muscle fructose-1,6-bisphosphatase cDNAs. In situ hybridization using a riboprobe for the liver isoform revealed a location of the hybridization signal mainly in neurons, while rat muscle fructose-1,6-bisphosphatase mRNA was detected in both neurons and astrocytes in the hippocampal formation and in layer I of the cerebral cortex. RT-PCR using RNA preparations of rat astrocytes, neurons, and adult whole brain demonstrated a localization of liver fructose-1,6-bisphosphatase mRNA isoform in neurons but not in astrocytes. The muscle fructose-1,6-bisphosphatase mRNA isoform could be detected by RT-PCR in total rat brain, astrocytic, and neuronal mRNA preparations. The isoforms of fructose-1,6-bisphosphatase mRNA seemingly demonstrate a distinct cellular expression pattern in rat brain suggesting a role of glyconeogenesis in both neurons and glial cells. © 2001 ISDN. Published by Elsevier Science Ltd. All rights reserved. Keywords: Gluconeogenesis; Glial fibrillary acidic protein; Rat brain; Riboprobes; In situ hybridization
1. Introduction Fructose-1,6-bisphosphatase (FBPase) is one of the key enzymes in the gluconeogenic pathway catalyzing the hydrolysis of fructose-1,6-bisphosphate into fructose-6-phosphate and inorganic phosphate. In mammals two isoforms of FBPase occur which are encoded by different genes and are denoted as liver and muscle FBPase according to the organs of preferential expression (Mizunuma and Tashima, 1986; Tillmann and Eschrich, 1998; Al-Robaiy and Eschrich, 1999). Gluconeogenesis has been described to occur in various organs including the liver, the kidney and the muscles. However, the brain was initially considered as an organ in which glucose synthesis pathways do not occur
* Corresponding author. Tel.: + 49-341-9725700; fax: +49-3412114492. E-mail address: [email protected] (R. Schliebs).
(Scrutton and Utter, 1968). Later, a number of studies convincingly demonstrated the presence of FBPase in the brain (Majumder and Eisenberg, 1977; Hevor and Gayet, 1978; Liu and Fonnum, 1988), but the functional importance of the gluco- or glyconeogenic pathway in the brain is still a matter of debate (Hevor, 1994). From cell culture studies and immunocytochemistry in brain slices it has been suggested that the FBPase in the brain is an astrocyte-specific enzyme playing a role in the metabolic interaction between neurons and glial cells (Hevor et al., 1986; Schmoll et al., 1995; Verge and Hevor, 1995). Analysis of steadystate mRNA levels by RT-PCR and Northern blotting revealed the presence of one or two FBPase mRNA isoforms in a number of rat tissues including brain (Al-Robaiy and Eschrich, 1999). Non-radioactive in situ hybridization histochemistry in combination with immunocytochemistry for astroglial marker, should allow to further elucidate the
0736-5748/01/$20 © 2001 ISDN. Published by Elsevier Science Ltd. All rights reserved. PII: S0736-5748(01)00011-9
280
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
cellular source of the expression of FBPase mRNA isoforms in rat brain. To demonstrate any functional role of gluco- or glyconeogenesis in the brain, experimental paradigms are required that may affect brain glucose metabolism. Previously, we have shown that basal forebrain cholinergic lesion produced by a single intracerebroventricular injection of the cholinergic immunotoxin 192IgG-saporin, leads to reactive astrogliosis within the lesion site, mostly pronounced 15 days following toxin injection (Lemke et al., 1999). As carbohydrate anabolism may significantly be different between various types of astrocytes (Verge et al., 1996), investigations of the FBPase mRNA expression around or within the cholinergic lesion site should allow to further elucidate functional roles of gluconeogenesis in the brain. The present report demonstrates the expression of the muscle FBPase mRNA in rat brain by both cortical neurons and hippocampal resting astrocytes, while reactive astrocytes induced in the basal forebrain of cholinergic lesioned rats do not express muscle FBPase mRNA. 2. Experimental procedures
2.1. Treatment of animals To selectively degenerate basal forebrain cholinergic neurons, adult male Wistar rats received under anaesthesia an intracerebroventricular injection of 4 mg of the cholinergic immunotoxin 192IgG-saporin (Chemicon International, Temecula, CA, USA) in 7 ml of phosphate buffered saline (PBS, pH 7.4) at a rate of 1.0 ml/min using a Hamilton syringe, as already described previously in more detail (Roßner et al., 1995). Control rats received injections of the vehicle solution only. Animals were analyzed 14 days post surgery. The animal experiments performed in this study have been approved by the Independent Ethical Committee for Animal Experiments of the Regierungspra¨ sidium Leipzig, licence no. TVV 8/97
2.2. Tissue preparation For in situ hybridization, adult Wistar rats were transcardially perfused under deep anaesthesia with saline, followed by 200 ml of fixative (4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4) using a peristaltic pump. Thereafter, the brains were removed from the skull and postfixed in the same fixative overnight at 4°C. After cryoprotection, they were cut on a microtome at 30 mm thickness. For RNA preparations, adult Wistar rats were sacrificed, the brain was rapidly dissected out of the skull and frozen in liquid nitrogen. The tissue was stored at − 80°C until RNA preparation.
Astrocytes were cultured from cerebral cortex of newborn rats with the method described by Hertz et al. (1989), while neuron cultures have been obtained from 15- to 17-day-old rat embryos with the procedure of Hansson and Ro¨ nnba¨ ck (1989). The cells were harvested after 3 weeks using 1× trypsin-EDTA solution (Gibco BRL, Grand Island, USA).
2.3. In situ hybridization of FBPase mRNA Cloned rat muscle and liver FBPase cDNAs were used as templates to generate digoxigenin-labeled cRNA probes by in vitro transcription. The plasmids are pCR-Script SK(+) (Stratagene, La Jolla, CA, USA) containing either an insert of 343 bp including the region 834– 1177 of rat muscle FBPase cDNA or a 422 bp DNA fragment including the region 744–1166 of rat liver FBPase cDNA (when position + 1 is the A of the start codon). Sections were acetylated and pre-hybridized for 6 h at room temperature. For hybridization, sections were covered with the hybridization solution containing 50% formamide, 5× Denhardt’s solution, 5× SSC (150 mM NaCl, 15 mM sodium citrate, pH 7.0), 250 mg/ml yeast tRNA, 500 mg/ml herring sperm DNA and the labeled cRNA probe, hybridized overnight in a humid chamber at 37°C for about 16 h. Following hybridization, the slides were subjected to a high-stringency washing in 0.2× SSC at 72°C for 2 h, followed by rinses at room temperature. The detection of digoxigenin-labeled hybridization products was performed using an anti-digoxigenin antibody conjugated to alkaline phosphatase. As substrate, 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium salt were used. FBPase mRNA positive cells were labeled dark blue. Hybridization with digoxigenin-labeled sense cRNA was used as control. Following in situ hybridization sections were subjected to immunocytochemistry for glial fibrillary acidic protein (GFAP) to visualize astrocytes. Sections were incubated with a rabbit antiserum raised against GFAP from cow brain (Dako, 1:2000) for 14–16 h at 4°C. GFAP positive cells were visualized with the streptavidin/biotin-technique described elsewhere (Bru¨ ckner et al., 1992). Slides were coverslipped, sealed, and analyzed on a phase-contrast microscope (Zeiss).
2.4. RNA preparation and synthesis of the first strand cDNA Total RNA was isolated from 100 mg of rat brain tissue according to Chomczynski and Sacchi (1987) using Trizol (Gibco BRL, Grand Island, USA). Total RNA from astrocytes and neurons was prepared from the cell pellet obtained from one plate using the method described above.
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
2.5. RT-PCR One microgram of total RNA was reverse transcribed with the 1st Strand cDNA Synthesis Kit For RT-PCR (AMV, Roche Diagnostics, Mannheim, Germany) in a final volume of 20 ml using random primer p(dN)6. For the PCR 2 ml of the obtained cDNAs was added to 48 ml of the reaction mixture containing 1× buffer, 1.5 mM MgCl2, 20 pmol of each primer, 0.2 mM of each deoxyribonucleoside triphosphate and 2.5 U Taq polymerase (TaKaRa Biochemicals, Shiga, Japan). Primers used for the amplification are listed in Table 1. For amplification 35 cycles of 94°C for 55 s, 53°C for 45 s and 72°C for 70 s, preceded by 3 min denaturation at 94°C and followed by 3 min extension at 72°C, were used. The PCR products were analyzed on a 2% agarose gel.
3. Results The presence and anatomical distribution of muscle and liver FBPase mRNA isoforms in the rat brain was evaluated by in situ hybridization using digoxigenin-labeled riboprobes. High expression of muscle FBPase mRNA was detected in the hippocampal formation, in upper layers of the cerebral cortex and in the piriform cortex, while a less intense labeling was observed in subcortical and thalamic brain regions. The liver FBPase mRNA demonstrated a similar regional distribution pattern (data not shown). In order to identify those cell populations expressing FBPase mRNA isoforms, in situ hybridization in combination with immunocytochemistry for GFAP to label astrocytes was performed. Representative examples of in situ hybridization to label muscle FBPase mRNA in sections of normal rat brain are shown in Fig. 1. Hybridization with the antisense probe for muscle FBPase mRNA revealed very intense staining in the Table 1 Primer pairs used for the amplification of rat liver (A) and muscle (B) specific FBPase cDNA fragments from the reverse transcribed total RNA of rat brain, cultured astrocytes, and neurons Name
A prRlFo1 prRlRe1 B
prRMFDPFo prRMFDPRe
Sequence
Product length (bp)
CAC CTG CCT GCA CCT TTA GC CAG TTG ACG CCA CAA TTC A
608
GAAGTTTCACTTGCCAC AATG CCTCATCCCCTGTCACA TTC
200
281
hippocampal formation, cerebral cortex and basal forebrain nuclei, whereas hybridization with the sense probe in adjacent sections does not produce any hybridization signals. The pyramidal cells, hilar interneurons, and granular cells of the dentate gyrus display very strong hybridization signals. An analysis of these hippocampal sections at higher magnification discloses that muscle FBPase mRNA is expressed in both neurons and astrocytes. In order to examine this in more detail, in situ hybridization for muscle FBPase mRNA was combined with immunocytochemistry for GFAP, a specific marker for astrocytes, revealing a location of the hybridization signal in both neurons and astrocytes (Fig. 1C, E). A neuronal location of muscle FBPase mRNA was observed in the hippocampus, cerebral cortex, and basal forebrain nuclei, while labeled astrocytes were found in the hippocampal formation and in layer I of the cerebral cortex only (Fig. 1A, C). Liver FBPase mRNA was exclusively expressed in neurons of rat cerebral cortex and hippocampal formation but not in astrocytes, as representatively displayed in Fig. 2. In order to further determine whether the FBPase mRNA isoforms are distinctly expressed by glial and/or neuronal cells, RT-PCR was performed in total RNA isolated from astrocytic and neuronal primary cell cultures as well from the total brain of adult rats using primer pairs specific for muscle and liver FBPase, respectively. As shown in Fig. 3 the RT-PCR experiments demonstrated the localization of liver FBPase mRNA in neurons but not in astrocytes. On the other hand, muscle FBPase mRNA could be detected by RT-PCR in total rat brain, astrocytes, and neurons. In order to reveal whether reactive astrocytes demonstrate altered FBPase expression, a specific cholinergic lesion by intracerebroventricular administration of the cholinergic immunotoxin 192IgG-saporin, was placed in rat basal forebrain, an experimental approach known to produce reactive astrocytes within the lesion site mostly pronounced 15 days after toxin application (Lemke et al., 1999). In situ hybridization to label muscle FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in medial septal sections from rats 15 days after cholinergic basal forebrain lesion by intracerebroventricular 192IgG-saporin demonstrated the presence of reactive astrocytes in the lesion site but they did not express muscle FBPase mRNA (Fig. 4).
4. Discussion The presence of FBPase in the brain has been known for a long time (for historical review, see Hevor, 1994). Cell culture and immunocytochemical studies provided evidence that FBPase represents an astrocyte-specific
282
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
Fig. 1. Representative examples of in situ hybridization to label muscle FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in sections of normal rat brain containing the cerebral cortex (A), medial septum (C), and CA1 subfield of hippocampal formation (E) using digoxygenin-labeled riboprobes. Hybridization with digoxigenin-labeled sense riboprobes in adjacent sections (B, D, F) did not result in any hybridization signals. Magnification: 70 × .
enzyme (Hevor et al., 1986; Schmoll et al., 1995; Verge and Hevor, 1995), indicating neuronal glial interactions with respect to energy metabolism. Recently, the existence of three different FBPase transcripts that might correspond to three genes has been suggested (Cloix et al., 1997). However, the brain-specific fragment was shown to represent an artifact (Stein et al., 2001). Investigations revealing the cellular source of FBPase mRNA expression in the in vivo brain are still lacking. Based on the sequence of recently cloned rat liver and muscle FBPase cDNAs (Al-Robaiy and Eschrich, 1999), we have synthesized digoxigenin-labeled riboprobes that were used to perform in situ hybridization histochemistry in rat brain sections to achieve a cellular
resolution of the FBPase mRNA hybridization signal. Moreover, the hybridization protocol can easily be combined with immunocytochemistry to stain for cellular markers such as the astrocyte-specific glial fibrillary acidic protein (GFAP). The combination of both techniques has demonstrated that a major part of the hippocampal astrocytes as well as astrocytes localized in layer I of the cerebral cortex express muscle FBPase mRNA. We further observed a predominant neuronal expression of the muscle FBPase in most brain regions in the normal, adult rat with higher expression in the cerebral cortex, hippocampus, and basal forebrain nuclei but to a less extent in thalamic and subcortical regions. The expression pattern of the liver FBPase was
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
283
Fig. 2. Representative examples of in situ hybridization to label liver FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in sections of normal rat brain containing the parietal cortex (A), medial septum (B), and CA3 subfield of hippocampal formation (C) using digoxigenin-labeled riboprobes. Hybridization with the digoxigenin-labeled sense riboprobe in adjacent cortical section (D) did not result in any hybridization signals.
similar to that of the muscle FBPase, but contrasting to muscle FBPase, there was no expression of liver FBPase by astrocytes. The findings by in situ hybridization were confirmed by RT-PCR with RNA preparations from total brain and astrocytic as well as neuronal cell populations. Liver FBPase mRNA was detected in RNA preparations of the total brain and in neurons. Muscle FBPase was detectable in RNA preparations from the total brain, astrocytic and neuronal cells. Carbohydrate anabolic difference between various types of astrocytes as observed in cultured astrocytes (Verge et al., 1996), may explain the regionally distinct astrocytic expression pattern of the muscle FBPase detected by in situ hybridization. Our observations that the expression of muscle FBPase is restricted to astrocytes localized in the hippocampal formation and layer I of the cerebral cortex suggest a functional heterogeneity of brain astrocytes with respect to gluco- or glyconeogenesis. Pathological situations in the brain such as neurodegenerative events may certainly affect the cellular energy metabolismas recently described in a cholinergic lesion paradigm (Mehlhorn et al., 1998). A single intracerebroventricular injection of the cholinergic immunotoxin 192IgG-saporin results in selective and specific loss of basal forebrain cholinergic cells in rat basal forebrain that is accompanied by a massive mi-
cro-and astrogliosis (Schliebs et al., 1996). In situ hybridization in brain sections from cholinergic lesioned rats did not reveal any muscle FBPase RNA in reactive astrocytes present in the lesion site. This finding may suggest that in reactive astrocytes the glycolytic cascade predominates presumably due to the high energy demand. While the astrocytic expression pattern of muscle FBPase in rat brain detected in the present report is
Fig. 3. Demonstration of FBPase mRNA isoforms by RT-PCR in rat astrocytes, neurons and brain. RT-PCR was performed using total RNA of rat astrocytes (lane 1), neurons (lane 2) and brain (lane 3) with primers specific for rat liver FBPase cDNA (A) and rat muscle FBPase cDNA (B). PCR controls (lane 4) were performed with the same primers but without cDNA addition. M represents the 123 bp DNA marker.
284
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285
Fig. 4. Representative examples of in situ hybridization to label muscle FBPase mRNA combined with immunocytochemistry for GFAP to label astrocytes in coronal sections through the medial septum from untreated control (A, C) and experimental rats 15 days after cholinergic basal forebrain lesion by intracerebroventricular 192IgG-saporin (B, D) at different magnifications (A, B: 220 × ; C, D: 140× ). Hybridization signals are indicated by blue precipitates, while GFAP-immunoreactive astrocytes are labeled by brown staining (arrows). Note the presence of reactive astrocytes in basal forebrain sections from cholinergic lesioned rats (B) that do not express muscle FBPase mRNA.
partly consistent with the immunocytochemical staining pattern for the FBPase enzyme using a poloclonal antiserum that did not differentiate between isoforms (Schmoll et al., 1995), the neuronal expression pattern observed here for both FBPase mRNA isoforms is in contradiction to the immunocytochemical observations by Schmoll et al. (1995) who did not find FBPase-immunoreactive neurons in rat brain. It could be speculated that muscle FBPase mRNA is present at low concentrations in both neurons and astrocytes in normal rat brain, but the corresponding gene product is preferentially produced in astrocytes that are gluco- or glyconeogenetically active. The presence of FBPase mRNA in neurons may exhibit a reservoir to rapidly induce FBPase enzyme production under extreme metabolic situations such as ischaemia or lack of energy substrates. However, regardless of any speculations, our data clearly demonstrate the expression of FBPase mRNA isoforms in the normal adult brain and a localization in both neurons and astrocytes.
Acknowledgements This work was supported by grants from the Interdisziplina¨ res Zentrum fu¨ r Klinische Forschung (IZKF) at the University of Leipzig, 01KS9504, project C6. We
thank Dr. M. Bru¨ ckner, Paul-Flechsig-Institute for Brain Research, University of Leipzig, for culturing of astrocytes and neurons. T.L. gratefully acknowledges the receipt of a stipend from the Graduiertenkolleg fu¨ r Neurowissenschaften ‘INTERCELL’ at the Medical Faculty of the University of Leipzig.
References Al-Robaiy, S., Eschrich, K., 1999. Rat muscle fructose-1,6-bisphosphatase: cloning of the cDNA, expression of the recombinant enzyme, and expression analysis in different tissues. Biol. Chem. 380, 1079 – 1085. Bru¨ ckner, G., Schober, W., Ha¨ rtig, W., Osterman-Latif, C., Webster, H.H., Dykes, R.W., Rasmusson, D.D., Biesold, D., 1992. The basal forebrain cholinergic system in the raccoon. J. Chem. Neuroanat. 5, 441 – 452. Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation by guanidinium thiocyanat-phenol-chloroform extraction. Anal. Biochem. 162, 156 – 159. Cloix, J.-F., Beaulieu, E., Hevor, T., 1997. Various fructose-1,6-bisphosphatase mRNAs in mouse brain, liver, kidney and heart. Neuroreport 8, 617 – 622. Hansson, E., Ro¨ nnba¨ ck, L., 1989. Primary cultures of astroglia and neurons from different brain regions. In: A Dissection and Tissue Culture Manual of the Nervous System. Alan R. Liss, New York, pp. 92– 104. Hertz, L., Juurlink, B.H.J., Hertz, E., Fosmark, H., Schousboe, A., 1989. Preparation of primary cultures of mouse (rat) astrocytes.
T. Lo¨ ffler et al. / Int. J. De6l Neuroscience 19 (2001) 279–285 In: A Dissection and Tissue Culture Manual of the Nervous System. Alan R. Liss, New York, pp. 105 –108. Hevor, T.K., 1994. Some aspects of carbohydrate metabolism in the brain. Biochimie 76, 111 –120. Hevor, T.K., Gayet, J., 1978. Fructose-1,6-diphosphatase and phosphofructokinase activities in the brain of mice submitted to methionine sulfoxime. Brain Res. 150, 210 –215. Hevor, T.K., Delorme, P., Beauvillain, J.C., 1986. Glycogen synthesis and immunocytochemical study of fructose-1,6-bisphosphatase in methionine sulfoxime epileptogenic rodent brain. J. Cerebr. Blood Flow Metab. 6, 292 –297. Lemke, R., Hartlage-Ru¨ bsamen, M., Schliebs, R., 1999. Differential injury-dependent glial expression of interleukins-1a,b and interleukin-6 in rat brain. Glia 27, 75 –87. Liu, F., Fonnum, H.J., 1988. Purification and characterization of fructose-1,6-bisphosphatase from bovine brain. Arch. Biochem. Biophys. 260, 609 – 615. Majumder, A.L., Eisenberg, F., 1977. Unequivocal demonstration of fructose-1,6-bisphosphatase in mammalian brain. Proc. Natl. Acad. Sci. USA 74, 3222 –3225. Mehlhorn, G., Lo¨ ffler, T., Apelt, J., Roßner, S., Urabe, T., Hattori, H., Nagamatsu, S., Bigl, V., Schliebs, R., 1998. Glucose metabolism in cholinoceptive cortical rat brain regions after basal forebrain cholinergic lesion. Int. J. Devl. Neurosci. 16, 675 – 690. Mizunuma, H., Tashima, Y., 1986. Characterization of rat muscle fructose-1,6-bisphosphatase. J. Biochem. (Tokyo) 99, 1781 – 1788.
285
Roßner, S., Ha¨ rtig, W., Schliebs, R., Bru¨ ckner, G., Brauer, K., Perez-Polo, J.R., Wiley, R.G., Bigl, V., 1995. 192IgG-saporin immunotoxin-induced loss of cholinergic cells differentially activates microglia in rat basal forebrain nuclei. J. Neurosci. Res. 41, 335 – 346. Schliebs, R., Roßner, S., Bigl, V., 1996. Immunolesion by 192IgG-saporin of rat basal forebrain cholinergic system — a useful tool to produce cortical cholinergic dysfunction. Progr. Brain Res. 109, 253 – 264. Schmoll, D., Cesar, M., Fuhrmann, E., Hamprecht, B., 1995. Colocalization of fructose-1,6-bisphosphatase and glial fibrillary acidic protein in rat brain. Brain Res. 677, 341 – 344. Scrutton, M.C., Utter, M.F., 1968. The regulation of glycolysis and gluconeogenesis in animal tissues. Annu. Rev. Biochem. 37, 249 – 302. Stein, S., Liehr, Th., Eschrich, K., 2001. Characterization of the mouse fructose-1,6-bisphosphatase gene. Gene (in press). Tillmann, H., Eschrich, K., 1998. Isolation and characterization of an allelic cDNA for human muscle fructose-1,6-bisphosphatase. Gene 212, 295 – 304. Verge, V., Hevor, T.K., 1995. Regulation of fructose-1,6-bisphosphatase activity in primary cultured astrocytes. Neurochem. Res. 20, 1049 – 1056. Verge, V., Legrand, A., Hevor, T., 1996. Isolation of carbohydrate metabolic clones from cultured astrocytes. Glia 18, 244 – 254.
. .