Hearing Research, 60 (1992) 34-44 © 1992 Elsevier Science Publishers B.V. All rights reserved 0378-5955/92/$05.00
34
HEARES 01738
Morphology of HRP-labelled cochlear nerve axons in the dorsal cochlear nucleus of the developing hamster Laura Schweitzer and Tina Cecil Department of Anatomical Sciences and Neurobioiog),, Unicersitv of Louiscille School oJ"Medicbu" Louiscille Kentucky, USA (Received 12 October 1991; accepted 20 January 1992)
To study the development of the central terminal arbors of the cochlear nerve fibers in the dorsal cochlear nucleus, horseradish peroxidase-labelled axons in young and adult hamsters were analyzed morphometrically. Brainstem slices with whole cochlear nuclei were maintained in a slice chamber and the cochlear nerve root was injected with a mixture of wheat germ agglutinin-horseradish peroxidase, horseradish peroxidase and poly-L-ornithine. The poly-L-ornithine was added to keep the injection site small: small injections resulted in only a few axons being labelled and permitted reconstruction of individual fibers. Axons underwent an initial period of ingrowth that was completed prior to the onset of hearing (postnatal day 16). After this time the morphology and area of influence of the axons remained unchanged but the nucleus continued to increase in size. Since no additional cochlear nerve axons grow into the nucleus during this period of nuclear growth, the existing axons necessarily become more widely spaced as development proceeds. These anatomical changes may contribute to the progressive narrowing of auditory cell tuning curves.
Auditory: Axons: Development: Cochlear nerve: Cochlear nucleus: Tonotopic; Tuning
Introduction
As sensory systems develop the ability of the neonate to discriminate between sensory stimuli increases. In the auditory system, for example, hatchling chicks arc less able to distinguish tones than their 3 to 4-day-old counterparts (Kerr et al., 1979); from 3 to 7 years of age children show an increasing ability to discriminate between pure tones (Jensen et ai., 1987). This has been referred to as 'perceptual sharpening' (Kerr et al., 1979). In auditory cells the physiological counterpart of perceptual sharpening is the maturation of frequency selectivity. Auditory tuning curves develop gradually during postnatal maturation of the mammalian cochlea (Romand, 1983; Waish and McGee, 1990), cochlear nucleus (Woolf and Ryan, 1985) and inferior colliculus (Aitkin and Moore, 1975; Willott and Shnerson, 1978; Moore and Irvine, 1979). This process may be thought of as consisting of two components. The first is a fall in threshold at all frequencies to which a cell responds. The second is a selective fall in threshold for frequencies at or near that which is to become a cell's 'best
Correspondence to: Laura Schweitzer, Anatomical Sciences and Neurobiology, University of Louisville School of Medicine, Louisville, KY 40292, USA.
frequency', narrowing the range of frequencies to which a cell is maximally responsive and thus sharpening the tuning curve. In the cochlear nucleus this sharpening occurs gradually after the onset of hearing (especially for frequencies below 4 kHz) and improves steadily until mat~.,rity is reached at postnatal day 60 (Woolf and Ryan, 1985). In many neural systems, anatomical development of axonal inputs is characterized by initial axonal ingrowth followed by expansion of axonal domains and finally the elimination of parts of axonal arbors or whole collaterals 'axonal regression' (see Cowan et al., 1984; Purves and Lichtman, 1985; Easter et al., 1985). One of the earliest demonstrations of axonal regression related to the segregation of inputs from the lateral geniculate nucleus to the visual cortex during the establishment of ocular dominance columns (Levay et al., 1978). Concurrent with axonal changes of the sort referred to, layer IV cells in visual cortex that initially respond equally to both eyes come to be dominated by input from only one eye. In other neural systems it has been demonstrated that cells postsynaptic to regressing axons become responsive to progressively fewer inputs. This has been shown in systems as diverse as the neuromuscular junction (Dennis, 1981), the autonomic ganglia (Johnson and Purves, 1981; Lichtman and Purves, 1980), and, in the central nervous system, the cerebellum
35 (Mason and Gregory, 1984) and sensory nuclei (Jackson and Parks, 1982; Sachs et al., 1986). Innervation by fewer axons may be the anatomical basis for the development of stimulus selectivity or 'tuning' in these systems. In the auditory system, Jackson and Parks (1982) demonstrated in the chick that as branches from cochlear nerve axons are eliminated, target cells of the axons in the nucleus magnocellularis (cochlear nucleus analog) become responsive to fewer afferents. In the adult, these cells respond to input from only one or two axons, whereas in the hatchling, they respond to multiple inputs. Thus in the avian auditory system it is clear that the narrowing of a cell's effective inputs correlates with the elimination of axonal branches. In this study, using horseradish peroxidase (HRP) labelling, we sought to determine whether regression of cochlear nerve axons plays a significant part in the maturation of the mammalian cochlear nucleus. Since such labelling is difficult in very young live animals, we used a brain slice preparation. Reconstructions of individual axons revealed that after initial ingrowth has occurred in the dorsal cochlear nucleus of the hamster, the area that an axon occupies and its branching pattern are remarkably stable. This stability must be interpreted, however, in light of concurrent changes in the cochlear nucleus which are also described in the following paper.
Methods
NaHCO 3 23 and glucose 10 (Schurr et al., 1987). Time from decapitation to suspension in aCSF was approximately 120 s. The brain was divided in the midsagittal plane and an approximately 2 mm thick (mediolaterally) parasagittal brainstem slice was then cut from each half brain with a surgical blade. One was placed in a holding chamber and the other was placed in the slice chamber on its medial surface exposing the lateral brainstem with the cochlear nerve stump and the cochlear nucleus on the top of the slice. The aCSF in both the slice and holding chambers was maintained at 31°C and bubbled with a 95% 02-5% CO 2 mixture. Pulled glass pipettes were broken so that the tips were approximately 20 # m in diameter. These were filled with 10% horseradish peroxidase (HRP, Sigma: P-6782), 1% wheat germ agglutinated-HRP (WGAHRP, Sigma: L-7017), and 0.2% poly-L-ornithine (Sigma: P-4638), in 0.3 M KCI/0.1 M Tris buffer (pH 7.2-7.4). The tip of the electrode was lowered into the nerve root and the HRP mixture was iontophoretically delivered with 0.75/~A continuous current for 3.5 min. The electrode tip was retracted and this slice was then exchanged with the one in the holding chamber. After each half was injected, the slices were maintained for 4 h in the holding chamber. The slices were then fixed by immersion in 2.5% paraformaldehyde/1.25% glutaraldehyde in phosphate buffer (pH 7.4) for 1 h and then placed in 30% sucrose in phosphate buffer overnight. The brains were frozen-sectioned the next day into 100/.tm transverse sections, treated with nickel and cobalt chloride (Adams, 1981) and reacted with 3,3'-diaminobenzidine (DAB).
Animals Adult Syrian hamsters were obtained from Harlan Sprague-Dawley (Indianapolis, IN) and a breeding colony was maintained. Infants were housed with their mothers until postnatal day (PND) 21, at which time they were weaned and housed with no more than three littermates of the same sex. Hamsters from PND 2 to 87 were used; animals from PND 60 through 87 were combined into one group called 'adults', and designated simply as PND 60 + . In vitr, ~ experiments Hamsters at PND 2 (3), 4 (1), 5 (1), 6 (1), 7 (2), 8 (5), 9 (1), 10 (4), 11 (3), 12 (3), 14 (2), 16 (2), 19 (3), 21 (1), 23 (3), 30 (1), 31 (2), 35 (3), 37 (2), 60 + (3) IN shown in parentheses] were successfully used in the in vitro experiments. Each hamster was deeply anesthetized with ether and decapitated. The brain was removed and placed immediately in cold artificial cerebrospinal fluid (aCSF) at 5°C. The millimolar concentrations of the components of the aCSF were as follows: NaCl 124, KC! 5, NaH2PO 4 3, CaC! 2 2.5, MgSO4 2.4,
Morphometric analysis of labelled axons One axon from each brain slice was chosen for reconstruction with the Eutechtics Neuron Tracing System (Raleigh, NC). To ensure consistent comparisons across ages, axons chosen for analysis were always from the ventrolateral third of the nucleus. This region of the nucleus contains axons from the cochlear apex (Collinge and Schweitzer, 1991) which were labelled when the electrode tip was placed in the lateral aspect of the nerve stump. These axons transmit low frequency auditory information (Kaltenbach and Lazor, 1991; Yajima and Hayashi, 1989). The tuning curves of cells with low characteristic frequencies show the greatest change during development in the gerbil (Woolf and Ryan, 1985). Axons were traced with an 100 x oil-immersion objective. They were reconstructed in three dimensions and the following measures were taken from their point of entrance into the deep layer of the dorsal cochlear nucleus: 1) the number of branches, 2) the curvilinear length of the axon, 3) the width of the axonal field from ventrolateral to dorsomedial (perpendicular to the isofrequency planes; this dimension
36 is also parallel to the dorsal acoustic stria and thus has been called 'strial' by Blackstad et al., 1984) and 4) the width of the axonal field from rostral to caudal (parallel to the isofrequency plane, also called 'transtrial' by Blackstad et al., 1984). Axonal field width was measured with respect to the long axis of the axon. Axons are tilted congruent with the isofrequency planes, and thus their long axis corresponds to the isofrequency plane in which the axon is located. The isofrequency planes in the hamster have been defined in physiologic experiments by Kaltenbach and Lazor (1991) and in anatomical experiments by Collinge and Schweitzer (1991).
Nissl studies For a developmental reference series, the brains of four PND 1, 5, 10, 15, 25, 40, and 60 + hamsters each were stained with a Nissl stain. The animals were deeply anesthetized with sodium pentobarbital and perfused intracardially with 2.5% paraformaldehyde and 1.25% glutaraldehyde in phosphate buffer (pH 7.4). The brains remained in the fixative for 4 h and were then placed in 30% sucrose in phosphate buffer overnight. Frozen sections 50 /.~m thick were cut through the cochlear nucleus and were stained with cresyl violet.
Morphometric analysis of 'reference' Nissl-stained sections Cochlear nucleus dimensions. Each section through the dorsal cochlear nucleus was drawn with a microscope (10 × objective) equipped with a drawing tube (16 × ). The drawings were measured using morphometric analysis software (Sigma Scan, Jandel Inc., Corte Madera, CA). Three-dimensional reconstructions of the dorsal cochlear nucleus were done as well (PC3D, Jandel Inc.). Particular attention was focused on the dorsal cochlear nucleus volume and its span in the dimension perpendicular to the isofrequency planes (ventrolateral to dorsomedial dimension) and parallel to the isofrequency planes (rostrocaudal dimension). 'Span' was defined as the curvilinear distance along the inferior contour of the dorsal cochlear nucleus.
Morphometric analysis of brain slice sections Dorsal cochlear nucleus dimensions. To determine if flattening, deformation or expansion of the cochlear nucleus had occurred during preparation or processing of the brain slice material, the dorsal cochlear nuclei in brain slices were measured and compared to those in the perfusion-fixed reference Nissl-stained sections. Brain slices from PND 15 (N - 4) and PND 25 (N = 4) hamsters were treated as the HRP-injected brain slices had been handled except after immersion fixation they were cut and stained with cresyl violet. These 'brain
slice Nissl-stained sections' were analyzed as the reference Nissl-stained sections had been (described above).
Statistical analysis Morphometric parameters (e.g., axon length, branch points, axonal field width, dorsal cochlear nucleus volume, dorsal cochlear nucleus span) were analyzed with one-way analyses of variance with age as the independent variable. For statistical analysis of the brain slice data, functional age groups were defined as 'ingrowth stage' (PND 2-9, N = 14), 'prehearing' (PND 10-16, N = 14), 'hearing' (PND 17-23, N = 7), 'adolescent' (PND 31-37, N = 8) and, adult, (PND 60 + , N - 3 ) . The actual age groups were used for the statistical analysis of the Nissl data (PND 1, 5, 10, 15, 25, 40, 60 + , all N's = 4). To compare the dimensions of the dorsal cochlear nucleus in the reference Nissl-stained and brain slice Nissl-stained material a two-way analysis of variance was employed (age x preparation) with subsequent one-way analyses to test for main effects. When significant F ratios were obtained, Fisher's post hoc tests (Siegel, 1956) were used to assess differences between individual age groups (alpha = 0.05).
Results
Quality of the results In typical sections, several axons were labelled. For this reason axons were chosen for study that were entirely confined to the 100/zm thick section. This was not a problem since axonai field width averaged no greater than 33/~m (maximum standard error = 12.42 /~m) at any age, and axons with their parent stem centered in the section rarely traversed the cut edges of the sectic~a. However unlikely, it is possible that this restriction caused a sampling error with axons with very large fields eliminated consistently from the analysis. Although several axons were labelled in each section, individual axons were easily traced under oil immersion at 100 x . The brain surrounding the axons and the axons themselves appeared normal microscopically. That is, the morphology of the axons in the dorsal cochlear nucleus was similar to that in Golgi-impregnated material (Schweitzer and Cant, 1984). The HRP-filled axons were not swollen or broken. Axonal diameters ranged from 0.5 to 1.1/.~m and were similar to those measured for Golgi-impregnated cochlear nerve axons. Necrosis, seen as a darkened area deep in the brainstem of older animals ( > PND 16) did not encroach on the cochlear nucleus in the cases used. At the oldest ages (PND 60 + ) necrosis was more diffuse and only three cases could be used.
37
Development of the cochlear nerve axons Qualitative results In the newborn hamster, HRP-labelled cochlear nerve axons can be traced into the ventral cochlear nucleus where they divide into an ascending and a
DAY 3
descending branch. Individual nerve fibers in the descending branch can be traced to the border between the posteroventral cochlear nucleus and dorsal cochlear nucleus. On PND 3 cochlear fibers course deep to, but do not make the upward turn into, the dorsal cochlear nucleus proper (Fig. 1). By PND 6 fibers enter the
7
J
&5
12 ~
85
r
/'
J 23
Fig. 1. Drawings of HRP-filled cochlear axons in the dorsal cochlear nucleus. The ages in postnatal days are noted. 'Hair-like' appendages that are much thinner than the axon's main arbor, are represented by the finer ~ines. On the right a smaller version of each axon is shown with the outline of the nucleus to demonstrate that while the axon's area of influence does not change much after PND 10, the nucleus continues to expand. Dashed line, deep border of the molecular layer; dorsal (d) is up, lateral (l) is to the left.
38
deep region of the ventrolateral dorsal cochlear nucleus and sometimes branch near their terminal tips. Expansions resembling growth cones can be seen on some axons at this age (Fig. 2). Over the next four days fibers g~'ow to the superficial limit that they will reach in the older hamsters, the deep border of the molecular layer. Even at this age, when many closely aligned axons are labelled, they are often organized into sheets as described for the adult (Collinge and Schweitzer, 1991). During the period between postnatal days 6 and 10 the axons branch and are also adorned with hair-like appendages (Fig. 1). The appendages are much less common on axons from hamsters after PND 15, whereas swellings at the distal tips of the axons become more common (Fig. 2). After PND 15 new, immature axons, just entering the deep aspect of the nucleus are not seen. The axons change little in appearance after the second postnatal week. What does change is the dimensions of the nucleus, which is expanding, while the dimensions of the axons remain about the same (right column, Fig. 1).
Quantitatice results The dorsal cochlear nuclei in brain slices were smaller than those in the Nissi-stained material (Fig. 3, F = 24.88; d.f. - 1,12; P < 0.01). Reduced volume was due to shrinkage predominantly in the dorsomedial to ventrolateral dimension (perpendicular to the isofrequency planes, F = 19.38; d.f.= 1,12; P P > 0.10). The analysis o f branch points failed to reveal differences between age groups ( F = 1.15; d.f. = 4,45; P > 0.3) although a post hoc c o m p a r i s o n revealed an increase in branch points p e r axon b e t w e e n the ingrowth ( P N D 0 - 9 ) and prehearing ( P N D 10-16) periods (p < 0.05). Axonal fields did not widen significantly in the plane parallel to the isofrequency p l a n e s of the nucleus ( F = .93; d.f. = 4,45; P > 0.4). H o w e v e r the axonal fields did widen significantly in the p l a n e p e r p e n d i c u l a r to t h e isofrequency planes (F = 3.05; d.f. - 4,45; P < 0.03). Differences were found b e t w e e n axons just growing into the nucleus ( P N D 0 - 9 , m e a n = 16.5 # m ) a n d those in ' p r e h e a r i n g ' animals ( P N D 10-16, m e a n = 3 2 . 3 6 / z m ) ( P < 0.05). Additional expansion in t h e ventrolateral to dorsomedial ( p e r p e n d i c u l a r ) dimension after the onset of hearing was n o t f o u n d (all P ' s > 0.05).
Changes in the dimensions of the dorsal cochlear nucleus based on the reference Nissl-stained sections T h e v o l u m e of the dorsal cochlear nucleus increases nearly eight times between birth a n d a d u l t h o o d (Fig. 3). T h e r e was no significant increase between P N D 1 and 5 ( P > 0.05) but a significant increase was ob-
served at each subsequent age studied up to P N D 25 (all P ' s < 0.05). A f t e r P N D 25 no further expansion
was seen. To account for the increase in volume between PND 5 and 25, the span of the dorsal cochlear nucleus parallel to the isofrequency planes and perpendicular to the isofrequency planes was measured. Both of these dimensions showed significant increases. The span of the dorsal cochlear nucleus parallel to the isofrequency planes nearly doubled between birth and adulthood ( F = 7.35; d.f.=6,27; P
34
HEARES 01738
Morphology of HRP-labelled cochlear nerve axons in the dorsal cochlear nucleus of the developing hamster Laura Schweitzer and Tina Cecil Department of Anatomical Sciences and Neurobioiog),, Unicersitv of Louiscille School oJ"Medicbu" Louiscille Kentucky, USA (Received 12 October 1991; accepted 20 January 1992)
To study the development of the central terminal arbors of the cochlear nerve fibers in the dorsal cochlear nucleus, horseradish peroxidase-labelled axons in young and adult hamsters were analyzed morphometrically. Brainstem slices with whole cochlear nuclei were maintained in a slice chamber and the cochlear nerve root was injected with a mixture of wheat germ agglutinin-horseradish peroxidase, horseradish peroxidase and poly-L-ornithine. The poly-L-ornithine was added to keep the injection site small: small injections resulted in only a few axons being labelled and permitted reconstruction of individual fibers. Axons underwent an initial period of ingrowth that was completed prior to the onset of hearing (postnatal day 16). After this time the morphology and area of influence of the axons remained unchanged but the nucleus continued to increase in size. Since no additional cochlear nerve axons grow into the nucleus during this period of nuclear growth, the existing axons necessarily become more widely spaced as development proceeds. These anatomical changes may contribute to the progressive narrowing of auditory cell tuning curves.
Auditory: Axons: Development: Cochlear nerve: Cochlear nucleus: Tonotopic; Tuning
Introduction
As sensory systems develop the ability of the neonate to discriminate between sensory stimuli increases. In the auditory system, for example, hatchling chicks arc less able to distinguish tones than their 3 to 4-day-old counterparts (Kerr et al., 1979); from 3 to 7 years of age children show an increasing ability to discriminate between pure tones (Jensen et ai., 1987). This has been referred to as 'perceptual sharpening' (Kerr et al., 1979). In auditory cells the physiological counterpart of perceptual sharpening is the maturation of frequency selectivity. Auditory tuning curves develop gradually during postnatal maturation of the mammalian cochlea (Romand, 1983; Waish and McGee, 1990), cochlear nucleus (Woolf and Ryan, 1985) and inferior colliculus (Aitkin and Moore, 1975; Willott and Shnerson, 1978; Moore and Irvine, 1979). This process may be thought of as consisting of two components. The first is a fall in threshold at all frequencies to which a cell responds. The second is a selective fall in threshold for frequencies at or near that which is to become a cell's 'best
Correspondence to: Laura Schweitzer, Anatomical Sciences and Neurobiology, University of Louisville School of Medicine, Louisville, KY 40292, USA.
frequency', narrowing the range of frequencies to which a cell is maximally responsive and thus sharpening the tuning curve. In the cochlear nucleus this sharpening occurs gradually after the onset of hearing (especially for frequencies below 4 kHz) and improves steadily until mat~.,rity is reached at postnatal day 60 (Woolf and Ryan, 1985). In many neural systems, anatomical development of axonal inputs is characterized by initial axonal ingrowth followed by expansion of axonal domains and finally the elimination of parts of axonal arbors or whole collaterals 'axonal regression' (see Cowan et al., 1984; Purves and Lichtman, 1985; Easter et al., 1985). One of the earliest demonstrations of axonal regression related to the segregation of inputs from the lateral geniculate nucleus to the visual cortex during the establishment of ocular dominance columns (Levay et al., 1978). Concurrent with axonal changes of the sort referred to, layer IV cells in visual cortex that initially respond equally to both eyes come to be dominated by input from only one eye. In other neural systems it has been demonstrated that cells postsynaptic to regressing axons become responsive to progressively fewer inputs. This has been shown in systems as diverse as the neuromuscular junction (Dennis, 1981), the autonomic ganglia (Johnson and Purves, 1981; Lichtman and Purves, 1980), and, in the central nervous system, the cerebellum
35 (Mason and Gregory, 1984) and sensory nuclei (Jackson and Parks, 1982; Sachs et al., 1986). Innervation by fewer axons may be the anatomical basis for the development of stimulus selectivity or 'tuning' in these systems. In the auditory system, Jackson and Parks (1982) demonstrated in the chick that as branches from cochlear nerve axons are eliminated, target cells of the axons in the nucleus magnocellularis (cochlear nucleus analog) become responsive to fewer afferents. In the adult, these cells respond to input from only one or two axons, whereas in the hatchling, they respond to multiple inputs. Thus in the avian auditory system it is clear that the narrowing of a cell's effective inputs correlates with the elimination of axonal branches. In this study, using horseradish peroxidase (HRP) labelling, we sought to determine whether regression of cochlear nerve axons plays a significant part in the maturation of the mammalian cochlear nucleus. Since such labelling is difficult in very young live animals, we used a brain slice preparation. Reconstructions of individual axons revealed that after initial ingrowth has occurred in the dorsal cochlear nucleus of the hamster, the area that an axon occupies and its branching pattern are remarkably stable. This stability must be interpreted, however, in light of concurrent changes in the cochlear nucleus which are also described in the following paper.
Methods
NaHCO 3 23 and glucose 10 (Schurr et al., 1987). Time from decapitation to suspension in aCSF was approximately 120 s. The brain was divided in the midsagittal plane and an approximately 2 mm thick (mediolaterally) parasagittal brainstem slice was then cut from each half brain with a surgical blade. One was placed in a holding chamber and the other was placed in the slice chamber on its medial surface exposing the lateral brainstem with the cochlear nerve stump and the cochlear nucleus on the top of the slice. The aCSF in both the slice and holding chambers was maintained at 31°C and bubbled with a 95% 02-5% CO 2 mixture. Pulled glass pipettes were broken so that the tips were approximately 20 # m in diameter. These were filled with 10% horseradish peroxidase (HRP, Sigma: P-6782), 1% wheat germ agglutinated-HRP (WGAHRP, Sigma: L-7017), and 0.2% poly-L-ornithine (Sigma: P-4638), in 0.3 M KCI/0.1 M Tris buffer (pH 7.2-7.4). The tip of the electrode was lowered into the nerve root and the HRP mixture was iontophoretically delivered with 0.75/~A continuous current for 3.5 min. The electrode tip was retracted and this slice was then exchanged with the one in the holding chamber. After each half was injected, the slices were maintained for 4 h in the holding chamber. The slices were then fixed by immersion in 2.5% paraformaldehyde/1.25% glutaraldehyde in phosphate buffer (pH 7.4) for 1 h and then placed in 30% sucrose in phosphate buffer overnight. The brains were frozen-sectioned the next day into 100/.tm transverse sections, treated with nickel and cobalt chloride (Adams, 1981) and reacted with 3,3'-diaminobenzidine (DAB).
Animals Adult Syrian hamsters were obtained from Harlan Sprague-Dawley (Indianapolis, IN) and a breeding colony was maintained. Infants were housed with their mothers until postnatal day (PND) 21, at which time they were weaned and housed with no more than three littermates of the same sex. Hamsters from PND 2 to 87 were used; animals from PND 60 through 87 were combined into one group called 'adults', and designated simply as PND 60 + . In vitr, ~ experiments Hamsters at PND 2 (3), 4 (1), 5 (1), 6 (1), 7 (2), 8 (5), 9 (1), 10 (4), 11 (3), 12 (3), 14 (2), 16 (2), 19 (3), 21 (1), 23 (3), 30 (1), 31 (2), 35 (3), 37 (2), 60 + (3) IN shown in parentheses] were successfully used in the in vitro experiments. Each hamster was deeply anesthetized with ether and decapitated. The brain was removed and placed immediately in cold artificial cerebrospinal fluid (aCSF) at 5°C. The millimolar concentrations of the components of the aCSF were as follows: NaCl 124, KC! 5, NaH2PO 4 3, CaC! 2 2.5, MgSO4 2.4,
Morphometric analysis of labelled axons One axon from each brain slice was chosen for reconstruction with the Eutechtics Neuron Tracing System (Raleigh, NC). To ensure consistent comparisons across ages, axons chosen for analysis were always from the ventrolateral third of the nucleus. This region of the nucleus contains axons from the cochlear apex (Collinge and Schweitzer, 1991) which were labelled when the electrode tip was placed in the lateral aspect of the nerve stump. These axons transmit low frequency auditory information (Kaltenbach and Lazor, 1991; Yajima and Hayashi, 1989). The tuning curves of cells with low characteristic frequencies show the greatest change during development in the gerbil (Woolf and Ryan, 1985). Axons were traced with an 100 x oil-immersion objective. They were reconstructed in three dimensions and the following measures were taken from their point of entrance into the deep layer of the dorsal cochlear nucleus: 1) the number of branches, 2) the curvilinear length of the axon, 3) the width of the axonal field from ventrolateral to dorsomedial (perpendicular to the isofrequency planes; this dimension
36 is also parallel to the dorsal acoustic stria and thus has been called 'strial' by Blackstad et al., 1984) and 4) the width of the axonal field from rostral to caudal (parallel to the isofrequency plane, also called 'transtrial' by Blackstad et al., 1984). Axonal field width was measured with respect to the long axis of the axon. Axons are tilted congruent with the isofrequency planes, and thus their long axis corresponds to the isofrequency plane in which the axon is located. The isofrequency planes in the hamster have been defined in physiologic experiments by Kaltenbach and Lazor (1991) and in anatomical experiments by Collinge and Schweitzer (1991).
Nissl studies For a developmental reference series, the brains of four PND 1, 5, 10, 15, 25, 40, and 60 + hamsters each were stained with a Nissl stain. The animals were deeply anesthetized with sodium pentobarbital and perfused intracardially with 2.5% paraformaldehyde and 1.25% glutaraldehyde in phosphate buffer (pH 7.4). The brains remained in the fixative for 4 h and were then placed in 30% sucrose in phosphate buffer overnight. Frozen sections 50 /.~m thick were cut through the cochlear nucleus and were stained with cresyl violet.
Morphometric analysis of 'reference' Nissl-stained sections Cochlear nucleus dimensions. Each section through the dorsal cochlear nucleus was drawn with a microscope (10 × objective) equipped with a drawing tube (16 × ). The drawings were measured using morphometric analysis software (Sigma Scan, Jandel Inc., Corte Madera, CA). Three-dimensional reconstructions of the dorsal cochlear nucleus were done as well (PC3D, Jandel Inc.). Particular attention was focused on the dorsal cochlear nucleus volume and its span in the dimension perpendicular to the isofrequency planes (ventrolateral to dorsomedial dimension) and parallel to the isofrequency planes (rostrocaudal dimension). 'Span' was defined as the curvilinear distance along the inferior contour of the dorsal cochlear nucleus.
Morphometric analysis of brain slice sections Dorsal cochlear nucleus dimensions. To determine if flattening, deformation or expansion of the cochlear nucleus had occurred during preparation or processing of the brain slice material, the dorsal cochlear nuclei in brain slices were measured and compared to those in the perfusion-fixed reference Nissl-stained sections. Brain slices from PND 15 (N - 4) and PND 25 (N = 4) hamsters were treated as the HRP-injected brain slices had been handled except after immersion fixation they were cut and stained with cresyl violet. These 'brain
slice Nissl-stained sections' were analyzed as the reference Nissl-stained sections had been (described above).
Statistical analysis Morphometric parameters (e.g., axon length, branch points, axonal field width, dorsal cochlear nucleus volume, dorsal cochlear nucleus span) were analyzed with one-way analyses of variance with age as the independent variable. For statistical analysis of the brain slice data, functional age groups were defined as 'ingrowth stage' (PND 2-9, N = 14), 'prehearing' (PND 10-16, N = 14), 'hearing' (PND 17-23, N = 7), 'adolescent' (PND 31-37, N = 8) and, adult, (PND 60 + , N - 3 ) . The actual age groups were used for the statistical analysis of the Nissl data (PND 1, 5, 10, 15, 25, 40, 60 + , all N's = 4). To compare the dimensions of the dorsal cochlear nucleus in the reference Nissl-stained and brain slice Nissl-stained material a two-way analysis of variance was employed (age x preparation) with subsequent one-way analyses to test for main effects. When significant F ratios were obtained, Fisher's post hoc tests (Siegel, 1956) were used to assess differences between individual age groups (alpha = 0.05).
Results
Quality of the results In typical sections, several axons were labelled. For this reason axons were chosen for study that were entirely confined to the 100/zm thick section. This was not a problem since axonai field width averaged no greater than 33/~m (maximum standard error = 12.42 /~m) at any age, and axons with their parent stem centered in the section rarely traversed the cut edges of the sectic~a. However unlikely, it is possible that this restriction caused a sampling error with axons with very large fields eliminated consistently from the analysis. Although several axons were labelled in each section, individual axons were easily traced under oil immersion at 100 x . The brain surrounding the axons and the axons themselves appeared normal microscopically. That is, the morphology of the axons in the dorsal cochlear nucleus was similar to that in Golgi-impregnated material (Schweitzer and Cant, 1984). The HRP-filled axons were not swollen or broken. Axonal diameters ranged from 0.5 to 1.1/.~m and were similar to those measured for Golgi-impregnated cochlear nerve axons. Necrosis, seen as a darkened area deep in the brainstem of older animals ( > PND 16) did not encroach on the cochlear nucleus in the cases used. At the oldest ages (PND 60 + ) necrosis was more diffuse and only three cases could be used.
37
Development of the cochlear nerve axons Qualitative results In the newborn hamster, HRP-labelled cochlear nerve axons can be traced into the ventral cochlear nucleus where they divide into an ascending and a
DAY 3
descending branch. Individual nerve fibers in the descending branch can be traced to the border between the posteroventral cochlear nucleus and dorsal cochlear nucleus. On PND 3 cochlear fibers course deep to, but do not make the upward turn into, the dorsal cochlear nucleus proper (Fig. 1). By PND 6 fibers enter the
7
J
&5
12 ~
85
r
/'
J 23
Fig. 1. Drawings of HRP-filled cochlear axons in the dorsal cochlear nucleus. The ages in postnatal days are noted. 'Hair-like' appendages that are much thinner than the axon's main arbor, are represented by the finer ~ines. On the right a smaller version of each axon is shown with the outline of the nucleus to demonstrate that while the axon's area of influence does not change much after PND 10, the nucleus continues to expand. Dashed line, deep border of the molecular layer; dorsal (d) is up, lateral (l) is to the left.
38
deep region of the ventrolateral dorsal cochlear nucleus and sometimes branch near their terminal tips. Expansions resembling growth cones can be seen on some axons at this age (Fig. 2). Over the next four days fibers g~'ow to the superficial limit that they will reach in the older hamsters, the deep border of the molecular layer. Even at this age, when many closely aligned axons are labelled, they are often organized into sheets as described for the adult (Collinge and Schweitzer, 1991). During the period between postnatal days 6 and 10 the axons branch and are also adorned with hair-like appendages (Fig. 1). The appendages are much less common on axons from hamsters after PND 15, whereas swellings at the distal tips of the axons become more common (Fig. 2). After PND 15 new, immature axons, just entering the deep aspect of the nucleus are not seen. The axons change little in appearance after the second postnatal week. What does change is the dimensions of the nucleus, which is expanding, while the dimensions of the axons remain about the same (right column, Fig. 1).
Quantitatice results The dorsal cochlear nuclei in brain slices were smaller than those in the Nissi-stained material (Fig. 3, F = 24.88; d.f. - 1,12; P < 0.01). Reduced volume was due to shrinkage predominantly in the dorsomedial to ventrolateral dimension (perpendicular to the isofrequency planes, F = 19.38; d.f.= 1,12; P P > 0.10). The analysis o f branch points failed to reveal differences between age groups ( F = 1.15; d.f. = 4,45; P > 0.3) although a post hoc c o m p a r i s o n revealed an increase in branch points p e r axon b e t w e e n the ingrowth ( P N D 0 - 9 ) and prehearing ( P N D 10-16) periods (p < 0.05). Axonal fields did not widen significantly in the plane parallel to the isofrequency p l a n e s of the nucleus ( F = .93; d.f. = 4,45; P > 0.4). H o w e v e r the axonal fields did widen significantly in the p l a n e p e r p e n d i c u l a r to t h e isofrequency planes (F = 3.05; d.f. - 4,45; P < 0.03). Differences were found b e t w e e n axons just growing into the nucleus ( P N D 0 - 9 , m e a n = 16.5 # m ) a n d those in ' p r e h e a r i n g ' animals ( P N D 10-16, m e a n = 3 2 . 3 6 / z m ) ( P < 0.05). Additional expansion in t h e ventrolateral to dorsomedial ( p e r p e n d i c u l a r ) dimension after the onset of hearing was n o t f o u n d (all P ' s > 0.05).
Changes in the dimensions of the dorsal cochlear nucleus based on the reference Nissl-stained sections T h e v o l u m e of the dorsal cochlear nucleus increases nearly eight times between birth a n d a d u l t h o o d (Fig. 3). T h e r e was no significant increase between P N D 1 and 5 ( P > 0.05) but a significant increase was ob-
served at each subsequent age studied up to P N D 25 (all P ' s < 0.05). A f t e r P N D 25 no further expansion
was seen. To account for the increase in volume between PND 5 and 25, the span of the dorsal cochlear nucleus parallel to the isofrequency planes and perpendicular to the isofrequency planes was measured. Both of these dimensions showed significant increases. The span of the dorsal cochlear nucleus parallel to the isofrequency planes nearly doubled between birth and adulthood ( F = 7.35; d.f.=6,27; P
