HZPPOCAMPUS, VOL. 2, NO. 3, PAGES 247-268, JULY 1992

Lesion-induced Synapse Reorganization in the Hippocampus of Cats: Sprouting of Entorhinal, Commissural/Associational, and Mossy Fiber Projections After Unilateral Entorhinal Cortex Lesions, With Comments on the Normal Organization of These Pathways Oswald Steward Departments of Neuroscience a n d Neurosurgery, University of Virginia Health Sciences Center, Charlottesville, VA, U.S.A.

ABSTRACT This study evaluates whether three forms of sprouting occur in the hippocampus of the cat following unilateral entorhinal cortex (EC) lesions: (1) sprouting of projections from the EC contralateral to the lesion; ( 2 ) sprouting of the commissuraUassociationa1 system; and (3) sprouting of mossy fibers. Tract tracing techniques were used to define the normal organization of the entorhinal cortical projection system, the commissuraliassociational (CIA) systems, and the mossy fiber projections in normal cats. The same techniques were then used to evaluate whether there were changes in these projections in animals with long-standing unilateral EC lesions. The projections from the entorhinal cortex were evaluated autoradiographically following injections of 3H proline into the entorhinal area. The projections of the CIA system were traced using the Fink-Heimer technique after lesions of the hippocampal commissures, and by using autoradiographic techniques after injections of 3H proline into the hippocampus. The distribution of mossy fibers was evaluated using the Timm’s stain. The results reveal that unilateral lesions of thc EC in cats lead to the same sorts of sprouting that have been described in rats. There is: (1) an increase in the density of the crossed projection from the surviving EC to the contralateral dentate gyrus that had been deprived of its normal EC inputs; (2) an expansion of the terminal field of the C/A projection system into portions of the molecular layer of the dentate gyrus normally occupied by EC projections; and (3) an increase in supragranular mossy fibers in some animals. The mossy fiber sprouting was especially prominent when the lesions encroached upon the hippocampus. The studies also reveal additional details about the normal organization of hippocampal pathways in cats. The most important points are: (1) there is a crossed projection from the entorhinal cortex to the contralateral dentate gyrus; and (2) there is a complex laminar organization of the commissural and associational terminal fields in the molecular layer of the dentate gyrus that appears to be related to the point of origin of the projections along the septotcmporal axis of the hippocampus. This heretofore unrecognized aspect of the laminar organization of CIA terminations has important implications for the temporal competition hypothesis, which has been advanced to account for the development of these afferent systems. Key words: sprouting, reinnervation, lesions, dentate gyrus, tract tracing, entorhinal cortex, hippocampal comrnissure

Synapse reorganization following lesions has now been documented in a variety of brain regions (for a recent review, Comespondence and requests to Oswald Steward, D ~ ments of Neuroscience and Neurosurgery, University of Virginia Health Sciences Center, Charlottesville,VA 22908.

see Steward, 1989). Such reorganization is thought to play an important role in adaptation of the nervous system to injury. In particular, it has been demonstrated that synapse reorganization ~ ~ ~ can. lead to the reinnervation of neurons that have lost their normal input, and may lead to the reestablishment of some aspects of normal physiological processing by

247

248

HIPPOCAhfPUS VOL. 2, NO. 3, JULY 1992

these neurons (Steward, 1982). Moreover, in some cases, synapse reorganization seems to contribute to recovery of function (Steward, 1982). For these reasons, an understanding of the processes of synapse reorganization may provide important insights into processes that contribute to nervous system repair following injury. The hippocampal formation has been an especially valuable site in which to study synapse reorganization. Lesion-induced alterations in pathways can be easily detected because of the precise laminar organization of hippocampal projection systems, and a number of examples of synapse growth and reorganization have been characterized following various types of pathological insult to the hippocampus or its afferents. Although the nature of the growth processes and the extent of participation of different afferent systems varies depending upon the type of lesion, essentially all known hippocampal pathways seem to be capable of postlesion growth (for a review, see Steward, 1986). Despite an extensive literature, however, there are relatively few studies of synapse reorganization in species other than rodents. This lack of information is significant for two reasons. First, it is, of course, presumed that information derived from studies of rodents is applicable to other species including man. However, certain features of the synapse reorganization that has been so extensively characterized in rodents may be different than occurs in higher mammalian species. If this is the case, then studies in the rodent may not provide information that would be directly applicable to understanding nervous system repair following CNS injury in humans. Second, differences in the nature of lesion-induced synapse growth across species could help to elucidate cellular mechanisms underlying the processes. However, this approach requires a direct comparison of the given types of reorganization following comparable injuries. In this regard, the hippocampus offers advantages because of the extensive data base on reorganization of the different systems, and because it is possible to produce comparable lesions in different species by selectively eliminating one hippocampal afferent system. Studies of synapse reorganization in the hippocampus in species other than rodents have, so far, been limited to pathways that can be traced using histochemical techniques. The growth of the cholinergic septohippocampal pathway following entorhinal cortex lesions has been assessed in cats and humans using a histochemical stain for AChE (Grady et al., 1989; Steward and Messenheimer, 1978). Similarly, mossy fiber sprouting has been described in humans with temporal lobe epilepsy using the Timm’s method for staining heavy metals (Babb et al., 1991). Changes in the distribution of neurotransmitter receptor binding sites have also provided indirect evidence for synapse reorganization in humans with Alzheimer’s disease (for a review, see Cotman et al., 1990). However, some of the more dramatic examples of lesioninduced synapse reorganization have not yet been documented using anatomical techniques in species other than rodents. The present study examines three well-characterized examples of synapse reorganization that occur in the hippocampus following damage to the entorhinal coatex: (1) the sprouting of projections from the contralateral entorhinal cor-

tex following unilateral entorhinal lesions (Steward et al., 1974, Zimmer and Hjorth-Simonsen, 1975); (2) the sprouting of the commissuraliassociational system after EC lesions (Lynch et al., 1973; Zimmer, 1973); and (3) the sprouting of mossy fibers, which occurs, to some extent, following EC lesions, but is more extensive following damage to the hippocampus proper (Laurberg and Zimmer, 1981; Frotscher and Zimmer, 1983; West and Dewey, 1984). These examples of synapse reorganization were evaluated in cats using the same types of tract-tracing techniques that have been used in rats. We report that, in general, lesions of hippocampal circuitry in cats induce the same sorts of postlesion afferent reorganization that have been described in rats. However, there are some minor differences in the pattern of sprouting of the different systems. The studies also provide additional information about the normal organization of hippocampal pathways in cats. First, our results confirm previous reports that prominent commissural projections to both the dentate gyrus and hippocampus exist in the cat (Van Groen and Wyss, 1988). This is in contrast to the situation in primates, where commissural projections are very limited in comparison to lower mammals (Amaral et al., 1984; Demeter et al., 1985). Second, the results demonstrate the presence of a sparse crossed projection from the entorhinal cortex to the contralateral dentate gyrus in cats. This pathway is important as a potential source of reinnervating fibers following unilateral injury. Third, our results indicate the existence of a more complex laminar organization of the commissural and associational projections to the dentate gyrus than has been described to date in other species. This organization cannot be easily explained by the temporal competition hypothesis that has been advanced to account for the differential laminar distribution of commissural and associational projections in rats.

METHODS The observations of this study were based upon a total of 30 adult mixed breed cats of both sexes: ranging in age from 6 months to approximately 2 years at the time of the initial surgery. Tables 1-4 list all experimental animals and describe the experimental procedures used on each. All operations, including those involving injections of tracers, were carried out using sterile techniques.

Entorhinal cortex lesions The entorhinal area and part of the surrounding temporal neocortex was destroyed by thermal coagulation in a total of 22 animals, as described previously (Steward and Messenheimer, 1978). Animals were anesthetized with sodium pentobarbital (Nembutal), and placed in a stereotaxic apparatus. A temporal craniotomy was performed, and a heated flat probe about 4 mm wide was inserted between the floor of the skull and the temporal cortex. After completion of the lesion, the skull defect was filled with gelfoam, and the scalp was sutured. The animals did not experience postoperative complications, and were thus maintained under routine conditions for the duration of the postoperative survival intervals (see Tables 2-4 for the duration of the survival intervals). Animals

LESION-INDUCED SYNAPSE REORGANIZATION i Steward Table 1. Experimental Animals (Tract Tracing of Normal Pathways) Animal No. Tract Tracing 0128a Coordinates 0128b Coordinates 0404 0510 05 19a Coordinates’

0519b Coordinates* 0520b Coordinates3 0317 Coordinates

249

Processing

SOpCi (Ip1) 3H proline in EC 3.OA, lO.OL, 3.5 from skull base (6 day survival) SOkCi (Ipl) 3H proline in EC 3.OA, lO.OL, 3.5 from skull base (6 day survival) Commissure transection (3 day survival) Commissure transection ( 3 day survival) 25pCi ( I pI) 3H proline in hippocampus 3.OA, 9.0L, 14 deep (rostral/dorsal site) (6 day survival) 25pCi (lpl) 3H proline in hippocampus 4.OA, 13L, 17 deep (mid rostro-caudal site) (6 day survival) 25pCi (1pI) 3H proline in hippocampus 4.OA, 6L, I 1 deep (caudaliventral site) (6 day survival)

Autoradiography

Commissure transection AND 12.5~Ci(0.5pl) 3H proline in / p s i hippocampus 4.OA, 13L, 17 deep (6 day survival)

Autoradiography Fink-Heimer *Injection on one side misplaced. Data are only for associational projections on one side and commissure projections by Fink-Heimer staining

Autoradiography Fink-Heimer Fink-Heimer Autoradiography Autoradiography Autoradiography

The first column indicates the number assigned to each animal. The second column indicates the operative procedures, coordinates for lesions and injections, amounts of 3H-leucine injected, and postoperative survival interval. The third column indicates the histological procedure employed. Comments (*) indicate special circumstances (i.e., misplaced injections). I ,2,3; These injections were targeted to the rostraVdorsal (septal), middle, and caudaliventral (temporal) portions of the hippocampus, respectively. EC, entorhinal cortex.

were allowed to survive for 2 weeks t o 12 months to allow for reorganization of projections to occur (see Tables 2-4). Tract tracing techniques

The projections of the entorhinal cortex of one hemisphere were evaluated in two normal animals, and five animals with long standing unilateral EC lesions by injecting 3H proline into the entorhinal area, and tracing the projections from cells

in the injected area using autoradiographic techniques (see Tables 1 and 2 for further details). Animals were anesthetized with Nembutal, placed in a stereotaxic apparatus, and the scalp was opened. Injections of 3H-proline were made via a Hamilton 1 p L syringe using stereotaxic techniques. A 1 mm burr hole was made in the skull at the point of entry of the microliter syringe. Coordinates for the injections were 3.0 anterior to the intra-aural line, 10 mm lateral, and 3.5 mm

Table 2. Experimental Animals Long-standing EC Lesions-Secondary Tract Tracing (for Crossed Entorhinal Cortical Projections) Lesion and Animal No.

Survival

0130 Coordinates

ECL (2.5 mo.)

0131 Coordinates

ECL (2.5 mo.)

0727a Coordinates

ECL (2 mo.)

0727b Coordinates

ECL (2 mo.)

0407 Coordinates

ECL (2.5 mo.)

Tract Tracing 50pCi 3H proline in EC 3.OA, IO.OL, 3.5 from skull base (7 day survival) 50pCi 3H proline in EC 3.OA, lO.OL, 3.5 from skull base (7 day survival) 40pCi 3H proline in EC 3.OA, IO.OL, 3.5 from skull base (7 day survival) 40pCi 3H proline in EC 3.OA, IO.OL, 3.5 from skull base (7 day survival) 20pCi 3H proline in EC 3.OA, IO.OL, 3.5 from skull base (6 day survival)

Processing

Autoradiography Timm’s Autoradiography Timm’s Autoradiography Autoradiography Autoradiography Timm’s

The first column indicates the number assigned to each animal. The second column indicates the survival interval following the EC lesion. The third column indicates the coordinates for injections, amounts of 3H-leucine injected, and post-injection survival interval. The fourth column indicates the histological procedure employed. EC, entorhinal cortex; ECL, entorhinal cortex lesion.

250 HIPPOCAMPUS VOL. 2, NO. 3, JULY 1992 Table 3. Experimental Animals Long-standing EC Lesions-Secondary Tract Tracing (for Commissural/Associational Projections) Animal No. Lesion & Survival Tract Tracing I

I

~

02 17

ECL (3 mo.)

0223

ECL (7 mo.)

0415

ECL (4.5 mo.)

0422

ECL (12 mo.)

1114c Coordinates

ECL (4 mo.)

0505

ECL (2.5 mo.)

Coordinates 1114a

ECL (4 mo.)

Coordinates 1114b

Commissure transection 3 day survival Commissure transection 2 day survival Commissure transection 3 day survival Commissure transection 3 day survival 2SpCi ( 1 ~ 1 3H ) Proline 1 .OA, 9.0L, 14 deep 3 day survival Commissure transection AND 12.5uCi (0.SpI) 3H Proline Bilaterally in the hippocampus 3.OA, 9.0L, 14 deep (3 day survival) Commissure transection AND 25uCi ( 1 ~ 1 )3H Proline Bilaterally in the hippocampus l.OA, 9.0L, 14 deep (3 day survival) Commissure transection AND 25uCi ( 1 ~ 1 3H ) Proline Bilaterally in the hippocampus 1.OA. 9.0L, 14 deep (3 day survival)

ECL (4 mo.)

Coordinates

Fink-Heimer Fink-Heimer Fink-Heimer Autoradiography

Fink-Heimer Autoradiography

Fink-Heimer Autoradiography

* Commissure transection was incomplete and injection on contra side was misplaced. Data are only for ipsilateral projections Fink-Heimer Autoradiograp hy * Both injections were misplaced. Data are only for FinkHeimer staining of commissure projections

The first column indicates the number assigned to each animal. The second column indicates the survival interval following the EC lesion. The third column indicates the operative procedures used for producing the secondary lesions and for the injections of tracers, coordinates for lesions and injections, amounts of 3H-leucine injected, and survival interval following the lesions or injections. The fourth column indicates the histological procedure employed. Comments (*) indicate special circumstances (i.e., misplaced injections). ECL, entorhinal cortex lesion.

Table 4. Experimental Animals (Processed for Timm’s Stai ni na)

Animal No. 0407 0706 0710

Lesion and Survival

0925

ECL (2.5 mo.) ECL (2 mo.) ECL (3 mo.) (involved Hippocampus) ECL (2 weeks)

0818 1023 1024 1106 1112

ECL ECL ECL ECL ECL

(1 mo.)

(1.5 mo.) (3 weeks) (1.5 mo.) (2 weeks)

Mossy Fiber Sprouting

None None Substantial

Substantial (see Fig. I I) None Slight None None Slight

The first column indicates the number assigned to each animal. The second column indicates the survival interval following the EC lesion. The third column indicates the extent of mossy fiber sprouting (as revealed by the existence of a supragranular band of mossy fiber terminals. ECL, entorhinal cortex lesion.

above the floor of the cranium. Three of the animals that were prepared for autoradiography were perfused for Timm’s staining. The remainder were perfused with 10%formalin in saline. For details on the survival time after the primary lesion, the amount of radioactivity injected, the postinjection survival interval, and the method of histological preparation, see Tables 1 and 2 . The projections of the commissuraliassociational pathways were assessed using autoradiographic and Fink-Heimer techniques. The distribution of the commissural pathway was assessed in three normal control animals and seven animals with long-standing unilateral entorhinal cortex lesions by transecting the hippocampal commissures and then using the Fink-Heimer technique to reveal the distribution of terminal degeneration (see Tables 1 and 2 for further details). To transect the hippocampal commissures, animals were anesthetized and prepared for surgery as described above. The skull overlying the midsaggital sinus on one side was removed, the dura was opened, and a surgical knife made from a scalpel blade was lowered at the midline to a depth of 17 mm. The surgical knife was moved from its insertion point (about 15 mm anterior to the intra-aural line) in a posterior direction to the intra-aural line, thus transecting the hippocampal commissures and overlying corpus callosum. As with the entorhinal cortex lesions, these animals exhibited no postsurgical complications. The animals were allowed to survive for 2 or

LESION-INDUCED SYNAPSE REORGANIZATION / Steward

3 days and were then prepared as described below for histology using the Fink-Heimer technique. Commissural and associational projections were evaluated autoradiographically in four normal control animals, and three animals with long-standing unilateral entorhinal cortex lesions. Animals were anesthetized, placed in a stereotaxic apparatus, and 3H-proline was injected into the hippocampus using stereotaxic techniques, as described above. For details on the survival time after the primary lesion, the amount of radioactivity injected, the coordinates for the injections, and the postinjection survival interval, see Tables 1 and 3. Four animals (one normal and three with long-standing unilateral entorhinal cortex lesions) received simultaneous injections and commissure transections. In this way, the projections of the commissural system were traced using the Fink-Heimer technique, and ipsilateral associational projections were traced using autoradiography. In several of these cases, one or both injections missed their targets or the transections were incomplete, as noted in the respective tables; thus, the combined procedures revealed only some of the projections that were targeted.

HISTOLOGY Autoradiography For cases in which the only tract tracing technique used was autoradiography, animals were euthdnized with Nembutal, and perfused transcardially with 10% formalin in saline. The brains were removed, blocked, as described in ReinosoSuarez (1961), and placed in 30% sucrose/lO% formalin until the brains sank. The brains were then sectioned at 40 pm on a freezing microtome. A 1-in-20 series of sections was mounted on microscope slides, defatted in xylene, rehydrated through graded alcohols, and then coated with Kodak NTB2 photographic emulsion. Autoradiographs were exposed in the dark for 3-5 weeks, developed using Kodak D-19, and stained with cresyl violet. The three animals that were prepared for both autoradiography and Timm’s staining were deeply anesthetized and perfused as described below in preparation for Timm’s staining. The brains were sectioned on a freezing microtome, and two 1-in-20 series of sections were collected. One series was prepared for autoradiography as described above; the other was stained by the Timm’s procedure as described below.

The Fink-Heimer Technique Animals were euthanized with Nembutal, and perfused transcardially with 10% formalin in saline. The brains were removed, blocked, as described in Reinoso-Suarez (1961), and placed in 30% sucrose/lO% formalin until the brains sank. The brains were then sectioned at 40 km on a freezing microtome. A l-in-20 series of sections was stained using a modification of the Fink-Heimer technique (Fink and Heimer, 1967).

Timm’s staining Animals prepared for staining with the Timm’s method were euthanized with an overdose of Nembutal, and perfused transcardially in two stages, as described by West et al.,

251

(1981). The animals were first perfused with 1.0 liter of a solution containing 1 1.7 g of Na2S in 0.lM phosphate buffer (pH 7.6). The perfusion was then continued with 1.0 liter of 1% paraformaldehyde/l.25% glutaraldehyde in the same buffer. The brains were removed, placed in 30% sucrose until they sank, and sectioned using a sliding microtome. A l-in20 series of sections were mounted on microscope slides and stained according to the Timm’s procedure of Haug, as modified by West et al. (1981).

RESULTS Organization of the hippocampal formation of cats Prior to describing the experimental results, a brief review of the organization of the cat hippocampal formation is in order. As in other species, the subdivisions of the hippocampal formation can be readily distinguished in sections stained using the Timm’s stain for heavy metals (see Fig. 1). The boundary between the CAI and CA2-CA3 regions of the hippocampus can be easily distinguished because of the presence of darkly stained mossy fiber terminals in stratum lucidum of CA3. These mossy fiber terminals originate from dentate granule neurons, and terminate on the most proximal apical dendrites of the pyramidal neurons. Mossy fibers are also present throughout the hilus of the dentate gyrus (CA4). In a few cats, we did observe some “straying” mossy fibers in the CA1 region. It was our impression that some of these cats had features suggesting a Siamese heritage. This is consistent with the report of Laurberg and Zimmer (1981) of straying mossy fiber5 in Siamese cats. Even in these cats, it was still quite easy to distinguish the boundary of the normal mossy fiber zone because the straying fibers, when present, were sparse. The Timm’s stain also marks the terminal fields of the other major afferent systems in the hippocampus and dentate gyrus. Terminal fields in the stratum oriens and stratum radiatum of the hippocampus were heavily stained, whereas stratum lacunosum-moleculare was lightly stained. The darkly-stained laminae correspond to the terminal fields of the commissural and associational projections to the hippocampus, and the lightly-stained zone in the distal dendritic regions corresponds to the terminal field of the projection from the entorhinal cortex. In the molecular layer of the dentate gyrus, there was a trilaminar pattern of staining. The three laminae correspond to the site of termination of afferents from the lateral and medial entorhinal cortex (outer and middle zones) and the commissuraliassociational projection (inner zone). The darkest staining was in the outer molecular layer, the middle molecular layer was light, and the inner molecular layer exhibited intermediate staining intensity. This is somewhat different than the pattern in rats, in which the inner molecular layer is the most heavily stained (Zimmer, 1974). In terms of the relative densities of staining of the three laminae, the cat appears to be more similar to the domestic pig (Holm and Geneser, 1991). In the following sections, we will describe the reorganization of the projections of the entorhinal cortex itself, the expansion of the terminal field of the commissurallassocia-

252

HIPPOCAMPUS VOL. 2, NO. 3, JULY 1992

Fig. 1 . Organization of the hippocampus of the cat. (A) Nissl stained section. (B) Section stained by the Timm’s method. Note the trilaminar pattern of staining in the molecular layer of the dentate gyrus. CAI and CA3 indicate the major subdivisions of the hippocampus. DG, dentate gyrus; Sub, subiculum (magnification x 32).

tional pathway, and the sprouting of the mossy fibers after unilateral lesions of the entorhinal cortex. In order to describe these changes in pathways, we will first define the basic organization of each pathway in normal cats, as revealed by our techniques.

Projections from the entorhinal cortex to the hippocampus and dentate gyrus The normal projections from the entorhinal cortex to the hippocampus and dentate gyrus have been extensively described in rats and rabbits, and there is also some information on these projections in cats (Steward and Messenheimer,

1978; Witter and Groenewegen, 1984; Van Groen and Lopes da Silva, 1985). There are two major projection systems: (1) a projection to the outer molecular of the dentate gyrus that is predominantly ipsilateral; and ( 2 ) a bilateral projection to the stratum lacunosum moleculare of CA1, which contains the distal dendrites of hippocampal pyramidal neurons. As illustrated in Fig. 2, these projections were heavily labeled following injections of 3H proline into the entorhinal cortex. The injection site for this case was centered in the medial entorhinal cortex about 0.5 mm anterior to the section illustrated. One projection that is present in other species but has not

Fig. 2. Projections from the entorhinal cortex to the hippocampus of a normal cat (animal #0128c, see Table 1 for details). A and €3 illustrate the pattern of termination of entorhinal projections in the hippocampus and dentate gyrus ipsilateral (A) and contralateral (B) to an injection of 3H proline. Note the transsynaptic labeling of mossy fibers ipsilateral to the injection. C and D illustrate the very sparse labeling in the molecular layer of the dentate gyrus contralateral to the injection indicating the presence of a sparse crossed temporodentate (CTD) projection. CTA, the terminal field of the crossed temporoammonic pathway in stratum lacunosum-molecdare of CA1; sg, stratum granulosum. Other abbreviations are as for Figure 1. Calibration bars: A and B, 0.5 mm; C and D, 200 pm.

254

HZPPOCAMPUS VOL. 2, NO. 3, JULY 1992

been described in previous studies in cats is the crossed projection from the entorhinal cortex to the contralateral dentate gyrus. Although this “crossed temporodentate” (CTD) pathway is very sparse, it is important in the present context since it is an important potential source of fibers that reinnervate the dentate gyrus after unilateral entorhinal cortex lesions (Zimmer and Hjorth-Simonsen, 1975; Steward et al., 1976; Steward, 1976b). In other species, the CTD is difficult to detect using silver staining techniques (Goldowitz et al., 1975; Hjorth-Simonsen and Zimmer, 1975). However, the projection can be reliably demonstrated using orthograde tract tracing techniques (Steward et al., 1976; 1980). Thus, it was of considerable interest whether a CTD projection would be revealed by the autoradiographic techniques used in this study. As illustrated in Fig. 2D, there was light labeling in the molecular layer of the dentate gyrus contralateral to the injection (arrows), consistent with the presence of a CTD pathway. The labeling was most dense in temporal regions of the hippocampus at a site homologous with the area that was most heavily labeled ipsilateral to the injection. This projection was evident in both of the normal animals prepared for autoradiography following injections of proline into the EC, although all projections were much more lightly labeled in the second animal.

Reinnervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesions The lesions in the present study were comparable to those we have described in a previous study (Steward and Messenheimer, 1978). In most animals, the lesions Completely destroyed the entorhinal cortex and parasubiculum, along with portions of the neocortex lateral to the rhinal fissure (see for example Figs. 3 and 4).In most animals, the presubiculum was destroyed, and in some animals there was damage to the subiculum. To determine whether unilateral lesions of the EC induced a sprouting of projections from the surviving contralateral EC, we used tract-tracing techniques to determine whether there was an increase in the density of the CTD pathway in the animals with long-standing unilateral EC lesions. As illustrated in Figure 3, there was clear evidence of an increase in the density of labeling of the CTD projection to the molecular layer of the dentate gyrus that had been deprived of its normal input from the ipsilateral entorhinal cortex. It was noteworthy that the labeling was most prominent in the suprapyramidal blade of the dentate gyms immediately adjacent to the terminal field of the crossed temporoammonic (CTA) pathway in CA1. In fact, in the animal illustrated in Figure 3, there was no detectable labeling in the infrapyramidal

blade. The pattern was similar in a different animal with a similar lesion and injection, in that labeling was most dense in the portion of the dentate gyrus adjacent to the hippocampal fissure. However, in this second animal, there was also light labeling in the infrapyramidal blade (data not shown). As is evident in Figure 3, there was a considerable shrinkage of the molecular layer of the denervated dentate gyrus. It is important to consider the possibility that the apparent increase in labeling in the CTD terminal field reflects the compression of existing elements into a smaller volume. Two features of the data argue against this possibility. First, the extent of the increase in labeling appeared greater than would result from shrinkage. However, we did not prepare a sufficient number of animals to evaluate this issue using the quantitative autoradiographic analyses that have been used in rats (Steward et al., 1976). Second, the increases in labeling occurred selectively in the suprapyramidal blade, whereas shrinkage occured in both blades. Taken together, these two facts suggest that the increase in labeling reflects a sprouting response similar to that which has been described in rats.

The commissural/associational systems The commissural/associational (C/A) projections to the dentate gyrus were evaluated using both Fink-Heimer and autoradiographic tract-tracing techniques. The complete projection system of the commissural system can be revealed by the pattern of degeneration in the hippocampus following transection of the hippocampal commissures. This approach offers the advantage that one can compare the distribution of the commissural system on the two sides after unilateral entorhinal cortex lesions. A potential disadvantage of the lesioning technique is that other projections can be damaged (for example, the septohippocampal projections). However, this is probably not a large concern because degeneration resulting from damage to septa1 projections is sparse even when the projection is completely destroyed, and the present lesions would be expected to damage only the medial-most septohippocampal fibers. Nevertheless, it is important to use autoradiographic techniques to corroborate the results of the experiments using silver staining. Moreover, the autoradiographic technique is useful for tracing the ipsilateral associational system, and for defining the discrete topographic organization of projections from particular locations.

Distribution of the terminal field of the commissural pathway as revealed by FinkHeimer staining Examples of the knife-cut lesions of the commissures are illustrated in Figure 4.The cuts transected most of the corpus callosum, slightly damaged the dorsal portion of the septum,

Fig. 3. Projections from the surviving entorhinal cortex contralateral to an EC lesion (animal #0131, see Table 2 for details). A and B illustrate the pattern of termination of entorhinal projections in the hippocampus and dentate gyrus ipsilateral (A) and contralateral (B) to the injection. C and D illustrate the labeling in the molecular layer of the dentate gyrus contralateral to the injection that had been deprived of its normal innervation from the ipsilateral EC. Note that the CTD projection is especially dense in the suprapyramidal blade (arrows in D) immediately across the hippocampal fissure from the terminal field of the CTA pathway in CA1. The dashed line indicates the location of the hippocampal fissure. There is no detectable labeling in the infrapyramidal blade (C). Abbreviations are as for Figure 2 . Calibration bars: A and B, 0.5 mm; C and D,

200 wn.

LESION-INDUCED SYNAPSE REORGANIZATION / Steward

255

256

HIPPOCAMPUS VOL. 2, NO. 3, JULY 1992

and completely transacted both the dorsal and ventral hippocampal commissures. As illustrated in Figure 5 , transection of the hippocampal commissures led to terminal degeneration in the stratum oriens and radiatum of the hippocampus, and in the inner molecular layer of the dentate gyrus. The bands of degeneration ceased abruptly at the boundary between stratum radiatum and stratum lacunosum-moleculare in the hippocampus proper, and at the boundary between the inner one-third and the outer two-thirds of the molecular layer of the dentate gyrus. In the case illustrated in Fig. 5, the density of terminal degeneration is greatest in the rostral hippocampus (compare C and D). In other animals, however, the density of degeneration was comparable in rostral and caudal regions. An important feature of the organization of the C/A system in rats is that the terminal fields of the commissural and associational pathways are not entirely coextensive. In particular, in the lateral portion of the suprapyramidal blade, the commissural projections are limited to the inner-most portion of the CIA zone, whereas the associational projections terminate throughout the zone. The difference in the distribution of the two projections is thought to be due to difference9 in the time of ingrowth of commissural vs. associational axons (Gottlieb and Cowan, 1972). However, in the cat, the thickness of the terminal field of the commissural pathway (as revealed by Fink-Heimer staining) was similar in the two leaves, and appeared to occupy the full extent of the C/A zone. This was directly demonstrated in one animal by comparing the distribution of the commissural projection as revealed by Fink-Heimer staining, with the distribution of the associational pathway as revealed by autoradiography (Fig. 6). In the portion of the dentate gyrus where the associational projections were maximally labeled, the terminal fields of the commissural and associational pathways were virtually coextensive in both infrapyramidal and suprapyramidal leaves. Topographic organization of the commissural and ipsilateral associational pathways as revealed by orthograde tract-tracing techniques

Fig. 4. Examples illustrating the commissural transections and entorhinal lesions. (A) A section at the level of the septum (Sept). Note the complete transection of the corpus callosum (CC), and the slight damage to the dorsal septum. (B) A section at the level of the dorsal psalterium (psd). Note the complete transection of all crossing fibers. (C) An example of the extent of a typical entorhinal cortical lesion. Arrows indicate the lesion boundary. All sections were stained by the FinkHeimer method, thus degeneration debris in the terminal fields of the commissural pathway can be detected in B and C. DG, dentate gyrus.

Figures 7 and 8 illustrate some features of the topographic organization of the commissuraliassociational projections as revealed by the autoradiographic tract-tracing technique. These two figures compare the distribution of terminal labeling after injections into the rostralidorsal (Fig. 8) or caudal/ ventral (Fig. 9) hippocampus. These sites were chosen to evaluate differences in the pattern of projections along the septotemporal axis of the hippocampus. In general, both the contralateral (commissural) and ipsilateral (associational) projections were heaviest at the same septotemporal level as the injection site. This was especially true in the case of the more rostrddorsal injection. For example, the injection centered in the rostralidorsal dentate gyrus (Fig. 7B) produced heavy labeling of the the commissural projections to the homologous portion of the contralateral dentate gyrus (Fig. 7A), while the projections to more caudal regions on the contralateral side were very light (Fig. 7C). In contrast, the injection centered in the caudaliventral hippocampus produced labeling of commissural projections to both rostral and caudal portions of the contralateral dentate

LESION-INDUCED SYNAPSE REORGANIZATION / Steward

25 7

Fig. 5. Commissural projections in the cat as revealed by Fink-Heimer staining (animal #0404, see Table 1). A illustrates the pattern of degeneration in the rostralidorsal hippocampus after a complete transection of the commissures. B reveals the distribution of terminal degeneration across the hippocampal laminae. alv, alveus; 0, stratum oriens; p, stratum pyramidale; 1, stratum lucidum; r, stratum radiatum; m, stratum moleculare; hf, hippocampal fissure; sg, stratum granulosum. C and D illustrate the distribution of degeneration debris in the caudalhentral hippocampus. a, b, and c indicate the positions at which high magnification photomicrographs were taken in 9 and 10. Calibration bars: A and C, 0.5 mm; B and D, 200 km.

258

HIPPOCAMPUS VOL. 2, NO. 3, JULY 1992

Fig. 6. Co-localization of the terminal fields of the comrnissural and associational projections in the molecular layer of the dentate gyrus. Commissural and associational projections were traced in an individual animal using a combination of FinkHeimer and autoradiographic techniques (animal #0317, see Table 1 and Methods section for details). A and C illustrate the width of the terminal field of the commissural pathway in the lateral portion of the suprapyramidal blade as revealed by Fink-Heimer staining (dark field illumination). B and D illustrate the width of the terminal field of the associational system as revealed by autoradiography. Calibration bars: A and B, 200 km; C and D, 100 pm. Abbreviations are as in Figure 2.

gyrus (Fig. 8A and C). However, the labeling at the homologous level was still somewhat higher (Fig. 8C). The projections to the hippocampus proper exhibited a more restricted rostrocaudal distribution than the projections to the dentate gyrus. For example, in Figure 7, the commissural projection to the more temporal portion of the contralateral dentate gyrus was labeled, whereas there was no labeling in the hippocampus proper at this level (Fig. 7C). Similarly, in Figure 8, labeling of the commissural projections to the hippocampus was greatest in the caudauventral region

homologous to the site of the injection (Fig. 8C). In confiimation of previous autoradiographic studies of commissural and associational projections in cats, labeling was present in both stratum radiatum and stratum oriens ipsilateral and contralateral to the injection (Van Groen and Wyss, 1988). This is in contrast to the situation in other species, in which ipsilateral projections are concentrated in stratum radiatum and crossed projections are concentrated in stratum oriens (Van Groen and Wyss, 1988). The autoradiographic preparations also revealed features

LESION-INDUCED SYNAPSE REORGANIZATION i Steward

259

Fig. 7 . Commissural and associational projections from a site in the rostral/dorsal hippocampus (animal #0519a, see Table

1). A illustrates the pattern of termination of commissural projections contralateral to the injection (which was centered in the rostraVdorsa1 hippocampus near the section illustrated in B). C and D illustrate the distribution of labeling in the caudal ventral hippocampus contralateral (C) and ipsilateral (D) to the injection. Abbreviations are as in Figure 2. Calibration bar, 0.5 mm.

260

HZPPOCAMPUS VOL. 2, NO. 3, JULY 1992

Fig. 8. Commissural and associational projections from a site in the caudaliventral hippocampus (animal #0520b, see Table

1). A and C illustrate the pattern of termination of commissural projections contralateral to the injection (which was centered in the candalhentral hippocampus near the section illustrated in D). B illustrates the distribution of labeling in the rostral/ dorsal dentate gyrus ipsilateral to the injection. Note the absence of labeling in the hippocampus proper in B despite the prominent labeling of the associational projection to the dentate gyrus. Abbreviations are as in Figure 2. Calibration bar, 0.5 mm.

Fig. 9. Expansion of the terminal field of the commissural pathway after unilateral EC lesions as revealed by Fink-Heimer staining (animal #0223, see Table 3). A, B, and C illustrate the width of the band of terminal degeneration on the control side of an animal with a long-standing unilateral EC lesion. A’, B’, and C’ illustrate the width of the band of degeneration in complementary positions on the side of the long-standing lesion.

262

HZPPOCAMPUS VOL. 2, NO. 3, JULY 1992

LESION-INDUCED SYNAPSE REORGANIZATION / Steward

of the laminar organization of the C/A projections to the dentate gyrus which, to the best of our knowledge, have not previously been described. Figure 7 reveals that following an injection into the rostraudorsal hippocampus, the band of labeling in the inner molecular layer of the dentate gyms occupied almost the full width of the CIA terminal field at the same septotemporal level on the contralateral side (Fig. 7A). There was only a slight thinning of the band of labeling in the lateral-most portion of the suprapyramidal blade. However, in caudal and ventral regions, the band of labeling was much thinner, and was restricted to the most distal portion of the CIA terminal field. Importantly, the thinning of the band of labeling was evident both contralateral and ipsilateral to the injection, and on both sides, the thin band of labeling appeared to occupy the same sublamina of the CIA terminal field. Following an injection into the caudalhentral hippocampus (Fig. 8), the band of labeling appeared to occupy almost all of the CIA terminal field ipsilateral and contralateral to the injection in the septal portions of the dentate gyms (Fig. 8A). At more temporal sites, the band of labeling appeared to occupy the full width of the CIA zone in the infrapyramidal blade, but was thinner than the total C/A zone in the suprapyramidal blade (Fig. 8C). It was especially noteworthy that the band of labeling on the contralateral side occupied a greater proportion of the CIA zone than was the case following the injection into the rostralldorsal dentate gyrus (compare Figs. 7C and 8C). This fact is significant, since the commissural fibers that originate from the caudalhentral hippocampus must be longer than the fibers originating from the rostraVdorsa1 hippocampus (because the commissural tibers cross at the most septal extreme of the hippocampus). The general conclusion suggested by these results is that fibers projecting from the injection site toward the septal pole terminate throughout the CIA terminal zone in the inner molecular layer, whereas fibers projecting from the injection site to more temporal regions terminate in only a portion of the CIA terminal field. This is true of both ipsilateral and contralateral projections from a given location. Thus, there is a complex laminar organization of the CIA projection system so that projections from different septotemporal levels terminate within different proximodistal laminae within the C/ A terminal field. A much more extensive analysis will be required to fully characterize the nature of this laminar organization. Expansion of the terminal fields of the C/A systems following lesions of the entorhinal cortex The distribution of the commissural system in animals with long-standing EC lesions was evaluated using the Fink-Heimer technique and orthograde tract-tracing techniques.

263

The studies that evaluated sprouting with the Fink-Heimer technique used the classical primary-secondary lesion paradigm. The “primary” lesions of the entorhinal cortex that induced the sprouting were unilateral, thus denervating the ipsilateral dentate gyrus. Animals were allowed to survive long enough to allow for removal of most of the degeneration debris from this primary lesion. The “secondary” lesions of the commissures then interrupted the commissural projections to both hemispheres. Figure 9 illustrates sample photomicrographs of the terminal fields of the commissural system on the “control” side with an intact EC, and the side on which the EC had been destroyed. These paired photographs illustrate comparable sites on the two sides. A and A’ illustrate the suprapyramidal blade of the rostraUdorsa1 portion of the dentate gyrus (at the level indicated as “a” in Fig. 5A). B and B’ illustrate the suprapyramidal blade in a more caudal zone, at approximately the level indicated by “b” in Figure 5C. C and C’ illustrate the infrapyramidal blade of the dentate gyrus at about the level indicated by “c” in Figure 5C. Although the molecular layer on the side with the long-standing EC lesion is shrunken, the terminal field of the commissural system was wider than on the control side in most locations. The increase in the width of the terminal field was especially striking in the infrapyramidal blade. It is important to note that the increases were in absolute width; thus, the distal most boundary of the band of degeneration was actually further from the granule cell layer on the side of the EC lesion. Therefore, the increases in the width of the terminal field of the commissural system cannot be explained by the shrinkage of the molecular layer. A very similar pattern of termination was observed in all but one of the animals with long-standing EC lesions that were prepared for staining by the Fink-Heimer method after transection of the commissures. In the exceptional case (#0415) two features of the pattern of degeneration were different. First, the density of degeneration in the molecular layer of the rostraUdorsa1 dentate gyrus was dramatically less than the density of degeneration in the hippocampus proper, or in the dentate gyrus in caudalhentral regions (Fig. 10). Nevertheless, in areas with heavy degeneration, the increase in the width of the terminal field of the commissural pathway was comparable to the other cases (see Figs. lOC,C’). The second exceptional feature was a prominent band of degeneration in the outer molecular layer of the suprapyramidal blade at sites “a” and “b” (Figs. IOA,B). There were no obvious differences in the nature of the secondary lesion that would account for the differences in the pattern of degeneration in this animal. The fact that this atypical projection was evident on the side of the long-standing EC lesion, and not on the control side, suggests that it represents a pathway whose distribution was altered as a result of the EC lesion.

f

Fig. 10. Atypical distribution of terminal degeneration following commissural transections in one animal with a long-standing EC lesion (animal #0415, see Table 3). Sites from which photomicrographs were taken are comparable to those of Figure 9. Note the very dense band of degeneration in the hippocampus proper above the hippocampal fissure, in A and A‘, and the light degeneration in the molecular layer of the dentate gyrus. Large arrows in A’ and B‘ indicate the atypical band of degeneration that was observed in this animal only. In the infrapyramidal blade, the increase in the width of the terminal field of the commissural pathway was comparable to that observed in the other animals (see Fig. 9).

264 HZPPOCAMPUS VOL. 2, NO. 3, JULY 1992

LESION-INDUCED SYNAPSE REORGANIZATION / Steward It is noteworthy that, in some respects, this atypical band of degeneration in the outer molecular layer had a distribution comparable to the CTD pathway as revealed by autoradiography (i.e., that it was most prominent in the suprapyramidal blade). The expansion of the terminal field of the C/A systems was also evident using autoradiographic techniques (Fig. 11). Again, the width of the terminal field of the commissural pathway on the side ipsilateral to the EC lesion was greater than the width of the associational pathway on the opposite side (compare Figs. 11C,D). Evidence for similar sprouting of ipsilateral associational projections was obtained in a different cat, in which the associational projections ipsilateral to the lesion were traced autoradiographically (data not shown). In this latter case, the terminal field of the associational pathway ipsilateral to the lesion was wider than the terminal field of the commissural pathway on the contralateral (control) side. Taken together, these data indicate that the C/A system expanded into a portion of the zone normally occupied by afferents from the ipsilateral entorhinal cortex, although the extent of the expansion vaned in different locations.

265

intensity in the inner molecular layer (the one corresponding to the terminal field of the CIA system). This expansion was difficult to appreciate, however, because of the general increase in staining intensity in the denervated zone.

DISCUSSION

This study demonstrates three forms of sprouting in the hippocampal formation of cats. In general, the responses of the different afferent systems to lesions of the entorhinal cortex were similar to those described in rats. Although this similarity is not surprising given other examples of postlesion growth in cats (Guillery, 1972; Goldberger and Murray, 1974; Murray and Goldberger, 1974; Nakamura et al., 1974), the present observations lay to rest any nagging suspicion that there may be qualitative species differences in the ability of hippocampal systems to grow following injury. This evidence supports the notion that synapse reorganization following injury is a ubiquitous phenomenon in mammalian species, differing only because of differences in the normal organization of the pathways involved. While the overall sprouting response in cats was similar to that described in rats, there appeared to be some minor difSprouting of mossy fibers ferences in the sprouting response. For example, the expanThe distribution of mossy fibers in intact cats was very sion of the terminal field of the commissural pathway was comparable to that described in other species (based on eval- most pronounced in the infrapyramidal blade. In many sites, uations of the control side of operated animals). Timm stain- little, if any, expansion occurred in the suprapyramidal blade. ing mossy fibers were prominent in the hilus of the dentate On the other hand, the sprouting of the CTD pathway was gyrus, and the stratum lucidum of CA3 of the hippocampus. evident only in the suprapyramidal blade, in contrast to the Occasionally, Timm staining elements resembling mossy- situation in rats, where the increases in the density of the fiber terminals were observed in the inner molecular layer of CTD pathway occur in all areas in which the CTD projections the dentate gyms immediately adjacent to the layer of granule are normally found (Le., in both suprapyramidal and infracells (supragranular mossy fibers). There was some variabil- pyramidal blades). The significance of this difference in disity in the extent of the supragranular mossy fibers on the tribution is not clear. However, it is interesting that the excontrol side in the different animals. pansion of the terminal field of the commissural pathway was Comparison of the Timm staining pattern ipsilateral and greatest in the same regions in which the sprouting of the contralateral to an EC lesion suggested that EC lesions did CTD pathway was minimal. induce some mossy-fiber sprouting, although the increases Taken together with previous studies, this study suggests were not consistently observed. Supragranular mossy fibers that the examples of synapse reorganization described in the were clearly increased on the side ipsilateral to the lesion in hippocampus of rats are likely to occur generally across mamfour of the cases (see for example Fig. 12). In no case were malian species. This supports the suggestion that similar postsupragranular mossy fibers more prominent on the control lesion reorganization is likely to occur in humans as well, side. Supragranular mossy fibers were especially evident in particularly because there is already evidence that two forms animals in which the lesion encroached upon the hippocam- of sprouting do occur in humans (Grady et al., 1989; Babb et pus proper. al., 1991). Thus synapse reorganization remains a viable canThere were other changes in the Timm’s staining pattern didate mechanism for the well-known examples of functional in the denervated dentate gyrus that presumably reflect the recovery that are known to occur following brain injury in loss of EC afferents. For example, the normal trilaminar humans. staining pattern was abolished on the side of the lesion. InAt the same time, it is important to note that the precise stead, there was a region of moderate staining intensity cor- nature of the sprouting response will depend upon the orresponding to the portion of the molecular layer normally ganization of normal pathways; thus, there may be important occupied by entorhinal afferents. In some cases, there ap- species differences in normal pathways that would result in peared to be an expansion of the zone of intermediate staining a somewhat different pattern of reorganized projections. For

< Fig. 11. Expansion of the terminal field of the commissural pathway after unilateral EC lesions as revealed by autoradiographic tract tracing techniques (animal #1114d, see Table 3). A and C illustrate the distribution of the terminal field of the commissural pathway on the side of the long-standing EC lesion. B and D illustrate the distribution of the terminal field of the associational pathway on the control side (ipsilateral to the injection). The brackets in A and B illustrate the location of the higher magnification photomicrographs in C and D.

266

HZPPOCAMPUS VOL. 2, NO. 3, JULY 1992

Fig. 12. Appearance of a supragranular band of mossy fibers after EC lesions (animal #0925, see Table 4). A control side; B ipsilateral to the lesion. Note the presence of a supragranular band of mossy fibers on the side of the EC lesion The dashed line indicates the approximate position of the hippocampal fissure. Other abbreviations are as in Figure 2 .

example, some portions of the dentate gyrus of primates receive essentially no commissural projections (Amaral et al., 1984; Demeter et al., 1985). In these regions, one would not expect commissural pathways to contribute to the reinnervation of granule cells following damage to the entorhinal cortex. Similar but less obvious restrictions may arise from the differential organization of other systems in different species. Normal organization of hippocampal pathways in cats

This study also provides additional information about the normal organization of hippocampal pathways in cats. The results confirm previous findings that crossed projections in cats are at least as extensive as in rats and rabbits (Van Groen and Wyss, 1988), and that this general conclusion is also true of the crossed projections from the entorhinal cortex to the contralateral hippocampus and dentate gyrus. This study also reveals the existence of a complex laminar organization of the C/A terminal fields that has not previously been recognized. As discussed below, this complex organization suggests a mechanism for establishing laminar domains different than those proposed based upon studies in rats.

Laminar organization of terminal fields of commissural and associational systems

The key observations of the present study regarding the laminar organization of C/A terminal fields in the dentate gyrus are as follows: (1) That the CIA terminal fields originating from different locations along the septotemporal axis terminate in different sublaminae within the total CIA terminal field; and ( 2 ) that both commissural and associational projections from a given rostrocaudal site terminate in the same sublaminae on the two sides of the brain. In the areas in which associational projections occupied the full width of the CIA terminal field, the commissural projections on the opposite side did as well. In areas in which the labeled terminal fields were restricted to a portion of the C/A terminal field, the bands of labeling on the two sides occupied similar sublaminae. These features of the laminar organization of CIA projections in cats are of interest because of the temporal competition hypothesis that has been advanced to account for the laminar organization in rats (Gottlieb and Cowan, 1972). This hypothesis holds that the laminar distribution of the commissural and associational pathways results from a difference

LESION-INDUCED SYNAPSE REORGANIZATION / Steward

in the time of axon ingrowth. The hypothesis is based upon the fact that axons of the commissural system have further to grow to reach their targets than axons of the associational system. Thus, it was suggested that early in development, associational axons would reach their targets first, and would be more successful in capturing available synaptic sites. Such competition would operate in the portions of the dentate gyrus that were present at these early stages (the suprapyramidal blade). Because the infrapyramidal blade develops late, a differential distribution based upon temporal factors would be expected only in the suprapyramidal blade. There are two features of the laminar organization of CIA systems in cats that are difficult to explain by the temporal competition hypothesis. First, comparison of Figures 7C and 8C reveals that the commissural terminal field in the temporal hippocampus is wider when the injection was centered in the temporal hippocampus contralaterally than when the injection was centered in more septal regions. This is important because the axons originating from temporal sites and projecting contralaterally would of necessity be longer than the axons originating from more septal sites. Thus, in this case, the axons that would have longer to grow are more successful in capturing synaptic territory. The second feature that is difficult to explain is the symmetrical distribution of commissural and associational terminal fields in homologous sublaminae on both sides of the brain. For example, in Figure 7, both commissural and associational projections to temporal sites terminate in a thin band that occupies only a fraction of the entire C/A terminal field, and on each side, the laminar distribution of labeling is comparable. Thus, the laminar pattern of termination of both ipsilateral and contralateral components of the CIA system appears to be related in part to the points of origin and termination of the particular projections along the septotemporal axis, irrespective of the distances involved. How such a selective pattern of termination could be achieved solely on the basis of temporal competition is difficult to imagine. This conclusion does not negate the temporal competition hypothesis, but does indicate that some other mechanism must also be operating.

267

the system (Witter and Groenewegen, 1984; Van Groen and Lopes da Silva, 1985). Although the CTD pathway is so sparse that one might question its functional significance in normal animals, it represents an important source for reinnervating fibers after unilateral injury. These connections are particularly important because they are homologous with the ones that had been destroyed, and thus may carry similar information. One wonders whether similar very sparse crossed projections remain undetected in other systems that are predominantly unilateral, and whether such pathways contribute to sparing or recovery of function after unilateral injury.

ACKNOWLEDGMENTS The authors wish to thank L. Smith, R. Ogle, and P. M. Falk for technical assistance. Supported by National Institute of Health Research Grant NS 12333. References

Amaral, D. G., R. Insausti, and W. M. Cowan (1984) The commissural connections of the monkey hippocampal formation. J. Comp. Neurol. 224:307-336. Babb, T. L., W. R. Kupfer, J. K. Pretorius, P. H. Crandall, and M. S. Levesque (1991) Synaptic reorganization by mossy fibers in human epileptic fascia dentata. Neurosci. 42:351-363. Cotman, C. W . ?J. W. Geddes, and J. S. Kahle (1990)Axon sprouting in the rodent and Alzheimer’s disease brain: a reactivation of developmental mechanisms? Prog. Brain Res. 83:427-434. Demeter, S., D. L. Rosene, and G. W. Van Hoesen (1985) Interhemispheric pathways of the hippocampal formation, presubiculum, and entorhinal and posterior parahippocampal cortices in the Rhesus monkey: The structureand organization of the hippocampal cornmissures. J. Comp. Neurol. 233:30-47. Fink, R. P., and L. Heimer (1967) Two methods for selective silver impregnation of degenerating axons and their synaptic endings in the central nervous system. Brain Kes., 4:369-374. Frotscher, M., and J. Zimmer (1983) Lesion-induced mossy fibers to the molecular layer of the fascia dentata: identification of postsynaptic granule cells by the Golgi-EM technique. J. Comp. Neurol. 215299-3 11. Goldberger, M. E., and M. Murray (1974) Restitution of function and collateral sprouting in the cat spinal cord: The deafferented animal. The crossed temporo-dentate projection J. Comp. Neurol. 158:37-54. The present study also revealed the existence of a sparse Goldowitz, D., W. F. White, 0. Steward, G. Lynch, and C. Cotman (1975) AndtOmiCdl evidence for a projection from the entorhinal crossed projection from the entorhinal cortex to the contracortex to the contralateral dentate gyrus of the rat. Exp. Neurol. lateral dentate gyrus (the CTD pathway). To the best of our 47:433-441. knowledge, this projection has not previously been described in cats, although it has been described in all other species Gottlieb, D. I., and W. M. Cowan (1972) Evidence for a temporal factor in the occupation of available synaptic sites during the dethat have been examined in detail, including primates (Golvelopment of the dentate gyrus. Brain Res. 41:452-456. dowitz et al., 1975; Hjorth-Simonsen and Zimmer, 1975; Grady, M. S., J. A. Jane, and 0. Steward (1989) Synaptic reorgaSteward, 1976b; Amaral et al., 1984). nization within the human central nervous system. J. Neurosurg. The distribution of the CTD projection in cats appears to 71 534-537. be somewhat different than the distribution in rats and rab- Guillery, R. W. (1972) Experiments to determine whether retino-geniculate axons can form trans-laminar collateral sprouts in the dorbits; in the latter two species, the CTD pathway is restricted sal lateral geniculate nucleus of the cat. J. Comp. Neurol. 146:407to the most rostra1 portion of the dentate gyrus (Hjorth-Si420. monsen and Zimmer, 1975; Steward et al., 1976). In the cat, Hjorth-Simonsen, A., and J. Zimrner (1975) Crossed pathways from the CTD pathway extended well into the temporal pole of the the entorhinal area to the fascia dentata I Normal in rabbits. J. hippocampus (Fig. 2). A complete evaluation of the rostroComp. Neurol 161(1):57-70. caudal distribution of the CTD pathway will require a series Holm, 1. E., and F. A. Geneser (1991) Histochemical demonstration of injections spaced throughout the rostrocaudal axis of the of zinc in the hippocampal region of the domestic pig: 111. The entorhinal cortex because of the topographic organization of dentate area. J. Cornp. Neurol. 308:409-417.

268 HZPPOCAMPUS VOL. 2, NO. 3, JULY 1992 Laurberg, S . , and J. Zimmer (1981) Lesion-induced sprouting of hippocampal mossy fiber collaterals to the fascia dentata in developing and adult rats. J. Comp. Neurol. 200, 433-459. Lorente de No, R. (1934) Studies on the structure of the cerebral cortex. I. Continuation of the study of the ammonic system. J. Psychol. Neurol. Lpz. 46:113-177. Lynch, G., B. Stanfield, and C. W. Cotman (1973) Developmental differences in post-lesion axonal growth in the hippocampus. Brain Res. 59:155-168. Murray, M., and M. E. Goldberger (1974) Restitution of function and collateral sprouting in the cat spinal cord: The partially hemisected animal. J. Comp. Neurol. 158:19-36. Nakamura, Y.,N. Mizuno, A. Konishi, and M. Sato (1974) Synaptic reorganization of the red nucleus after chronic deafferentation from cerebellorubral fibers: An electron microscopic study in the cat. Brain Res. 82:298-301. Reinoso-Suarez, F. (1961) Tapographischer Hirnatlas der Katz fur experimental-phyJiologische Untersuchengen. E. Merck, Darmstadt Germany. Steward, 0. (1976a) Reinnervation of dentate g y m by homologous afferents following entorhinal cortical lesions in adult rats. Science 194~426-428. Steward, 0.(1976b) Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. Comp. Neurol. 167:285-314. Steward, 0. (1980) Trajectory of contralateral entorhinal axons which reinnervate the fascia dentata of the rat following ipsilateral entorhinal lesions. Brain Res., 183:277-289. Steward, 0. (1982) Assessing the functional significance of lesioninduced neuronal plasticity. Int. Rev. Neurobiol. 23:197-254. Steward, 0 . (1986) Lesion-induced synapse growth in the hippocampus: In search of cellular and molecular mechanisms. In The Hippocnmpus R. L. Isaacson, and K. H. Pribam Eds. Plenum, N.Y. pp. 65-1 11. Steward, 0. (1989) Reorganization of neuronal connections following trauma: principles and experimental paradigmns. J. Neurotrauma 6:99-152. Steward, O., C. W. Cotman, and G . Lynch (1974) Growth of a new fiber projection in the brain of the adult rat: Reinnervation of the

dentate gyrus by the contralateral entorhind cortex following ipsilateral entorhinal lesions. Exp. Brain Res. 20:45-66. Steward, O., C. W. Cotman, and G. Lynch (1976) A quantitative autoradiographic and electrophysiological study of the reinnervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesions. Brain Res. 114:181-200. Steward, O., and J. A. Messenheimer (1978) Histochemical evidence for a postlesion reorganization of cholinergic afferents in the hippocampal formation of the mature cat. J. Comp. Neurol. 178:697710. Steward, O., and S. A. Scoville (1976) Cells of orgin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J . Comp. Neurol. 169:347-370. Van Groen, T., and F. H. Lopes da Silva (1985) Septotemporal distribution of entorhinal projections to the hippocampus in the cat: electrophysiological evidence. J. Comp. Neurol. 238:l-9. Van Groen, T., and J. M. Wyss (1988) Species differences in hippocampal commissural connections: Studies in rat, guinea pig, rabbit, and cat. J. Comp. Neurol. 267:322-334. West, J. R., C. A. Hodges, and A. C. Black (1981) Distal infrapyramidal granule cell axons possess typical mossy fiber morphology. Brain Res. Bull. 6:119-124. West, J. R., and S. L . Dewey (1984) Mossy fiber sprouting in the fascia dentata after unilateral entorhinal lesions: Quantitative analysis using computer-assisted image processing. Neurosci. 13:377384. Witter, M. P.? and H. J. Groenewegen (1984) Laminar origin and septotemporal distribution of entorhinal and perirhinal projections to the hippocampus in the cat. J. Comp. Neurol. 224:371-385. Zimmer, J. (1973) Extended commissural and ipsilateral projections in postnatally deentorhinated hippocampus and fascia dentata demonstrated in rats by silver impregnation. Brain Res. 64:293-311. Zimmer, J. (1974) Long term synaptic reorganization in rat fascia dentata deafferented at adolescent and adult stages: observations with the Timm methods. Brain Res. 76, 336-342. Zimmer, J. and A. Hjorth-Simonsen (1975) Crossed pathways from the entorhinal area to the fascia dentata I1 Provokable in Rats. J. Comp. Neurol. 161(1):71-101.