- Email: [email protected]
Selective reinnervation of hippocampal area CA1 and the fascia dentata after destruction of CA3-CA4 afferents with kainic acid
Brain Research, 182 (1980) 1-9 © Elsevier/North-Holland Biomedical Press
1
Research Reports
SELECTIVE R E I N N E R V A T I O N OF H I P P O C A M P A L A R E A CA1 A N D T H E FASCIA D E N T A T A A F T E R D E S T R U C T I O N OF CA3-CA4 A F F E R E N T S W I T H K A I N I C ACID
J. VICTOR NADLER*, BRUCE W. PERRY and CARL W. COTMAN Department of Psychobiology, University of California, Irvine, Calif. 92717 (U.S.A.)
(Accepted May 31st, 1979)
SUMMARY Intraventricular injections of kainic acid were used to destroy the hippocampal CA3-CA4 cells, thus denervating the inner third of the molecular layer of the fascia dentata and stratum radiatum and stratum oriens of area CA1. The responses of intact afferents to such lesions were then examined histologically. The hippocampal mossy fibers densely reinnervated the inner portion of the dentate molecular layer after bilateral destruction of CA4 neurons and to a lesser extent after unilateral destruction. Septohippocampal fibers replaced CA4-derived fibers in the dentate molecular layer only after particularly extensive bilateral CA4 lesions. Medial perforant path fibers showed no anatomical response to any of these lesions. Neither septohippocampal, temporoammonic nor mossy fibers proliferated in or grew into the denervated laminae of area CA 1. These results show a preferential ordering in the reinnervation of dentate granule cells which is not readily explained by proximity to the degenerating fibers and also that removal of CA3-CA4-derived innervation more readily elicits translaminar growth in the fascia dentata than in area CA1. These results may be relevant to clinical situations in which neurons of the hippocampal end-blade are lost. INTRODUCTION The pyramidal cells of the hippocampal end-blade (areas ha-h5 of Rose) are among the most sensitive neurons in the brain to a variety of deleterious conditions, including anoxia, senile dementia and severe epilepsy1,16. In all these conditions pyramidal cells degenerate and are replaced by a proliferation of glial cells, but * To whom all correspondence should be addressed. Present address: Department of Pharmacology, Duke University Medical Center, Durham, N.C. 27710, U.S.A.
afferent fibers and fibers of passage in the area of the lesion are not directly damaged. The consequences for hippocampal connectivity are unknown, but are likely to be of importance, given the widely acknowledged behavioral role of the hippocampus. Hence we sought an animal model of this pathology. We have reported that intraventricular administration of the potent convulsant, kainic acid (KA), reproduces in rats this common form of hippocampal damage19, 20. Low doses of KA (0.5-4 nmol) preferentially destroy the CA3-CA4 pyramidal cells, which are considered homologous to the pyramidal cells of the human end-blade. These neurons have been shown to innervate and interconnect all parts of the hippocampal area6,10,~l,25,26. Accordingly, we may conclude that they play a central role in the functions of this region and that loss of their connections, as occurs in pathological states, would severely disrupt these functions. It was therefore of interest to determine whether heterologous afferents replace the innervation provided by CA3-CA4 cells, when the latter are destroyed with KA. Heterologous reinnervation has previously been demonstrated in the molecular layer of the fascia dentata denervated by a perforant path lesion2, 3. In the present studies we have examined the terminal fields of CA3-CA4 fibers in stratum oriens and stratum radiatum of area CA1 and in the molecular layer of the fascia dentata. MATERIALS AND METHODS KA was administered intraventricularly to adult male Sprague-Dawley ratslL Either 0.5 or 0.8/zg (2.34 or 3.75 nmol) of KA dissolved in 1/tl of Elliott's artificial CSF 4 was injected over a 30-min period. In most cases the surgical and injection procedures were then repeated on the contralateral side. Animals were given atropine sulfate (0.5-1 mg/kg, i.p.) as often as hourly to alleviate any respiratory distress. Histological studies were performed after a survival period of at least 30 days. Projections from the ipsilateral entorhinal cortex to the hippocampal formation were traced autoradiographically 5, mossy fiber boutons were visualized by use of the Timm's stain for heavy metals 7 and acetylcholinesterase histochemistry14 was employed to demonstrate the cholinergic septohippocampal fibers la,14,17,21,z4. RESULTS Intraventricular injection of 0.8/zg of KA consistently destroyed the pyramidal cells of the ipsilateral CA3 and CA4 areas, as previously reported 19,2°. Most of the neurons in the 'cell-poor' layer of area CA4, which provide the predominant associational and commissural innervation of the fascia dentata 12,2°,26,27, were also killed. KA destroyed a higher percentage of these cells in the septal two-thirds of the hippocampal formation than in the temporal third. The CA1 pyramidal cells were usually spared, but occasionally one or more clusters of these neurons were found to have been destroyed. In the most severe cases, most CA1 pyramidal cells were destroyed along with the CA3-CA4 cells. Few pyramidal cells of the hz zone (area CA2 and the adjacent part of area CA3a) and apparently no dentate granule cells were destroyed by the drug. Bilateral injection of either 0.5 or 0.8 /,tg of KA produced
lesions identical to this in both hippocampi (Fig. 1A, B). While KA at the doses used here destroyed neurons in several other regions of the brain 1a,2°, the only markedly sensitive areas known to project to the hippocampal formation were the nucleus reuniens of the thalamus a and layer IIl of the entorhinal cortex 23. Ceils in these areas project to stratum lacunosum-moleculare of area CA1, a lamina that does not receive a CA3-CA4 projection. Only very few neurons (perhaps 1-2 ~) in layer I1 of the entorhinal cortex, the predominant or exclusive source of hippocampal perforant path fibers 2a, were found to degenerate. Fink-Heimer stains (Fig. 2), as well as electron microscopic studies, have shown extensive terminal degeneration in the inner third of the dentate molecular layer and in stratum radiatum and stratum oriens of area CA1 (see also ref. 20). Destruction of associational and commissural afferents to the fascia dentata evoked a selective growth of adjacent afferents into the denervated inner third of the dentate molecular layer. When the great majority of CA4 neurons on both sides of the brain were destroyed by KA (n --~ 11), septohippocampal fibers grew into the associational-commissural terminal zone, an area from which they are normally largely excluded (Fig. 1C). Septohippocampal fiber growth appeared most prominent along the superficial edge of this zone and in several cases was confined to it. Light (Fig. 2A, B) and electron microscopic studies have shown this narrow area to be the most extensively denervated portion of the dentate molecular layer. The latter studies have also shown that the synaptic density of the inner third of the external or dorsal leaf of the fascia dentata typically decreases by a greater fraction (about two-thirds) than that of the internal or ventral leaf (about one-half). In accordance with this observation, the septohippocampal fibers usually occupied a larger portion of the associational-commissural terminal zone on the external leaf than on the internal leaf. Thus the extent of translaminar growth correlated with the degree of denervation. We could not tell from the present material whether the anomalous cholinergic fibers originated from the more superficial part of the molecular layer or from the supragranular zone. Septohippocampal fibers were observed not to respond to 9 less extensive bilateral CA4 lesions (Fig. 1D), nor to extensive destruction of either the ipsilateral or contralateral area CA4 alone in 6 animals (unilateral injection of 0.8-1.1/~g of KA). In contrast, these fibers proliferate in response to even a very small lesion of the perforant path is. In sections from unoperated rats stained by the Timm's method, a few fiber-like projections can be seen to pass from the CA4 area through the granule cell layer of the fascia dentata and end just above the granule cell bodies (Fig. 3A). Since these projections display the black staining characteristic of hippocampal mossy fibers (granule cell axons) s and originate from the mossy fiber-rich infragranular plexus, they are presumed to be recurrent collaterals of the mossy fibers. Normally these collaterals are confined mainly to the ends and apex of the granule cell arch; they are but sparsely distributed through the intervening sections of the fascia dentata. In contrast, a dense plexus of these collaterals developed in the deep part of the associational-commissural terminal zone in all 5 animals with bilateral destruction of
Fig. 1. Coronal sections through the dorsal hippocampal formation after bilateral KA injection. Abbreviations: 1,2, 3 and 4, area CAI, area CA2, etc. ; FD, fascia dentata. Scale bar -: 1 ram. Sections A and C were cut from the same brain, as were B and D. A and B: cresyl violet stains to show extent of most (A) and least (B) extensive hippocampal lesions in the present series. In both cases virtually all neurons in area CA3 and most in area CA4 have been destroyed and dentate granule cells remain intact. C and D: acetylcholinesterase brains to demonstrate septohipl0ocampal fibers. Arrows indicate supragranular cholinergic bundle, which is expanded in C. In D the staining of fascia dentata and area CAI is normal.
the C A 4 area (Fig. 3B). A greater n u m b e r o f m o s s y fibers a p p e a r to have p e n e t r a t e d the granule cell layer in these animals. W e also detected some g r o w t h o f putative mossy fiber collaterals into the a s s o c i a t i o n a l - c o m m i s s u r a l terminal zone o f b o t h sides o f the b r a i n after unilateral destruction o f C A 4 cells (Fig. 3C, D ; n - - 4). This finding suggests t h a t the m o s s y fibers m o r e readily replace the a s s o c i a t i o n a l - c o m m i s s u r a l p r o j e c t i o n t h a n do s e p t o h i p p o c a m p a l fibers. The medial p e r f o r a n t p a t h fibers did n o t a p p e a r to enter the associationalc o m m i s s u r a l terminal zone d e n e r v a t e d by a bilateral lesion, even when s e p t o h i p p o c a m p a l fibers were seen to do so in the same a n i m a l (Fig. 4A, B; n ~ 8). I n c o n t r a s t to the t r a n s l a m i n a r reinnervation o f the d e n e r v a t e d fascia dentata, we saw no evidence o f this in the d e n e r v a t e d laminae o f a r e a C A l . There was no histochemical evidence o f s e p t o h i p p o c a m p a l proliferation (Fig. 1D), the mossy fibers did n o t grow into the a r e a f r o m their n o r m a l t e r m i n a t i o n in a r e a CA3 a n d the t e m p o r o a m m o n i c fibers d i d n o t cross the l a m i n a r b o u n d a r y which divides them f r o m the d e n e r v a t e d s t r a t u m r a d i a t u m (Fig. 4C, D). The failure o f the t e m p o r o a m m o n i c
5
Fig. 2. Fink-Heimer stains showing terminal degeneration 3 days after a bilateral KA lesion. A: external leaf of fascia dentata. Granule cell layer is at the bottom and hippocampal fissure (arrows) at the top. Note small grains of silver deposit concentrated toward the upper boundary of the inner third of the molecular layer. Terminal degeneration just above the hiplcocampal fissure most likely arose from the destruction of neurons in nucleus reuniens of the thalamus and layer III of the entorhinal cortex. B : internal leaf of the fascia dentata. C: stratum radiatum of area CA1. Pyramidal cell layer is at the bottom. Arrows indicate boundary between stratum radiatum, where degeneration products are heavily concentrated, and stratum lacunosum-moleculare, where they are much more sparsely distributed. A similarly dense concentration of degeneration products was observed in stratum oriens. Scale bar = 0.05 mm. fibers to grow into stratum radiatum may, however, have been influenced by destruction o f perhaps half their cell bodies o f origin. DISCUSSION This study confirms and extends previous reports that deafferented laminae in the rat hippocampal formation are extensively reinnervated with a considerable degree o f selectivity a. Bilateral destruction o f CA4-derived afferents to the dentate molecular layer elicited the formation o f a dense plexus o f fibers and boutons with heavy metal staining characteristic o f the mossy fiber projection. Thus granule cells deprived o f innervation from area C A 4 evidently make connections with each other. This result agrees with previous findings o f a similar mossy fiber response to complete isolation o f the dentate area from area CA3-CA4ZZ, 29. In those studies, however, afferents to the fascia dentata other than the associational-commissural fibers, including the septohippocampal fibers, were also transected, as were the mossy fibers themselves. These complications were avoided in the present study by use o f a selective lesioning technique, which spares the septohippocampal and mossy fibers. Clearly, the putative mossy fiber collaterals can penetrate into the associational-commissural terminal zone whether or not the main branch o f the axon is cut and regardless o f whether or not the
B
Fig. 3. High-power views of external leaf of the fascia dentata stained for the presence of heavy metals by Timm's method. Abbreviations: G, granule cell layer; C-A, commissural-associational (CA3 CA4) terminal zone; MPP, terminal zone of medial perforant path fibers. Arrows indicate individual mossy fiber boutons. Scale bar : 0.05 mm. A : untreated rat. B: bilateral KA lesion. C : unilateral KA lesion - - ipsilateral side. D: unilateral KA lesion - - contralateral side.
intervening septohippocampal fibers are intact. In contrast to a previous report 28, the mossy fibers even grew into the associational-commissural terminal zone to some extent after destruction of either ipsilateral or contralateral C A 4 cells alone. This finding suggests that the mossy fibers readily replace degenerated associationalcommissural fibers. A l t h o u g h the functional consequences o f this plasticity are as yet unknown, they are likely to be unfavorable, since an abnormally powerful pathway o f direct communication a m o n g granule cells appears to be created. Other dentate afferents responded less readily to destruction o f the C A 3 - C A 4 area. Septohippocampal fibers reinnervated the associational-commissural terminal zone only after especially extensive bilateral destruction o f CA4 neurons. Furthermore, autoradiographic studies showed no evidence of reinnervation by medial perforant path fibers. One cannot, however, conclude that the perforant path fibers are incapable o f translaminar growth, since in no case did we destroy all the CA4 neurons. Indeed Stanfield and C o w a n have reported that medial perforant path fibers do grow into the associational-commissural terminal zone from their adjacent lamina when all connections between the dentate granule cells and area CA4 are severed 22. In their study, however, mossy fiber and septohippocampal a x o t o m y might have contributed to the observed result.
Fig. 4. Autoradiographic tracing of projections from entorhinal cortex to hippocampus. [2,3-all]Pro line (30-40 Ci/mmol) was injected into the entorhinal cortex ipsilateral to the side shown. After 2 days survival, the animals were perfused with neutralized formalin-saline, and 20/~m thick sections of brain were cut, mounted on slides and processed as described elsewhere5. Abbreviations: CA1, area CA1 of hippocampus; FD, fascia dentata; G, granule cell layer of fascia dentata. Circles denote the hippocampal fissure and dashed lines the upper border of the granule cell layer. Sections A and C are from the brain of an uninjected rat and B and D are from the brain of a rat injected with 0.5 ktg of KA into each lateral ventricle. A and B" dark-field photomicrographs of the dentate molecular layer. Note only background grain density in the inner third of the layer. This indicates that medial perforant path fibers did not enter the denervated area after destruction of CA3-CA4 afferents with KA. Scale bar = 0.1 mm. C and D: low-power photomicrographs showing the autoradiographic localization of perforant path terminals in the dentate molecular layer and temporoammonic terminals in area CA1. Note in D that the temporoammonic projection did not spread into the adjacent stratum radiatum, which was denervated by the KA lesion. Scale bar = 0.1 mm. Thus o u r results suggest a degree of selectivity in the r e i n n e r v a t i o n of the associational-commissural l a m i n a of the fascia d e n t a t a ; the mossy fibers most readily replacing C A 4 afferents, the septohippocampal fibers replacing them somewhat less readily a n d the perforant path fibers r e i n n e r v a t i n g the area least readily. Proximity o f these fibers to the denervated zone appears n o t to be a m a j o r factor in this instance, since the mossy fibers must grow over a greater distance t h a n either of the other afferents to reach the target area. There m u s t be some selectivity a m o n g afferents either in terms of the ease with which the degeneration evokes the t r a n s l a m i n a r growth of n e i g h b o r i n g pathways or in terms of the chemospecificity o f the denervated p o r t i o n of the dendrite 3. N o n e o f the C A 1 afferents traced in this study - - septohippocampal, mossy fiber or t e m p o r o a m m o n i c - - either grew into or proliferated in s t r a t u m r a d i a t u m or s t r a t u m oriens. S t r a t u m r a d i a t u m is denervated by 80 ~ or more after a bilateral K A
lesion and is extensively reinnervated (ref. 20 and manuscripts in preparation). Translaminar growth evidently makes little or no contribution to this reinnervation process. Goldowitz et al. have obtained similar results in area CAI of rats after transection of afferents from the CA3-CA4 area s. These studies raise the possibility that axons of the few surviving CA3-CA4 pyramidal cells (or possibly o f C A l or CA2 pyramidal cells) selectively occupy the vacated synaptic sites in area CA1, thus preventing heterologous reinnervation. Our finding that the surviving CA4 afferents to the dentate molecular layer are somewhat less successful in doing this suggests a difference between CA 1 pyramidal cells and dentate granule cells in their acceptance of abnormally placed connections. A comparison of the present data with those from studies of perforant path lesions 2,3 indicates that the ease of heterologous reinnervation is influenced not only by the nature of the denervated neurons, but also by the identity of the degenerating pathway. For example, the septohippocampal fibers respond to destruction of even a small portion of the perforant path, but require extensive damage to the associationalcommissural projection. This differential responsiveness could be easily understood if the latter two projections innervate different cells, as has been suggested on electrophysiological grounds 15. The granule cells which receive perforant path innervation might more readily accept reinnervation by septohippocampal fibers. On the other hand, the degree of heterologous reinnervation could be related to the degree of topographic specificity in the degenerating system. The strict topography of the perforant path projection may limit the ability of these fibers to occupy synaptic territory vacated by adjacent perforant path fibers, whereas the less rigid topographic specificity of CA3-CA4 projections may encourage homologous reinnervation. In conclusion, our findings suggest that extensive and selective reinnervation of denervated laminae follows a type of hippocampal cell loss similar to that which occurs in man. In severe epilepsy, senile dementia and other conditions such changes in synaptic relationships might either ameliorate or contribute to the progressive behavioral disorder. ACKNOWLEDGEMENTS We thank Ms. B. Wire for technical assistance and Dr. S. Scheff for assistance with photography. This study was supported by NSF Grant BNS76-09973 (to J.V.N.) and N I H Grants NS 08597, M H 19691 and A G 00538 (to C.W.C.).
REFERENCES 1 Blackwood, W. and Corsellis, J. A. N. (Eds.), Greenfield's Neuropathology, Edward Arnold, London, 1976. 2 Cotman, C. W. and Lynch, G. S., Reactive synaptogenesisin the adult nervous system: the effects of partial deafferentation on new synapse formation. In S. Barondes (Ed.), Neuronal Recognition, Plenum, New York, 1976, pp. 69-108. 3 Cotman, C. W. and Nadler, J. V., Reactive synaptogenesisin the hippocampus. In C. W. Cotman (Ed.), Neuronal Plasticity, Raven, New York, 1978, pp. 227-271.
4 Elliott, K. A. C., The use of brain slices. In A. Lajtha (Ed.), Handbook ofNeurochemistry, Vol. 2, Plenum, New York, 1969, pp. 103-114. 5 Goldowitz, D., Scheff, S. W. and Cotman, C. W., The specificity of reactive synaptogenesis: a comparative study in the adult rat hippocampai formation, Brain Research, in press. 6 Gottlieb, D. I. and Cowan, W. M., Autoradiographic studies of the commissural and ipsilateral association connections of the hippocampus and dentate gyrus of the rat. I. The commissural connections, J. comp. Neurol., 149 (1973) 393422. 7 Haug, F.-M. ~., Heavy metals in the brain. A light microscope study of the rat with Timm's sulphide silver method. Methodological considerations and cytological and regional staining patterns, Advanc. Anat. Embryol. Cell Biol., 47 (1973) 1-71. 8 Haug, F.-M. ~., Blackstad, T. W., Hjorth-Simonsen, A. and Zimmer, J., Timm's sulfide silver reaction for zinc during experimental anterograde degeneration of hippocampal mossy fibers, J. comp. NeuroL, 142 (1971) 23-32. 9 Herkenham, M., The connections of the nucleus reuniens thalami, J. comp. Neurol., 177 (1978) 589-610. 10 Hjorth-Simonsen, A., Some intrinsic connections of the hippocampus in the rat: an experimental analysis, J. comp. NeuroL, 147 (1973) 145-162. 11 Hjorth-Simonsen, A. and Laurberg, S., Commissural connections of the dentate area in the rat, J. comp. NeuroL, 174 (1977) 591-606. 12 Laurberg, S., Commissural and intrinsic connections of the rat hippocampus, J. comp. Neurol., 184 (1979) 685-708. 13 Lewis, P. R., Shute, C. C. D. and Silver, A., Confirmation from choline acetylase analyses of a massive cholinergic innervation to the rat hippocampus, J. Physiol. (Lond.), 191 (1967) 215-224. 14 Matthews, D. A., Nadler, J. V., Lynch, G. S. and Cotman, C. W., Development of cholinergic innervation in the hippocampal formation of the rat. I. Histochemical demonstration of acetylcholinesterase activity, Develop. Biol., 36 (1974) 130-141. 15 McNaughton, B. L., Douglas, R. M. and Goddard, G. V., Synaptic enhancement in fascia dentata: cooperativity among coactive afferents, Brain Research, 157 (1978) 277-294. 16 Minckler, J. (Ed.), Pathology of the Nervous System, Vol. 2, McGraw-Hill, New York, 1971. 17 Mosko, S., Lynch, G. S. and Cotman, C. W., The distribution of septal projections to the hippocampus of the rat, J. comp. Neurol., 152 (1973) 163-174. 18 Nadler, J. V., Cotman, C. W., Paoletti, C. and Lynch, G. S., Histochemical evidence of altered development of cholinergic fibers in the rat dentate gyrus following lesions. 1I. Effects of partial entorhinal and simultaneous multiple lesions, J. comp. Neurol., 171 (1977) 589-604. 19 Nadler, J. V., Perry, B. W. and Cotman, C. W., Intraventricular kainic acid preferentially destroys hippocampal pyramidal cells, Nature (Lond.), 271 (1978) 676-677. 20 Nadler, J. V., Perry, B. W. and Cotman, C. W., Preferential vulnerability of hippocampus to intraventricular kainic acid. In E. G. McGeer, J. W. Olney and P. L. McGeer (Eds.), Kainic Acid As a Tool in Neurobiology, Raven, New York, 1978, pp. 219-237. 21 Shute, C C. D. and Lewis, P. R., Cholinesterase-containing systems of the brain of the rat, Nature (Lond.), 199 (1963) 1160-1166. 22 Stanfield, B. and Cowan, W. M., Evidence for the sprouting of entorhinal afferents within the dentate gyrus of the rat following removal of the associational and commissural projections, Anat. Rec., 190 (1978) 550. 23 Steward, O. and Scoville, S. A., Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat, J. comp. NeuroL, 169 (1976) 347-370. 24 Storm-Mathisen, J., Quantitative histochemistry of acetylcholinesterase in rat hippocampal region correlated to histochemical staining, J. Neurochem., 17 (1970) 739-750. 25 Swanson, L. W. and Cowan, W. M., An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat, J. comp. Neurol., 172 (1977) 49-84. 26 Swanson, L. W., Wyss, J. M. and Cowan, W. M., An autoradiographic study of the organization of intrahippocampal association pathways in the rat, J. comp. NeuroL, 181 (1978) 681-715. 27 West, J. R., Nornes, H. O., Barnes, C. L. and Bronfenbrenner, M., The cells of origin of the commissural afferents to the area dentata in the mouse, Brain Research, 160 (1979) 203-215. 28 Zimmer, J., Changes in the Timm sulfide silver staining pattern of the rat hippocarnpus and fascia dentata following early postnatal deafferentation, Brain Research, 64 (1973) 313-326. 29 Zimmer, J., Proximity as a factor in the regulation of aberrant axonal growth in the postnatally deafferented fascia dentata, Brain Research, 72 (1974) 137-142.
1
Research Reports
SELECTIVE R E I N N E R V A T I O N OF H I P P O C A M P A L A R E A CA1 A N D T H E FASCIA D E N T A T A A F T E R D E S T R U C T I O N OF CA3-CA4 A F F E R E N T S W I T H K A I N I C ACID
J. VICTOR NADLER*, BRUCE W. PERRY and CARL W. COTMAN Department of Psychobiology, University of California, Irvine, Calif. 92717 (U.S.A.)
(Accepted May 31st, 1979)
SUMMARY Intraventricular injections of kainic acid were used to destroy the hippocampal CA3-CA4 cells, thus denervating the inner third of the molecular layer of the fascia dentata and stratum radiatum and stratum oriens of area CA1. The responses of intact afferents to such lesions were then examined histologically. The hippocampal mossy fibers densely reinnervated the inner portion of the dentate molecular layer after bilateral destruction of CA4 neurons and to a lesser extent after unilateral destruction. Septohippocampal fibers replaced CA4-derived fibers in the dentate molecular layer only after particularly extensive bilateral CA4 lesions. Medial perforant path fibers showed no anatomical response to any of these lesions. Neither septohippocampal, temporoammonic nor mossy fibers proliferated in or grew into the denervated laminae of area CA 1. These results show a preferential ordering in the reinnervation of dentate granule cells which is not readily explained by proximity to the degenerating fibers and also that removal of CA3-CA4-derived innervation more readily elicits translaminar growth in the fascia dentata than in area CA1. These results may be relevant to clinical situations in which neurons of the hippocampal end-blade are lost. INTRODUCTION The pyramidal cells of the hippocampal end-blade (areas ha-h5 of Rose) are among the most sensitive neurons in the brain to a variety of deleterious conditions, including anoxia, senile dementia and severe epilepsy1,16. In all these conditions pyramidal cells degenerate and are replaced by a proliferation of glial cells, but * To whom all correspondence should be addressed. Present address: Department of Pharmacology, Duke University Medical Center, Durham, N.C. 27710, U.S.A.
afferent fibers and fibers of passage in the area of the lesion are not directly damaged. The consequences for hippocampal connectivity are unknown, but are likely to be of importance, given the widely acknowledged behavioral role of the hippocampus. Hence we sought an animal model of this pathology. We have reported that intraventricular administration of the potent convulsant, kainic acid (KA), reproduces in rats this common form of hippocampal damage19, 20. Low doses of KA (0.5-4 nmol) preferentially destroy the CA3-CA4 pyramidal cells, which are considered homologous to the pyramidal cells of the human end-blade. These neurons have been shown to innervate and interconnect all parts of the hippocampal area6,10,~l,25,26. Accordingly, we may conclude that they play a central role in the functions of this region and that loss of their connections, as occurs in pathological states, would severely disrupt these functions. It was therefore of interest to determine whether heterologous afferents replace the innervation provided by CA3-CA4 cells, when the latter are destroyed with KA. Heterologous reinnervation has previously been demonstrated in the molecular layer of the fascia dentata denervated by a perforant path lesion2, 3. In the present studies we have examined the terminal fields of CA3-CA4 fibers in stratum oriens and stratum radiatum of area CA1 and in the molecular layer of the fascia dentata. MATERIALS AND METHODS KA was administered intraventricularly to adult male Sprague-Dawley ratslL Either 0.5 or 0.8/zg (2.34 or 3.75 nmol) of KA dissolved in 1/tl of Elliott's artificial CSF 4 was injected over a 30-min period. In most cases the surgical and injection procedures were then repeated on the contralateral side. Animals were given atropine sulfate (0.5-1 mg/kg, i.p.) as often as hourly to alleviate any respiratory distress. Histological studies were performed after a survival period of at least 30 days. Projections from the ipsilateral entorhinal cortex to the hippocampal formation were traced autoradiographically 5, mossy fiber boutons were visualized by use of the Timm's stain for heavy metals 7 and acetylcholinesterase histochemistry14 was employed to demonstrate the cholinergic septohippocampal fibers la,14,17,21,z4. RESULTS Intraventricular injection of 0.8/zg of KA consistently destroyed the pyramidal cells of the ipsilateral CA3 and CA4 areas, as previously reported 19,2°. Most of the neurons in the 'cell-poor' layer of area CA4, which provide the predominant associational and commissural innervation of the fascia dentata 12,2°,26,27, were also killed. KA destroyed a higher percentage of these cells in the septal two-thirds of the hippocampal formation than in the temporal third. The CA1 pyramidal cells were usually spared, but occasionally one or more clusters of these neurons were found to have been destroyed. In the most severe cases, most CA1 pyramidal cells were destroyed along with the CA3-CA4 cells. Few pyramidal cells of the hz zone (area CA2 and the adjacent part of area CA3a) and apparently no dentate granule cells were destroyed by the drug. Bilateral injection of either 0.5 or 0.8 /,tg of KA produced
lesions identical to this in both hippocampi (Fig. 1A, B). While KA at the doses used here destroyed neurons in several other regions of the brain 1a,2°, the only markedly sensitive areas known to project to the hippocampal formation were the nucleus reuniens of the thalamus a and layer IIl of the entorhinal cortex 23. Ceils in these areas project to stratum lacunosum-moleculare of area CA1, a lamina that does not receive a CA3-CA4 projection. Only very few neurons (perhaps 1-2 ~) in layer I1 of the entorhinal cortex, the predominant or exclusive source of hippocampal perforant path fibers 2a, were found to degenerate. Fink-Heimer stains (Fig. 2), as well as electron microscopic studies, have shown extensive terminal degeneration in the inner third of the dentate molecular layer and in stratum radiatum and stratum oriens of area CA1 (see also ref. 20). Destruction of associational and commissural afferents to the fascia dentata evoked a selective growth of adjacent afferents into the denervated inner third of the dentate molecular layer. When the great majority of CA4 neurons on both sides of the brain were destroyed by KA (n --~ 11), septohippocampal fibers grew into the associational-commissural terminal zone, an area from which they are normally largely excluded (Fig. 1C). Septohippocampal fiber growth appeared most prominent along the superficial edge of this zone and in several cases was confined to it. Light (Fig. 2A, B) and electron microscopic studies have shown this narrow area to be the most extensively denervated portion of the dentate molecular layer. The latter studies have also shown that the synaptic density of the inner third of the external or dorsal leaf of the fascia dentata typically decreases by a greater fraction (about two-thirds) than that of the internal or ventral leaf (about one-half). In accordance with this observation, the septohippocampal fibers usually occupied a larger portion of the associational-commissural terminal zone on the external leaf than on the internal leaf. Thus the extent of translaminar growth correlated with the degree of denervation. We could not tell from the present material whether the anomalous cholinergic fibers originated from the more superficial part of the molecular layer or from the supragranular zone. Septohippocampal fibers were observed not to respond to 9 less extensive bilateral CA4 lesions (Fig. 1D), nor to extensive destruction of either the ipsilateral or contralateral area CA4 alone in 6 animals (unilateral injection of 0.8-1.1/~g of KA). In contrast, these fibers proliferate in response to even a very small lesion of the perforant path is. In sections from unoperated rats stained by the Timm's method, a few fiber-like projections can be seen to pass from the CA4 area through the granule cell layer of the fascia dentata and end just above the granule cell bodies (Fig. 3A). Since these projections display the black staining characteristic of hippocampal mossy fibers (granule cell axons) s and originate from the mossy fiber-rich infragranular plexus, they are presumed to be recurrent collaterals of the mossy fibers. Normally these collaterals are confined mainly to the ends and apex of the granule cell arch; they are but sparsely distributed through the intervening sections of the fascia dentata. In contrast, a dense plexus of these collaterals developed in the deep part of the associational-commissural terminal zone in all 5 animals with bilateral destruction of
Fig. 1. Coronal sections through the dorsal hippocampal formation after bilateral KA injection. Abbreviations: 1,2, 3 and 4, area CAI, area CA2, etc. ; FD, fascia dentata. Scale bar -: 1 ram. Sections A and C were cut from the same brain, as were B and D. A and B: cresyl violet stains to show extent of most (A) and least (B) extensive hippocampal lesions in the present series. In both cases virtually all neurons in area CA3 and most in area CA4 have been destroyed and dentate granule cells remain intact. C and D: acetylcholinesterase brains to demonstrate septohipl0ocampal fibers. Arrows indicate supragranular cholinergic bundle, which is expanded in C. In D the staining of fascia dentata and area CAI is normal.
the C A 4 area (Fig. 3B). A greater n u m b e r o f m o s s y fibers a p p e a r to have p e n e t r a t e d the granule cell layer in these animals. W e also detected some g r o w t h o f putative mossy fiber collaterals into the a s s o c i a t i o n a l - c o m m i s s u r a l terminal zone o f b o t h sides o f the b r a i n after unilateral destruction o f C A 4 cells (Fig. 3C, D ; n - - 4). This finding suggests t h a t the m o s s y fibers m o r e readily replace the a s s o c i a t i o n a l - c o m m i s s u r a l p r o j e c t i o n t h a n do s e p t o h i p p o c a m p a l fibers. The medial p e r f o r a n t p a t h fibers did n o t a p p e a r to enter the associationalc o m m i s s u r a l terminal zone d e n e r v a t e d by a bilateral lesion, even when s e p t o h i p p o c a m p a l fibers were seen to do so in the same a n i m a l (Fig. 4A, B; n ~ 8). I n c o n t r a s t to the t r a n s l a m i n a r reinnervation o f the d e n e r v a t e d fascia dentata, we saw no evidence o f this in the d e n e r v a t e d laminae o f a r e a C A l . There was no histochemical evidence o f s e p t o h i p p o c a m p a l proliferation (Fig. 1D), the mossy fibers did n o t grow into the a r e a f r o m their n o r m a l t e r m i n a t i o n in a r e a CA3 a n d the t e m p o r o a m m o n i c fibers d i d n o t cross the l a m i n a r b o u n d a r y which divides them f r o m the d e n e r v a t e d s t r a t u m r a d i a t u m (Fig. 4C, D). The failure o f the t e m p o r o a m m o n i c
5
Fig. 2. Fink-Heimer stains showing terminal degeneration 3 days after a bilateral KA lesion. A: external leaf of fascia dentata. Granule cell layer is at the bottom and hippocampal fissure (arrows) at the top. Note small grains of silver deposit concentrated toward the upper boundary of the inner third of the molecular layer. Terminal degeneration just above the hiplcocampal fissure most likely arose from the destruction of neurons in nucleus reuniens of the thalamus and layer III of the entorhinal cortex. B : internal leaf of the fascia dentata. C: stratum radiatum of area CA1. Pyramidal cell layer is at the bottom. Arrows indicate boundary between stratum radiatum, where degeneration products are heavily concentrated, and stratum lacunosum-moleculare, where they are much more sparsely distributed. A similarly dense concentration of degeneration products was observed in stratum oriens. Scale bar = 0.05 mm. fibers to grow into stratum radiatum may, however, have been influenced by destruction o f perhaps half their cell bodies o f origin. DISCUSSION This study confirms and extends previous reports that deafferented laminae in the rat hippocampal formation are extensively reinnervated with a considerable degree o f selectivity a. Bilateral destruction o f CA4-derived afferents to the dentate molecular layer elicited the formation o f a dense plexus o f fibers and boutons with heavy metal staining characteristic o f the mossy fiber projection. Thus granule cells deprived o f innervation from area C A 4 evidently make connections with each other. This result agrees with previous findings o f a similar mossy fiber response to complete isolation o f the dentate area from area CA3-CA4ZZ, 29. In those studies, however, afferents to the fascia dentata other than the associational-commissural fibers, including the septohippocampal fibers, were also transected, as were the mossy fibers themselves. These complications were avoided in the present study by use o f a selective lesioning technique, which spares the septohippocampal and mossy fibers. Clearly, the putative mossy fiber collaterals can penetrate into the associational-commissural terminal zone whether or not the main branch o f the axon is cut and regardless o f whether or not the
B
Fig. 3. High-power views of external leaf of the fascia dentata stained for the presence of heavy metals by Timm's method. Abbreviations: G, granule cell layer; C-A, commissural-associational (CA3 CA4) terminal zone; MPP, terminal zone of medial perforant path fibers. Arrows indicate individual mossy fiber boutons. Scale bar : 0.05 mm. A : untreated rat. B: bilateral KA lesion. C : unilateral KA lesion - - ipsilateral side. D: unilateral KA lesion - - contralateral side.
intervening septohippocampal fibers are intact. In contrast to a previous report 28, the mossy fibers even grew into the associational-commissural terminal zone to some extent after destruction of either ipsilateral or contralateral C A 4 cells alone. This finding suggests that the mossy fibers readily replace degenerated associationalcommissural fibers. A l t h o u g h the functional consequences o f this plasticity are as yet unknown, they are likely to be unfavorable, since an abnormally powerful pathway o f direct communication a m o n g granule cells appears to be created. Other dentate afferents responded less readily to destruction o f the C A 3 - C A 4 area. Septohippocampal fibers reinnervated the associational-commissural terminal zone only after especially extensive bilateral destruction o f CA4 neurons. Furthermore, autoradiographic studies showed no evidence of reinnervation by medial perforant path fibers. One cannot, however, conclude that the perforant path fibers are incapable o f translaminar growth, since in no case did we destroy all the CA4 neurons. Indeed Stanfield and C o w a n have reported that medial perforant path fibers do grow into the associational-commissural terminal zone from their adjacent lamina when all connections between the dentate granule cells and area CA4 are severed 22. In their study, however, mossy fiber and septohippocampal a x o t o m y might have contributed to the observed result.
Fig. 4. Autoradiographic tracing of projections from entorhinal cortex to hippocampus. [2,3-all]Pro line (30-40 Ci/mmol) was injected into the entorhinal cortex ipsilateral to the side shown. After 2 days survival, the animals were perfused with neutralized formalin-saline, and 20/~m thick sections of brain were cut, mounted on slides and processed as described elsewhere5. Abbreviations: CA1, area CA1 of hippocampus; FD, fascia dentata; G, granule cell layer of fascia dentata. Circles denote the hippocampal fissure and dashed lines the upper border of the granule cell layer. Sections A and C are from the brain of an uninjected rat and B and D are from the brain of a rat injected with 0.5 ktg of KA into each lateral ventricle. A and B" dark-field photomicrographs of the dentate molecular layer. Note only background grain density in the inner third of the layer. This indicates that medial perforant path fibers did not enter the denervated area after destruction of CA3-CA4 afferents with KA. Scale bar = 0.1 mm. C and D: low-power photomicrographs showing the autoradiographic localization of perforant path terminals in the dentate molecular layer and temporoammonic terminals in area CA1. Note in D that the temporoammonic projection did not spread into the adjacent stratum radiatum, which was denervated by the KA lesion. Scale bar = 0.1 mm. Thus o u r results suggest a degree of selectivity in the r e i n n e r v a t i o n of the associational-commissural l a m i n a of the fascia d e n t a t a ; the mossy fibers most readily replacing C A 4 afferents, the septohippocampal fibers replacing them somewhat less readily a n d the perforant path fibers r e i n n e r v a t i n g the area least readily. Proximity o f these fibers to the denervated zone appears n o t to be a m a j o r factor in this instance, since the mossy fibers must grow over a greater distance t h a n either of the other afferents to reach the target area. There m u s t be some selectivity a m o n g afferents either in terms of the ease with which the degeneration evokes the t r a n s l a m i n a r growth of n e i g h b o r i n g pathways or in terms of the chemospecificity o f the denervated p o r t i o n of the dendrite 3. N o n e o f the C A 1 afferents traced in this study - - septohippocampal, mossy fiber or t e m p o r o a m m o n i c - - either grew into or proliferated in s t r a t u m r a d i a t u m or s t r a t u m oriens. S t r a t u m r a d i a t u m is denervated by 80 ~ or more after a bilateral K A
lesion and is extensively reinnervated (ref. 20 and manuscripts in preparation). Translaminar growth evidently makes little or no contribution to this reinnervation process. Goldowitz et al. have obtained similar results in area CAI of rats after transection of afferents from the CA3-CA4 area s. These studies raise the possibility that axons of the few surviving CA3-CA4 pyramidal cells (or possibly o f C A l or CA2 pyramidal cells) selectively occupy the vacated synaptic sites in area CA1, thus preventing heterologous reinnervation. Our finding that the surviving CA4 afferents to the dentate molecular layer are somewhat less successful in doing this suggests a difference between CA 1 pyramidal cells and dentate granule cells in their acceptance of abnormally placed connections. A comparison of the present data with those from studies of perforant path lesions 2,3 indicates that the ease of heterologous reinnervation is influenced not only by the nature of the denervated neurons, but also by the identity of the degenerating pathway. For example, the septohippocampal fibers respond to destruction of even a small portion of the perforant path, but require extensive damage to the associationalcommissural projection. This differential responsiveness could be easily understood if the latter two projections innervate different cells, as has been suggested on electrophysiological grounds 15. The granule cells which receive perforant path innervation might more readily accept reinnervation by septohippocampal fibers. On the other hand, the degree of heterologous reinnervation could be related to the degree of topographic specificity in the degenerating system. The strict topography of the perforant path projection may limit the ability of these fibers to occupy synaptic territory vacated by adjacent perforant path fibers, whereas the less rigid topographic specificity of CA3-CA4 projections may encourage homologous reinnervation. In conclusion, our findings suggest that extensive and selective reinnervation of denervated laminae follows a type of hippocampal cell loss similar to that which occurs in man. In severe epilepsy, senile dementia and other conditions such changes in synaptic relationships might either ameliorate or contribute to the progressive behavioral disorder. ACKNOWLEDGEMENTS We thank Ms. B. Wire for technical assistance and Dr. S. Scheff for assistance with photography. This study was supported by NSF Grant BNS76-09973 (to J.V.N.) and N I H Grants NS 08597, M H 19691 and A G 00538 (to C.W.C.).
REFERENCES 1 Blackwood, W. and Corsellis, J. A. N. (Eds.), Greenfield's Neuropathology, Edward Arnold, London, 1976. 2 Cotman, C. W. and Lynch, G. S., Reactive synaptogenesisin the adult nervous system: the effects of partial deafferentation on new synapse formation. In S. Barondes (Ed.), Neuronal Recognition, Plenum, New York, 1976, pp. 69-108. 3 Cotman, C. W. and Nadler, J. V., Reactive synaptogenesisin the hippocampus. In C. W. Cotman (Ed.), Neuronal Plasticity, Raven, New York, 1978, pp. 227-271.
4 Elliott, K. A. C., The use of brain slices. In A. Lajtha (Ed.), Handbook ofNeurochemistry, Vol. 2, Plenum, New York, 1969, pp. 103-114. 5 Goldowitz, D., Scheff, S. W. and Cotman, C. W., The specificity of reactive synaptogenesis: a comparative study in the adult rat hippocampai formation, Brain Research, in press. 6 Gottlieb, D. I. and Cowan, W. M., Autoradiographic studies of the commissural and ipsilateral association connections of the hippocampus and dentate gyrus of the rat. I. The commissural connections, J. comp. Neurol., 149 (1973) 393422. 7 Haug, F.-M. ~., Heavy metals in the brain. A light microscope study of the rat with Timm's sulphide silver method. Methodological considerations and cytological and regional staining patterns, Advanc. Anat. Embryol. Cell Biol., 47 (1973) 1-71. 8 Haug, F.-M. ~., Blackstad, T. W., Hjorth-Simonsen, A. and Zimmer, J., Timm's sulfide silver reaction for zinc during experimental anterograde degeneration of hippocampal mossy fibers, J. comp. NeuroL, 142 (1971) 23-32. 9 Herkenham, M., The connections of the nucleus reuniens thalami, J. comp. Neurol., 177 (1978) 589-610. 10 Hjorth-Simonsen, A., Some intrinsic connections of the hippocampus in the rat: an experimental analysis, J. comp. NeuroL, 147 (1973) 145-162. 11 Hjorth-Simonsen, A. and Laurberg, S., Commissural connections of the dentate area in the rat, J. comp. NeuroL, 174 (1977) 591-606. 12 Laurberg, S., Commissural and intrinsic connections of the rat hippocampus, J. comp. Neurol., 184 (1979) 685-708. 13 Lewis, P. R., Shute, C. C. D. and Silver, A., Confirmation from choline acetylase analyses of a massive cholinergic innervation to the rat hippocampus, J. Physiol. (Lond.), 191 (1967) 215-224. 14 Matthews, D. A., Nadler, J. V., Lynch, G. S. and Cotman, C. W., Development of cholinergic innervation in the hippocampal formation of the rat. I. Histochemical demonstration of acetylcholinesterase activity, Develop. Biol., 36 (1974) 130-141. 15 McNaughton, B. L., Douglas, R. M. and Goddard, G. V., Synaptic enhancement in fascia dentata: cooperativity among coactive afferents, Brain Research, 157 (1978) 277-294. 16 Minckler, J. (Ed.), Pathology of the Nervous System, Vol. 2, McGraw-Hill, New York, 1971. 17 Mosko, S., Lynch, G. S. and Cotman, C. W., The distribution of septal projections to the hippocampus of the rat, J. comp. Neurol., 152 (1973) 163-174. 18 Nadler, J. V., Cotman, C. W., Paoletti, C. and Lynch, G. S., Histochemical evidence of altered development of cholinergic fibers in the rat dentate gyrus following lesions. 1I. Effects of partial entorhinal and simultaneous multiple lesions, J. comp. Neurol., 171 (1977) 589-604. 19 Nadler, J. V., Perry, B. W. and Cotman, C. W., Intraventricular kainic acid preferentially destroys hippocampal pyramidal cells, Nature (Lond.), 271 (1978) 676-677. 20 Nadler, J. V., Perry, B. W. and Cotman, C. W., Preferential vulnerability of hippocampus to intraventricular kainic acid. In E. G. McGeer, J. W. Olney and P. L. McGeer (Eds.), Kainic Acid As a Tool in Neurobiology, Raven, New York, 1978, pp. 219-237. 21 Shute, C C. D. and Lewis, P. R., Cholinesterase-containing systems of the brain of the rat, Nature (Lond.), 199 (1963) 1160-1166. 22 Stanfield, B. and Cowan, W. M., Evidence for the sprouting of entorhinal afferents within the dentate gyrus of the rat following removal of the associational and commissural projections, Anat. Rec., 190 (1978) 550. 23 Steward, O. and Scoville, S. A., Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat, J. comp. NeuroL, 169 (1976) 347-370. 24 Storm-Mathisen, J., Quantitative histochemistry of acetylcholinesterase in rat hippocampal region correlated to histochemical staining, J. Neurochem., 17 (1970) 739-750. 25 Swanson, L. W. and Cowan, W. M., An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat, J. comp. Neurol., 172 (1977) 49-84. 26 Swanson, L. W., Wyss, J. M. and Cowan, W. M., An autoradiographic study of the organization of intrahippocampal association pathways in the rat, J. comp. NeuroL, 181 (1978) 681-715. 27 West, J. R., Nornes, H. O., Barnes, C. L. and Bronfenbrenner, M., The cells of origin of the commissural afferents to the area dentata in the mouse, Brain Research, 160 (1979) 203-215. 28 Zimmer, J., Changes in the Timm sulfide silver staining pattern of the rat hippocarnpus and fascia dentata following early postnatal deafferentation, Brain Research, 64 (1973) 313-326. 29 Zimmer, J., Proximity as a factor in the regulation of aberrant axonal growth in the postnatally deafferented fascia dentata, Brain Research, 72 (1974) 137-142.