THE JOURNAL OF COMPARATIVE NEUROLOGY 303:75-100 (1991)

Dendritic andAxonal Morphologyof HRP-hjectedNeurom in the Inferior Colliculw of the Cat DOUGLAS L. OLIVER, SJXIGEYUKI KCTWADA, TOM C.T. YIN, LEWIS B. HABERLY, AND CRAIG K. HENKEL Department of Anatomy and Center for Neurological Sciences, The University of Connecticut Health Center, Farmington, Connecticut 06032 (D.L.O., S.K.); Departments of Neurophysiology (T.C.T.Y.) and Anatomy (L.B.H.), University of Wisconsin, Madison, Wisconsin 53706; Department of Anatomy, Bowman Gray Medical School, Winston-Salem, North Carolina 27103 (C.K.H.)

ABSTRACT The dendritic and axonal morphology of neurons in the inferior colliculus of the cat was investigated after intracellular injection of HRP, in vivo. All injected axons gave off local collaterals, and most showed a widespread distribution and lacked a specific orientation. In contrast, the dendrites of injected neurons were distinguished by their degree of orientation and the direction of the longest axis of orientation. Dendrites showed a high, moderate, or low degree of orientation. Most highly oriented cells had their longest axis in the rostrocaudal direction with fewer in the mediolateral direction. In the central nucleus, only the rostrocaudally oriented cells correspond to the disc-shaped cells identified in Golgi preparations. Unlike most cells in our sample, the two cells that were disc-shaped had axons that were parallel to the orientation of the dendritic tree. In the dorsal cortex, rostrocaudally oriented cells also were found, but they had unoriented axons. In both the central nucleus and dorsal cortex, cells with a mediolateral axis of orientation or no specific orientation correspond to stellate cells and had axons with widespread local collaterals. These results suggest that an extensive network of local axon collaterals may contribute to neural processing within the inferior colliculus. In the central nucleus, local axons may establish connections within or across the fibrodendritic laminae. In the dorsal cortex, the local and afferent axons may form a complex reticular network. Finally, some injected cells had axons terminating locally and also entering the brachium of the inferior colliculus. This suggests that cells in the inferior colliculus may function as both interneurons and projection neurons. Key words: auditorypathways,binaural hearing, neural networks

Nearly all the auditory structures in the lower brainstem have ascending projections that converge in the inferior colliculus (IC). The organization of the inputs may provide an important structural basis for the responses of neurons in the IC. For example, lemniscal afferents often form sharp, well-defined, tonotopically organized bands within the IC (e.g., Oliver, '84b, '87; Shneiderman and Henkel, '87; Shneiderman et al., '88). Some of these aKerents may carry monaural information from the cochlear nucleus (van Noort, '69; Osen, '72; Roth et al., '78; Adams, '79; BrunsoBechtold et al., '81; Willard and Martin, '83). In other cases, they may convey binaural information from the superior olivary complex (Elverland, '78; Glendenning and Master-

o 1991 WILEY-LISS, INC.

ton, '83) or the lateral lemniscal nuclei (Goldberg and Moore, '67; Kudo, '81; Whitley and Henkel, '84; Tanaka et al., '85; Covey and Casseday, '86). Although the organization of afferents may be an important component of the morphological substrate for information processing, little is known of the organization of local axons in the IC. Intrinsic axons could influence nearby neurons or those in other subdivisions. Since all parts of the colliculus do not receive the same inputs, local connections could play a major role in distributing information within the IC. Accepted September 10,1990.

D.L. OLIVER ET AL.

76

Thus far, local axons in the IC have been studied only in Golgi-impregnated material (Ramon y Cajal, '09; Geniec and Morest, '71; Rockel and Jones, '73a,b; FitzPatrick, '75; Oliver and Hall, '78a; Morest and Oliver, '84; Faye-Lund and Osen, '85; Meininger et al., '86). This technique has both the advantage and the disadvantage that hundreds of cells and axons may be impregnated. Although this abundance is useful for deriving neuronal organization, it is nearly impossible to trace single axons though serial sections and relate their morphology to dendritic fields. However, these technical problems can be overcome by staining single axons by intracellular injections of a tracer. In the present study, the dendritic and axonal morphology of cells in the IC were studied after intracellular injection with HRP. This report will focus on the anatomical findings from these experiments. A second report will detail the physiological results and the correlations with the anatomy. Preliminary data were presented previously (Haberly et al., '79; Kuwada et al., '80; Oliver et al., '86).

lMETHoDS Experimental preparation The data were obtained from 34 adult cats with clean external and middle ears. Barbiturate anesthetized cats (Nembutal35 mg/kg i.p.1 underwent a tracheotomy, femoral vein catheterization, and removal of both pinnae. Supplemental doses of Nembutal were delivered through the i.v. cannula to maintain areflexia throughout the experiment. The dorsal and lateral surface of the right IC were exposed by aspiration of the occipital cortex and removal of the bony tentorium. Sometimes a small stainless steel foot-plate was attached to the skull via a stainless steel wire and used to retract the anterior surface of the cerebellum. A Plexiglas chamber centered over the IC was cemented to the skull, filled with mineral oil, and sealed with a glass plate. Since the microelectrode was inserted through a sleeve in the glass plate, this apparatus allowed the visual placement of the microelectrode onto the exposed surface of the IC. A Trent-Wells or Burleigh microdrive was used to advance the electrode. The microelectrodes were filled with 4% HRP in 0.5M KC1 and 0.04M Tris buffer (pH 8.6). A jet-stream beveler (Ogden et al., '78) was used to lower the resistance of the electrode from 200-400 MO to 30-100 MO. Cells were impaled by advancing the electrode in small steps (1-2 pm) or by delivering positive (5-100 nA) current pulses (100 msec) through the microelectrode. Cells were stained with HRP by delivering positive (5-15 nA) current pulses (100 msec on, 100 msec of!$ for 2-3 minutes into the impaled neuron. After an injection, the electrode was slowly retracted. Successive electrode penetrations were spaced 1-2 mm apart, and no more than 5 penetrations were made per animal.

Histologicalprocedures At least 1 hour after the last injection, while still under deep barbiturate anesthesia, the cat was perfused with a saline or buffer wash followed by aldehyde fixatives in neutral phosphate buffer. Earlier experiments used a single fixative (1,000 ml, 1%paraformaldehyde, 1.25%glutaraldehyde), while later experiments used a two stage fixative (500 ml, 0.5% paraformaldehyde, 1%glutaraldehyde; fol-

lowed by 1,000 ml, 1%paraformaldehyde, 3% glutaraldehyde; see Oliver, '84a,b). After dissection of the brain, the block containing the IC was sectioned at 70-100 pm on a freezing microtome or a Vibratome in the transverse or sagittal plane. After processing the sections with HRP histochemistry (Adams, '811, they were either mounted on glass slides and lightly counterstained with cresyl violet or prepared for electron microscopy. The latter method included postfixation in 1% osmium tetroxide, block staming in uranyl acetate, and Epon embedding in 1" x 3" molds (Cast-a-slide) under a dacron sheet (Aclar). Both methods allowed for complete analysis at the light microscopic level.

Analysis Injected cells were examined with a 63 x or 100 x long working-distance objectives (N.A. = 1.25) and drawn with a camera lucida at 700-1,000 x . In addition, three-dimensional reconstructions of injected cells were made on a IIEC, LSI/ll-73 computer interfaced to a Numonics 2200 digitizing tablet and an AED 512 graphics terminal. Custom software was developed at the University of Connecticut Health Center to digitize, align, annotate, connect, and display the dendritic segments of injected neurons in adjacent serial sections. The display of reconstructed 3D data used proprietary software (Visions, Digital Eflects, Inc.). The locations of the injected cells are illustrated in the transverse plane. To illustrate cells originally cut in the sagittal plane, the cells were plotted on a standard series of sagittal section, the series was reconstructed on the computer and resliced. This rendered computed images of the positions of the cells in the transverse plane. To quantify the degree of dendritic orientation, the perimeter of the dendritic field of each neuron was measured in several anatomical planes after rotation on the computer. Deviation from roundness (Oliver, '85; Shneiderman et al., '88) in each plane was used to establish the degree of dendritic orientation. This metric is independent of size and reflects the length of a dendritic field compared with its width in the plane that reveals the highest degree of orientation. For example, a perfectly spherical denclritic field has a round profile in every plane of cut and measures zero. A deviation score of 25 shows the length is twice the width while a score of 35 indicates that the length is three times the width.

RESULTS Forty neurons and axons were injected in the IC. Of these, 18 cells or axons appeared to be well filled and were reconstructed from serial sections. Most of the injected cells were located in two subdivisions of the IC, the dorsal cortex and central nucleus. Some injected axons appeared to originate outside of the IC.

Neurons in the IC have orienteddendrites andaxonalcollaterals There are two primary results. First, all impregnated axons gave off local axon collaterals with diverse morphology. Second, the dendritic trees of injected neurons could be distinguished by their degree and direction of orientation. For this reason, we divided the cells into three groups

ANATOMY OF HRP-INJECTED CELLS IN THE INFERIOR COLLICULUS (highly oriented, moderately oriented, or unoriented dendritic fields) on the basis of the deviation of the dendritic tree from roundness (see Methods), and examined the relationship of this factor to the axonal morphology. For each group, we shall first present the dendritic features followed by the axonal anatomy. Finally, we shall present the morphology of injected axons whose cell bodies are outside the IC proper.

77

diameter of the cell bodies could be large (85-9, 35 pm), medium-sized (78-349, 24 pm; 79-61, 26 pm), or small (79-314, 17 pm; 78-341, 19 pm). The long axes of the dendritic fields of 79-341 and 79-61 were about 600 pm long, while that of cell 79-314 was just over 300 pm. Cells that were highly oriented were not always in a rostrocaudal plane. Figure 7 shows a transverse view of three simultaneously injected neurons located in the middle part of the dorsal cortex. Each showed a long axis in the mediolateral direction. To simplify comparisons with rostroNeurons with highly oriented dendriticfields caudally oriented cells, Figure 8 shows these three cells Neurons with the highest degree of orientation had (85-10A-C) in the horizontal plane. This view showed that dendritic fields in which the longest axis was at least three the longest axis could be somewhat oblique to the frontal times the width (deviation from roundness >35). Seven plane, more mediocaudal-to-lateroventral (e.g., cells 85highly oriented cells had their longest axis in the rostrocau- 10C,B), a direction that is orthogonal to the rostrocaudally dal plane, while four cells had their longest axis in the oriented cells. Another cell (83-37, Fig. 8) also showed a mediolateral plane. mediolateral direction of orientation but was located in the Rostrocaudally oriented cells appeared to parallel the central nucleus. orientation of the lemniscal afferents. An example of this Like rostrocaudally oriented cells (Fig. 3), mediolaterally relationship is shown in Figure 1.A cell with rostrocaudally oriented cells also showed a mixed morphology. For examoriented dendrites is illustrated next to an injected axon, ple, cells 85-10A and B were covered with dendritic appendboth in the transverse plane. The cell and axon were from ages, while the other cell in Figure 7 was not. Likewise, the separate injections in the same animal. Both the long axis of soma sizes varied from small (12-17 pm, 85-10A,B,C) to the dendritic field and the long axis of the axon were parallel medium (21 pm, 83-37), although the long axis of the to each other. The axon was probably lemniscal in origin dendritic fields were all in the same range (320-386 pm). since lemniscal axons typically have dense clusters of Axons of these highly oriented cells exhibited two types of boutons and a dorsomedial to ventrolateral orientation morphology. They were either highly oriented and paral(Oliver and Morest, '84). Because of their planar geometry, lelled the dendritic tree, or they lacked a specific orientarostrocaudally oriented cells tended to have two long axes. tion. The local collaterals were concentrated in the same This feature can be seen by comparison of cells illustrated subdivision as the dendritic field. in different planes (Fig. 2). Neuron 78-349 (Fig. 2, left) is Neurons whose axons parallelled the long axis of the shown in the sagittal plane and has a dendritic field with its dendritic tree were found in the central nucleus. For longest axis in the rostrocaudal orientation. The second example, the axon of neuron 85-9 (Fig. 2) showed the same long axis was directed dorsomedial to ventrolateral and is orientation as the dendrites when viewed in the transverse seen in cells 85-9 (Fig. 2, upper right) and 83-25 (Fig. 2, plane. Scores of axon collaterals extended several hundred microns away from the dendritic field. Both the branches lower right), illustrated in the transverse plane. Cells with rostrocaudally oriented dendrites were located and the parent axon were very fine (less than 1 pm in in the central nucleus and dorsal cortex. The dendritic diameter). Terminal boutons were small and difficult to geometry of these cells can be compared directly with distinguish; but synaptic contacts were visible with the three-dimensional computer graphics (Fig. 3). Computed electron microscope (Oliver, unpublished observations). The axon from cell 78-349 also paralleled the dendritic images of four highly oriented cells in the horizontal plane each display a rostrocaudal orientation. In the central tree. In the sagittal plane (Fig. 21, this relationship was not nucleus, cells 85-9 and 78-349 had a rostrocaudal axis that obvious. However, computer rotation of the axon reveals is 15-30' from the sagittal plane. Although these neurons that it is highly oriented and in the same plane as the had a dendritic field less than 100 pm wide, cell 85-9 had a dendritic field (Fig. 9). This is seen by comparing the long axis (400 pm) about that of cell 78-349 (800 pm). In horizontal view of the axon (Fig. 9) with the horizontal view the dorsal cortex, cells 79-314 and 79-61 had a similar of the dendrites (Fig. 3). Since the plane of the axon was not alignment to those in the central nucleus (85-9, 78-349). exactly parallel to the sagittal plane, the optimal view, However, the longest axis of the cells in the dorsal cortex where the axonal orientation is sharpest, corresponds to a were about 45" to the sagittal plane as compared with the 15" rotation on the Y-axis (Fig. 9, OPTIMAL). One part of 15-30" for the cells in the central nucleus (Fig. 3). Rostro- the axon extended laterally in the direction of the brachium caudally oriented cells were found in both the superficial of the IC. Compared to cell 85-9, the axon of cell 78-349 (also see Figs. 4, 5) and deep part of the dorsal cortex (see (Fig. 2) extended even farther away from the dendritic tree, i.e., about a millimeter or more and was larger in diameter also Fig. 61, near the dorsal border of the central nucleus. Although rostrocaudal orientation was a primary feature (about 1.5 pm diameter) and less highly branched. In contrast to the cells from the central nucleus, the of these cells regardless of location, other morphological characteristics were more variable. Neurons varied in the axons of cells with rostrocaudally oriented dendrites in the extent of dendritic branching and in the number of den- dorsal cortex were not highly oriented (Figs. 4-6, 9). dritic appendages. Some cells had few dendritic appendages Although their axonal arbors were elongated in the rostro(Fig. 3, 85-9) while others exhibited a moderate number of caudal direction, they lacked a sharp planar orientation. pedunculated dendritic spines (Fig. 4, 79-314). Size of the For example, Figure 4 (top) shows the axon from cell 79-314 perikarya and dendritic fields also differ. The average in the sagittal plane. The optimal orientation (OPTIMAL,

D.L. OLIVER E'I' AL.

78

I

0

L

L

83-36

Fig. 1. Camera lucida drawing of an injected cell and separately injected axon in the central nucleus (inset).The neuron is a highly oriented type and is parallel to the adjacent axon. The arrow indicates

\the orientation of the axon parallel to the fibrodendritic laminae. Not,e the tight clusters of terminal boutons on the axon. Transverse plane. Scale = 100 bm.

ANATOMY OF HRP-INJECTED CELLS IN THE INFERIOR COLLICULUS

79

x

S A G ITTA L

TRANSVERSE

D

L, 78-349 D

L

L

Fig. 2. Three highly oriented cells in the central nucleus. The inset drawing of the inferior colliculus shows the location of each neuron (dot) in the transverse plane. Cell 78-349 in the sagittal plane has an axon ( a )that extends dorsally and ventrally to the dendritic tree. Cells

83-25 -

I

85-9 and 83-25 are in the transverse plane. Cell 85-9 has a richly arborized axon (a), drawn in thinner lines, that is parallel to the dendritic tree. Arrows indicate the orientation of the fibrodendritic laminae (also shown in the inset, lower right). Scale = 100 pm.

D.L. OLIVER ET AL.

80

M

78-349

79-61

Fig. 3. Computed images of four highly oriented cells with a rostrocaudal orientation. Although they were not cut in the same plane originally, they can be compared after a rotation to simulate their

appearance in the horizontal plane. Cells 85-9 and 78-349 are from the central nucleus. Cells 79-314 and 79-61 are from the dorsal cortex. Scale = 100 pm.

15"from the transverse plane along the X axis) showed that the axonal orientation from the dorsal cortex was broad compared with the narrow plane of the axon from the central nucleus. The same was true when the axons were viewed in the horizontal plane (HORIZONTAL).Moreover, by comparing the dendritic (Fig. 3) and axonal (Fig. 9) rotations in the horizontal plane, it is clear that the axonal orientation did not parallel the dendritic orientation. Axons from the highly oriented cells in the dorsal cortex were similar to each other. For example, the axon from 79-314 (Fig. 4) extended rostrocaudally from the cell body about 1 mm in each direction. It formed infrequent, simple branches as it coursed 1.8mm in the mediolateral direction. Clusters of terminal boutons were infrequent. Axons from two other cells showed similar features (Figs. 5,6).

However, after computer rotation, the horizontal view shows that it also had a second long axis in the mediolateral direction (Fig. 15). This cell had a large soma (32 pm) and a rich dendritic arbor with primary branches that formed equal-sized daughter branches. Higher order branches were unequal in size. The dendritic surfaces had sinuous undulations almost without dendritic appendages. A second moderately oriented cell (85-14, Figs. 1 2 , 15) had dendrites oriented in the mediolateral direction. This cell was found in the pars centralis of the central nucleus. In both the transverse and horizontal views, the dertdritic field was oriented perpendicular to the inferred axis of the fibrodendritic laminae (Fig. 12, dashed lines) and is about 500 pm in length. The rostrocaudal axis of the dendritic field was modest (around 200 pm). While the branching pattern of this cell was similar to the dorsoveritrally oriented cell (79-7), it had pedunculated and filamentous appendages that the other cell lacked. Its soma (17 p m ) was about half the diameter of the dorsoventrally oriented cell (79-7). Two moderately oriented cells i n the dorsal cortex showed dendrites parallel to the horizontal plane. One cell, located deep in the dorsal cortex, had its long axis in the metliolatera1 direction (84-25, Fig. 13).The dendritic field was wide and flat and roughly paralleled the dorsal surface of the IC. In the horizontal view (Fig. 15), the computer reconstructed field shows that the orientation of the longest axis was actually oblique and about 600 pm long. A second cell

Neurons with moderately oriented dendritic fields Some neurons (n = 4) had dendritic fields that were characterized by a moderate degree of orientation. These cells had dendritic fields where the width was roughly one-third to one-half of the longest axis (deviation from roundness 2 25 and I35). As in highly oriented cells, their dendritic fields showed more than one axis of orientation. Dendrites of two moderately oriented neurons i n central nwleus had dendrites parallel to the frontal plane. For example, the dendritic field of neuron 79-7 (Figs. 10, 15) had a long dorsoventral axis that extended over 700 pm.

ANATOMY OF HRP-INJECTED CELLS IN THE INFERIOR COLLICULUS

81

E, .

I

.

79-31 4 Fig. 4. Neuron 79-314 is highly oriented and located in the superficial dorsal cortex (see inset, transverse plane). The axon (a, upper panel) and the dendrites and cell body (lower panel) are illustrated in the sagittal plane. The dendritic field has its long axis in the rostrocaudal direction. Upper scale = 400 km; lower scale = 100 bm.

(79-250, Fig. 14) from the dorsal cortex was similar to 84-25. Displayed in the sagittal plane, the dendritic field of this cell also extended rostrocaudally about 450 pm. Although not quite horizontal, the axis of the dendritic field roughly paralleled the border of the central nucleus and the rostrodorsal surface of the IC. Both cells had frequent dendritic appendages. The cell bodies were medium (84-25, 26 pm) or small in average diameter (79-250, 18 pm). Axons of these moderately oriented cells lacked a specific orientation and terminated within a large area of the central nucleus. Most axons displayed extensive local axon collaterals. For example, the axon of cell 79-7 (Fig. 11)had numerous intrinsic axonal collaterals that terminated over a broad area at least 1.4 mm rostral to the cell body, nearly

1 mm above and below it, and about 2.5 mm in the mediolateral direction. Right-angled branches occurred frequently, and each branch was distinguished by clusters of 3-17 terminal boutons. Over 2,000 terminal boutons were found on this axon. Since clusters of boutons can be found in zones about the size of a neuronal cell body, they could make many contacts on a single cell. Besides the local terminals, the main stem of this axon (Fig. 11,arrowhead) coursed lateral and ventral to enter the brachium of IC. Thus, this neuron made many local connections and also appeared to project out of the IC. Axons of other moderately oriented cells shared the feature of widespread axonal terminations although the details of the branching patterns differ. The axon of the

D.L. OLIVER ET AL.

82

Fig. 5 . Neuron 78-341is from the superficial dorsal cortex (see inset, transverse plane). The dendritic field (lower panel) is highly oriented in the rostrocaudal direction. Axon (a,upper panel). Sagittal plane. Upper scale = 500 pm; lower scale = 100 pm.

mediolaterally oriented cell 85-14 (Fig. 12B) extended across the fibrodendritic laminae, medial to the cell body. Ventrolateral to the cell body, the axon followed the axis of a laminae. Small numbers of right-angle side branches and terminal boutons were observed. Another example (cell 84-25, Fig. 13) is illustrated in the transverse plane. It's axon was oriented in the mediolateral and rostrocaudal dimension and had long, infrequent collaterals that branched at right angles to the parent axon. Cell 79-250 also displayed an axon with widespread branching (Fig. 14). Illustrated in the sagittal plane, many of the long, infrequently branched segments ran parallel to the dendritic field. Other branches extended rostrally at almost right angles to the main axis. As the axon traversed over 2 mm in the mediolateral dimension, it displayed many boutons en passage.

UnorienMcells Some neurons (n = 3) had dendritic fields that lack a specific orientation. Typically, the width of the dendritic field is greater than half of the longest axis (deviation from

roundness

Dendritic and axonal morphology of HRP-injected neurons in the inferior colliculus of the cat.

The dendritic and axonal morphology of neurons in the inferior colliculus of the cat was investigated after intracellular injection of HRP, in vivo. A...
1MB Sizes 0 Downloads 0 Views