- Email: [email protected]
In situ hybridization analysis of c-fos and c-jun expression in the rat brain following transient forebrain ischemia
Brain Research, 567 (1991) 231-240 © 1991 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/91/$03.50
231
BRES 17254
In situ hybridization analysis of c-los and c-jun expression in the rat brain following transient forebrain ischemia Thomas C. Wessel, Tong H. Joh and Bruce T. Volpe Department of Neurology and Neuroscience, Cornell University Medical College, The Burke Medical Research Institute, White Plains, NY 106o5 (U.S.A.) (Accepted 30 July 1991)
Key words: Transient ischemia; c-los; e-jun; Four-vessel occlusion; In situ hybridization; Proto-oncogene; Hippocampus
Early induction of the mRNAs encoding the c-Fos and c-Jun nuclear proteins was examined in rat brain by in situ hybridization at various timepoints following global forebrain ischemia by the method of four-vessel occlusion. All animals were subjected to 20 min of transient ischemia. This produced a pattern of proto-oncogene activation that was most intense in the granule cells of the dentate gyrus 30 rain after ischemia, while the hilar cells in the dentate and the pyramidal cells of the CA3 region in the hippocampus showed a more delayed but robust expression of these immediate early genes at 1 h. The neurons of the CA1 region exhibited a more moderate hybridization signal at 1-2 h postischemia. Very little hybridization signal for either immediate early gene could be detected in animals perfused with fixative immediately following ischemia, suggesting that cellular energy levels may have to be restored to a certain level before efficient de novo mRNA synthesis can occur. In the cerebellum, a similar temporal pattern was observed: the granule cells exhibited a prompt but patchy expression of c-los and c-jun that was followed by a delayed signal in the Purkinje cells. Without exception c-los and c-jun appeared to be expressed in unison, although the time course of c-los and c-jun mRNA accumulation and decay was different in various brain regions: invariably the cerebellum returned rapidly to its baseline with virtually no remaining signal at 3 h postischemia, while c-los and c-jun activation in the hippocampus remained high at 3 h and returned to baseline by 6 h. Several other brain regions showed early production of c-los and c-jun mRNAs, such as the medial habenula, piriform cortex, the amygdala, the centromedian, lateral posterior, paracentral, intermediodorsal and reuniens nuclei of the thalamus and the ventromedial and dorsal nuclei of the hypothalamus; in the hrainstem, the trapezoid body and the noradrenergic neurons of the locus ceruleus as well as the adrenergic neurons in the ventrolateral medulla (C1 group) and nucleus tractus solitarius (C2 group) regions displayed slightly less intense hybridization signals. In addition, the ependyma of the lateral ventricles and the third ventricle showed a prompt albeit short-lived production of c-los and c-jun mRNAs. Sham-operated animals as well as animals that had survived to one week postischemia showed either no or only trace levels of hydbridization signal. Interestingly, a second peak of c-los and c-jun expression was observed in the CA1 pyramidal cells at 24-48 h postischemia when this population of cells undergoes a rapidly progressive decline in viable neurons. In contrast, the small- and medium-sized spiny neurons of the dorsolateral striatum that are also subject to an analogous selective neuronal degeneration displayed weak early and no late c-los or c-jun expression. No other brain areas with this type of delayed protooncogene activation were detected. These results indicate that the proto-oncogenes c-los and c-jun are rapidly activated in the brain following a transient ischemic insult in a temporal and spatial pattern that is distinct from the pattem seen in drug-induced seizure activity and other paradigms.
INTRODUCTION The expression of the proto-oncogene c-los is rapidly induced in the central nervous system by a variety of physiologic processes and experimental manipulations, such as peripheral nerve stimulation4'26'34, intracerebral electrical stimulation45 and kindling9'15, circadian rhythms 33, t r a u m a 16'3°'47, seizure activity 1xA4'36, pharmacologic responses 5'6'19'43 and ischemia 24'29'31'38'39. Of the studies that demonstrate c-los responses in the ischemic rat model, two have investigated the production of the specific m R N A for c-los by in situ hybridization and Northern blot analysis, respectively: JCrgensen et al. demonstrated the expression of c-los m R N A at 24-48 h
in the selectively vulnerable CA1 region of the hippocampus 29, while O n o d e r a et al. detected the presence of c-los m R N A in the postischemic rat brain on N o r t h e r n blot at several timepoints, starting at 30 min postischemia 39. Thus the precise localization of c-los expression during early ischemia in the rat is u n k n o w n . I n the gerbil, Nowak et al. examined the induction of 70 kDa heat shock protein m R N A and c-los m R N A by in situ hybridization, and demonstrated a strong signal over the granule cells of the dentate gyrus 30 min after transient ischemia of 5 min duration aS. Recently, Kindy et al. have investigated the temporal response of c-los and c-jun expression in the mongolian gerbil by Northern blot analysis31, and have found that the magnitude of m R N A
Correspondence: T.C. Wessel, Department of Neurology and Neuroscience, Cornell University Medical College, The Burke Medical Research Institute, 785 Mamaroneck Avenue, White Plains, NY 10605, U.S.A.
232 accumulation for these proto-oncogenes correlated with the duration of ischemia. The strongest signal was detected at the 30 min timepoint after cessation of ischemia and there was a rapid decay in m R N A levels back to baseline within 3 h. In order to visualize the neuronal and possibly glial, ependymal or endothelial involvement in proto-oncogene expression, we have focused on the dynamic changes in c-fos and c-jun expression that take place in the postischemic rat brain following 20 min of transient global ischemia. The protein products of these and related members of the fos and ]un proto-oncogene families are thought to form heterodimers 4t that directly contact D N A binding sites I and function as transcriptional regulators at AP-1 binding sites 7'42. The binding of these heterodimers to the AP-1 site is modulated by phosphorylation 37. Various homo- and heterodimer combinations can be formed and there is some evidence to suggest that the relative levels of these multiple homo- and heterodimers may have positive as well as negative transcriptional function s'2°'46. We have included c-jun in situ hybridization in the present study because both c-Fos and c-Jun phosphoproteins are required to form active heterodimers before binding to AP-1 sites and initiating transcription of target genes. The target genes for the Fos-Jun heterodimers in the hippocampus and other brain regions are unknown. It is conceivable that these proto-oncogenes set a cascade of events in action that leads to neuronal destruction in some cell groups, while it may contribute to the survival of other cells depending on their genetically preprogrammed repertoire. MATERIALS AND METHODS Animals were subjected to transient forebrain ischemia by a method previously described 4° and modified: on day 1, the rats were anesthesized by inhalation of 2% halothane mixed with 20% oxygen and 80% nitrogen, the carotid arteries were encircled with silastic ligatures and the vertebral arteries were permanently occluded by electrocautery. Sham-operated control rats also underwent halothane anesthesia, skin incisions and carotid manipulations. The rats were allowed to recover for 24 h. On day 2, forebrain ischemia was induced in awake animals by tightening the carotid ligatures. Rats that lose their righting reflex within 1 min of carotid clamping and for the subsequent 20 min have been shown to have their cerebral blood flow (CBF) reduced to <10 ml/100 g/rain,
approximately 10% of that in control rats. CBF of > 10 ml/100 g/min does not reproducibly impair the righting reflex and produces variable ischemic injury. The loss of the righting reflex was always associated with fixed and dilated pupils. Rats that lost their righting reflex for 20 min were classified as ischemic, whereas those animals that did not were eliminated from the experiment. During ischemia, body temperature was maintained at 37 °C by a heating lamp connected to a rectal thermistor. The carotid ligatures were released after 20 min, at which time the rats generally remained unresponsive and showed no righting reflex, while rapidly regaining temperature homeostasis. Animals were allowed to recover for variable periods and then anesthesized at 30 min, 1 h, 2 h, 3 h, 6 h, 24 h and 1 week. Rats in the 0 min time group were immediately perfused transcardially following the opening of the carotid ligatures. Animals that developed seizures during the ischemic period or during the first 3 h postoperatively (at which point they were returned to the animal room) were excluded from this study. For in situ hybridization, the postischemic rats were anesthesized with pentobarbital and rapidly perfused transcardially with 0.9% sodium chloride, containing 0.5% nitrite and 1000 U heparin/100 ml. This was followed by rapid perfusion with ice-cold 4% formaldehyde in 0.1 M sodium phosphate buffer. Brains were immediately removed, cut into blocks, and submerged in the ice-cold fixative where they remained for one hour. The blocks were then rinsed twice with phosphate buffer and cryoprotected by storing the tissue in 30% sucrose overnight at 4 °C. Tissue sections of 40/~m thickness were cut on a sliding microtome and stored in 20 ml glass vials filled with 2x SCC (0.3 M sodium chloride/0.03 M sodium citrate) with 10 mM D T r at 4 °C. This storage solution was then replaced with prehybridization buffer containing 50% formaldehyde, 10% dextran sulfate, 2x SSC, l x Denhardt's solution, 50 mM DTT, and 0.5 mg/ml sonicated and denatured salmon sperm DNA. Prehybridization was carried out for 60 rain at 48 °C. For the main experiments, labeled denatured c-los probe or c-jun antisense oligonucleotide (1 × 107 cpm) was then added to the vial in 100 /~l of the same prehybridization buffer. In control experiments, labeled sense and antisense c-los and c-jun oligonucleotide were used to (1) confirm the specificity of labeling for both c-los and c-jun antisense probes, (2) prove loss of hybridization under cold antisense probe excess and (3) verify that the hybridization patterns observed for the c-fos antisense oligonucleotide and cDNA probe are identical. Hybridization was performed overnight at 48 °C. After extensive washes in decreasing concentrations of SSC, 10 min steps of 1:1 dilutions starting at 2x SSC and ending with 0.1× SSC, tissue sections were mounted onto gelatin-subbed slides, air-dried and dehydrated through graded ethanols (70, 90, 100%). Finally the slides were dipped into undiluted Kodak NTB-2 in the darkroom. After storage in light-tight boxes at 4 °C for 7-21 days, the slides were developed in Kodak D-19, counterstained with Cresyl violet and cover-slipped with Permount. A great advantage of this particular technique is that all tissue sections from a particular timepoint and animal can be hybridized with a given radiolabeled probe under exactly the same conditions, as all floating sections are exposed to the same hybridization buffer, probe concentrations and washing procedure. The probes used in these experiments were (1) sense and antisense rat c-fos oligonucleotides corresponding to bases 138-173 (36-mer) of the published
Fig. 1. Darkfield photomicrographs showing time course of c-los expression in the rat hippocampus after 20 min of ischemia. A: shamoperated control animal; open arrows indicate pyramidal cell layer of the CA1 region which is selectively vulnerable to ischemia. B: faint outline of granule cells in the dentate gyms (DG) in an ischemic animal perfused immediately following opening of the carotid ligatures. C: after 30 min of recovery, strong expression of c-los is apparent in the DG and medial habenula as well as the ependyma. D: 60 min recovery, involvement of the hilar neurons of the DG, CA3 neurons and to a lesser extent the CA1 region. E: 2 h recovery, some decrease in D G labeling. F: 3 h recovery, weak residual signal over the DG which vanishes by 6 h (not shown). G: 24 h after ischemia, CA1 region is selectively outlined (arrows). H: 1 week after ischemia, note near complete degeneration of CA1 pyramidal neurons (open arrows). Bar = 500 #m.
233
4~
235 Fig. 2. Photomicrographs showing time course of c-jun expression in the rat hippocampus in adjacent sections to those in Fig. 1. A: sham-operated control animal. B: 0 rain. C: 30 rain. D: 1 h. E: 2 h. F: 3 h. G: 24 h. H: 1 week. Note identical induction of e-jun in the CA1 region of the hippocampus 24 h after ischemia. Bar = 500/~m.
sequence x°, (2) sense and antisense oligonucleotides to mouse c-]un base number 297-332 (36-mer) of the published sequence44 and (3) a mouse v-los probe (ATCC, Rockville, MD 20585) digested with Psfl to yield a 1.0 kb fragment for comparison to the antisense e-fos oligoprobe. The oligonucleotides were synthesized on a Biosearch 8650 DNA synthesizer and purified in a standard fashion. Labeling efficiency varied between 0.2-0.5 x 10~ cpm/ng for the oligonucleotides and 0.8-1.6 × 106 epm/ng for the c-los eDNA probe.
RESULTS Twenty minutes of forebrain ischemia leads to a r a p i d and transient expression of c-los and c-jun in several regions of the rat brain. Figs. 1 and 2 illustrate the t e m p o r a l p a t t e r n o b s e r v e d in the h i p p o c a m p u s for c-los
and c-jun, respectively. Control animals showed no or only minimal labeling of the granule o r p y r a m i d a l cells in the hippocampus. A n i m a l s that were peffused with fixative i m m e d i a t e l y following 20 rain of ischemia, t h e r e b y p r e s u m a b l y preventing any meaningful reperfusion o r recovery (Figs. 1B and 2B), d e m o n s t r a t e d very low hybridization signal for c-los and c-jun in the d e n t a t e gyrus and no signal in the r e m a i n d e r of the brain. T h e subsequent p h o t o g r a p h s (Figs. 1 C - F and 2 C - F ) depict the dynamic changes that take place during the early reperfusion period. It appears that the hybridization signal is already quite p r o n o u n c e d over the granule cells of the d e n t a t e gyrus at 30 rain (Figs. 1C and 2C), while the hilar neurons of the d e n t a t e gyrus as well as the C A 3 and C A 2
Fig. 3. Photomicrographs of cerebellar c-los expression postischemia. A: control animal, gr, granule cell layer; m, molecular layer; wm, white matter tracts. B: patchy and partially confluent areas of c-los expression in the granule cell layer 30 rain after ischemia. C: rapid resolution of c-los expression in the granule cell layer associated with a moderate density of grains over the molecular layer and an interrupted line of hybridization signal over the Purkinje cells (arrow). D: 3 h after ischemia faint signal is seen only over white matter tracts in the cerebellum. Bar = 500/zm.
236
Fig. 4. Induction of c-los in the rat brain following 20 rain of forebrain ischemia. A: hypothalamus and amygdala, strong hybridization signal over the posterodorsal part of the medial amygdala (MA) and adjacent intercalated nuclei, the ventromedial hypothalamus (VMH) and the dorsal hypothalamus (DH) show more moderate signal levels 1 h after ischemia; 3, third ventricle. B: piriform cortex (PirCx) 1 h after ischemia. C: trapezoid body (TB) in the brainstem 1 h after ischemia. D: medial habenula and overlying ependyma ventral to dentate gyri show strong hybridization signal 30 min after ischemia. E: cortex (Cx) and dorsal caudate putamen (CP) show a diffuse hybridization pattern 1 h after ischemia. F: medial geniculate (MG), superficial gray layer of the superior colliculus (SC) and dorsal dentate gyrus (DG) with intense hilar labeling 1 h after ischemia. G: thaiamic nuclei 1 h after ischemia include the intermediodorsal (imd), centromedian (cm), reuniens (re), paracentral (pc) and lateral posterior (lp) nuclei. Bar = 500/zm, all same magnification.
237
Fig. 5. Involvement of catecholaminergic cell groups 1 h postischemia. A: no induction of c-fos was detected in the dopaminergic neurons of the substantia nigra compacta (SNc) below the medial lemniscus (ml) or the ventral tegmental area below the aqueduct (aq). B: the noradrenergic neurons of the locus ceruleus (LC) show a pronounced response. C: hybridization signal is also observed in the medulla oblongata in a distribution conforming to that of the adrenergic neurons in the ventrolateral medulla (C1 cell group) and the nucleus tractus solitarius (C2 cell group). Bars = 500 #m. regions of the hippocampus proper show greater c-los and c-jun m R N A levels at i h postischemia (Figs. 1D and 2D). Interestingly, the ependyma of the lateral ventricles, especially the ependyma overlying the medial habenula (Fig. 4D) and the ependyma of the third ventricle show a pronounced but short-lived expression of both c-los and c-jun that vanishes by 1 h of recovery. Furthermore, there appear to be an increased number of grains over the dentate gyrus and the CA3 region of the hippocampus at 1 h that are diffuse in their distribution and not as densely packed as over easily identified neuronal layers and the presumed hilar neurons; the origin of this increased 'background' is unknown but may reflect proto-oncogene expression in glial cells and is not observed at later timepoints. The CA1 region of the hippocampus appears to be fully involved at 2 and 3 h after ischemia, as seen in panels E and F in Figs. 1 and 2. A reduction of hybridization signal for c-los and c-jun to near baseline levels was seen at 6 h after reperfusion (data not shown). Photographs in G and H in Figs. 1 and 2 depict the changes in the hippocampus at later timepoints: at the 24 h timepoint in panels G, moderate hybridization signal for c-los and c-jun is seen over the CA1 region of the hippocampus; one week after ischemia in panels H, a profound loss of CA1 neurons is visible without any detectable hybridization signal for either proto-oncogene. In contrast, the evolution of events in the cerebellum
occurs at a quicker pace: the granule cell layer of the cerebellum shows a strong but patchy induction of the c-los and c-jun genes at 30 min that dissipates more rapidly (Fig. 3B). One hour after ischemia, disseminated regions of high signal are seen in the granule cell layer and appear associated with local Purkinje cells, moderate levels of hybridization signal are apparent in the molecular layer (Fig. 3C). Three hours after ischemia, very weak hybridization is detectable in the white matter of the cerebellum (Fig. 3D). Here again the time course of c-los and c-jun gene activation are virtually identical. Only weak hybridization signal could be detected in the caudate putamen at any of the timepoints examined, despite the fact that the dorsolateral spiny neurones in this brain region are known to be prone to a similar neuronal vulnerability as their counterparts in the dorsal hippocampus. Fig. 4E shows the weak hybridization pattern in the dorsal caudate and overlying cortex at 1 h postischemia. The cortex has a diffuse hybridization pattern that seems to involve all neuronal layers. Several other brain regions displayed robust, simultaneous c-los and c-jun expression. The most dramatic responses were seen in the medial habenula (Figs. 1C and 2C as well as 4D in greater detail), piriform cortex (Fig. 4B), amygdala (Fig. 4A) and medial geniculate (Fig. 4C), all areas with known interconnections to the hippocampus. The thalamus and hypothalamus had weaker and more delayed responses (Fig. 4A,G). Inter-
238 estingly, the ependyma of the lateral ventricles and third ventricle exhibited an impressive and early expression of both proto-oncogenes (Figs. 1C and 2C, also 4D). In the brainstem, strong hybridization signal was observed over the trapezoid body, while the locus ceruleus (Fig. 5A) and the adrenergic neurons of the ventrolateral medula (Fig. 5B) and nucleus tractus solitarius (Fig. 5C) showed a more moderate hybridization signal. DISCUSSION These experiments indicate that there is a stereotypic pattern of c-los and c-]un gene activation in several parts of the brain following transient forebrain ischemia. Because of our interest in the early behavior of those neuronal groups that are the most vulnerable to ischemia, we focused first on the changes in the hippocampus and striatum. It is well known that this particular method of inducing transient forebrain ischemia causes gradually evolving and reproducible neuronal damage and cell death that is most pronounced and measurable in the CA1 region of the hippocampus and dorsolateral caudate 27'32'40 and causes impaired functional outcome 12'49'5°. There is also clear evidence of severe damage in another hippocampal region, namely the hilus of the dentate gyrus 28. This latter phenomenon precedes the neuronal degeneration observed in the CA1 region and in the dorsolateral caudate. These differential temporal kinetics may provide therapeutic windows of opportunity and may be mediated by activation of immediate early genes. However, the present results demonstrate that c-los and c-jun are expressed in a variety of neurons as well as choroid and ependymal cells and possibly glial cells of which the vast majority survive transient forebrain ischemia. Expression of c-los has been demonstrated immunohistochemically in glial cells in vivo following ischemia 24, brain injury 16, heat shock 17, pharmacologic elevation of cAMP TM and hypoxia-ischemia in infant rats 23 as well as in cultured astrocytes exposed to neurotransmitters 2. Thus the expression of these protooncogenes does not appear to be linked to susceptibility for ischemic injury in the vast majority of neurons and other constituent cells of the brain. The expression of c-los and c-]un in the hippocampus occurs early during the early re-perfusion phase and peaks in a fashion that may reflect the trisynaptic circuitry of the hippocampus (entorhinal cortex --~ dentate gyrus (DG) --* CA3 --~ CA1). This pattern has been previously observed by Morgan et al. in penetylenetetrazole-induced seizures in mice 36. These authors demonstrated a stepwise increase in c-Fos immunoreactivity which was barely detectable at 30 min and pronounced at 1-2 h in the granule ceils of the dentate
gyrus and at 2-3 h in the pyramidal cells of the hippocampus proper. This temporal sequence in c-Fos protein production in the hippocampus seems compatible with the more rapidly occurring accumulation of c-fos m R N A observed here. Both c-fos and c-]un are expressed in a bitemporal pattern in the selective vulnerable CA1 region of the hippocampus. The second peak occurs at 24-48 h and coincides with the period of neuronal degeneration. The basis for this phenomenon is unknown but may reflect calcium ion influx which has been shown to be a potent trigger of c-los gene transcription in tissue culture systems 35. Accumulation of calcium and progression to ischemic cell death has been well documented in the brain 13'25. In fact, the hilar cells of the dentate gyrus which display early vulnerability to ischemia 28 show a rapid influx of radiolabeled calcium after ischemia 3. This matches the early induction of c-los and c-jun mRNAs in the dentate hilus observed here. Similarly, the second peak in c-los and c-]un expression in the CA1 region may be due to calcium pump failure in degenerating pyramidal cells which could result in the repeated activation of these immediate early genes. Surprisingly, the caudate nucleus exhibited a much less distinct response to ischemia than expected in that only moderate hybridization signal could be discerned over individual neurons that were widely distributed throughout the caudate nucleus. No second peak of c-fos or c-]un expression, as occurs in the hippocampus, was observed in the dorsolateral caudate where large numbers of spiny neurons degenerate. The cortex appeared to follow a similar pattern to the caudate nucleus with a diffuse, slow induction of both proto-oncogenes in all neuronal layers of the cortex. In fact, grain density appeared greater in the cortex than in the caudate putamen at any of the timepoints examined here. Another example of how different brain regions display varying time courses of c-fos and c-jun production and attrition is the cerebellum. Here a rapid and patchy induction of both proto-oncogenes in the granule cell layer, particularly over the median aspect of the cerebellum, was observed which peaked early and rapidly returned to baseline levels. It is unclear whether these differences in magnitude and time course of immediate early gene expression are based on variable gene transcription rates, different m R N A degradation rates or a combination of both, as in situ hybridization signal will only reflect net states at various timepoints. Interestingly, it appears that the Purkinje cells that are in close proximity to those areas of the granule cell layer that exhibit the most pronounced induction of c-fos and c-jun show a somewhat delayed response -- in analogy to the situation in the hippocampus -- but return much more
239 quickly to pre-ischemia levels. One possibility that may contribute to this speedy return to baseline in the cerebellum is that during four-vessel occlusion this brain region may be exposed to less severe ischemic injury: small collateral vessels could provide for marginal tissue perfusion thereby blunting proto-oncogene induction. This does not seem very likely, however, as great care was taken to suppress any collateral blood flow through an additional dorsal neck ligature which was placed during vertebral artery occlusion and tightened simultaneously to carotid artery occlusion. Furthermore, virtually no c-los or c-jun expression was detected in the cerebellum at the zero timepoint, exactly as observed in other brain areas, which suggests that re-perfusion must occur for a sufficient period of time for neurons to recover and regain the ability to mount an immediate early gene response. A more probable reason for the swift return of the cerebellum to baseline levels may be that the more rapid improvement of regional blood flow during the re-perfusion phase puts this structure at a relative advantage compared to the rest of the brain 22, so that cerebeUar perfusion may be much more rapidly optimized and any cellular derangements, such as elevated cytosolic calcium, more rapidly reversed. At present, it is unknown whether any of the target genes of the c-Fos-c-Jun dimer are involved in mediating ischemic injury but a particularly intriguing possibility is that a sudden surge in glutamate receptor production could initiate excitotoxic injury in the postischemic brain
by leading to a vicious cycle of receptor activation, calcium influx, proto-oncogene activation and consequent increased receptoi" synthesis. There is some evidence to suggest that c-fos expression is linked to N M D A glutamate receptor activation 21'48. Accordingly, any therapeutic intervention would have to be aimed at either (1) receptor activation and calcium influx, (2) the initial proto-oncogene activation (both of which would be difficult because of the short time involved) or (3) the response of target genes which may include the formation of functional as well as dysfunctional receptor and channel proteins. In conclusion, this study delineates the time course and localization of c-los and c-jun gene expression. Both c-los and c-jun are induced in a virtually identical pattern in the hippocampus and the rest of the brain. These results will help to define the temporal characteristics of proto-oncogene activation, and may be important in future studies that investigate whether the m R N A s for these genes are translated into functional proteins and what the possible target genes for the F o s - J u n dimer might be. Further combined neurobiologic and molecular biologic research will be necessary to establish the involvement of these proto-oncogenes in transcriptional responses after ischemia.
Acknowledgements. This research was supported by PHS Grants MH 40090 to B.T.V. and MH 24245 to T.H.J.
REFERENCES 1 Abate, C., Luk, D., Gentz, R., Rauscher, E and Curran, T., Expression and purification of the leucine zipper and DNAbinding domains of Fos and Jun: both Fos and Jun contact DNA directly, Proc. Natl. Acad. Sci. U.S.A., 87 (1990) 1032-1036. 2 Areander, A., de VeUis, J. and Herschman, H., Induction of c-los and TIS genes in cultured rat astrocytes by neurotransmitters, J. Neurosci. Res., 24 (1989) 107-114. 3 Beneviste, H. and Diemer, N.H., Early postischemic 45Ca accumulation in the rat dentate gyms, J. Cereb. Blood Flow Metab., 8 (1988) 713-719. 4 BuUitt, E., Induction of c-los-like protein within the lumbar spinal cord and thalamus of the rat following peripheral stimulation, Brain Research, 493 (1989) 391-397. 5 Ceccatelli, S., Villar, M., Goldstein, M. and H6kfelt, T., Expression c-Fos immunoreactivity in transmitter-characterized neurons after stress, Proc. Natl. Acad. Sci. U.S.A., 86 (1989) 9569-9573. 6 Chang, S., Squinto, S. and Harlan, H., Morphine activation of c-los in rat brain, Biochem. Biophys. Res. Commun., 157 (1988) 698-704. 7 Chiu, R., Boyle, W., Meek, J., Smeal, T., Hunter, T. and Karin, Y., The c-los protein interacts with c-jun/AP-1 tostimulate transcription of AP-l-responsive genes, Cell, 54 (1988) 541-552. 8 Chiu, R., Angel, P. and Karin, Y., Jun-B differs in its biological properties from, and is a negative regulator of, c-jun, Cell, 59 (1989) 979-986. 9 Cole, A., Saffen, D., Baraban, J. and Worley, P., Rapid in-
10
11 12
13 14 15 16 17 18
crease of an immediate early gene messenger RNA in hippocampal neurons by synaptic NMDA receptor activation, Nature, 340 (1989) 474-476. Curran, T., Gordon, M.B., Rubino, K.L. and Sambucetti, L.C., Isolation and characterization of the c-los (rat) cDNA and analysis of post-translational modification in vitro, Oncogene, 2 (1987) 79-84. Daval, J., Nakajima, T., Gleiter, C., Post, R. and Narangos, P., Mouse brain c-los mRNA distribution following a single electroconvulsive shock, J. Neurochem., 52 (1989) 1954-57. Davis, H.P. and Volpe, B.T., Memory performance after ischemia or neurotoxin damage in the hippocampus. In L.R. Squine and E. Lindenlaub (Eds.), The Biology of Memory, Schattauer, New York, 1990, pp. 477-505. Dienel, G., Regional accumulation of calcium in post-ischemic brain, J. Neurochem., 43 (1984) 913-925. Dragunow, M. and Robertson, H., Generalized seizures induce c-los protein(s) in mammalian neurons, Neurosci. Lett., 82 (1987) 157-161. Dragunow, M. and Robertson, H., Kindling stimulation induces c-los protein(s) in granule cells of the rat dentate gyrus, Nature, 329 (1987) 441-442. Dragunow, M. and Robertson, H., Brain injury induces c-los protein(s) in nerve and glial-like cells in adult mammalian brain, Brain Research, 455 (1988) 295-299. Dragunow, M., Currie, R., Robertson, H. and Faull, R., Heat shock induces c-fos protein-like immunoreactivity in glial cells in adult rat brain, Exp. Neurol., 106 (1989) 105-109. Dragunow, M. and Faull, R.L.M., Rolipram induces c-los
240 protein-like immunoreactivity in ependymal and glial-like cells in the adult rat brain, Brain Research, 501 (1989) 382-388. 19 Dragunow, M. and Faull, R.L.M., MK-801 induces c-los protein in thalamic and neocortical neurons of rat brain, Neurosci. Lett., 111 (1990) 39-45. 20 Dragunow, M., Presence and induction of Fos B-like immunoreactivity in neural, but not non-neural, cells in the adult rat brain, Brain Research, 533 (1990) 324-328. 21 Dragunow, M., Goulding, M., Faull, R.L.M., Ralph, R., Mee, E. and Frith, R., Induction of c-los mRNA and protein in neurons and glia after traumatic brain injury: pharmacological characterization, Exp. Neurol., 107 (1990) 236-248. 22 Ginsburg, M.D., Graham, D.I. and Busto, R,, Regional blood flow following forebrain ischemia in the rat, Ann. Neurol., 18 (1982) 470-481. 23 Gunn, A.J., Dragunow, M., Faull, R.L.M. and Gluckman, P.D., Effects of hypoxia-ischemia and seizures on neuronal and glial-like c-los protein in the infant rat, Brain Research, 531 (1990) 105-116. 24 Herrera, D.G. and Robertson, H.A., Unilateral induction of c-los protein in cortex following cortical devascularization, Brain Research, 503 (1989) 205-213. 25 Hossmann, K.A., Paschen, W. and Csiba, L., Relationship between calcium accumulation and recovery of cat brain after prolonged cerebral ischemia, J. Cereb. Blood Flow Metab., 3 (1983) 346-353. 26 Hunt, S., Pnin, A. and Evan, G., Induction of c-los-like protein in spinal cord neurons following sensory stimulation, Nature, 328 (1987) 632-634. 27 Ito, U., Spatz, J.T., Walker, J. and Klatzo, I., Experimental cerebral ischemia in mongolian gerbils. I. Light microscopic observations, Acta Neuropathol., 32 (1975) 209-223. 28 Johansen, F.F., Zimmer, J. and Diemer, N.H., Early loss of somatostatin neurons in dentate hilus after cerebral ischemia in the rat precedes CA-1 pyramidal loss, Acta Neuropathol., 73 (1987) 110-114. 29 Jergensen, M.B., Deckert, J., Wright, D.C. and Gehlert, D.R., Delayed c-los proto-oncogene expression in the rat hippocampus induced by transient global cerebral ischemia: an in situ hybridization study, Brain Research, 484 (1989) 393-8. 30 Kaczmarek, L., Siedlecki, J. and Danysz, W., Proto-oncogene c-los induction in rat hippocampus, Mol. Brain Res., 3 (1988) 183-186. 31 Kindy, M.S., Carney, J.P., Dempsey, R.J. and Carney, J.M., Ischemic induction of protooncogene expression in gerbil brain, J. Mol. Neurosci., 2 (1991) 217-228. 32 Kirino, T., Delayed neuronal death in the gerbil hippocampus following ischemia, Brain Research, 239 (1982) 57-69. 33 Koistinaho, J. and Yang, G., Induction of c-los protein-like immunoreactivity in the rat and hamster pineal gland after the onset of darkness, Histochemistry, 95 (1990) 73-76. 34 Men6trey, D., Gannon, A., Levine, J.D. and Basbaum, A.I., Expression of c-los protein in intemeurons and projection neurons of the rat spinal cord in response to noxious somatic, articular, and visceral stimulation, J. Comp. Neurol., 285 (1989)
177-195. 35 Morgan, J. and Curran, T., Role of ion flux in the control of c-los expression, Nature, 322 (1986) 552-555. 36 Morgen, J., Cohen, D., Hempstead, J. and Curran, T., Mapping patterns of c-los in the central nervous system after seizure, Science, 237 (1987) 192-197. 37 Muller, R., Bravo, R., Muller, D., Kurz, C. and Reiz, M., Different types of modification of c-los and its associated protein p39: modulation of DNA binding by phosphorylation, Oncogene Res., 2 (1987) 19-32. 38 Nowak Jr., T.S., Ikeda, J. and Nakajima, T., 70-kDa heat shock protein and c-los gene expression after transient ischemia, Stroke, Suppl. 21 (1990) 107-111. 39 Onodera, H., Kogure, K., Ono, Y., Igarashi, K., Kiyota, Y. and Nagaoka, A., Proto-oncogene c-los is transiently induced in the rat cerebral cortex after forebrain ischemia, Neurosci. Lett., 98 (1989) 101-104. 40 Pulsinelli, W,A., Brierley, J.B. and Plum, F., Temporal profile of neuronal damage in a model of transient ischemia, Ann. Neurol., 11 (1982) 491-98. 41 Rauscher III, F.J., Cohen, D.R., Curran, T., Bos, T.J., Vogt, P.K., Bohman, D., Tjian, R. and Franza, B.R., Fos-associated protein p39 is the product of the jun proto-oncogene, Science, 240 (1988) 1010-1016. 42 Rauscher III, F.J., Sambucetti, L., Curran, T., Distel, R. and Spiegelman, B., Common DNA binding site for los protein complexes and transcription factor AP-1, Cell, 52 (1988) 471480. 43 Robertson, H.A., Peterson, M.R., Murphy, K. and Robertson, G.S., Dl-dopamine receptor agonists selectively activate striatal c-los independent of rotational behavior, Brain Research, 503 (1989) 346-349. 44 Ryder, K. and Nathans, D., Induction of protooncogene c-jun by serum growth factors, Proc. Natl. Acad. Sci. U.S.A., 85 (1988) 8464-8467. 45 Sagar, S., Sharp, F. and Curran, T., Expression of c-los protein in the brain: metabolic mapping at the cellular level, Science, 240 (1988) 1328-1331. 46 Schutte, J., Viallet, J., Nan, M., Segal, S., Fedorko, J. and Minna, J., Jun-B inhibits and c-Fos stimulates the transforming and trans-activating activities of c-jun, Cell, 59 (1989) 987-997. 47 Sharp, E, Gonzales, M., Kisanaga, K., Mobley, W. and Sagar, S., Induction of the c-fos gene product in rat forebrain following cortical lesions and NGF injections, Neurosci. Left., 100 (1989) 117-122. 48 Uemura, Y., Kowall, N.W. and Beal, M.E, Global ischemia induces NMDA receptor-mediated c-los expression in neurons resistant to injury in gerbil hippocampus, Brain Research, 542 (1991) 343-347. 49 Volpe, B.T., Pulsinelli, W.A., Tribuna, J. and Davis, H.P., Behavioral performance of rats following transient forebrain ischemia, Stroke, 15 (1984) 558-562. 50 Volpe, B.T., Waczek, B. and Davis, H.P., Modified T-maze training demonstrates dissociated memory loss in rats with ischemic hippocampal injury, Behav. Brain Res., 27 (1988) 259-268.
231
BRES 17254
In situ hybridization analysis of c-los and c-jun expression in the rat brain following transient forebrain ischemia Thomas C. Wessel, Tong H. Joh and Bruce T. Volpe Department of Neurology and Neuroscience, Cornell University Medical College, The Burke Medical Research Institute, White Plains, NY 106o5 (U.S.A.) (Accepted 30 July 1991)
Key words: Transient ischemia; c-los; e-jun; Four-vessel occlusion; In situ hybridization; Proto-oncogene; Hippocampus
Early induction of the mRNAs encoding the c-Fos and c-Jun nuclear proteins was examined in rat brain by in situ hybridization at various timepoints following global forebrain ischemia by the method of four-vessel occlusion. All animals were subjected to 20 min of transient ischemia. This produced a pattern of proto-oncogene activation that was most intense in the granule cells of the dentate gyrus 30 rain after ischemia, while the hilar cells in the dentate and the pyramidal cells of the CA3 region in the hippocampus showed a more delayed but robust expression of these immediate early genes at 1 h. The neurons of the CA1 region exhibited a more moderate hybridization signal at 1-2 h postischemia. Very little hybridization signal for either immediate early gene could be detected in animals perfused with fixative immediately following ischemia, suggesting that cellular energy levels may have to be restored to a certain level before efficient de novo mRNA synthesis can occur. In the cerebellum, a similar temporal pattern was observed: the granule cells exhibited a prompt but patchy expression of c-los and c-jun that was followed by a delayed signal in the Purkinje cells. Without exception c-los and c-jun appeared to be expressed in unison, although the time course of c-los and c-jun mRNA accumulation and decay was different in various brain regions: invariably the cerebellum returned rapidly to its baseline with virtually no remaining signal at 3 h postischemia, while c-los and c-jun activation in the hippocampus remained high at 3 h and returned to baseline by 6 h. Several other brain regions showed early production of c-los and c-jun mRNAs, such as the medial habenula, piriform cortex, the amygdala, the centromedian, lateral posterior, paracentral, intermediodorsal and reuniens nuclei of the thalamus and the ventromedial and dorsal nuclei of the hypothalamus; in the hrainstem, the trapezoid body and the noradrenergic neurons of the locus ceruleus as well as the adrenergic neurons in the ventrolateral medulla (C1 group) and nucleus tractus solitarius (C2 group) regions displayed slightly less intense hybridization signals. In addition, the ependyma of the lateral ventricles and the third ventricle showed a prompt albeit short-lived production of c-los and c-jun mRNAs. Sham-operated animals as well as animals that had survived to one week postischemia showed either no or only trace levels of hydbridization signal. Interestingly, a second peak of c-los and c-jun expression was observed in the CA1 pyramidal cells at 24-48 h postischemia when this population of cells undergoes a rapidly progressive decline in viable neurons. In contrast, the small- and medium-sized spiny neurons of the dorsolateral striatum that are also subject to an analogous selective neuronal degeneration displayed weak early and no late c-los or c-jun expression. No other brain areas with this type of delayed protooncogene activation were detected. These results indicate that the proto-oncogenes c-los and c-jun are rapidly activated in the brain following a transient ischemic insult in a temporal and spatial pattern that is distinct from the pattem seen in drug-induced seizure activity and other paradigms.
INTRODUCTION The expression of the proto-oncogene c-los is rapidly induced in the central nervous system by a variety of physiologic processes and experimental manipulations, such as peripheral nerve stimulation4'26'34, intracerebral electrical stimulation45 and kindling9'15, circadian rhythms 33, t r a u m a 16'3°'47, seizure activity 1xA4'36, pharmacologic responses 5'6'19'43 and ischemia 24'29'31'38'39. Of the studies that demonstrate c-los responses in the ischemic rat model, two have investigated the production of the specific m R N A for c-los by in situ hybridization and Northern blot analysis, respectively: JCrgensen et al. demonstrated the expression of c-los m R N A at 24-48 h
in the selectively vulnerable CA1 region of the hippocampus 29, while O n o d e r a et al. detected the presence of c-los m R N A in the postischemic rat brain on N o r t h e r n blot at several timepoints, starting at 30 min postischemia 39. Thus the precise localization of c-los expression during early ischemia in the rat is u n k n o w n . I n the gerbil, Nowak et al. examined the induction of 70 kDa heat shock protein m R N A and c-los m R N A by in situ hybridization, and demonstrated a strong signal over the granule cells of the dentate gyrus 30 min after transient ischemia of 5 min duration aS. Recently, Kindy et al. have investigated the temporal response of c-los and c-jun expression in the mongolian gerbil by Northern blot analysis31, and have found that the magnitude of m R N A
Correspondence: T.C. Wessel, Department of Neurology and Neuroscience, Cornell University Medical College, The Burke Medical Research Institute, 785 Mamaroneck Avenue, White Plains, NY 10605, U.S.A.
232 accumulation for these proto-oncogenes correlated with the duration of ischemia. The strongest signal was detected at the 30 min timepoint after cessation of ischemia and there was a rapid decay in m R N A levels back to baseline within 3 h. In order to visualize the neuronal and possibly glial, ependymal or endothelial involvement in proto-oncogene expression, we have focused on the dynamic changes in c-fos and c-jun expression that take place in the postischemic rat brain following 20 min of transient global ischemia. The protein products of these and related members of the fos and ]un proto-oncogene families are thought to form heterodimers 4t that directly contact D N A binding sites I and function as transcriptional regulators at AP-1 binding sites 7'42. The binding of these heterodimers to the AP-1 site is modulated by phosphorylation 37. Various homo- and heterodimer combinations can be formed and there is some evidence to suggest that the relative levels of these multiple homo- and heterodimers may have positive as well as negative transcriptional function s'2°'46. We have included c-jun in situ hybridization in the present study because both c-Fos and c-Jun phosphoproteins are required to form active heterodimers before binding to AP-1 sites and initiating transcription of target genes. The target genes for the Fos-Jun heterodimers in the hippocampus and other brain regions are unknown. It is conceivable that these proto-oncogenes set a cascade of events in action that leads to neuronal destruction in some cell groups, while it may contribute to the survival of other cells depending on their genetically preprogrammed repertoire. MATERIALS AND METHODS Animals were subjected to transient forebrain ischemia by a method previously described 4° and modified: on day 1, the rats were anesthesized by inhalation of 2% halothane mixed with 20% oxygen and 80% nitrogen, the carotid arteries were encircled with silastic ligatures and the vertebral arteries were permanently occluded by electrocautery. Sham-operated control rats also underwent halothane anesthesia, skin incisions and carotid manipulations. The rats were allowed to recover for 24 h. On day 2, forebrain ischemia was induced in awake animals by tightening the carotid ligatures. Rats that lose their righting reflex within 1 min of carotid clamping and for the subsequent 20 min have been shown to have their cerebral blood flow (CBF) reduced to <10 ml/100 g/rain,
approximately 10% of that in control rats. CBF of > 10 ml/100 g/min does not reproducibly impair the righting reflex and produces variable ischemic injury. The loss of the righting reflex was always associated with fixed and dilated pupils. Rats that lost their righting reflex for 20 min were classified as ischemic, whereas those animals that did not were eliminated from the experiment. During ischemia, body temperature was maintained at 37 °C by a heating lamp connected to a rectal thermistor. The carotid ligatures were released after 20 min, at which time the rats generally remained unresponsive and showed no righting reflex, while rapidly regaining temperature homeostasis. Animals were allowed to recover for variable periods and then anesthesized at 30 min, 1 h, 2 h, 3 h, 6 h, 24 h and 1 week. Rats in the 0 min time group were immediately perfused transcardially following the opening of the carotid ligatures. Animals that developed seizures during the ischemic period or during the first 3 h postoperatively (at which point they were returned to the animal room) were excluded from this study. For in situ hybridization, the postischemic rats were anesthesized with pentobarbital and rapidly perfused transcardially with 0.9% sodium chloride, containing 0.5% nitrite and 1000 U heparin/100 ml. This was followed by rapid perfusion with ice-cold 4% formaldehyde in 0.1 M sodium phosphate buffer. Brains were immediately removed, cut into blocks, and submerged in the ice-cold fixative where they remained for one hour. The blocks were then rinsed twice with phosphate buffer and cryoprotected by storing the tissue in 30% sucrose overnight at 4 °C. Tissue sections of 40/~m thickness were cut on a sliding microtome and stored in 20 ml glass vials filled with 2x SCC (0.3 M sodium chloride/0.03 M sodium citrate) with 10 mM D T r at 4 °C. This storage solution was then replaced with prehybridization buffer containing 50% formaldehyde, 10% dextran sulfate, 2x SSC, l x Denhardt's solution, 50 mM DTT, and 0.5 mg/ml sonicated and denatured salmon sperm DNA. Prehybridization was carried out for 60 rain at 48 °C. For the main experiments, labeled denatured c-los probe or c-jun antisense oligonucleotide (1 × 107 cpm) was then added to the vial in 100 /~l of the same prehybridization buffer. In control experiments, labeled sense and antisense c-los and c-jun oligonucleotide were used to (1) confirm the specificity of labeling for both c-los and c-jun antisense probes, (2) prove loss of hybridization under cold antisense probe excess and (3) verify that the hybridization patterns observed for the c-fos antisense oligonucleotide and cDNA probe are identical. Hybridization was performed overnight at 48 °C. After extensive washes in decreasing concentrations of SSC, 10 min steps of 1:1 dilutions starting at 2x SSC and ending with 0.1× SSC, tissue sections were mounted onto gelatin-subbed slides, air-dried and dehydrated through graded ethanols (70, 90, 100%). Finally the slides were dipped into undiluted Kodak NTB-2 in the darkroom. After storage in light-tight boxes at 4 °C for 7-21 days, the slides were developed in Kodak D-19, counterstained with Cresyl violet and cover-slipped with Permount. A great advantage of this particular technique is that all tissue sections from a particular timepoint and animal can be hybridized with a given radiolabeled probe under exactly the same conditions, as all floating sections are exposed to the same hybridization buffer, probe concentrations and washing procedure. The probes used in these experiments were (1) sense and antisense rat c-fos oligonucleotides corresponding to bases 138-173 (36-mer) of the published
Fig. 1. Darkfield photomicrographs showing time course of c-los expression in the rat hippocampus after 20 min of ischemia. A: shamoperated control animal; open arrows indicate pyramidal cell layer of the CA1 region which is selectively vulnerable to ischemia. B: faint outline of granule cells in the dentate gyms (DG) in an ischemic animal perfused immediately following opening of the carotid ligatures. C: after 30 min of recovery, strong expression of c-los is apparent in the DG and medial habenula as well as the ependyma. D: 60 min recovery, involvement of the hilar neurons of the DG, CA3 neurons and to a lesser extent the CA1 region. E: 2 h recovery, some decrease in D G labeling. F: 3 h recovery, weak residual signal over the DG which vanishes by 6 h (not shown). G: 24 h after ischemia, CA1 region is selectively outlined (arrows). H: 1 week after ischemia, note near complete degeneration of CA1 pyramidal neurons (open arrows). Bar = 500 #m.
233
4~
235 Fig. 2. Photomicrographs showing time course of c-jun expression in the rat hippocampus in adjacent sections to those in Fig. 1. A: sham-operated control animal. B: 0 rain. C: 30 rain. D: 1 h. E: 2 h. F: 3 h. G: 24 h. H: 1 week. Note identical induction of e-jun in the CA1 region of the hippocampus 24 h after ischemia. Bar = 500/~m.
sequence x°, (2) sense and antisense oligonucleotides to mouse c-]un base number 297-332 (36-mer) of the published sequence44 and (3) a mouse v-los probe (ATCC, Rockville, MD 20585) digested with Psfl to yield a 1.0 kb fragment for comparison to the antisense e-fos oligoprobe. The oligonucleotides were synthesized on a Biosearch 8650 DNA synthesizer and purified in a standard fashion. Labeling efficiency varied between 0.2-0.5 x 10~ cpm/ng for the oligonucleotides and 0.8-1.6 × 106 epm/ng for the c-los eDNA probe.
RESULTS Twenty minutes of forebrain ischemia leads to a r a p i d and transient expression of c-los and c-jun in several regions of the rat brain. Figs. 1 and 2 illustrate the t e m p o r a l p a t t e r n o b s e r v e d in the h i p p o c a m p u s for c-los
and c-jun, respectively. Control animals showed no or only minimal labeling of the granule o r p y r a m i d a l cells in the hippocampus. A n i m a l s that were peffused with fixative i m m e d i a t e l y following 20 rain of ischemia, t h e r e b y p r e s u m a b l y preventing any meaningful reperfusion o r recovery (Figs. 1B and 2B), d e m o n s t r a t e d very low hybridization signal for c-los and c-jun in the d e n t a t e gyrus and no signal in the r e m a i n d e r of the brain. T h e subsequent p h o t o g r a p h s (Figs. 1 C - F and 2 C - F ) depict the dynamic changes that take place during the early reperfusion period. It appears that the hybridization signal is already quite p r o n o u n c e d over the granule cells of the d e n t a t e gyrus at 30 rain (Figs. 1C and 2C), while the hilar neurons of the d e n t a t e gyrus as well as the C A 3 and C A 2
Fig. 3. Photomicrographs of cerebellar c-los expression postischemia. A: control animal, gr, granule cell layer; m, molecular layer; wm, white matter tracts. B: patchy and partially confluent areas of c-los expression in the granule cell layer 30 rain after ischemia. C: rapid resolution of c-los expression in the granule cell layer associated with a moderate density of grains over the molecular layer and an interrupted line of hybridization signal over the Purkinje cells (arrow). D: 3 h after ischemia faint signal is seen only over white matter tracts in the cerebellum. Bar = 500/zm.
236
Fig. 4. Induction of c-los in the rat brain following 20 rain of forebrain ischemia. A: hypothalamus and amygdala, strong hybridization signal over the posterodorsal part of the medial amygdala (MA) and adjacent intercalated nuclei, the ventromedial hypothalamus (VMH) and the dorsal hypothalamus (DH) show more moderate signal levels 1 h after ischemia; 3, third ventricle. B: piriform cortex (PirCx) 1 h after ischemia. C: trapezoid body (TB) in the brainstem 1 h after ischemia. D: medial habenula and overlying ependyma ventral to dentate gyri show strong hybridization signal 30 min after ischemia. E: cortex (Cx) and dorsal caudate putamen (CP) show a diffuse hybridization pattern 1 h after ischemia. F: medial geniculate (MG), superficial gray layer of the superior colliculus (SC) and dorsal dentate gyrus (DG) with intense hilar labeling 1 h after ischemia. G: thaiamic nuclei 1 h after ischemia include the intermediodorsal (imd), centromedian (cm), reuniens (re), paracentral (pc) and lateral posterior (lp) nuclei. Bar = 500/zm, all same magnification.
237
Fig. 5. Involvement of catecholaminergic cell groups 1 h postischemia. A: no induction of c-fos was detected in the dopaminergic neurons of the substantia nigra compacta (SNc) below the medial lemniscus (ml) or the ventral tegmental area below the aqueduct (aq). B: the noradrenergic neurons of the locus ceruleus (LC) show a pronounced response. C: hybridization signal is also observed in the medulla oblongata in a distribution conforming to that of the adrenergic neurons in the ventrolateral medulla (C1 cell group) and the nucleus tractus solitarius (C2 cell group). Bars = 500 #m. regions of the hippocampus proper show greater c-los and c-jun m R N A levels at i h postischemia (Figs. 1D and 2D). Interestingly, the ependyma of the lateral ventricles, especially the ependyma overlying the medial habenula (Fig. 4D) and the ependyma of the third ventricle show a pronounced but short-lived expression of both c-los and c-jun that vanishes by 1 h of recovery. Furthermore, there appear to be an increased number of grains over the dentate gyrus and the CA3 region of the hippocampus at 1 h that are diffuse in their distribution and not as densely packed as over easily identified neuronal layers and the presumed hilar neurons; the origin of this increased 'background' is unknown but may reflect proto-oncogene expression in glial cells and is not observed at later timepoints. The CA1 region of the hippocampus appears to be fully involved at 2 and 3 h after ischemia, as seen in panels E and F in Figs. 1 and 2. A reduction of hybridization signal for c-los and c-jun to near baseline levels was seen at 6 h after reperfusion (data not shown). Photographs in G and H in Figs. 1 and 2 depict the changes in the hippocampus at later timepoints: at the 24 h timepoint in panels G, moderate hybridization signal for c-los and c-jun is seen over the CA1 region of the hippocampus; one week after ischemia in panels H, a profound loss of CA1 neurons is visible without any detectable hybridization signal for either proto-oncogene. In contrast, the evolution of events in the cerebellum
occurs at a quicker pace: the granule cell layer of the cerebellum shows a strong but patchy induction of the c-los and c-jun genes at 30 min that dissipates more rapidly (Fig. 3B). One hour after ischemia, disseminated regions of high signal are seen in the granule cell layer and appear associated with local Purkinje cells, moderate levels of hybridization signal are apparent in the molecular layer (Fig. 3C). Three hours after ischemia, very weak hybridization is detectable in the white matter of the cerebellum (Fig. 3D). Here again the time course of c-los and c-jun gene activation are virtually identical. Only weak hybridization signal could be detected in the caudate putamen at any of the timepoints examined, despite the fact that the dorsolateral spiny neurones in this brain region are known to be prone to a similar neuronal vulnerability as their counterparts in the dorsal hippocampus. Fig. 4E shows the weak hybridization pattern in the dorsal caudate and overlying cortex at 1 h postischemia. The cortex has a diffuse hybridization pattern that seems to involve all neuronal layers. Several other brain regions displayed robust, simultaneous c-los and c-jun expression. The most dramatic responses were seen in the medial habenula (Figs. 1C and 2C as well as 4D in greater detail), piriform cortex (Fig. 4B), amygdala (Fig. 4A) and medial geniculate (Fig. 4C), all areas with known interconnections to the hippocampus. The thalamus and hypothalamus had weaker and more delayed responses (Fig. 4A,G). Inter-
238 estingly, the ependyma of the lateral ventricles and third ventricle exhibited an impressive and early expression of both proto-oncogenes (Figs. 1C and 2C, also 4D). In the brainstem, strong hybridization signal was observed over the trapezoid body, while the locus ceruleus (Fig. 5A) and the adrenergic neurons of the ventrolateral medula (Fig. 5B) and nucleus tractus solitarius (Fig. 5C) showed a more moderate hybridization signal. DISCUSSION These experiments indicate that there is a stereotypic pattern of c-los and c-]un gene activation in several parts of the brain following transient forebrain ischemia. Because of our interest in the early behavior of those neuronal groups that are the most vulnerable to ischemia, we focused first on the changes in the hippocampus and striatum. It is well known that this particular method of inducing transient forebrain ischemia causes gradually evolving and reproducible neuronal damage and cell death that is most pronounced and measurable in the CA1 region of the hippocampus and dorsolateral caudate 27'32'40 and causes impaired functional outcome 12'49'5°. There is also clear evidence of severe damage in another hippocampal region, namely the hilus of the dentate gyrus 28. This latter phenomenon precedes the neuronal degeneration observed in the CA1 region and in the dorsolateral caudate. These differential temporal kinetics may provide therapeutic windows of opportunity and may be mediated by activation of immediate early genes. However, the present results demonstrate that c-los and c-jun are expressed in a variety of neurons as well as choroid and ependymal cells and possibly glial cells of which the vast majority survive transient forebrain ischemia. Expression of c-los has been demonstrated immunohistochemically in glial cells in vivo following ischemia 24, brain injury 16, heat shock 17, pharmacologic elevation of cAMP TM and hypoxia-ischemia in infant rats 23 as well as in cultured astrocytes exposed to neurotransmitters 2. Thus the expression of these protooncogenes does not appear to be linked to susceptibility for ischemic injury in the vast majority of neurons and other constituent cells of the brain. The expression of c-los and c-]un in the hippocampus occurs early during the early re-perfusion phase and peaks in a fashion that may reflect the trisynaptic circuitry of the hippocampus (entorhinal cortex --~ dentate gyrus (DG) --* CA3 --~ CA1). This pattern has been previously observed by Morgan et al. in penetylenetetrazole-induced seizures in mice 36. These authors demonstrated a stepwise increase in c-Fos immunoreactivity which was barely detectable at 30 min and pronounced at 1-2 h in the granule ceils of the dentate
gyrus and at 2-3 h in the pyramidal cells of the hippocampus proper. This temporal sequence in c-Fos protein production in the hippocampus seems compatible with the more rapidly occurring accumulation of c-fos m R N A observed here. Both c-fos and c-]un are expressed in a bitemporal pattern in the selective vulnerable CA1 region of the hippocampus. The second peak occurs at 24-48 h and coincides with the period of neuronal degeneration. The basis for this phenomenon is unknown but may reflect calcium ion influx which has been shown to be a potent trigger of c-los gene transcription in tissue culture systems 35. Accumulation of calcium and progression to ischemic cell death has been well documented in the brain 13'25. In fact, the hilar cells of the dentate gyrus which display early vulnerability to ischemia 28 show a rapid influx of radiolabeled calcium after ischemia 3. This matches the early induction of c-los and c-jun mRNAs in the dentate hilus observed here. Similarly, the second peak in c-los and c-]un expression in the CA1 region may be due to calcium pump failure in degenerating pyramidal cells which could result in the repeated activation of these immediate early genes. Surprisingly, the caudate nucleus exhibited a much less distinct response to ischemia than expected in that only moderate hybridization signal could be discerned over individual neurons that were widely distributed throughout the caudate nucleus. No second peak of c-fos or c-]un expression, as occurs in the hippocampus, was observed in the dorsolateral caudate where large numbers of spiny neurons degenerate. The cortex appeared to follow a similar pattern to the caudate nucleus with a diffuse, slow induction of both proto-oncogenes in all neuronal layers of the cortex. In fact, grain density appeared greater in the cortex than in the caudate putamen at any of the timepoints examined here. Another example of how different brain regions display varying time courses of c-fos and c-jun production and attrition is the cerebellum. Here a rapid and patchy induction of both proto-oncogenes in the granule cell layer, particularly over the median aspect of the cerebellum, was observed which peaked early and rapidly returned to baseline levels. It is unclear whether these differences in magnitude and time course of immediate early gene expression are based on variable gene transcription rates, different m R N A degradation rates or a combination of both, as in situ hybridization signal will only reflect net states at various timepoints. Interestingly, it appears that the Purkinje cells that are in close proximity to those areas of the granule cell layer that exhibit the most pronounced induction of c-fos and c-jun show a somewhat delayed response -- in analogy to the situation in the hippocampus -- but return much more
239 quickly to pre-ischemia levels. One possibility that may contribute to this speedy return to baseline in the cerebellum is that during four-vessel occlusion this brain region may be exposed to less severe ischemic injury: small collateral vessels could provide for marginal tissue perfusion thereby blunting proto-oncogene induction. This does not seem very likely, however, as great care was taken to suppress any collateral blood flow through an additional dorsal neck ligature which was placed during vertebral artery occlusion and tightened simultaneously to carotid artery occlusion. Furthermore, virtually no c-los or c-jun expression was detected in the cerebellum at the zero timepoint, exactly as observed in other brain areas, which suggests that re-perfusion must occur for a sufficient period of time for neurons to recover and regain the ability to mount an immediate early gene response. A more probable reason for the swift return of the cerebellum to baseline levels may be that the more rapid improvement of regional blood flow during the re-perfusion phase puts this structure at a relative advantage compared to the rest of the brain 22, so that cerebeUar perfusion may be much more rapidly optimized and any cellular derangements, such as elevated cytosolic calcium, more rapidly reversed. At present, it is unknown whether any of the target genes of the c-Fos-c-Jun dimer are involved in mediating ischemic injury but a particularly intriguing possibility is that a sudden surge in glutamate receptor production could initiate excitotoxic injury in the postischemic brain
by leading to a vicious cycle of receptor activation, calcium influx, proto-oncogene activation and consequent increased receptoi" synthesis. There is some evidence to suggest that c-fos expression is linked to N M D A glutamate receptor activation 21'48. Accordingly, any therapeutic intervention would have to be aimed at either (1) receptor activation and calcium influx, (2) the initial proto-oncogene activation (both of which would be difficult because of the short time involved) or (3) the response of target genes which may include the formation of functional as well as dysfunctional receptor and channel proteins. In conclusion, this study delineates the time course and localization of c-los and c-jun gene expression. Both c-los and c-jun are induced in a virtually identical pattern in the hippocampus and the rest of the brain. These results will help to define the temporal characteristics of proto-oncogene activation, and may be important in future studies that investigate whether the m R N A s for these genes are translated into functional proteins and what the possible target genes for the F o s - J u n dimer might be. Further combined neurobiologic and molecular biologic research will be necessary to establish the involvement of these proto-oncogenes in transcriptional responses after ischemia.
Acknowledgements. This research was supported by PHS Grants MH 40090 to B.T.V. and MH 24245 to T.H.J.
REFERENCES 1 Abate, C., Luk, D., Gentz, R., Rauscher, E and Curran, T., Expression and purification of the leucine zipper and DNAbinding domains of Fos and Jun: both Fos and Jun contact DNA directly, Proc. Natl. Acad. Sci. U.S.A., 87 (1990) 1032-1036. 2 Areander, A., de VeUis, J. and Herschman, H., Induction of c-los and TIS genes in cultured rat astrocytes by neurotransmitters, J. Neurosci. Res., 24 (1989) 107-114. 3 Beneviste, H. and Diemer, N.H., Early postischemic 45Ca accumulation in the rat dentate gyms, J. Cereb. Blood Flow Metab., 8 (1988) 713-719. 4 BuUitt, E., Induction of c-los-like protein within the lumbar spinal cord and thalamus of the rat following peripheral stimulation, Brain Research, 493 (1989) 391-397. 5 Ceccatelli, S., Villar, M., Goldstein, M. and H6kfelt, T., Expression c-Fos immunoreactivity in transmitter-characterized neurons after stress, Proc. Natl. Acad. Sci. U.S.A., 86 (1989) 9569-9573. 6 Chang, S., Squinto, S. and Harlan, H., Morphine activation of c-los in rat brain, Biochem. Biophys. Res. Commun., 157 (1988) 698-704. 7 Chiu, R., Boyle, W., Meek, J., Smeal, T., Hunter, T. and Karin, Y., The c-los protein interacts with c-jun/AP-1 tostimulate transcription of AP-l-responsive genes, Cell, 54 (1988) 541-552. 8 Chiu, R., Angel, P. and Karin, Y., Jun-B differs in its biological properties from, and is a negative regulator of, c-jun, Cell, 59 (1989) 979-986. 9 Cole, A., Saffen, D., Baraban, J. and Worley, P., Rapid in-
10
11 12
13 14 15 16 17 18
crease of an immediate early gene messenger RNA in hippocampal neurons by synaptic NMDA receptor activation, Nature, 340 (1989) 474-476. Curran, T., Gordon, M.B., Rubino, K.L. and Sambucetti, L.C., Isolation and characterization of the c-los (rat) cDNA and analysis of post-translational modification in vitro, Oncogene, 2 (1987) 79-84. Daval, J., Nakajima, T., Gleiter, C., Post, R. and Narangos, P., Mouse brain c-los mRNA distribution following a single electroconvulsive shock, J. Neurochem., 52 (1989) 1954-57. Davis, H.P. and Volpe, B.T., Memory performance after ischemia or neurotoxin damage in the hippocampus. In L.R. Squine and E. Lindenlaub (Eds.), The Biology of Memory, Schattauer, New York, 1990, pp. 477-505. Dienel, G., Regional accumulation of calcium in post-ischemic brain, J. Neurochem., 43 (1984) 913-925. Dragunow, M. and Robertson, H., Generalized seizures induce c-los protein(s) in mammalian neurons, Neurosci. Lett., 82 (1987) 157-161. Dragunow, M. and Robertson, H., Kindling stimulation induces c-los protein(s) in granule cells of the rat dentate gyrus, Nature, 329 (1987) 441-442. Dragunow, M. and Robertson, H., Brain injury induces c-los protein(s) in nerve and glial-like cells in adult mammalian brain, Brain Research, 455 (1988) 295-299. Dragunow, M., Currie, R., Robertson, H. and Faull, R., Heat shock induces c-fos protein-like immunoreactivity in glial cells in adult rat brain, Exp. Neurol., 106 (1989) 105-109. Dragunow, M. and Faull, R.L.M., Rolipram induces c-los
240 protein-like immunoreactivity in ependymal and glial-like cells in the adult rat brain, Brain Research, 501 (1989) 382-388. 19 Dragunow, M. and Faull, R.L.M., MK-801 induces c-los protein in thalamic and neocortical neurons of rat brain, Neurosci. Lett., 111 (1990) 39-45. 20 Dragunow, M., Presence and induction of Fos B-like immunoreactivity in neural, but not non-neural, cells in the adult rat brain, Brain Research, 533 (1990) 324-328. 21 Dragunow, M., Goulding, M., Faull, R.L.M., Ralph, R., Mee, E. and Frith, R., Induction of c-los mRNA and protein in neurons and glia after traumatic brain injury: pharmacological characterization, Exp. Neurol., 107 (1990) 236-248. 22 Ginsburg, M.D., Graham, D.I. and Busto, R,, Regional blood flow following forebrain ischemia in the rat, Ann. Neurol., 18 (1982) 470-481. 23 Gunn, A.J., Dragunow, M., Faull, R.L.M. and Gluckman, P.D., Effects of hypoxia-ischemia and seizures on neuronal and glial-like c-los protein in the infant rat, Brain Research, 531 (1990) 105-116. 24 Herrera, D.G. and Robertson, H.A., Unilateral induction of c-los protein in cortex following cortical devascularization, Brain Research, 503 (1989) 205-213. 25 Hossmann, K.A., Paschen, W. and Csiba, L., Relationship between calcium accumulation and recovery of cat brain after prolonged cerebral ischemia, J. Cereb. Blood Flow Metab., 3 (1983) 346-353. 26 Hunt, S., Pnin, A. and Evan, G., Induction of c-los-like protein in spinal cord neurons following sensory stimulation, Nature, 328 (1987) 632-634. 27 Ito, U., Spatz, J.T., Walker, J. and Klatzo, I., Experimental cerebral ischemia in mongolian gerbils. I. Light microscopic observations, Acta Neuropathol., 32 (1975) 209-223. 28 Johansen, F.F., Zimmer, J. and Diemer, N.H., Early loss of somatostatin neurons in dentate hilus after cerebral ischemia in the rat precedes CA-1 pyramidal loss, Acta Neuropathol., 73 (1987) 110-114. 29 Jergensen, M.B., Deckert, J., Wright, D.C. and Gehlert, D.R., Delayed c-los proto-oncogene expression in the rat hippocampus induced by transient global cerebral ischemia: an in situ hybridization study, Brain Research, 484 (1989) 393-8. 30 Kaczmarek, L., Siedlecki, J. and Danysz, W., Proto-oncogene c-los induction in rat hippocampus, Mol. Brain Res., 3 (1988) 183-186. 31 Kindy, M.S., Carney, J.P., Dempsey, R.J. and Carney, J.M., Ischemic induction of protooncogene expression in gerbil brain, J. Mol. Neurosci., 2 (1991) 217-228. 32 Kirino, T., Delayed neuronal death in the gerbil hippocampus following ischemia, Brain Research, 239 (1982) 57-69. 33 Koistinaho, J. and Yang, G., Induction of c-los protein-like immunoreactivity in the rat and hamster pineal gland after the onset of darkness, Histochemistry, 95 (1990) 73-76. 34 Men6trey, D., Gannon, A., Levine, J.D. and Basbaum, A.I., Expression of c-los protein in intemeurons and projection neurons of the rat spinal cord in response to noxious somatic, articular, and visceral stimulation, J. Comp. Neurol., 285 (1989)
177-195. 35 Morgan, J. and Curran, T., Role of ion flux in the control of c-los expression, Nature, 322 (1986) 552-555. 36 Morgen, J., Cohen, D., Hempstead, J. and Curran, T., Mapping patterns of c-los in the central nervous system after seizure, Science, 237 (1987) 192-197. 37 Muller, R., Bravo, R., Muller, D., Kurz, C. and Reiz, M., Different types of modification of c-los and its associated protein p39: modulation of DNA binding by phosphorylation, Oncogene Res., 2 (1987) 19-32. 38 Nowak Jr., T.S., Ikeda, J. and Nakajima, T., 70-kDa heat shock protein and c-los gene expression after transient ischemia, Stroke, Suppl. 21 (1990) 107-111. 39 Onodera, H., Kogure, K., Ono, Y., Igarashi, K., Kiyota, Y. and Nagaoka, A., Proto-oncogene c-los is transiently induced in the rat cerebral cortex after forebrain ischemia, Neurosci. Lett., 98 (1989) 101-104. 40 Pulsinelli, W,A., Brierley, J.B. and Plum, F., Temporal profile of neuronal damage in a model of transient ischemia, Ann. Neurol., 11 (1982) 491-98. 41 Rauscher III, F.J., Cohen, D.R., Curran, T., Bos, T.J., Vogt, P.K., Bohman, D., Tjian, R. and Franza, B.R., Fos-associated protein p39 is the product of the jun proto-oncogene, Science, 240 (1988) 1010-1016. 42 Rauscher III, F.J., Sambucetti, L., Curran, T., Distel, R. and Spiegelman, B., Common DNA binding site for los protein complexes and transcription factor AP-1, Cell, 52 (1988) 471480. 43 Robertson, H.A., Peterson, M.R., Murphy, K. and Robertson, G.S., Dl-dopamine receptor agonists selectively activate striatal c-los independent of rotational behavior, Brain Research, 503 (1989) 346-349. 44 Ryder, K. and Nathans, D., Induction of protooncogene c-jun by serum growth factors, Proc. Natl. Acad. Sci. U.S.A., 85 (1988) 8464-8467. 45 Sagar, S., Sharp, F. and Curran, T., Expression of c-los protein in the brain: metabolic mapping at the cellular level, Science, 240 (1988) 1328-1331. 46 Schutte, J., Viallet, J., Nan, M., Segal, S., Fedorko, J. and Minna, J., Jun-B inhibits and c-Fos stimulates the transforming and trans-activating activities of c-jun, Cell, 59 (1989) 987-997. 47 Sharp, E, Gonzales, M., Kisanaga, K., Mobley, W. and Sagar, S., Induction of the c-fos gene product in rat forebrain following cortical lesions and NGF injections, Neurosci. Left., 100 (1989) 117-122. 48 Uemura, Y., Kowall, N.W. and Beal, M.E, Global ischemia induces NMDA receptor-mediated c-los expression in neurons resistant to injury in gerbil hippocampus, Brain Research, 542 (1991) 343-347. 49 Volpe, B.T., Pulsinelli, W.A., Tribuna, J. and Davis, H.P., Behavioral performance of rats following transient forebrain ischemia, Stroke, 15 (1984) 558-562. 50 Volpe, B.T., Waczek, B. and Davis, H.P., Modified T-maze training demonstrates dissociated memory loss in rats with ischemic hippocampal injury, Behav. Brain Res., 27 (1988) 259-268.