Hearing Research, 55 (1991) 117-132 0 1991 Elsevier Science Publishers B.V. All rights reserved 037%5955/91/$03.50

117

HEARES 01607

Structural features of neurons in whole grafts of the rat inferior colliculus Mark C. Zrull L* and James R. Coleman 1,2 Departments of I Psychology and ’ Physiology, Urkersity of South Carolina, Columbia, South Carolina, U.S.A

(Received 9 November 1990; accepted 2 March 1991)

The inferior colliculus (IC) is a midbrain structure that receives ascending auditory input from brainstem nuclei via the lateral lemniscus, sends efferent fibers to the medial geniculate body of thalamus and receives descending projections from auditory cortex. In the rat, the IC consists of dorsal and external cortices surrounding the central nucleus of IC (CNIC) which is populated by discoid and stellate neurons: the CNIC has a laminar appearance arising from organization of lemniscal fibers and processes of discoid cells. The IC of adult rats was chosen for implantation of whole grafts of E16-17 caudal tectum into unilateral lesion sites. Dendritic and somal architecture of grafr neurons was examined 1 to 4.5 months following implantation using rapid Golgi, HRP and Nissl methods. The CNIC of rat is dominated by principal neurons with relatively flattened dendritic fields. In grafts of caudal tectum the most common neuron class observed possesses flattened dendritic arbors which often parallel one another. These neurons also resemble CNIC neurons of host tissue adjacent to the graft border. Spine formations appear on both proximal and distal dendrites of this neural type in both normal and implanted tissues. In addition, comparable somal features of graft neurons include ovoid or fusiform shapes with regular nuclear membranes as found in the normal colliculus. In Golgi stained material fewer stellate class neurons appear as in the normal CNIC, although stellate cell classes are more abundant in the pericentral areas of normal tissue. Both neuron populations are retrogradely labelled in graft and normal IC after HRP injection into the medial geniculate body. These features suggest that the graft core typically consists of prototypic CNIC cells. Other features of neuron and glial cell density vary in graft material which also shows a complex network of vasculature. These results demonstrate that whole grafts of caudal tectum placed into the inferior colliculus can form organized neural architecture similar to the normal CNIC. The somal, dendritic and spine features of these neurons form a potential substrate for connectional and functional properties which establish this preparation as suitable for further investigation as a model for development and recovery of function in the central auditory system. Inferior colliculus; Tectum; Transplantation;

Rapid Golgi; Morphology

Introduction The wealth of available information on the neuroanatomical and neurophysiological substrates of sensory systems provides the basis for good models to study many aspects of neural grafting. Neural transplantation studies have focused largely upon nonsensory pathways, such as those involved in motor activity (Bjorklund et al., 19871, and this work is achieving success in human clinical research (Lindvall et al., 1990). However, it is the distinct organization of sensory systems that may offer optimal opportunities for structural and functional studies and permit precise testing of both host and implanted tissue after graft placement. In the visual system, neural grafts of fetal tectum into the midbrain develop structural and func-

Correspondence

to: James R. Coleman, Department of Psychology, University of South Carolina, Columbia, SC 29208, U.S.A. * Presenf address: Department of Neurophysiology, 273 Medical Sciences Center, University of Wisconsin, Madison, WI 53706, U.S.A.

tional relationships with host retina and brain tissue (Harvey et al., 1982; McLoon and Lund, 1983; Dyson et al., 1988; Golden et al., 1989). Sensorimotor cortex transplanted into analogous host cortex develops circuits which have connections with subcortical structures, including pathways that mediate responses to stimulation of the thalamus and the periphery (Levin et al., 1987; Neafsey et al., 1989). In each of these models the specificity of implanted neurons to the sensory system seems to be essential for integration of the graft into the host brain. Recently, several parameters required for neural implantation of dissociated tectum into the auditory midbrain were identified (Zrull and Coleman, 1989, 1990a,b). However, details of the structure and function of neural grafts placed into the auditory system requires further investigation. The inferior colliculus (10 is a prominent and well organized auditory structure in the mammalian midbrain. Much of the IC is divided into dorsal and external cortices which encapsulate the central nucleus (CNIC) with each division having known connections with brainstem and forebrain auditory nuclei (e.g., Adams, 1979; Brunso-Bechtold et al., 1981; Oliver,

11X

1984a; Ryugo and Willard, 1985; Faye-Lund, 1985, 1986; Coleman and Clerici, 1987). Principal cells of CNIC give rise to disc shaped dendritic fields which, in conjunction with lemniscal afferents, form the laminar architecture of the nucleus (Geniec and Morest, 1971; Morest and Oliver, 1984; Oliver and Morest, 1984; Faye-Lund and Osen, 1985). In rat, these laminae are in register with tonotopic contours (Coleman et al.. 1982; Faye-Lund and Osen, 1985; Clerici and Coleman, 1986; Huang and Fex, 1986) as in other mammals (e.g., gerbil; Ryan et al, 1982); thus, a specific relationship exists between structural and functional organization of the CNIC. Often crossing laminae, the dendritic fields of large and small stellate cells represent a second neuronal population of CNIC (Oliver and Morest, 1984). Dorsal and external cortices of IC (DCIC and ECIC) each have neuronal constituents and organization that are easily distinguished from CNIC (Merest and Oliver, 1984). In the auditory system, the IC provides a unique substrate for examination of structural features of graft tissue. In previous work (Zrull and Coleman, 1990a), we demonstrated cytoarchitectural features of neurons grafted as prelabelled cell suspensions; the morphology of these cells was readily identified in counterstained tissue under bright- and darkfield illumination and under fluorescence, with some characteristics comparable to host neurons. In the present study we employed the Golgi method using whole graft procedures to determine how the architecture of implanted neurons compares to the structure of host and control IC cells. Additionally, retrograde tracing of horseradish peroxidase and standard Nissl procedures were used to elucidate relationships between neuronal architecture and graft and host tissues. The results show excellent survival and growth of grafted neurons which display several features that characterize neuron structure in the CNIC.

Materials

and Methods

Subjects Host animals were 28 Long-Evans (Harlan) rats aged 60 to 90 days and weighing 200 to 350 g. A group of 3 conspecifics, with age and weight ranges similar to hosts, served as normals for the Golgi procedure. The animals were housed in pairs or triplets in plastic shoe-box cages and maintained on a 12 h light, 12 h dark illumination cycle (lights on at 0700 h). Throughout the study, food and water were available ad libiturn. Graft tissue Rat fetuses between 16 and 17 days gestation (E16E17; mating day defined as EO) were removed individ-

ually from anesthetized (ketamine:xylazine hydrochlorides 5O:lO mg/kg b.w., ip) pregnant Long-Evans rats by caesarian section and placed on sterile glass slides. Each fetal brain was removed and placed in an ice cold nutrient mixture (HAM F-10; Sigma). Bilaterally, the midbrain tectum was identified as a region on the dorsal surface of the fetal brain extending from the mesencephalic flexure to the rhombencephalon. The caudal two-thirds of each tectum was dissected away, the remaining meninges and blood vessels carefully removed, tissue bisected, and each fetal tecta transferred to fresh ice cold nutrient mixture. Grafting procedure Animals were anesthetized with ketamine and xylazine hydrochlorides (50 : 10 mg/kg b.w., ip) and placed in a stereotaxic frame. Unilateral (N = 25) or bilateral (N = 3) lesions in host IC were produced by passing 1.2 mA dc through a tungsten electrode for 15 s. Immediately following the lesion procedure, all host rats received unilateral grafts of 1 caudal fetal tecta. Grafts were placed stereotaxically by pressure injection using a 50 ~1 Hamilton syringe equipped with a 0.8 mm (0.D.) glass capillary. Following 5 min diffusion time, the capillary was raised and inspected to verify expulsion of fetal tissue into the host brain. After successful tissue implantation, burr holes in the skull were filled with Gelfoam (Upjohn) and bone wax, and the skin was sutured. Histology procedures Animals were sacrificed at various post-graft intervals, ranging from 30 to 135 days, with a lethal overdose of sodium pentobarbital and processed by Nissl (N = 121, horseradish peroxidase (HRP; N = 6) or rapid Golgi (N = 10; normals N = 3) protocols. Nissl procedures included intracardial perfusion with 0.9% saline followed by 10.0% phosphate buffered formalin. These brains were blocked in the frontal plane and postfixed in 30.0% sucrose-buffered formalin for 24 h at 4°C. Frozen sections were cut at 50 pm, mounted onto gel-coated slides, and the tissue air dryed. Sections were dehydrated in alcohol, stained with thionin, differentiated, cleared in toluene and coverslipped with Permount. Retrograde tracing of graft and IC projections followed the protocol used by Coleman and Clerici (1987). The medial geniculate body (MGB) was stereotaxically injected with 0.1 ~1 of HRP labelled wheat germ lectin (WGA-HRP) 24 h prior to sacrificing these rats. Intracardial perfusion included ice cold 0.9% saline followed by 3.0% glutaraldehyde and then 20.0% sucrose in 7.4 pH 0.1 phosphate buffer. The brains were postfixed 24 h at 4°C in sucrose buffer, 50 pm frozen sections were cut, and the tissue was reacted using a TMB (Sigma) protocol. Tissue was mounted onto gel-

119

coated slides with alternate sections air dryed and counterstained with Neutral Red; all sections were cleared and coverslipped with Permount. Rapid Golgi processing (e.g., Clerici and Coleman, 1990) included intracardial perfusion with 50 ml 0.9% saline, removal of the whole brain and cutting a 5 mm slab centered visually at the IC. The slabs were fixed in a 0.3% osmium tetroxide 2.0% potassium dichromate solution for 72 h and then transferred to 0.75% aqueous silver nitrate for 48 h infiltration. Tissue was embedded in paraffin and 150 pm sections cut on a rotary microtome. Sections were placed in 100% ethanol, cleared with 2 changes of xylenes, mounted on glass slides and coverslipped with Permount. Analysis

Tectal grafts were located and reconstructed from Nissl and counterstained WGA-HRP sections with a projecting microscope (Bausch and Lomb). In Golgi sections, graft location was determined under brightfield illumination and initial resconstructions made by gradually reducing magnification and using a drawing tube (Sankei) attached to a Nikon Optiphot system (Plan 20, 10 and 5 apochromat objectives). Final Golgi reconstructions were made subsequently in a manner similar to those from Nissl material. Detailed anaysis of thionin stained sections and photomicrographs were made under brightfield and phase contrast illumination at 50 X to 1000 X (Nikon Optiphot system). Discoid neurons were identified by somal orientation and relatively even nuclear envelopes, and stellate cells were classified by somal shape and uneven nuclear envelopes (Oliver, 1984b; Eyerly et al., 1989). Drawings of host and graft neurons were made using the drawing tube and oil immersion (Phase 4 illumination; Plan 100 apochromat objective, 1.25 n.a.>. Plots of labelled cells in grafts and host IC were made using the projecting microscope with magnification of 80 x ; injection sites were made at 20 x . The location and classification of WGA-HRP filled neurons was verified, and structural features observed, under brightfield illumination at 50 X to 630 x using the Nikon system. Drawings of retrogradely labelled neurons were made using the drawing tube (Plan 40 apochromat objective, 0.65 n.a.1. Discoid neurons were identified as having fusiform soma and exiting processes that aligned with the long somal axis of the parent and/or adjacent discoid cells. Neurons considered stellate had a rounded or polygonal soma, and exiting processes traveled in many distinct directions from the cell body. Golgi material was observed under brightfield illumination at 50 x to 1000 x (Plan 40 apochromat objective, 0.65 n.a.; Plan 100 apochromat objective, 1.25 n.a.1. Location of drawn neurons in graft and host

tissue was accomplished by tracking individual cells through decreasing magnification and placing them on drawings made at 50 x via the drawing tube. In Golgi sections, discoid neurons were identified as having major dendritic processes oriented in line with the longer somal axis. Distinguishing features of stellate cells included processes that exited from many aspects of the neural somata and showed many branches.

Results

Neural implants were successful in 78.6% of the cases sacrificed after at least 1 month in the host. Survival rates varied by approximately +4.0% to a low of 75.0% in rats sacrificed 3.5 or more months after grafting; Table I on cumulative graft survival shows no effect of post-implant duration (P = 0.98). Graft configuration varied from a small volume of tissue occupying a unilateral dorsal cavity to larger tissue volumes extending across the midline (Fig. 1). The dorsal surface of the graft may approximate the normal lines of the IC, including shallowing at the midline commissure, and in other cases displays contour irregularities. In frontal view, the graft can replace ablated portions of the midbrain either medially or laterally. Implanted tissue also fills cavities in and dorsal and rostra1 to the IC as viewed in the sagittal plane (Fig. 2B). Cytoarchitecture

In Nissl-stained material the morphology and packing density of graft neurons can readily be compared to cells of host or control IC. In at least a portion of each graft there were neuron populations which resemble neurons that constitute the normal adult CNIC. The most common neuron observed in normal material is one with an oval or fusiform soma, usually characterized by a relatively smooth nuclear membrane (Oliver, 1984b). In graft segments these neurons are often oriented along a single dimension (e.g., medial-lateral or dorsomedial directed; Fig. 2D,E). In both host and control CNIC many subpopulations of neurons are similarly oriented (Fig. 2C,E). Graft cells occasionally enter a swirling pattern, usually close to the margin of the donor tissue (Fig. 3B). Less common in both nor-

TABLE

I

CUMULATIVE SURVIVAL IOR COLLICULUS LESION

OF TECTAL SITES

Months

Number of host rats Percent surviving grafts a Survival

of 3.5 months

GRAFT’S

IN INFER-

after implantation

1.0

1.5

2.0

2.5

3.0

3.5+

28 78.6

20 80.0

17 82.4

14 78.6

6 83.3

4 75.0

or longer.

a

Fig. I. Photomicrograph

of 60 day post-implant

graft adjacent to undamaged host inferior normal adult IC and commissure.

ma1 and grafted tissues are more rounded, asymmetrical or triangular neurons which are often identified by an irregular nuclear membrane. Soma of graft neurons generally approximate and occasionally exceed normal cells in size. Neuron cell density in grafts is variable (Fig. 31, but sometimes higher than in comparable adult tissues, and there are more glial cells in the neuropil, particularly at the graft-host interface. Architecture of projecting neurons Of 6 animals receiving WGA-HRP injections into MGB, 5 had surviving tectal grafts. In all of these cases both graft and host IC contained labelled neural cells with injection sites restricted to ventral and dorsal divisions of MGB (MGv, MGd). The predominant neural type labelled in remaining host CNIC has an elongated or fusiform cell body, and the visible processes of these neurons tend to align forming dorsomedial to ventrolateral rows in the host tissue (Fig. 4). Intermixed with the principal cells are small and moderately sized stellate neurons with visible processes exiting from many aspects of rounded somata. Graft projections to MGv and MGd of the auditory thalamus arise from neurons with characteristics simi-

colliculus

(IO.

Such grafts

showed

conformation

like

lar to those of principal and stellate cells of host CNIC (Fig. 4). Neurons with elongated cell bodies are common; however, processes of these cells often suggest an intrinsic organization of graft tissue which may not mimic contours of host IC. Other retrogradely labelled graft neurons are rounded and resemble the stellate cells of the normal rat IC. Labelled neurons populate central regions of graft tissue and those areas in contiguity with host IC. With the exception of the largest grafts found in Nissl cases, implanted tissue in WGAHRP cases tended to be very similar in size, location and shape to grafts found in Nissl and Golgi processed brains. Dendritic architecture The most distinctive neuron class of the normal central nucleus is a cell with flattened dendritic arborizations which give a discoid appearance to each dendritic field. This organization is obvious in sections viewed in the frontal plane (Fig. 5). In mid-sectors of the normal central nucleus most dendritic fields of discoid neurons characteristically are oriented ventrolateral to dorsomedial. Neurons with larger dendritic trees usually have larger soma; soma of discoid neu-

Fig. 2. (k) Frontal view of a tectal graft 60 days after implantation into lesion cavity. Large grafts often filled the region of ablated host IC extending from cerebral cortex to contralateral host IC. Scale is 500 pm. (B) Sagittal drawing shows that some grafts lodged between the remaining portion of damaged host IC and cortex rostrally, and a contiguous border with the dorsal surface of host SC was formed; 60 days post-implant. Scale is 500 Frn. (C) Undamaged host right CNIC showing fusiform neurons oriented along dorsomedial to ventrolateral gradients (left top to bottom the origin of individual cells in host-graft right). (D) Cells from the center of a tectal graft with morphology and density similar to those of the host CNIC. (E) Often clear at low magnification, D., E., Oil immersion, Plan 100 apochromat interface regions (thick line) is difficult to discern at high magnification due to orientation and apparent maturation of grafted tectal neurons. (C., objective, 1.25 na.; Cbl, cerebellum; CNIC, central nucleus of inferior colliculus; f, fusiform neurons; G, graft; IC, inferior colliculus; s, stellate neurons; SC, superior coliicuhts.)

Fig. 3. (A) Fusiform and stellate neurons of the normai rat CNIC are prominent in thionin stained sections. (Dorsal is up, medial to the left). (B) Neural populations in graft tissue may form heterogeneous patterns. Here densely packed lateral (left) and dorsal (top) cells swirl into a center region containing many neurons with fusiform morphology. (0 In the sagittal plane, the caudoventral portion of an implant from a Nissl case reveals fusiform and rounded graft neurons with packing density similar to normal IC.

123

rons commonly orient in a plane close to that of the dendritic fields. In combination with similarly oriented incoming afferents from the lateral lemniscus, the central nucleus shows a clear laminated fiber alignment in frontal view. This alignment is altered at the margins of the central nucleus including the ventral border (Fig. 5). Other neurons in the normal central nucleus pos-

sess more spherical dendritic fields. Fewer such stellate neurons are observed, although they are intermingled among the principal neuron population. In addition, neurons with pyramidal shaped somal and dendritic features sometimes appear near the ventral margin of the central nucleus. Neurons localized to grafted tissue and to surround-

Fig. 4. Graft and IC neurons are retrogradely labelled following WGA-HRP injection that includes MGv and MGd. In this tectal graft, labelled cells are located both distal and proximal to the border with host tissue. Lower left shows mostly labelled host fusiform principal cells organized along a dorsomedial to ventrolateral contour. Elongated fusiform and rounded stellate neurons were among cells labelled in the graft (lower right); intercellular organization of graft neurons may not align with adjacent CNIC. (G, graft; IC, inferior colliculus; MGd, dorsal medial geniculate; MGv, ventral medial geniculate; scale is 20 pm.)

I24

ing host neuropil are readily identified in Golgi material. As in normal neurons, dendrites of grafted neurons most commonly show flattened arborizations.

Graft neurons with larger dendritic fields and soma are easily comparable in size and organization to discoid neurons of the normal central nucleus (Fig. 6). Smaller

cl

m

I

t

v

Fig. 5. Neurons located at the ventrolateral aspect of normal rat central nucleus of inferior colticulus. Processes of discoid principal ceils form the lamina that characterizes CN; large stellate cells with defined polygonal morphology are more common in EC. In this figure and Figs. 7, 8 and 10 dendritic spines are not shown as the focus of these analyses was general morphology of implanted neurons; typical spine formations are shown in Figs. 6 and 9. (CN, cenIra1 nucleus; EC, external cortex; Plan 40 objective, 0.65 na.; scale is 20 pm.)

125

graft neurons with these features are also routinely stained. Typically, one to three dendritic trunks appear at or near the somatic poles of these neurons. Secondary branches from these trunks are often observed, while further bifurcations occur in several neurons. In many cases, the orientation of most processes of graft neurons follow a parallel course which approximates the lamination gradient expected in the host colliculus (see Figs. 6 and 9). At the border with host tissue, dendrites of graft neurons sometimes orient along the axis of the graft-host interface (Fig. 6). In addition to

graft neurons with various degrees of flattening of dendritic fields, occasional cells show more stellate-like features similar to normal and host neurons (Figs. 6, 7 and 8). Similar to normal neurons with flattened dendritic fields, many graft cells show spine formations on both proximal and distal graft segments (Fig. 6 and 9). Spine size ranges from 1 pm to 3 pm with shapes varying from stubs to knobby spines. The longer knobby spines are observed on more dendrites of normal discoid neurons than those in grafts. Encrustacians also appear

d

m

J

Fig. 6. Rapid Golgi sections revealed flattened (straight arrows) graft neurons, with characteristics like discoid cells (d) of the normal central nucleus of inferior colliculus (CNIC), and neurons with stellate (s) morphology. Fibers of host origin (curved arrows) follow ventrolateraldorsomedial contours expected in normal CNIC; however, upon crossing into graft tissue host dendrites align with processes emanating from graft neurons. While not always matching the host CNIC, flattened cells and processes within grafts maintain an intrinsic organization. Varicosities and swellings are more common to processes of graft neurons but are found on host fibers as well (star). (70 days post-implant; Plan 40 objective, 0.65 n.a.; d, dorsal, m, medial; orientation bars are 20 Km.)

126 on both normal and graft neurons. In contrast, the less populous graft neurons with stellate features usually have fewer spines on dendrites and regions of more complex dendritic branching (see Fig. 8) than are found in most comparable normal cells. On the other hand, some graft neurons show thickenings along both primary and secondary dendrites which are less common in neurons of the normal adult centra1 nucleus. These features are similar to growth cones exhibited during the course of neural development (Pinto-Lord and Caviness, 1979; DiFiglia et al., 1988). Silver staining methods also highlighted the vasculature of graft material. The course of blood vessels within the graft often nearly parallels the graft-host interface (Fig. 10). This organization of graft vasculature may be more pronounced than in normal central nucleus, although there are considerable numbers of approximately orthagonal vessels in both tissues.

Discussion

The present study demonstrated that neurons of tectai grafts placed into the damaged inferior colliculus show a high survival rate (78.6%) and cellular growth comparable to normal inferior colliculus. Healthy grafts ranged in size from those replacing small lesions to ones extending bilaterally above the remnant colliculus. In most cases the graft material closely followed the host tissue contours at a readily identifiable interface of fibers and some glial cells (DiFiglia et al., 1988). As judged by cytoarchitecture and dendritic organization the material using a whoIe grafting procedure appears optimal even compared to suspension grafts into inferior colliculus (Zrull and Coleman, 1990a). The whole graft procedure yields somal features which are identifiable as ovoid or round cells with smooth nuclear membranes, as well as cells with more rounded

h Fig. 7. Further evidence of interface between located in host tissue crossed into the neuropil

graft and host tissue at 105 days survival. Processes of both flattened of adjacent graft tissue. Small cells with stellate architecture frequent (Plan 40 objective, 0.65 ma.; scale is 20 pm.1

(f) and steltate ts) neurons the host-graft border (line).

Fig. 8. The processes of flattened (f) neurons often were oriented according to location of the tectal graft extending toward host tissue (upper most cell). Some complex stellate (s) cells were found in grafts. Note extensive dendritic branching of the upper stellate neuron as compared to the lower stellate cell more common to tectal grafts. (60 days post-implant; Plan 40 objective, 0.65 n.a.; scale is 20 pm.)

or irregular nuclei (Oliver, 1984b). Furthermore, the observation that graft neurons were typically as large or larger than collicular neurons in the host or in

control animals is supported by similar comparisons between neuronal soma of graft and host tissues in other material (Perry et al., 1985; Harvey and Whar-

Fig. 9. Large graft neurons show disc shaped dendrsttc tIcIds typica ot normal CNIC principal cells; note spines on dendrites (bracket) similar to, but fewer in number than, those found on normal principal ceils. While the discoid neurons Cd) represent a typical constituent of most grafts, the large: size of the stellate (s) celf is somewhat unusual. Note that groups of spines(bracket) are more common on the processes of the stellate cell which is positioned near a graft border. Arrows denote varicositics mare common on grafted neurons than in host or normal tissue. These dendritic swellings or varicosities are most prominent on processes observed near other neurons. (70 day post-implant; oil immersion, Plan 100 objective, 1.25 n.a.; scale is 10 km.)

Lm

d

129

Fig. 10. The presence of stained vasculature in Golgi material reveals further qualities neurons at the interface betweenare host and ingraft tissuewith (line). Large in graft tissue located regions abundant blood vessels course along the graft-host border with branches entering both regions; vascularization. (Plan 40 objective, 0.65 n.a.; scale is 20 Wm.1

130

ton, 1986; DiFiglia et al., 1988). These results suggest that graft neurons are able to express several characteristic structural features found in neurons of normal adult colliculus and that conditions in the host tissue favor growth of graft neurons. Survival and final structural development of grafted neurons is likely influenced by host-graft interconnections and intragraft cell interactions. Differences in survival of graft neuron subpopulations requires further delineation of graft neuron classes (e.g., flattened cell types) and examination of the presence or absence of normal connectional patterns by viable graft neurons. This process would also be facilitated by study of cell differentiation and classification present in fetal tectum at the time of transplantation. The dendritic features of a common neuron class in collicular grafts resembled the organization of dendrites of cells intrinsic to central nucleus of inferior colliculus. The dendrites with flattened fields routinely observed in Golgi preparations of graft material were often comparable in orientation and length to the discoid neurons which characterize the central nucleus of many mammals (Geniec and Morest, 1971; Oliver and Morest, 1984; Faye-Lund and Osen, 1985; Meininger et al., 1986). In particular, neurons with one to three dendritic trunks giving rise to various degrees of compressed fields is common in the central nucleus of mammals. Furthermore, neurons with these features were located near the border and in the core of grafts, suggesting that the morphology of graft neurons was not a function of interactions occurring only at the graft-host interface. The orientations of flattened dendritic trees of many graft neurons may resemble a form of laminar organization. A laminated or stratified appearance has been observed in fetal grafts of cerebellum (Kramer et al., 19831, hippocampus (Kramer et al., 1983; Nilsson et al., 19881 and neocortex (Jaeger and Lund, 1981); tectal grafts placed above the intact superior colliculus result in near normal cell classes, but without lamination (Harvey and Warton, 1986). Since implanted cells of the inferior colliculus in the present study were derived from the fetal caudal tectum it is assumed that primary dendritic formation reflects characteristics of donor tissue expression as in other systems (Frotscher and Zimmer, 1987; Nilsson et al., 1988; Tonder et al., 1989). In the present study, graft material is selected to maximize the presence of presumptive CNIC neurons (Altman and Bayer, 1981). There is currently no information on perinatal architectural organization of these cells, although there is clear orientation preference of CNIC neurons by postnatal day 8 (Dardennes et al., 1984). Initial orientation selectivity of dendrites in these cells prior to arrival of afferent fibers would reveal

epigenic influences of cellular programs which would likely be observed in graft cells as well. Further evidence for normative growth of implanted IC cells is the appearance of spine formations on some neurons, particularly those with discoid features. Spine density is low on both primary and secondary dendrites of graft neurons as is the case in cells of the normal central nucleus (Oliver and Morest, 19841, although fewer Golgi stained graft neurons appeared to show the larger spines. Spine and encrusted formations in graft neurons likely reflect influences of afferent systems as in normal tissues (e.g., Pinto-Lord and Caviness, 1979). This idea is further supported by recent work which shows functional activation of tectal graft neurons by sound, and by lemniscal and brachial fiber interrelations with the graft (Coleman and Zrull, 1990; Zrull and Coleman, 1990b). The present study shows that many graft neurons exhibiting morphological similarities to host cell types form connections with the normal thalamic target of CNIC efferents. Occasional varicosities on dendrites and enlargements of endings suggest that certain graft neurons have not reached the full stage of maturation at the time of sacrifice (see also DiFiglia et al., 19881. The present results show that several anatomical features of graft neurons from the caudal tectum can mimic the characteristics of neurons normally found in the inferior colliculus including those of the central nucleus. Electrophysiological recordings and other functional measures are required to establish whether identified graft neurons in the inferior colliculus acquire typical monaural and binaural response features to acoustic stimuli. Recent accumulation of data on immunohistochemical (e.g., Roberts and Ribak, 1987; Eyerly et al., 1989; Coleman et al., 1991) and neuropharmacological (Faingold et al., 1989) properties of neurons of the central nucleus provide additional criteria for identification of successful grafts into the rat inferior colliculus. For example, the presence of neurons immunoreactive for calbindin or parvalbumin can be used to identify structural characteristics of specific populations of graft and host collicular neurons (Coleman et al., 1990). Examination of neurochemical properties of afferent and efferent systems of graft tissue is essential for understanding physiological performance of neurons and behavioral recovery (Gage and Bjiirklund, 1986; Nilsson et al., 1988; Daszuto te al., 1989; Lindvall et al., 1990). Since the transmitter systems and receptor properties of collicular neurons are becoming better known, the acoustic response characteristics of graft neurons can ultimately be studied for glutaminergic, glycinergic, GABAergic and other neurochemical influences.

131

Acknowledgements We wish to thank Luba Novak for her help preparing Golgi material. This work was supported by NIH Grant NS 20785, BRSG Grant 507-RR07160 and a Deafness Research Foundation Grant to James R. Coleman.

References Adams, J.C. (1979) Ascending projections to the inferior colliculus. J. Comp. Neurol. 183,519-538. Altman, J. and Bayer, S.A. (1981) Time origin of neurons of the rat inferior colliculus and the relations between cytogenesis and tonotopic order in the auditory pathway. Exp. Brain Res. 42, 411-423. Bjorklund, A., Lindvall, 0. Isacson, O., Brundin, P., K. Wictorin, K., Strecker, R.E., Clarke, D.J. and Dunnett, S.B. (1987) Mechanisms of action of intracerebral neural implants: studies on nigral and striatal grafts to the lesioned striatum. TINS 10, 509-516. Brunso-Bechtold, J.K., Thompson, G.C. and Masterton, R.B. (1981) HRP study of the organization of auditory afferents ascending to the central nucleus of inferior colliculus. J. Comp. Neurol. 197, 705-722. Clerici, W.J. and Coleman, J.R. (1986) Resting and high-frequency evoked 2-deoxyglucose uptake in the rat inferior colliculus: developmental changes and effects of short-term conduction blockade. Dev. Brain Res. 27, 127-137. Clerici, W.J., McDonald, A.J., Thompson, R. and Coleman, J.R. (1990) Anatomy of the rat medial geniculate body: II. Dendritic morphology. J. Comp. Neurol. 296, l-23. Coleman, J.R. and Clerici, W.J. (1987) Sources of projections to subdivisions of the inferior colliculus in the rat. J. Comp. Neurol. 262, 215-226. Coleman, J.R., Campbell, M. and Clerici, W.J. (1982) Observations on tonotopic organization within the rat inferior colliculus using 2-deoxy-D(l-‘H) glucose. Otolaryngol. Head Neck Surg. 90, 795800. Coleman, J.R., Zrull, M.C., Pinek, B. and McDonald A.J. (1991) The inferior colliculus: calbindin and parvalbumin immunoreactivity in neural grafts. Exp. Neurol. (in press). Coleman, J.R. and Zrull, MC. (1990) Structural features of fetal tectum grafted into the adult inferior colliculus. Assoc. Res. Otolaryngol. 13th Annual Midwinter Meeting. Daszuta, A., Kalen, P., Strecher, R.E. Brundin, P. and Bjorklund, A. (1989) Serotonin neurons grafted to the adult rat hippocampus. II. 5-HT release as studied by intracerebral microdialysis. Brain Res. 498, 323-332. Dardennes, R., Jarreau, P.H. and Meininger, V. (1984) A quantitative Golgi analysis of the post-natal maturation of dendrites, in the central nucleus of the inferior colliculus of the rat. Devel. Brain Res. 16, 159-169. DiFiglia, M., Schiff, L. and Deckel, A.W. (1988) Neuronal organization of fetal striatal grafts in kainateand sham- lesioned rat caudate nucleus: light- and electron-microscopic observations. J. Neurosci. 8, 1112-1130. Dyson, S.E., Harvey, A.R., Stone, A.F. and Bell, G.A. (1988) Metabolic activity in rat tectal grafts is influenced by host sensory innervation. J. Neurosci. 8, 1822-1829. Eyerly, M.R., Coleman, J.R., Zrull, M.C. and McDonald, A.J. (1989) Glutamate and aspartate: localization and colocalization in the rat inferior colliculus. Sot. Neurosci. Abstr. 1.5, 747.

Faingold, C.L., Gehlbach, G. and Caspary, D.M. (1989) On the role of GABA as an inhibitory transmitter in the inferior colliculus: iontophoretic studies. Brain Res. 500, 302-312. Faye-Lund, H. (1985) The neocortical projection to the inferior colliculus in the albino rat. Anat. Embryol. 173, 53-70. Faye-Lund, H. (1986) Projection from the inferior colliculus to the superior olivary complex in the albino rat. Anat. Embryo]. 175, 35-52. Faye-Lund, H. and Osen, K.K. (1985) Anatomy of the inferior colliculus in rat. Anat. Embryol. 171, l-20. Frotscher, M. and Zimmer, J. (1987) GABAergic nonpyramidal neurons in intracerebral transplants of the rat hippocampus and fascia dentata: a combined light and electron microscopic immunocytochemical study. J. Comp. Neurol. 259, 266-276. Gage, F.H. and Bjdrklund, A. (1986) Cholinergic septal grafts into the hippocampal formation improve spatial learning and memory in aged rats by an atropine-sensitive mechanism. J. Neurosci. 6. 2837-2847. Geniec, P. and Morest, D.K. (1971) The neuronal architecture of the human posterior colliculus. A study with the Golgi method. Acta Otolaryngol. Suppl. (Stockh.) 295, l-33. Golden, G.T., Reyes, P.F. and Fariello, R.G. (1989) Morphological and functional studies of rat tectal transplants. Brain Res. Bull. 22, 461-468. Harvey, A.R., Golden, G.T. and Lund, R.D. (1982) Transplantation of tectal tissue in rats. III. Functional innervation of transplants by host afferents. Exp. Brain Res. 47, 437-445. Harvey, A.R. and Warton, S.S. (1986) The morphology of neurons in rat tectal transplants as revealed by Golgi-Cox impregnation. Anat. Embryol. 174, 361-367. Huang, C-M. and Fex, J. (1986) Tonotopic organization in the inferior colliculus of the rat demonstrated with the 2-deoxyglucase method. Exp. Brain Res. 61, 506-512. Jaeger, C.B. and Lund, R.D. (1981) Transplantation of embryonic occipital cortex to the brain of newborn rats: a golgi study of mature and developing transplants. J. Comp. Neurol. 200. 213230. Kromer, L.F.. Bjorklund, A. and Stenevi, U. (1983) Intracephalic neural implants in the adult rat brain. I. Growth and mature organization of brain stem, cerebellar and hippocampal implants. J. Comp. Neurol. 218, 433-459. Levin, B.E., Dunn-Meynell, A. and Sved. A.F. (1987) Functional integration of fetal cortical grafts into the afferent pathway of the rat somatosensory cortex (SmI). Brain Res. Bull. 19, 723-734. Lindvall, 0.. Brundin, P., Widner, H.. et al., (1990) Grafts of fetal dopamine neurons survive and improve motor function in Parkinson’s disease. Science 247, 574-577. McLoon, S.C. and Lund, R.D. (1983) Development of fetal retina, tectum, and cortex transplanted to the superior colliculus of adult rats. J. Comp. Neurol. 217, 376-389. Morest, D.K. and Oliver, D.L. (1984) The neuronal architecture of the inferior colliculus in the cat: defining the functional anatomy of the auditory midbrain. J. Comp. Neural. 222. 209-236. Neafsy, E.J., Sorensen, J.C.. Tonder, N. and Castro, A.J. (1989) Fetal cortical transplants into neonatal rats respond to thalamic and peripheral stimulation in the adult, An electrophysiological study of single-unit activity. Brain Res. 493, 33-40. Nilsson. D.G.. Clarke, D.J., Brundin, P. and Bjorklund, A. (1988) Comparison of growth and reinnervation properties of cholinergic neurons from different brain regions grafted to the hippocampus. J. Comp. Neurol. 268, 204-222. Oliver, D.L. (1984a) Dorsal cochlear nucleus projections to the inferior colliculus in the cat: a light and electron microscopic study. J. Comp. Neurol. 224, 155-172. Oliver, D.L. (1984b) Neuron types in the central nucleus of the inferior colliculus that project to the medial geniculate body. Neuroscience 1 I, 409-424.

132 Oliver. D.L. and Morest, D.K. (1984) The central nucleus of the inferior colliculus in the cat. J. Comp. Neurol. 222, 237-264. Perry. V.H.. Lund, R.D. and McLoon, S.C. (1985) Ganglion cells in retinae transplanted to newborn rats. J. Comp. Neurol. 231. 353-363. Pinto-Lord, M.C. and Caviness, V.S. (1979) Determinants of cell shape and orientation: a comparative Golgi analysis of ceil axon interrelationships in developing neocortex of normal and Reeler mice. J. Comp. Neurol. 187, 49-70. Roberts, R.C. and Ribak, C.E. (1987) An electron microscopic study of GABAergic neurons and terminals in the central nucleus of the inferior colliculus of the rat. J. Neurocytol. 16, 333-345. Ryan, A.F., Woolf, N.K. and Sharp, F.R. (1982) Tonotopic organization in the central auditory pathway of the mongolian gerbil: a 2-deoxyglucose study. J. Comp. Neurol. 207. 369-380.

Ryugo, D.K. and Willard, F.H. (1985) The dorsal cochlear nucleus of the mouse: a light microscopic analysis of neurons that project to the inferior colliculus. J. Comp. Neural. 242, 381-396. Tonder, N., Sorensen, T., Zimmer, J., Jorgensen, M.B., Johansen, F.F. and Diemer, N.H. (1989) Neural grafting to ischemic lesions of the adult rat hippocampus. Exp. Brain Res. 74, 512-526. Zrull, M.C. and Coleman, J.R. (1989) Adult-like structure of fetal tectum grafted to inferior colliculus. Sot. Neurosci. Abstr. 15. 747. Zrull, M.C. and Coleman, J.R. (1990a) Fetal tectum grafted as a cell suspension into the adult rat inferior colliculus. Hear Res. 45, 237-246. Zrull, M.C. and Coleman, J.R. (1990b) Properties of fetal tectum grafted to adult and neonatal rat inferior colliculus. Sot. Neurosci. Abstr. 16. 722.