Hypoglycemia-Elicited Immediate Early Gene ... - SAGE Journals

Journal of Cerebral Blood Flow alld Metabolism 15:989-1001 © 1995 The International Society of Cerebral Blood Flow and Metabolism Published by Lippincott-Raven Publishers. Philadelphia

Hypoglycemia-elicited Immediate Early Gene Expression in Neurons and Glia of the Hippocampus: Novel Patterns of FOS, JUN, and KROX Expression Following Excitotoxic Injury

Peter Gass, *Ken-ichiro Katsura, tWerner Zuschratter, *Bo Siesjo, and Marika Kiessling Institute of Neuropathology, University of Heidelberg, Heidelberg, Germany, *Laboratory for Experimental Brain Research, Experimental Research Center, University of Lund, Lund, Sweden, and tFederal Institute for Neurobiology, Magdeburg, Germany

Summary: In the hippocampus there is a graded vulner

glucose replenishment, differential temporospatial induc tion of IEG-coded transcription factors of the FOS, JUN and KROX families were observed in moderately injured neuronal subpopulations, including the majority of den tate granule cells and CA3 neurons. At later time points, however, a delayed and persistent c-JUN expression was found in severely, but reversibly, injured CAl neurons and in neurons in the immediate vicinity of irreversibly damaged neurons in the crest of the dentate gyrus. Sim ilar to the results with experimental models of central and peripheral axotomy, selective c-JUN induction in these neurons may represent an initial event in the regeneration process of sublethally injured neurons. In contrast to other models of excitotoxic injury such as ischemia and epilepsy, marked glial c-FOS expression was restricted to astrocytes, as assessed by confocal laser scanning mi croscopy. Key Words: Hypoglycemia-Immediate early genes-c-JUN.

ability of neuronal subpopulations to hypoglycemia induced degeneration, most likely due to excitotoxic ac tivation of glutamate receptors. The present study was conducted to investigate whether the induction of tran scription factors of the immediate early gene (lEG) family after hypoglycemia reflects these different grades of neu ronal vulnerability. We studied the expression profile of seven lEG-coded proteins in the rat hippocalJlPus follow ing severe insulin-induced hypqglycemia with 30 min of EEG isoelectricity and various survival periods for up to 42 h after glucose replenishment. Immunocytochemistry was performed on vibratome sections with specific poly clonal antisera directed against c-FOS, FOS B, c-JUN, JUN B, JUN D, KROX-24, and KROX-20. To unequiv ocally define the type of glial cells showing IEG induc tion, we investigated coexpression of c-FOS and glial marker proteins (glial fibrillary acid protein [GFAP], OX42) by confocal laser scanning microscopy. Up to 3 h after

Hypoglycemia, global ischemia, and generalized seizures severely affect the metabolism of wide spread CNS areas (Auer and Siesj6, 1988). Neuro nal degeneration, however, is restricted to a few specific cell populations. This phenomenon has been termed "selective neuronal vulnerability." In

the hippocampus, the CAl pyramidal cells are par ticularly vulnerable. The vulnerability is most pro nounced following ischemia (Kirino, 1982; Pul sinelli et al., 1982), but CA1 damage is also usually prominent following prolonged hypoglycemia coma (Auer et al., 1984a,b), and it is frequently observed after long periods of status epilepticus (Sloviter, 1983; Nevander et al., 1985). Cells in the dentate gyrus and CA3 pyramidal neurons are relatively re sistant to ischemic insults, but cells in the crest of the dentate gyrus are particularly vulnerable in hy poglycemic coma (Auer et al., 1984a,b), whereas CA3 damage is often observed in seizure disorders (Lundgren et al., 1994). The molecular basis of this

Received October 1 9, 1 994; final revision rec eived January 24 , 1 995; accepted January 24, 1 995. Addre s s corre spondence and reprint reque sts to Dr. P. Gass at Institute of Neuropathology, 1m Neuenheimer Feld 220, 69120 Heidelberg, Germany. Abbreviations used: GFAP, glial fibrillary acidic protein; I EG, immediate early gene; IR, immunoreactivity; PBS, phosphate buffered saline; PB ST, PBS and Triton X-IOO.

989

990

P. GASS ET AL.

selective neuronal vulnerability is still an enigma. According to current concepts, excitotoxic activa tion of postsynaptic glutamate receptors seems to be a key event, triggering a cascade of intracellular events that finally results in irreversible neuronal injury (Rothman and Olney, 1987; Choi and Roth man, 1990; Teichberg, 1992). In support of this con cept are results showing that glutamate antagonists may protect against ischemic and hypoglycemic brain damage (Sheardown et aI., 1990; Nellgard and Wieloch, 1992; Pulsinelli et aI., 1993; Papagapiou and Auer, 1994). Recent evidence suggests that activation of glu tamate receptors may also lead to profound changes in neuronal gene expression. Thus, glutamate re ceptor stimulation causes the rapid induction of transcription factors that belong to the class of immediate early genes (lEGs) (for review see Kiessling and Gass, 1993; 1994). These genes par ticipate in a fundamental biological control mecha nism, the regulation of the rate of gene transcrip tion, and are thought to be involved in coupling neuronal excitation to changes of target gene ex pression (Herschman, 1989; Bravo, 1990; Sheng and Greenberg, 1990). Indeed, IEG-coded proteins are thought to act as third messengers in an intra cellular signal transduction cascade between cell surface receptors (first messengers), cytoplasmic second messenger systems, and the nucleus, a pro cess for which the term "stimulus-transcription coupling" has been coined (Morgan and Curran, 1989; 199 1). In a recent study our group could dem onstrate that, following transient global ischemia, expression of IEG-coded proteins identifies hippo campal target populations of glutamate receptor mediated neurotoxicity that are destined to survive, i.e., dentate gyrus and CA3 neurons (Kiessling et ai., 1993). In contrast, irreversibly damaged CAl neurons demonstrate strong transcription, but no translation of various IEGs of the FOS, JUN and KROX families (Kiessling et ai., 1993). However, it is not known whether the failure to synthesize IEG coded proteins is a causal event, in the selective and delayed degeneration of CAl neurons, or rather a valuable early marker of neurons destined to die. In hypoglycemia, there is also a gradient of vul nerability among subsets of hippocampal neurons. The most vulnerable ones are the dentate crest granule cells, followed by pyramidal cells in the most medial part of the CA I sector (Auer et aI., 1985b). The lateral parts of the CAl sector contain cells that are more resistant; the most resistant neu rons are those in the CA3 sector and granule cells that are not at the crest or are far from the cerebral spinal fluid pathways (Auer et aI., 1984a,b). We J

Cereb Blood Flow Metab, Vol. 15, No.6, 1995

wanted to investigate whether lEG expression re flects neuronal vulnerability after hypoglycemia as it does after global ischemia. Therefore the present study was designed to examine the expression pro file of seven individual IEG-coded proteins in the hippocampus following insulin-induced hypoglyce mic coma with 30 min of EEG isoelectricity. To investigate the expression of lEG-coded proteins in glial cells, double labeling for IEG-coded proteins and glial marker proteins (glial fibrillary acidic pro tein [GFAP] and the microglial marker OX-42) was performed, and analyzed by confocal laser scanning microscopy to unequivocally define the type of glial cells demonstrating posthypoglycemic lEG induc tion. MATERIAL AND METHODS Animal experiments Male wistar rats of an SPF stain (Mollegaard's Breed

ing Center, Copenhagen, Denmark), weighing 280--340 g,

were starved overnight but had free access to tap water.

One-half hour before the operation, the animals were lightly anesthetized and injected i.p. with insulin (Ac

trapid, 2 IU/kg; Novo Intlustri A/S, Denmark). After in

duction of anesthesia with 3% halothane in 70% N2 0 and 30% 07, the rats were intubated and artificially ventilated with a mixture of 1.5% halothane in 70% N2 0 and 30% O2

during the subsequent operation. Halothane anesthesia was maintained at 0. 5% during the whole experiment. All

procedures followed the guidelines of the National Insti

tutes of Health (Guide for the Care and Use of Labora

tory Animals) and were approved by the Lund University Animal Ethics Committee. The tail artery, tail vein, and

right jugular vein were cannulated for blood sampling,

monitoring and control of blood pressure, and injection of

glucose solution during the recovery phase. The rats were heparinized (180 Iu/kg) and paralyzed with continuous

infusion of vecuronium bromide (Norcuron, 2 mg/h)

through the tail vein. EEG was continuously recorded by bipolar needle electrodes inserted in the muscle lateral to the skull bone. The rats were maintained at a rectal tem

perature of 37°C with the help of a thermistor-controlling

heating blanket.

Sham control rats were treated similarly to those sub

jected to hypoglycemic coma. Insulin was injected and the animals were artificially ventilated (some of them were recovered from anesthesia and allowed free move

ments in the cage), but plasma glucose levels were main

tained between 6 and 10 mM by continuous infusion of 20% glucose solution for 3 to 6 h (n 5). Physiological parameters of experimental animals are shown in Table 1. =

Apart from the sham rats, all experimental animals

were subjected to a 30 min period of hypoglycemic coma,

defined in terms of cessation of EEG activity (Auer et aI.,

1984a,h). Recovery was induced by administration of a continuous infusion of 10% glucose solution, plasma glu cose levels being maintained between 6 and 10 mM. Rats were extubated and allowed to recover for 1 to 42 h. All rats were reanesthetized and artificially ventilated for perfusion fixation with 4% paraformaldehyde in phos

phate buffer solution at 4°C. The brains were removed

IMMEDIATE EARLY GENE EXPRESSION AFTER HYPOGLYCEMIA

991

TABLE 1. Physiological parameters of experimental animals

BP

No.

Temp 36.9 ± 0.2 36.9 ± 0. 1

1 18 ± 5 1 00 ± 7

3 h rec 6 h rec

5 3 4 7

3 6.8 ± 0.2 3 6.8 ± 0. 1

1 00 ± 9

18 h rec 42 h rec 3 h rec for double labelling

7 4 4

�- ---�--------

Sham C ontrol I h rec

Pao2

pH

PI. Glubefore"

PI. Glub after

\18 ± 15 1 23 ± 9

7.48 ± 0.0 7.45 ± 0.0

1 .2 ± 0. 1

8.2 ± 4.5 8.2 ± 3.5

1 3 1 ± 26 1 25 ± 2 1

7.49 ± 0. 1

1 .3 ± 0.2 1 .3 ± 0.2

9.6 ± 0.4 8.9 ± 1 .8

l.l ± 0.2 0.9 ± 0. 1 1 .3 ± 0. 1

8.9 ± 3 . 1

Paco2 -

---,--

3 6.7 ± 0.9 3 8.5 ± 3.6 36.5 ± 2.7 35.7 ± 2.4

3 7.0 ± 0.3

1 09 ± 9 1 03 ± II

3 8.6 ± 4.5

1 18 ± 14

7.48 ± 0. 1

3 7.0 ± 0.2 3 6.9 ± 0.3

1 15 ± 13 99 ± 2

37.6 ± 2.2 36.4 ± 4.7

1 23 ± 5 \12 ± 15

7.50 ± 0.0 7.47 ± 0. 1

7.49 ± 0. 1

8.5 ± 2.0 9. 1 ± 0.9

" Plasma glucose levels j u s t before hypoglycemic coma. b Plasma glucose levels at the end of glucose infu sion (�6 h after injection of insulin). There was no significant difference between groups.

and postfixed in the same solution for 48 h. Recovery periods were I h (n 4), 3 h (n 7), 6 h (n 7), 18 h (n 7), and 42 h (n 4) from the beginning of glucose infusion. In separate animals, after 3 h of recovery from hypoglycemic coma, double labeling-for c-FOS and OFAP or c-FOS and OX-42-was performed (n 4). =

=

=

=

=

=

Immunocytochemistry

All antibodies were generated in rabbits immunized with bacterially expressed fusion proteins (Kovary and Bravo, 1991a,b). All cDNAs used to construct the fusion proteins were of mouse origin. The specificity of the an tibodies was proven in irnmunoprecipitation and Western blot experiments (Kovary and Bravo, 1991a,b). To fur ther prove the specificity, preabsorption of the antisera with different antigens was performed (Herdegen et al., 1991a). For preabsorption, 1 and 10 mM solutions of each fusion protein (in 12.5 mM Tris base, 12.5 mM glycine, 0.01% SDS, and 30 f.LM Phenylmethylsulfonyl Fluoride were incubated for 24 h with antisera diluted as for im munocytochemistry: anti-c-FOS 1:40,000, anti-FOS B 1:2,000, anti-c-JUN 1:10,000, anti-JUN B 1:5,000, anti JUN D 1: 10,000, anti-KROX-24 1:20,000, and anti KROX-20 1:2,000. Thereafter, the complex of antiserum and fusion protein was processed for immunocytochem istry. In every case, immunoreactivity (IR) was abolished by the 1 nM concentration of the corresponding antigen. Incubation with 10 nM solutions of other, related, fusion proteins did not affect IR. These experiments virtually eliminated the possibility of antibody cross-reaction with other known members of the FOS/JUN or KROX fami lies. Immunocytochemistry was performed on free-floating 50 f.Lm (vibratome) coronal sections at the level of the dorsal hippocampus. Sections were incubated at 4°C in normal swine serum (5% in phosphate-buffered saline [PBS] and 0.2% Triton X-100 [ PBST]) for I h, followed by the primary antisera for 36 h. The primary antisera were diluted as described above. IR was visualized by the avidin biotin complex method (Vectastain, Vector Labo ratories, U.S.A.), as described previously (Gass et al., 1992a; 1993). Sections were developed in 0.02% diamino benzidine with 0.02% hydrogen peroxide, and a subset counterstained with hemalum.

(Boehringer Mannheim, Mannheim, F.R.G.) or OX-42 (SeraLab, Crawley Down, England). Free-floating sec tions were incubated with the monoclonal antibody for GFAP (1:20 in PBST) or OX-42 (I :500 in PBST) for 36 h at 4°C. Subsequently, sections were washed in PBST for 1 h, with 3 changes, then incubated with biotinylated anti mouse antibodies (Vectastain, Vector Laboratories) for 30 min. Sections were then washed in PBST and incu bated in Texas Red-labeled avidin biotin complex (Sigma, Deisenhofen, Germany) for 30 min. After being washed in PBST, sections were incubated in anti-c-FOS antiserum, diluted 1:40,000, for 36 h at 4°C. Subsequently, sections were washed in PBST, incubated with FITC-Iabeled anti rabbit antibodies (Sigma) for 30 min, washed again in PBST, and mounted on glass slides with a mixture of 90% glyceroll\O% PBS. In a subset of sections the order of double labeling was reversed (i.e., incubation with anti c-FOS antiserum first, followed by incubation with the monoclonal antibody for GFAP or OX-42) without affect ing the results. Sections were analyzed with a confocal laser scanning microscope (Leica TCS4D) equipped with a Krypton-Argon-Ion laser (488/568 nm). The configura tion of the system (excitation beam splitter: DD [488/568], detector beam splitter: RSP 580, barrier filters: BP 535 [channel I] and 00 590 [channel 2]) allowed a rapid si multaneous detection of the FITC fluorescence (indica tive of c-FOS) and Texas Red fluorescence (indicative of GFAP or OX-42) without crosstalk. For data acquisition at different magnifications we used one of the following objectives: PC Fluotar 25x OillNA 0.75, PC Fluotar 40x OillNA 1.0-0.5, or PC Apo 63x OillNA 1.4. Areas of in terest were scanned in a 1024 x \024 pixel format in the x/y direction using a 4 x framescan averaging routine. Along the z-axis, 25 to 50 thin optical sections were scanned through a specimen with a stepping size between 250 and 500 nm (depending on the magnification and the z-range). The depth (z-range) of the scanned areas varied between 15 and 40 f.Lm. The composite images and, if necessary, single scans of a stack were displayed on a graphic monitor, analyzed for double labeling, and subsequently stored on an optical disk for further analysis.

RESULTS Confocal laser scanning microscope analysis

In a subset of animals surviving for 3 h, double labeling was performed with the c-FOS antiserum and mouse monoclonal antibodies directed against either GFAP

Neuronal lEG expression c-FOS, FOS B. The IR for c-FOS-coded protein and all other lEG-coded proteins investigated was

] Cereb Blood Flow Metab, Vol.

IS,

No.6, 1995

992

P. GASS ET AL.

restricted to nuclei as demonstrated by counter staining with hemalum and GFAP. In untreated an imals without handling and injections, only a few scattered neuronal nuclei immunopositive for c-FOS protein were found in the hippocampus. Sham operated control animals, treated with insulin but kept normoglycemic by glucose infusion, did not demonstrate a higher number of immunoreac tive nuclei (Fig. 1). Similarly, no difference in hip pocampal lEG expression was observed between untreated animals and sham operated controls for all other transcription factors investigated. Follow ing hypoglycemia, a specific distribution pattern and time course of lEG expression was observed. The patterns obtained within a group of animals at each time point exhibited almost no variability. Representative time courses of c-FOS and FOS B expression are presented in Fig. I . At I h after termination of hypoglycemic coma, strong c-FOS induction was found in the large ma jority of dentate granule cells (Fig. I ). However, a specific subset of granule cells, i.e., those of the crest of the dentate gyrus, were devoid of any c-FOS protein IR at 1 h and all later time points. In most animals, this lack of IR was restricted to only a few of the 50 !Lm serial sections at the level of the dorsal hippocampus, whereas crest cells anterior or posterior to this level demonstrated strong IR. This finding was true for all other lEG-coded proteins investigated. At 1 h after hypoglycemia, CA3 pyra midal neurons also exhibited distinct c-FOS protein labeling, but CAl neurons did not. In addition to neuronal c-FOS induction, marked IR was ob served in nuclei of glial cells localized in the vicinity of the hippocampal fissure. At 3 h, strong c-FOS expression was present in dentate gyrus and CA3 neurons, as well as in numerous glial cell nuclei along the pyramidal cell layer of the hippocampus, but no IR was found in CAl neurons. At 6 h, intense c-FOS expression was restricted to CA4 neurons in the hilus; IR in other neuronal populations and in glial cells was significantly lower than it was at 3 h. At 18 h-apart from single immunopositive neurons in the hilus-c-FOS expression had returned to baseline levels (Fig. I). In sham operated control animals, only a few nu clei positive for FOS B protein were detected in the dentate gyrus. Compared with c-FOS expression, hypoglycemia-induced FOS B expression was clearly delayed and prolonged and was almost com pletely restricted to neurons of the dentate gyrus (Fig. 1). FOS B expression in the dentate gyrus was first observed at 3 h and exhibited peak levels at 6 h, with weak IR persisting up to 18 h. At 6 h, some scattered CA3 neurons demonstrated weak FOS B J

Cereb Blood Flow Metab, Vol. 15, No. 6, 1995

protein lR. In contrast to the c-FOS results, FOS B expression was not found in glial cells at any time point investigated (Fig. 1). c-fUN, fUN B, fUN D. Individual JUN proteins demonstrated different extents of constitutive ex pression (Fig. 2). c-JUN exhibited distinct constitu tive expression in dentate granule cells and in some scattered CA3 neurons compared with c-FOS (Fig. 1), the onset of c-JUN induction was delayed, with only moderately elevated lR levels in the dentate gyrus at I h after termination of hypoglycemic coma (Fig. 2). At 3 h, however, strong c-JUN expression was found in dentate granule cells, combined with a distinct expression in CA3, but not in CAL At 6 h, a delayed strong c-JUN induction was observed in CAl pyramidal cells; this induction selectively per sisted through 42 h after hypoglycemia. Moderate c-JUN protein IR was also observed in nuclei of glial cells, in particular those in the strata oriens and radiatum of CA3 and CA I , but also along the hip pocampal fissure (Fig. 2). Compared with c-FOS results, however, a significantly lower number of glial cells were immunostained for c-JUN protein. At 6 h, peak levels of IR were found in the dentate gyrus and CA3 (Fig. 2). Most dentate granule cells showed baseline expression at 18 h; however, a few granule cells localized in the immediate vicinity of degenerated crest neurons exhibited a high c-JUN expression, similar to the persistent expres sion in CAl neurons. Apart from these neurons and CAl pyramidal cells, c-JUN protein IR in all other hippocampal subpopulations had returned to control levels by 18 h after hypoglycemia termination. Constitutive JUN B expression in control animals was observed in scattered neurons of CAl (Fig. 2). JUN B induction was significantly faster than c-JUN induction, but also more transient. At 1 h after hypoglycemia, distinct expression of JUN B occurred in CA3, and marked expression, in den tate granule cells; induction in both areas peaked at 3 h and decreased at 6 h (Fig. 2). In contrast to c-JUN results, CAl neurons did not show increased JUN B expression at any time point investigated. Furthermore, no JUN B protein IR was found in glial cells. After 18 h, JUN B expression had re turned to baseline levels in all hippocampal cell populations (not shown). Unlike our observations of c-JUN and JUN B proteins, marked constitutive expression of JUN D was observed in the hippocampus of control ani mals (not shown). After hypoglycemia, induction of JUN D was more restricted and delayed than that of the other members of the JUN family. Prominent induction was restricted to dentate granule cells, at 3 h and 6 h following hypoglycemia. Apart from a


c-FOS Co

993

FOS B Co

-

1h

CA1

1h

3h

------6h

6h

18h

18h

'

.

FIG, 1. Hypoglycemia-induced expression of FOS proteins in the hippocampus. Hippocampal sections were incubated with the appropriate primary antisera, and processed for visualization of immunoreactivity (IR). In control animals (CO), c-FOS and FOS B are present only in a few scattered nuclei. In dentate granule cells (DG), c-FOS protein expression is apparent as early as 1 h after termination of hypoglycemic coma, with peak levels at 3 h. A specific subset of granule cells, however, i.e., those at the crest of the dentate gyrus, are devoid of c-FOS IR. In CA3 and hilus neurons, but not in CA1 neurons, distinct c-FOS expression is visible between 1 and 6 h. In addition to neuronal c-FOS expression, marked expression is also present in numerous astrocytes localized mainly along the pyramidal cell layer and the hippocampal fissure (cf. Fig. 4). At 18 h, c-FOS expression is restricted to scattered hilus neurons. FOS B induction is distinctly slower, with peak levels in dentate granule cells beginning at 3 to 6 h, and a decline as late as 18 h after termination of hypoglycemic coma. As is the case with c-FOS, FOS B protein is absent in neurons of the dentate gyrus crest. In CA3 neurons only weak FOS B IR is present at 6 h. FOS B expression is not observed in glial cells at any time point investigated (50 flm vibratome sections, original magnification x 40).

994

P. GASS ET AL.

c-JUN

JUN B

Co

Co

1h

3h

6h

.-q-+

/�.

� ../-

42h '-

I�

FIG. 2. Temporal profile of c-JUN and JUN B induction after hypoglycemic coma. Distinct constitutive expression is localized in dentate granule cells for c-JUN, and mainly in CA1 neurons for JUN B. c-JUN induction in dentate granule cells and CA3 neurons is slower and somewhat prolonged than c-FOS induction (cf. Fig. 1) with peak levels at 6 h after termination of hypoglycemic coma. At 6 h, marked c-JUN expression is also visible in CA1 pyramidal cells, and persists there up to 42 h; at that time point, c-JUN protein levels have returned to baseline in all other hippocampal subfields-except in granule cells in the immediate vicinity of degenerated crest neurons. The pyramidal c-JUN IR is apparent in some scattered glial cells in addition to the pyramidal neurons. A more rapid and transient expression of JUN B is seen in dentate gyrus and CA3, with peak levels at 3 h, a marked decline at 6 h and a return to baseline at all subsequent time points. No JUN B induction is observed in CA1 neurons or glial cells (50 fJ,m vibratome sections, original magnification x 40).


few strongly immunopositive neurons in the hilus, JUN D expression had returned to baseline levels at 18 h. As was the case for JUN B, no JUN D protein IR was observed in glial cells (not shown). KROX-24, KROX-20. A marked constitutive ex pression of KROX-24 was apparent in CAl pyrami dal cells (Fig. 3). Following hypoglycemia, the ki netics of KROX-24 expression was similar to that of c-FOS and JUN B expression, with rapid induction and decline. Strong expression was found in the dentate gyrus at 1 h, a peak level was observed at 3 h, and a distinct decline, at 6 h (Fig. 3). Marked KROX-24 expression was also observed in CA3 neurons, at 3 h and 6 h after hypoglycemia (Fig. 3). The IR in CAl neurons remained unchanged for 6 h, then showed a marked decline at 18 h. At 42 h after hypoglycemia, KROX-24 protein levels had re turned to normal in all hippocampal subpopula tions. In contrast to c-FOS and c-JUN results, nei ther KROX-24 nor KROX-20 expression was ob served in glial cells. In control animals, KROX-20 was absent in the hippocampus (Fig. 3). Compared with expression of KROX-24, hypoglycemia-induced KROX-20 ex pression was restricted and delayed. The first evi dence of moderate induction in the dentate gyrus was at 3 h, and peak levels were at 6 h, persisting up to 18 h. At 42 h after hypoglycemia, KROX-20 pro tein levels had declined to baseline (Fig. 3). Confocal laser scanning analysis of lEG expression in glial cells

To unequivocally define the type of glial cells ex pressing IEG-coded proteins, double labeling, for c-FOS and for GFAP or OX-42, was performed (Fig. 4). To yield a high resolution of coexpressing glial cells in 50 fLm vibratome sections, immunoflu orescence-labeled sections were investigated with a confocal laser scanning microscope. Following dou ble labeling with antibodies against c-FOS and GFAP, confocal laser scanning microscopy demon strated numerous c-FOS-expressing, GFAP-pos itive astrocytes (Fig. 4A-D). These astrocytes are localized throughout the hippocampus, mainly in the hilus and along the pyramidal cell layer. Higher magnifications clearly demonstrated a close associ ation of intracytoplasmic GFAP-positive intermedi ate filaments with the c-FOS immunoreactive nu cleus. Numerous microglial cells were detected by OX42 immunocytochemistry in the hilus and along the pyramidal cell layer of the hippocampus (Fig. 4E). In contrast to GFAP-positive astrocytes, however, OX-42-positive microglial cells did not exhibit co expression with c-FOS protein (Fig. 4E,F).

995

DISCUSSION

The present study investigated the expression profile of seven lEG-coded proteins in the rat hippocampus following insulin-induced hypoglyce mic coma with 30 min of EEG isoelectricity and various survival periods for up to 42 h. The main findings can be summarized as follows: (a) Differ ential temporospatial induction of seven lEG-coded transcription factors of the FOS, JUN, and KROX families was observed in reversibly injured neuro nal subpopulations during the early posthypoglyce mic recovery period; (b) selectively delayed c-JUN expression was found in the particularly vulnerable CA 1 neurons as well as in the dentate granule cells in the immediate vicinity of degenerated crest neu rons; (c) IEG-coded proteins were not present in irreversibly damaged neurons of the crest of the dentate gyrus; (d) in addition to neuronal IEG in duction, marked c-FOS expression was found in glial cells that could be unequivocally identified as astrocytes by confocal laser scanning microscopy analysis. In order to discuss posthypoglycemic gene ex pression it is justified to recall characteristics of his topathological alterations (Auer et ai., 1984a,b; 1985b). Hypoglycemic coma of 30 min duration yields dense necrosis of the tip of the dentate gyrus but usually spares all other dentate granule cells as well as CA3 neurons. In the CAl sector, some de gree of neuronal degeneration is almost invariably observed. This is most pronounced in the medial part of the CA 1 sector, and tapers off toward the lateral parts. However, since degeneration only af fects parts of the neuronal populations, even the parts with pronounced degeneration will contain cells that are reversibly injured. As discussed be low, these are presumably the ones that will resume protein synthesis during recovery. Hypoglycemia-induced early changes of neuronal lEG expression

During early posthypoglycemic recovery periods up to 3 h, marked lEG induction was observed in moderately and reversibly injured hippocampal neurons, i.e., dentate granule cells other than de generated crest neurons, and CA3 pyramidal neu rons (Figs. 1-3). In contrast, no expression of IEG coded proteins was observed in the more severely affected CAl neurons (Figs. 1-3). Hippocampal sUbpopulations with early IEG expression (dentate gyrus, CA3) correspond to areas with a rapid recov ery of overall protein synthesis following severe hy poglycemia (Kiessling et aI., 1986). Restoration of protein synthesis represents an early marker for re versible neuronal injury, whereas persistent postleJ

Cereh Blood Flow Metab, Vo/. 15, No.6, 1995

P. GASS ET AL.

996

KROX-24 Co

KROX-20 Co

"

1h

6h

18h

18h

42h

42h

FIG. 3. Temporospatial expression pattern of KROX proteins in the hippocampus after hypoglycemia. In control animals, KROX24 expression is localized in CA1 neurons, whereas no KROX-20 expression is detectable. Following hypoglycemia, the kinetics of KROX-24 expression was similar to those of c-FOS and JUN B (cf. Figs. 1,2), with a rapid induction and decline in dentate gyrus (apart from crest neurons) and CA3 between 1 and 6 h. In CA1, the IR levels remain unchanged from baseline up to 6 h after termination of hypoglycemic coma, after which the levels are reduced at 18 h, and subsequently return to normal. In contrast to KROX-24, KROX-20 induction of clearly delayed, and restricted to the dentate gyrus: Moderate levels of KROX-20 IR are found in granule cells between 3 and 18 h after hypoglycemia (50 fLm vibratome sections, original magnification x 40).


sional inhibition of protein synthesis is a hallmark of impeding neuronal death in both hypoglycemia and ischemia (Thilmann et aL, 1 986). In global ischemia, graded neuronal vulnerability (CAl> CA3> den tate gyrus) is inversely correlated with a staggered recovery of postischemic overall protein synthesis, first in dentate gyrus and subsequently in CA3, whereas irreversibly damaged CAl neurons fail to resume protein synthesis (Thilmann et aL, 1986). Similar to the events following hypoglycemia, lEG expression after global ischemia occurred first in the least vulnerable hippocampal sUbpopulation (dentate gyrus), then in CA3 and CA4 neurons, re spectively (Kiessling et aL, 1 993). Apart from different recovery periods of overall protein synthesis for different hippocampal sectors, individual transcription factors demonstrated a staggered induction, that could be best defined in the dentate gyrus (Figs. 1 -3). Thus, c-FOS, JUN B, and KROX-24 were already strongly expressed within the first hour after termination of hypoglyce mic coma, whereas expression of FOS B, c-JUN, JUN D, and KROX-20 exhibited a somewhat slower temporal profile (Figs. 1 -3). FOS and JUN proteins exert their transcriptional control at specific DNA-binding elements such as activator protein 1 (AP-l) by dimer formation be tween individual members of the FOS and JUN pro tein families via a specific domain called leucine zipper (Landschulz et aL, J 988; Kouiarides and Ziff, 1 988; Gentz et aL, 1 989; Ransone et aL, 1 989). Whereas FOS proteins can only form heterodimers, with members of the JUN family, all JUN proteins are able to form homodimeric complexes as well as heterodimeric complexes, with FOS proteins or members of other leucine zipper families, such as cyclic AMP/Ca2 + -response element binding protein 1 (Nakabeppu et aL, 1 988; Cohen et aL, 1 989; Mac gregor et aL, 1 990). Dimer formation increases the variety of protein combinations as well as the num ber of usable DNA motifs. This combinatorial mul tiplicity is of functional relevance, since different AP- 1 complexes have different molecular proper ties, including DNA binding affinity and half-life, as well as up- or down-regulatory effects on target genes (Ryseck and Bravo, 1 99 1 ). For example, while c-FOS/c-JUN dimers are efficient activators of promotors containing AP-l sites, c-JUN/JUN B complexes seem to repress them under certain cir cumstances (Chiu et aL, 1 989; Schutte et aL, 1 989). KROX-24 (also termed NGFl-A [Milbrandt, 1 987], EGR-l [Sukhatme et aL, 1 988], Zif/268 [Christy et aL, 1 988]) and KROX-20 (also termed EGR-2 [Jo seph et al., 1 988]) belong to a family of DNA binding proteins with zinc finger motifs (Klug and

997

Rhodes, 1 987; Evans and Hollenberg, 1988). Simi lar to different FOS/JUN dimers, these zinc finger proteins compete for a specific DNA site, the so called early growth response (EGR) consensus se quence (Christy and Nathans, .1989). The staggered induction of individual members of the lEG family may imply that different transcriptional mecha nisms occur in the respective hippocampal subpop ulations at different posthypoglycemic time points. Hypoglycemia-induced late changes of lEG expression

A selective and marked c-JUN expression in the highly vulnerable CAl neurons, and in neurons in the immediate vicinity of degenerated dentate gyrus crest cells, started at 6 h and persisted for at least up to 42 h after termination of hypoglycemic coma (Fig. 2). These neurons are more vulnerable and more likely to undergo hypoglycemia-elicited neu ronal degeneration than CA3 or dentate granule cells apart from crest neurons, but many of them survive, as demonstrated by acid fuchsin and cresyl violet staining (Auer et aL, 1 984a,b; 1 985a). In agreement with their high vulnerability, the surviv ing neurons demonstrated a delayed but complete restoration of global protein synthesis after 30 min of hypoglycemia-induced EEG isoelectricity and re covery periods for up to 7 d (Kiessling et aL, 1986). There are basically two mechanisms that can elicit cellular c-JUN expression. On the one had, there is transynaptic activation, most likely by glu tamate receptors, that causes early expression of c-JUN and a panel of related lEG-coded proteins in target cell populations of glutamate excitation after hypoglycemia, ischemia, and epilepsy (for review see Kiessling and Gass, 1 993; 1 994). On the other hand, there is a selective and delayed c-JUN ex pression in neurons that are more severely injured than the aforementioned cell popUlations. In the lat ter type of c-JUN inducibility, a certain threshold of vulnerability obviously needs to be surpassed. This resembles the conditions for stress protein induc tion that is similarly found in reversibly damaged neuronal subpopulations, above a specific threshold of neuronal injury (for review see Brown, 1990; Nowak, 1 990); stress protein expression is even more delayed than c-JUN expression is. A selective and persistent induction of c-JUN has similarly been observed in the nuclei of the mammillary body, substantia nigra, and motoneurons or dorsal root ganglion neurons, after central or peripheral axotomy (Herdegen et aL, 1 99 1 b; 1993a,b; Leah et aI., 1 99 1 ). In these models, c-JUN expression rep resents the earliest change of the neuronal genetic program that coincides with the cell body response, J

Cereb Blood Flow Metab, Vol. 15. No.6, 1995

J

Cereb Blood Flol!' Me/ah, Vol, 15, No, 6, 1995


and sprouting of axotomized ganglion cells. Indeed it has been postulated that c-JUN expression rep resents the initial event in neuronal regeneration programs (Herdegen et aI., 1993b). In contrast to other transcription factors, e.g., c-FOS, c-JUN can form homodimers and, does not need other molec ular partners for the transcriptional control of target genes (Halazonetis et aI., 1988). In view of the dif ferent molecular characteristics of differently com posed AP-1 complexes (see above), it can be spec ulated that in the absence of other inducible leucine zipper proteins, c-JUN homodimers have particular functions in, and effects on, the transcriptional pro gram of the affected neurons. Between 6 and 42 h after hypoglycemic coma, a transient, reversible decrease of constitutive KROX-24 expression was observed in CAl pyrami dal cells (Fig. 3). Constitutive KROX-24 expression is thought to be maintained by NMDA-receptor mediated physiological neuronal activity (Worley et aI., 199 1). Thus, the transient reduction of KROX24 expression levels in CAl hippocampus may re flect an intermittent decrease of neuronal activity in the hippocampal circuit during the posthypoglyce mic recovery period. A similar transient reduction of KROX-24 expression levels that correlated with a concomitant decrease in metabolic activity was observed in the neocortex after limbic seizures in rats (Gass et aI., 1993). Confocal laser scanning analysis of lEG expression in glial cells

To unequivocally define the type of glial cells with lEG induction, coexpression of nuclear c-FOS and cytoplasmic glial marker proteins was analyzed by laser scanning microscopy (Fig. 4). Severe hy poglycemia elicited marked c-FOS expression and distinct c-JUN expression in GFAP-positive hippo campal astrocytes, mainly localized within the hilus and along the pyramidal cells layer (Fig. 4A-D). However, no c-FOS expression was detected in OX-42-positive microglial cells (Fig. 4E,F). In con trast to c-FOS and c-JUN proteins, all other IEG-

999

coded proteins investigated were not induced in GFAP-positive-astrocytes. Similar to the hypogly cemia-induced late changes of neuronal lEG ex pression (see above), the selective induction of a subset of transcription factors in astrocytes may re flect the activation of specific cellular programs, regulating the expression of GFAP (see below) or other potential target genes. In contrast to these results with hypoglycemia, lEG induction was not found in astrocytes after other excitotoxic lesions such as ischemia or epi lepsy (Gass et aI., 1992a,b; 1993; Kiessling et aI., 1993). Similarly, experimental models for Parkin son's disease (Paul et aI., 1992; Jenkins et aI., 1993), axotomy (Herdegen et aI., 1993a,b), noxious stim ulation (Herdegen et aI., 199 1a), water deprivation (Giovanelli et aI., 1992; Ding et aI., 1994), or immo bilization stress (Harbuz et aI., 1993) did not cause lEG induction in glial cells. So far, glial c-FOS ex pression has only been found in astrocytes localized within the wound margins following brain injury by stab wounds; it has been postulated to represent a molecular correlate for astroglial cell activation in the regeneration process (Dragunow and Robert son, 1988; Dragunow et aI., 1990). Pennypacker and colleagues ( 1994), however, using gel shift assays and immunocytochemistry, observed a prolonged increase in hippocampal AP-1 binding activity and a marked increase in astrocytic GFAP expression, re spectively, after kainic acid-induced limbic sei zures. Since the GFAP gene contains an AP-I site (Masood et aI., 1993) that readily complexes with nuclear extracts following kainic acid-induced sei zures, Pennypacker et al ( 1994) postulated a poten tial regulation of GFAP by lEG-coded proteins. In contrast to the rather restricted inducibility of lEGs in astrocytes in vivo, a broad range of factors have been demonstrated to evoke lEG induction in cultured astrocytes. These include vasoactive intes tinal peptide, phorbol ester, dibutyryl cyclic AMP, epidermal growth factor, and basic fibroblast growth factor (Hisanaga et aI., 1993). Pechan et al. (1992) described the induction of c-FOS by free rad-

FIG. 4. Confocal laser scanning analysis of c-FOS expression in glial cells. Following double labeling with antibodies against c-FOS (green fluorescence) and glial fibrillary acid protein (GFAP [red fluorescence]), confocal laser scanning microscopy demonstrates numerous c-FOS-expressing GFAP-positive astrocytes (A-D). These astrocytes are localized throughout the hip pocampus, mainly in the molecular layer and the hilus of the dentate gyrus (A), and in the stratum oriens and radiatum along the pyramidal cell layer (8). Higher magnification clearly demonstrates a close association of intracytoplasmic GFAP-positive inter mediate filaments surrounding the c-FOS immunoreactive nucleus (C,D), In 0, a c-FOS-positive as well as a c-FOS-negative astrocyte are depicted, Following double labeling with antibodies against c-FOS (green fluorescence) and the microglial marker protein OX-42 (red fluorescence), confocal laser scanning microscopy fails to demonstrate coexpression of OX-42 and c-FOS within the same cell (E,F). Numerous microglial cells are visible in the hilus (F) and along the pyramidal cell layer, but higher magnification does not show a double labeling for the dense OX-42-positive microglial processes and c-FOS-positive nuclei (G) (dg dentate gyrus; arrowheads in 8 indicate CA1 pyramidal cell layer; 50 fLm vibratome sections, original magnification A, 8, E: x 250; C, 0, F: original magnification x 630). =

J

Cereh Blood Flow Me/ab, Vol. 15, No.6, 1995

P. GASS ET AL.

1000

icals in cultured astrocytes, a mechanism that may also play a role in excitotoxic injury in vivo. In view of the restricted inducibility of lEGs in astrocytes in vivo, these in vitro findings suggest that a relaxation of control over lEG expression occurs in vitro (Arenander and deVellis, 1992). At present, the mechanism and function of lEG induction in astrocytes following hypoglycemia re main to be defined. An interaction between neurons and astrocytes, caused by a rise of intracellular cal cium in astrocytes, has recently been demonstrated in vitro and has been suggested to affect signal pro cessing between neurons and astrocytes, both lo cally and at distance (Parpura et aI., 1 994; Neder gaard, 1994). As part of the neuronal-glial interde pendency, glial lEG expression may encode and integrate environmental signals associated with neuronal injury, and contribute to the restoration of neuronal function (Arenander and de Vellis, 1 992). The hypoglycemia-induced changes of lEG expres sion in astrocytes may thus be part of an interactive molecular response to excitotoxic injury. Acknowledgment: The authors wish to acknowledge the

excellent technical assistance of Claudia Kammerer. This work was supported by a grant from the Deutsche Fors chungsgemeinschaft (SFB 317).

REFERENCES Arenander A, deVellis J (1992) Early response gene induction in astrocytes as a mechan i s m for encoding and integrating neu ronal signals. Progr Brain Res 94:177-188 Auer RN, Olsson Y, Sie sj ii B K (l984a) H ypoglycemic brain in jury i n the rat: correlation of density of brain damage with the EEG is oelectric time. A quantitative study. Diabetes 33:1,090--1,098 Auer RN, S iesj ii B K (1988) B i ological difference s between i sch emia. h ypoglyce m i a . and epilepsy. Ann Nellrol 24:699-707 Auer RN, Wieloch T, Olsson Y, S i e sj ii B K (I984b) The di stri bution of h ypoglycemic brain damage. Acta Neuropathol 64:177-191 Auer RN, Kalimo H, O ls son Y, Sie sj ii B K (I985a) The te mporal evolution of hypoglycemic brain damage. I. Light- and elec tron-microscopic findings in the rat cerebral cortex. Acta Neuropathol 67:13-24 Auer RN, Kalimo H, Olsson Y, Wieloch T (l985b) The dentate gyrus in hypoglycemia. Pathology implicating excitotoxin medicated neuronal necrosis. Acta NeuropathoI67:279-288 B ravo R (1990) Growth factor re sponsive genes in fibroblasts. Cell Growth Diff 1:305-309 B rown IR (1990) Induction of heat shock (stre s s ) genes in the mammalian brain b y hyperthermia and other traumatic events: a current perspective. J Neurosci Res 27:247-255 C h iu R, Angel P, Karin M (1989) J U N B differs in its biological properties from, and is a negative regulator of, c-J U N. Cell 59:979-986 Choi D, Rothman SM (1990) The role of glutamate neurotoxicity i n hypoxic-ischemic neuronal death. Annu Rev Neurosci 13: l71-182 Christy B, Nathans D (1989) D N A binding sites of the growth

J

Cereb Blood Flow Melab, Vol. 15, No.6, 1995

factor-inducible protein Zif268. Proc Natl Acad Sci USA 86:8,737-8,741 Christy B A . Lau L F, Nathans D (1988) A gene activated i n m o u s e 3T3 cell s b y s e r u m growth factors encodes a protei n w i t h zinc fi nger s e q uence s . Proc Natl Acad S c i USA 85:7.857-7.861 Cohen D R, Ferreira P C P . Gentz R, Franza J r. B R . C u rran T (1989) The product of a fos-related gene, fra-I, binds coop eratively to the AP- I site with jun: transcription factor AP-I is compri sed of multiple protein complexes. Gene Develop 3:l73-184 D in g J M . Carver W C , Te rracio L . B u ggy J (1994) Proto oncogene c-fos and the regulation of vasopre s s i n gene ex pression during dehydration. Mol Brain Res 21:247-255 B ragunow M, Robertson HA (1988) B rain injury induces c-fos protein ( s ) in nerve and glia-like ce l l s in adult mammalian brain. Brain Res 455:295-299 Dragunow M . Goulding M . Faull RLM, Ralph R, Mee E, Frith R (1990) I nduction of c-fos mRNA and protein in neurons and glia after traumatic brain i nj ury; pharmacological character ization. Exp Neurol 107:136--148 Evans R . H o llenberg SM ( 1 988) Zinc fingers: gilt by association. Cell 52:1-3 Gass P . Herdegen T . B ravo R . K ies s ling M (l992a) Induction of immediate early gene (lEG) products in the rat h ippocampus after bicucu lline-induced seizures: differential e xpression of K ROX-24 . FOS and J U N proteins. Neuroscience 48:315324 Gass P. Spranger M . Herdegen T, Bravo R . Kiick P. H acke W . Kies s ling M (]992b) Induction of F O S and J U N protei n s after focal i schemia in the rat: differential effect of the N - methyl-D-as partate receptor antagonist M K-80 1 . Acta Nell,.opatlU?184:545-553 Gass p. Herdegen T . Bravo R . Kiessling M (1993) Spatiotempo ral induction of immediate early genes in the rat brain after limbic seizures: effect s of N M DA-receptor antagonist M K801. Ell,. J Nelll'Osci 5:933-943 Gentz R . Rauscher I I I FJ . Abate C . Curran T (1989) Parallel association of Fos and J u n leucine zipper j u xtapos e s DNA binding domains. Science 243:1.695-1,699 Giovanelli L. S hiromani PJ, J i rikowski GF, Bloom FE (1992) Expres sion of c-fos protein by immunohistochemically iden tified oxytocin neu rons in the rat hypothalamu s upon os motic stimulation. Brain Res 588:41-48 H alazonet i s T D . Georgo poulos K, Greenberg M E . Leder P ( 1 988) c-J un dimerizes with itself and with c- Fos . forming complexes of different D N A-binding affinities. Cell 55:9l7924 Harbuz M S . Chal mers 1. De Souza L . Lightman S L (1993) Stre s s - i nduced activation of CR F and c-fos mRNAs in the paraventricular nucleus are not affected by serotonin deple tion. Brain Res 609:167-l73 Herdegen T . Kovary K . Leah J D . B ravo R (1991a) Specific tem poral and s patial distribution of J U N . F O S and KROX-24 proteins in spi nal neurons following noxious transynaptic stimulation. J Comp NellroI313:l78-191 Herdegen T . Kummer W . Fiallos C . Leah J D . B ravo R (I991b) Expre s sion of c-J U N . J U N B and J U N D in the rat nervous sy stem following transection of the vagus nerve and cervical s y mpathetic trunk. Neuroscience 45:413-422 H erdegen T . Bastmeyer M . Bahr M . B ravo R, S turmer CAO . Zimmerman M (l993a) Expre s sion of J U N, KROX and CREB transcription factors and nitric oxide s ynthase i n rat central neurons following axotomy. J NeurobioI24:528-543 H e rdegen T . B recht S . Kummer W, Mayer B, Leah J, B ravo R, Zimmermann M (1993b) Persisting expre s s i o n of JUN and KROX transcription factors and increase of nitric oxide syn thase immunoreactivity i n rat central neurons following ax otomy. J Neurosci 13:4,130-4.146 H erschman HR (1989) E xtracellular signals, transcriptional re sponses and cellular s pecificity. TIBS 14:435-438 H i s anaga K, S agar SM, Koistinaho J, Hicks KJ, S h arp FR (1993) VIP-induced stellation and i mmediate early gene e xpre ssion


1001

in astrocyte s : effects of dexamethasone . Neurosci Lett 1 56 : 57--60

lepticus in well-oxygenated rats causes neuronal necro s i s . Ann Neurol 1 8 : 28 1 -290

Jenkins R, O ' Shea R, Thomas K L , Hunt SP ( 1 99 3 ) c-j u n expres sion i n substantia nigra neurons following striatal 6-hydroxy

Nowak Jr. TS ( 1 990) Protein s ynthe s i s and heat shock/stre s s response after ischemia. Cerebrovascular Brain Metab Rev 2 : 345-366 Papagapiou MP, Auer RN ( 1 994) Regional neuroprotective ef fe c t s of the N M D A receptor antag o n i s t M K-8 0 1 ( d i z o

dopamine lesions i n the rat . Neuroscience 5 3 : 447--455 Joseph LJ , L e B eau M M , Jamieson GA, Acharya S, Shows T B , Rowley J D , S ukhatme V P ( 1 98 8 ) Molecular cloning s e quencing an d mapping of EGR2 , a human early growth re sponse gene encoding a protein with zinc-binding finger

cilpine) in hypogl ycemic brain damage . J Cereb Blood Flow Metab 1 0 : 270--276

s tructure . Proc Natl Acad Sci USA 85 : 7 , 1 64-7 , 1 68 Kiessling M , Auer RN , Kleihues P, S i e sj ii B K ( 1 986) Cerebral

Parpura V, Basarsky TA , Liu F, Jeftinij a K, Jeftinij a S , H aydon PG ( 1 994) G lutamate-mediated astrocyte-neuron signalling .

protei n synthesis during long-term recovery from severe hy poglycemia. J Cereb Blood Flow Metab 6 : 42-5 1 Kiessling M , Gass P ( 1 99 3 ) Immediate early gene expre s sion in experimental epilep s y . Brain Pathol 3 : 3 8 1 -393 Kiessling M, Gass P ( 1 994) Stimulus-transcription coupling i n focal i s c h e m i a . Brain Pat/wi 4 : 77-83

Nature 369: 744-747 Paul M L , Graybiel A M , David J C , Robertson HA ( 1 992) DI-like and D2-like dopamine receptors synergistically activate ro

Kiessling M, Stumm G, Xie Y, Herdegen T, Aguzzi A, B ravo R ,

tation and c-fos e xpre s sion in the dopamine-depleted stria tum in a rat model of Parkin s o n ' s d i s e a s e . J Neurosci 1 2 : 3 , 729-3 , 742 Pechan PA , Chowdhury K, Seifert W ( 1 992) Free radicals induce

G a s s P ( 1 99 3 ) Differential tran scription and tran slation of

gene expre s sion of NGF and b FG F in rat astrocyte culture .

immediate early genes in the gerbil hippocampus after tran sient global ischemia. J Cereb Blood Flow Metab 1 3 : 9 1 4-924

Pennypacker K R , Thai L, Hong J S , M c M i l lian MK ( 1 994) Pro

Kirino T ( 1 982) Delayed neuronal death in the gerb i l hippocam pus following ischemia. Brain Res 2 3 9 : 5 7--69 Klug A, Rhodes D ( 1 987) Zinc finge r s : a novel protein motif for nucleic acid recognition. TIBS 1 2 : 464--469 Kouzarides T, Ziff E ( 1 988) The role of the leucine zipper in the fos-ju n interaction . Nature 3 3 6 : 646--6 5 1 Kovary K, B ravo R ( 1 99 1 a ) Expression of diffe rent J U N and FOS protein s during the GO to G 1 tran sition in mouse fibro blasts : i n v itro and in vivo a s s ociation s . Mol Cell Bioi 1 1 : 2 , 45 1 -2 ,459 Kovary K, B ravo R ( l 99 1 b) The J U N and FOS protein famil i es are b oth required for c e l l cycle progre s s ion in fibroblas t s . Mol Cell Bioi 1 1 : 4 , 466--4 ,472 Landschulz W H , Johnson PF. Mc Knight S L ( 1 98 8 ) The leucine zipper: a hypothetical structure common to a new class of DNA binding proteins . Science 240 : I ,759- 1 ,764 Leah J D , Herdegen T. Kovary K, B ravo R ( 1 99 1 ) Selective e x p res sion of J U N proteins following peripheral axotomy and axonal tran sport block in the mt : evidence for a role in the regeneration proce s s . Brain Res 566: 1 98-207 Lundgren J, Smith M L , Blennow G. Sie sj ii BK ( 1 994) H y per thermia aggravates and hypothermia ameliorates epileptic brain damage . Exp Brain Res 99 : 43-55 Macgregor PF, Abate C, Curran T ( 1 990) Direct c loning of leu cine zipper prote i n s : jun binds cooperatively to the C R E with C R E - B PI . Oncogene 5 : 45 1 --45 8 Masood K , B e s nard F , Su Y, Brenner M ( 1 99 3 ) Analysis o f a segment of the human glial fibrillary acidic prote in gene that directs astrocyte specific transcription. J Neurochem 6 1 : 1 60-- 1 66 Milbrandt J ( 1 987) A nerve growth factor-induced gene encodes a possible tran scriptional regulatory factor. Science 2 3 8 : 797-799 Morgan J l , Curran T ( 1 989) Stimulus-tran scription coupling in neurons : rol e of c e llular immediate-early gene s . Trends in Neurosciences 1 2 : 459--462 Morgan J l , C u rran T ( 1 99 1 ) Stimulus-transcription coupling in the nervous s y s t e m : involvement of the inducible proto oncogenes fos and j u n . Annu Rev Neurosci 1 4 : 42 1 --4 5 1 Nakabeppu Y, Ryder K . Nathans D ( 1 988) D N A binding activ ities of three different jun prote i n s : stimulation b y Fos . Cell 5 5 : 907-9 1 5 Nedergaard M ( 1 994) Direct signalling from astrocytes t o neu rons i n c u l tures of mammalian brain c e l l s . Science 263 : 1 , 768- 1 ,77 1 Nellgard B , Wieloch T ( 1 992) Cerebral protection b y AMPA- and N M D A-receptor antagonists administered after severe insu lin-induced hypogl ycemia. Exp Brain Res 92 : 259-266 Nevander G, Ingvar M , Auer RN , S i e sj ii BK ( 1 98 5 ) Status epi-

Neuroreport 3 : 469--472 longed e x pre s s ion of AP- I transcription factors in the rat hippocampus after systemic kainate treatment. J Neurosci 1 4 : 3 , 998--4 , 006 Pulsinelli W, B rierly I B , Plum F ( 1 982) Temporal profile of neu ronal damage i n a model tran sient forebrain ischemia. Ann Neurol I I : 49 1 --498 Pu l s inelli W, Sarokin A, B u c han A ( 1 99 3 ) Antagonism of the NMDA and non-NMDA receptors in global versus focal brain ischemia. I n : Neurobiology oj Ischemic Brain Damage ( Kogure K , H o s smann K A , S iesj ii B , eds ) , A m s terdam , E l sevier, pp 1 25- 1 3 5 Ransone LJ , V i s vader I , Sas sone-Corsi P, Verma 1M ( 1 989) Fos l u n interaction : mutational anal y s i s of the leucine zipper domain of both prote ins . Gene Del'elop 3 : 770--78 1 Rothman S M , Olney JW ( 1 98 7 ) Excitotoxicity a n d the N M DA receptor. Trends in Neurosciences 1 0 : 299-302 Ryseck RP, B ravo R ( 1 99 1 ) c-J U N , l U N B, and J U N D differ in their binding affinities to the AP- 1 and C R E consensus se quence s : effect of FOS prote i n s . Oncogene 6 : 5 3 3-542 Sc hutte J . Viallet J, N au M, Segal S, Fedorko 1, Minna J ( 1 989) Jun-B inhibits and c-FOS stimulates the transforming and tran s-activating activities of c-j u n . Cell 5 9 : 987-997 S heardown M , N ie l s e n E, Hansen AJ , Jacobsen P . Honore T ( 1 990) 2 , 3 - D i h y droxy-6-nitro - s u lfamo yl-benzo( F ) q u i n o xa line : a neuroprotectant for cerebral ischemia. Science 247 : 57 1 -574 S heng M, Greenberg M E ( 1 990) The regulation and function of c-fos and other i m mediate early genes in the nervous sys t e m . Neuron 4 : 477--485 Sioviter R ( 1 98 3 ) Epileptic brain damage in rats induced b y sus tained electrical stimulation of the perforant path . I . Ac ut e electroph y s iological and light micr o s co p ic studi e s . Brain Res Bull 1 0 : 675-697 S u khatme V P , Cao X, Chang LC , T s ai -Morris C , Stamenkovich D, Ferreira PCP. Cohen DR, Edwards S t A, Shows T B , Curran T , L e Beau M M , Adamson E D ( 1 988) A zinc finger encoding gene coregulated with c-fo s during growth and dif ferentiation , and after cellular depolarization . Cell 5 3 : 3 7-43 Teichberg V I ( 1 992) Excitotoxin s , glutamate receptors , and ex citotoxicity . I n : Handbook oJ Experimental Pharmacology , Vol 120: Selective Neurotoxicity (Herken H , Hucho F , e d s ) , Berli n , S pringer pp 483-506 Thilmann R , Xie y, Kleihues P, K i e s s ling M ( 1 986) Persistent inhibition of protein synthe s i s precedes delayed neuronal death in postischemic gerbil h i ppocampus Acta Neuropathol 7 1 : 88-93 Worley P F , Christy B A , Nakabeppu Y, B hat RV , Cole AJ , B ara ban 1M ( 1 99 1 ) Constitutive expres s io n of zif268 in neocortex i s regulated by synaptic activity . Proc Natl Acad Sci USA 8 8 : 5 , 1 06--5 , 1 1 0

J

Cereh Blood Flow Metab, Vol. 15 , No . 6, 1995