High Vocal Center Growth and Its Relation to Neurogenesis, Neuronal Replacement and Song Acquisition in Juvenile Canaries Arturo Alvarez-Buylla, Chang-Ying Ling, and Fernando Nottebohm Rockefeller University, New York, New York 10021
SUMMARY It is generally thought that most circuits of the adult central nervous system (CNS) are sculpted, in part at least, by selective elimination of some of the neurons present in an initial overabundant set. In this scenario, the birth of neurons precedes the period when brain functions, such as learning, first occur. In contrast to this form of brain assembly, we describe here the delayed development of the high vocal center (HVC) and one of its efferent pathways in canaries. The retrograde tracer Fluoro-Gold (FG) was injected into one of HVC’s two efferent targets, the nucleus robustus archistriatalis (RA),to define the boundaries of HVC. The HVC grows markedly between 1 and 4 months, invading neighboring territories of the caudal telencephalon. During this same period, 0.43%-0.64% of the HVC neurons present at I year of age are labeled per day of [ 3H]-thymidine injection. 13H]-Thymidine labeling is a marker of cell birth, and during the first 4 months HVC neuron number increases, probably accounting for part of the HVC growth
observed. Thereafter, the number of HVC neurons remains constant, but neuronal birth persists. We infer from this that neuronal replacement starts as early as 4 months after hatching and perhaps before then. About half of the neurons born after posthatching day 10 grow an axon to RA to form the main efferent pathway exiting from HVC. HVC growth, neurogenesis, axogenesis, and the observed replacement of neurons happen during the period of juvenile vocal learning. However, the recruitment of neurons that are still present at I year shows no particular inflections corresponding to the various stages in song learning, and continues at essentially the same rate after the more stereotyped adult song has been acquired. We suggest that a combination of neurogenesis and neuronal replacement provides unique advantages for learning. Keywords: neurogenesis, neuronal replacement, songbirds, plasticity.
INTRODUCTION
sculpted. There may be several reasons for this phenomenon. An excess of neurons may have to be available to ensure that when some cells fail to make the right connections, there will still be others to constitute the circuits needed. Neurons may also be discarded to adjust the relation between parts of the central nervous system (CNS) and their target organs when the relative size of these components is independently programmed (Purves, 1988). The situation that we describe here, in a nucleus of the developing song system of birds, is different. In this case, net growth of neuronal numbers continues during a long period after hatching, while the behavior controlled by these cells is acquired.
The development of many vertebrate brain systems includes a stage during which there is an excess of neurons, dendrites, or synapses, which are subsequently pared down (reviewed in Jacobson, 1991; Cowan, Fawcett, O’Leary, and Stanfield, 1984; Oppenheim, 1981;Finley and Pallas, 1989). A surplus of raw materials seems to be provided, out of which the final configuration of circuits is Received January 28, 1992, accepted March 16, 1992 Journal of Neurobiology, Vol. 23, No. 4, pp. 396-406 (1992) 0 1992 John Wiley & Sons, Inc. CCC 0022-3034/92/040396-11$04.00
396
Late Development of Canary Song Nucleus
We will argue in the Discussion section that this form of ontogeny may be well suited for learned behaviors that are acquired late in development. The high vocal center (HVC) of songbirds plays an important role in the acquisition and production of learned song (Nottebohm, Stokes, and Leonard, 1976; McCasland, 1987; Simpson and Vicario, 1990). The HVC of songbirds grows markedly after hatching, as shown by Nissl staining (Nottebohm, Nottebohm, and Crane, 1986). This growth in volume is accompanied by an increase in cell number (Bottjer, Miesner, and Arnold, 1986; Bottjer, Glaessner, and Arnold, 1985; Nordeen, Marler, and Nordeen, 1989; Burek, Nordeen, and Nordeen, 1991; Kirn and DeVoogd, 1989; Nordeen and Nordeen, 1988; Alvarez-Buylla, Theelen, and Nottebohm, 1988). While these studies indicate that many neurons are added to HVC during juvenile and adult life, there is no systematic study of neuronal incorporation throughout juvenile development or of how this process relates to the growth of HVC. This information is important to the hypothesis on the role of postnatal neurogenesis in song learning (Nottebohm, 1985; AlvarezBuylla et al., 1988; Nordeen and Nordeen, 1988). Specifically,is neurogenesis of HVC cells restricted or predominantly represented during only a particular stage of song acquisition? Or, to put it differently, how many of the HVC neurons present at 1 year of age, when canaries are in stable adult song, were recruited at various early stages of song development? Previous studies determined the size of HVC using cresyl violet, a Nissl staining technique that can disguise the true size of HVC (Gahr, 1990). In this paper we use Fluoro-Gold (FG) backfills to determine the size of HVC, and [ 3H]-thymidine to document DNA synthesis and, by inference, cell birth. By combining these methods, we present a detailed morphometric description of the relation between neuronal birth dates, neuronal numbers, and the size of HVC during juvenile development. The HVC includes two types of projection neurons: those that project to area X of lobus parolfactorius and those that project to nucleus robustus archistriatalis (RA). HVC neurons that project to area X are born before hatching or soon thereafter; those that project to RA (HVC-RA neurons) are born mainly after hatching and continue to be produced in adulthood ( Alvarez-Buylla et d., 1988, 1990a). HVC-RA neurons form the main efferent output from HVC and constitute an important part of the motor circuit for song production (Alvarez-Buylla et al., 1990a; Kim, Alvarez-Buylla,
397
and Nottebohm, 1991; Nottebohm et al., 1976). However, the data on the time of birth of HVCRA neurons during juvenile development are, at present, fragmentary (Nordeen and Nordeen, 1988; Alvarez-Buylla et al., 1988). For this reason, our present analysis includes a more comprehensive investigation of the time of birth of HVC-RA neurons. We report here that the HVC of male canaries reaches adult volume at 4 months of age, well before sexual maturity. The birth of new HVC neurons, and, in particular, of new HVC-RA neurons, is fairly constant during the first 130 days, with no major peaks that correlate with particular phases of song learning. This rate of addition is comparable to that seen at 240 days (sexual maturity), when song is much more stable (Alvarez-Buylla et al., 1988) . Neuronal replacement is already present at 4 months, well before adult song has acquired its stable form. The combination of late neurogenesis and neuronal replacement during juvenile development and their persistence in adulthood may offer unique opportunities for learning. MATERIALS AND METHODS Male Waterschlager canaries kept under the natural photoperiod of New York state were bred during the spring and summer in individual cages and the hatching dates of young birds were noted. Birds hatched from mid-April to mid-August ( 1987-1990) and were separated from their parents into flight cages at 21 days of age. Birds treated with [ 3H]-thymidine were housed in individual cages after this age. Younger birds treated with [3H]thymidine were housed with their parents in their breeding cage.
Treatment with [ 3H]-Thymidine Nineteen canaries were used in this study. Birds received daily injections of [ 3H]-thymidine (6.7 Ci/mM, New England Nuclear) on days 1-5 ( n = 3), 6-10 ( n = 3), I 1-20 ( n = I ), 2 1-30 ( n = 3), 3 1-40 ( n = 2 ) , 41-50 ( n = 2 ) , 91-100 ( n = 2), and 121-130 ( n = 3 ) ofage. All birds were sacrificed during the following breeding season, when they were 10-13 months old (referred to as I-year-old). The dose of [ 3H]-thymidinefor young birds was based on their weight at first injection (2.5 pCi/g body weight). Canaries received the following amounts of [ 3H]-thymidine per injection: 10 pCi at 1-5 days, 25 pCi for birds 6-10 days old, and 50 pCi for birds oIder than 10 days. All birds received bilateral micro injections of retrograde tracer Fluoro-Gold into nucleus RA 4 days before sacrifice. This procedure has been described in detail before (Alvarez-Buylla et al., 1988; Kirn et al.,
398
Alvarez-Buyllu et al.
1991 ). Birds were killed by an overdose of anesthesia (Nembutal) and perfused intracardially with 0.9% saline followed by 3% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. Brains were post-fixed in the same solution for 1-3 days at 4°C. Sagittal sections (6-pm thick) were cut and mounted onto gelatinized glass slides after embedding in polyethylene glycol ( Alvarez-Buylla, Buskirk, and Nottebohm, 1987; Clayton and AlvarezBuylla, 1989). Sections were covered with NTB2 (Kodak) nuclear track emulsion and incubated at 4°C for 30 days. The autoradiography was developed in D19 (Kodak) for 3 min at 17°C. Sections were counterstained with fluorescent cresyl violet as described before (Alvarez-Buylla, Ling, and Kirn, 1990b), and coverslipped with Krystalon (Harleco).
Development of HVC Twenty-four canaries were used for this study: four (ages 34-38 days old) for the 1-month group: five (ages 56-66 days old) for the 2-month group: seven (ages 1 14-1 39 days) for the 4-month group; three (ages 308-3 10 days old) for the 10-month group, and four (ages 373-398 days old) for the 13-month group. The HVC was backfilled bilaterally with FG injected (40 nl) into nucleus RA. These birds were killed 4 days later, and their brains were sectioned as described above. The amount of FG injected into RA was larger in the younger birds ( 1 and 2 months old, 80 nl) to ensure that enough of this substance reached their relatively smaller RA or was picked up by axons from HVC that had not yet grown into RA (Konishi and Akutagawa, 1985). Sections were stained with fluorescent cresyl violet ( Alvarez-Buylla et al., 1990b) and analyzed.
the number of neurons counted, we calculated the number of neurons of each category per millimeter square. The nuclear diameter was measured for 50 cells of each type on each brain. From nuclear sizes and cells/mm2, we calculated cell densities ( cells/mm3) following the correction for cell splitting ( Weibel, 1979: Clark, Cynx, Alvarez-Buylla, O’Loughlin, and Nottebohm, 1990). The corrected densities were used to calculate total neuronal numbers. For purposes of documenting the birthdates of HVC neurons present in the 1-year-old birds, we used 10 evenly spaced sections for each HVC. We counted the number of [ 3H]-thymidine-labeled HVC neurons and [ 3H]-thymidine-labeled HVC neurons backfilled with FG (HVC-RA neuron) in each of these sections. Cells were identified under the 63X objective (lox oculars) and considered [ 3H]-thymidine labeled if their nuclei were overlaid by more than seven exposed silver grains (>20X background). In order to calculate the proportion of HVC neurons labeled with [)HI -thymidine, we also counted unlabeled HVC neurons (distinguishing between those that projected to RA and those that did not) in three sections per HVC. The position of all [3H]labeled HVC neurons was mapped to determine whether neurons born during any particular period were concentrated in specific subregions of HVC. Values obtained for left and right hemispheres were averaged for each brain. In cases where the FG injection missed RA in one side, only the side of the brain with a good backfill was used. No systematic right-left differences were observed. Data were analyzed with a one-way analysis of variance (ANOVA). Comparisons between pairs of data sets were analyzed with a Student’s t test; p values >0.05 were rejected as not significant.
Microscopic Analysis A computer-yoked microscope described before ( Alvarez-Buylla and Vicario, 1988) was used for all quantification described here. For volume calculations, the perimeter of HVC in one of every 16 sections was traced into the computer, and its area was calculated. The borders of the HVC were determined by the boundaries of fluorescence (neuropil and somata) produced by FG injections into RA (Fig. 3). The sum of areas was multiplied by 96 pm (interval between sections) to determine the volume. To estimate the total number of neurons in HVC at the different developmental ages studied, all neurons within the area defined by the FG backfill were counted in three evenly distributed sections for each HVC. Our counts distinguished between neurons with FG-backfilled somata (HVC-RA neurons) and those without FG (other neurons). HVC-RA were labeled by FG in their cytoplasm, and neurons not backfilled with FG were identified based on Nissl staining as described before (Goldman and Nottebohm, 1983). Only neurons with a clear neuronal morphology were included in our quantification. From the area of these HVC sections and
RESULTS Birth of Neurons in HVC
Figure 1 shows the rates at which HVC neurons present at 1 year of age were labeled with [3H]thymidine during the period between 1 and 130 days after hatching. About 1% of all neurons present in the 1-year-old HVC were labeled per day during the first 5 days after hatching. The actual number born may have been higher since the Sphase preceding mitosis is likely to be shorter than 24 hours (Korr, 1980); that is, during this 5-day period at least 5% of the adult HVC neurons were generated. This daily rate dropped to about half after day 5 and remained between 0.4% and 0.6% for the rest of the 130 days posthatching period (Fig. 1 ). The fraction of these new neurons that project to nucleus RA increased with age. Thus be-
Lute Development 0fCanur.v Song Nucleus
HVC Growth after Hatching
1.2 8
e
2 1 A
’
399
OTotal Neurons HVC-RA Neurons
0.8 0.6
Days
.-
Age at fH]-Thymidine Treatment
Figure 3 shows the change in volume of the HVC during posthatching development. The size of HVC increased dramatically between 1 and 4 months of age ( p < 0.005). At 4 months of age, HVC reached its 1-year-old size. The growth of HVC is illustrated in the photomicrographs of Fig. 4. HVC encompasses more of the caudal neostriatum at 4 months than it does at 1 month. Since the number of sections containing HVC increased and the area of HVC per section was larger, the growth occurred in both saggital and frontal planes. Bulging of the HVC into the overlying ventricle cannot account for the growth observed between 1 and 4
Figure 1 Proportion of neurons in the 10- to 13month-old HVC of male canaries labeled with injections of [ 3H]-thymidine at different ages after hatching. The total height of each histogram bar indicates the average percentage of HVC neurons labeled during that treatment period per day of [ 3H]-thymidine injection; the shaded part of each bar represents the fraction of the total neurons labeled during that period that projected to RA. Filled circles and open squares indicate the values for individual birds from which the means were derived.
tween 1 and 5 days only 19.6% of the new HVC neurons were HVC-RA neurons. By 6-10 days, this value was 33.8% and reached 40.1% between days 1 1 and 20. Between 2 1 and 30 days and thereafter more than half (52%-57%) of the new HVC neurons born became RA-projecting neurons. These rates and trends did not seem to be affected by whether the counting of HVC neurons and characterization of cells types were done at 10 or 13 months or at the ages in between. Since [ 3H]-thymidine is a reliable marker of cell division, our data indicate that many of the neurons in the 1-year-old HVC were born after hatching. The HVC-RA neurons were an important fraction of these new neurons. We found no particular peak or trough in the rate of labeling of HVCRA neurons during the developmental period studied (Fig. 1 ), with this caveat: our counts of labeled neurons at I year of age do not tell us about neurons that may have been born at that time and died before sampling time. New neurons born at each of the developmental ages sampled were found throughout the adult HVC (Fig. 2). We infer that the new neurons were interspersed amongst others born earlier and that this was accompanied by a restructuring of the spatial relations between cells within HVC.
Figure 2 Position of labeled neurons in the HVC of four canaries that received [3HJ-thymidine at the different ages indicated as days after hatching. Maps from three sagittal sections (48 pm between sections) were superimposed and the position of labeled neurons marked by circles; HVC-RA neurons are shown filled. Rostra1 is to the left and dorsal is up. The position of the lateral ventricle is indicated by a heavier line dorsal to each HVC.
400
Alvarez-Biiylla et al. 0.3
HVC Volume
Age (months)
Figure 3 HVC volume in mm3 in male canaries at different ages after hatching. Volumes were determined by backfills of FG injected into RA, and correspond to one hemisphere only. Error bars are standard deviations of the mean (ANOVA F = 12.17; p < 0.0001 ).
months. Since adult brain volume is achieved as early as 1 month after hatching (Nottebohm et al., 1986), we infer that HVC grows by encroaching on other adjacent tissue. Many new neurons are added to HVC during the first 4 months after hatching. Figure 5 shows that this addition results in a net increase in HVC neuronal numbers ( p < 0.05). After 4 months, the number of HVC neurons decreased slightly but this decrease was not significant. Since the density of neurons in HVC did not change significantly between 1 and 4 months (Fig. 5 ) , the new neurons plus the supporting vasculanzation and glia presumably account for the observed increase in HVC volume. The number of RA-projecting neurons was 1607 ( k 3 9 0 S.D.) at 1 month, 766 1 ( & 1250 S.D.) at 4 months, and 11,287 (21160 S.D.) at 13 months ( p < 0.02) (Fig. 2). Unlike the case for all neurons (Fig. 5 ) , the packing density of RA-projecting neurons increased between 1 and 13 months of age ( p < 0.04) (Fig. 4, and Fig. 6 ) . In addition, the HVC neuropil of birds that received FG injections into RA at a younger age was less fluorescent than that of older birds. We do not know if this is because of the lower packing densities of HVC-RA neurons in the younger birds, or because the processes of RA-projecting cells are less we11 in younger than in Older birds, or for both these reasons (Fig. 4). DISCUSSION
The present results show that the HVC of canaries triples its volume between 1 and 4 months of age.
Figure 4 Photomicrographs of sagittal sections through the HVC of birds 1, 2, 4, and 13 months old. HVC-RA neurons and neuropil in the HVC are fluorescently labeled by retrograde transport of FG injected into the ipsilateral RA. Notice how the number of backfilled axons forming discrete bundles increases with age. Photographs were taken of the sections showing the largest HVC area for each brain. Dorsal is up and caudal is right. Scale bar = 200 l m .
401
Late Development of Canary Song Nucleus 15
--0
T
Neuronal Density
2-
x
Figure 5 Density (top panel) and total number of neurons (bottom panel) in the HVC of birds at different ages after hatching. Error bars indicate standard deviations of the mean (ANOVA for top panel, F = 0.92; p < 0.5; bottom panel, F = 10.70; p < 0.0001 ).
This increase in volume is paralleled by an increase in neuronal number. Many new HVC neurons are born during this period including new RA-projecting neurons.
Posthatching Birth of HVC Neurons [ 3H]-Thymidine data show that HVC neurons present at 1 year of age are produced at a relatively constant rate during the period between 5 and 130 days after hatching and that during this period new neurons are intercalated amongst the older ones. This pattern of growth must lead to a constant rearrangement of the structure of this nucleus. Our count of labeled cells requires a comment. The labeling index observed (i.e., the percentage of labeled neurons per day of [ 3H]-thymidine treatment) cannot be equated to the number of neurons born per day. There are several reasons for this: 1 ) We do not know if [ 3H]-thymidine was metabolized at the same rate at all ages; 2 ) The S-phase preceding division in warm-blooded vertebrates usually lasts from 8 to 12 h (Korr, 1980). By giving one injection of [ 3H]-thymidine per day, the numhers Of that were labeled per is probably somewhere between one-third to one-half of the number that could have been labeled if [3H]thymidine had been available throughout the 24-h period. 3 ) Ventricular-zone stem cells divide at least once ( Alvarez-Buylla, Theelen, and Nottebohm, 1990c), and perhaps several times, before one of the daughter cells becomes a young migratory neuron. A labeled precursor cell may divide and give birth to a neuron several days after the last
10
HVC Volume HVC is small but clearly distinguishable at 1 month of age (Fig. 4). It achieves its 1-year old size by 4 months. These observations, using backfills from RA to define the boundaries of HVC, differ from those of an earlier study that used Nissl staining (Nottebohm et al., 1986). That study reported that HVC grew until the age of 7.5 months. Thus the present results confirm the posthatching growth of this nucleus, but show that HVC achieves full adult volume earlier than previously thought. Gahr ( 1990) has recently shown that HVC volume measured by Nissl staining can differ from that revealed by retrogradely transported substances. We agree with Gahr that when there is a disagreement between the two methods, the backfill method provides the more reliable definition of boundary. Nissl staining may reflect the metabolic state of cells (Peters, Palay, and Webster de F., 199 1 ) . Perhaps the metabolic status of neurons in the periphery of a nucleus can make this nucleus look larger or smaller (Gahr, 1990).
1
Density of HVC-RA Neurons
1
Total HVC-RA Neurons
I
Age (months) Figure 6 Density (top panel) and total number of RAprojecting neurons (bottom panel) in the HVC of birds at different ages after hatching. Error bars indicate standard deviations of the mean (ANOVA for top panel, F = 4 . 8 3 ;< ~ 0.009; bottom panel, F = 9 . 8 8 ; < ~ 0.0001).
402
Alvurez-Buyllu et al.
[ 3H]-thymidine treatment. For these reasons we have avoided indicating exact birth dates, and for the same reasons we have avoided using the phrase “number of neurons born per day.” Despite these caveats, our counts tell us about the relative numbers of neurons produced at different times and that survive to the 1-year mark. Had HVC-RA neurons present at 1 year of age been born only during a restricted period of the 130 days sampled, then our results should have shown this; [ 3H]-thymidine administration before or after this restricted period would have resulted in no labeling of these neurons. Thus, the 1-year old HVC is made up of neurons born throughout juvenile development.
Neuronal Replacement in Juveniles
At 4 months of age, HVC has reached the volume and neuronal density seen at 1 year. Since the neurons labeled between 120 and 130 days would not have appeared yet in the neuronal counts at 4 months, we suggest that those HVC neurons and others born after this time could not be part of a process of net neuronal addition, but are replacements for other HVC neurons born earlier. Thus the process of neuronal replacement is well under way by 4 months of age and perhaps even earlier, well before canaries reach sexual maturity and develop stable adult song. Neuronal replacement may start before HVC stops growing but our data does not address this point. Since neurogenesis continues at a comparable rate at 240 days (Alvarez-Buylla et al., 1988), we suggest that neuronal replacement is an ongoing process that affects all of the juvenile period. This inference offers a new outlook on the cell dynamics of HVC at the time song is first acquired. Pycnotic cells have been described in the HVC of juvenile zebra finches (Kim and DeVoogd, 1989) and, therefore, the replacement of HVC neurons might start before the occurrence of stable adult song in this species also. The number of [ 3H]-thymidine-labeled neurons that we observed at 1 year of age reflect both production and replacement. We do not know if neurons labeled at the various times during which we injected [3H]-thymidine were more or less likely to be present when sampling occurred at 1 year of age. Neurons born earlier might be of a kind that is not replaced, or they might have been replaced to a greater extent because the survival period was longer for these neurons than for neurons labeled later.
Birth of HVC-RA Projection Neurons
Many of the new neurons added to the juvenile HVC grow an axon that reaches the archistriatum and innervates RA. There is a sevenfold increase in the number of HVC-RA neurons between 1 and 13 months of age, and this is the most numerous type of neuron in the adult HVC. The total number of HVC-RA neurons continues to increase between lOand 13 monthsofage(p
SUMMARY It is generally thought that most circuits of the adult central nervous system (CNS) are sculpted, in part at least, by selective elimination of some of the neurons present in an initial overabundant set. In this scenario, the birth of neurons precedes the period when brain functions, such as learning, first occur. In contrast to this form of brain assembly, we describe here the delayed development of the high vocal center (HVC) and one of its efferent pathways in canaries. The retrograde tracer Fluoro-Gold (FG) was injected into one of HVC’s two efferent targets, the nucleus robustus archistriatalis (RA),to define the boundaries of HVC. The HVC grows markedly between 1 and 4 months, invading neighboring territories of the caudal telencephalon. During this same period, 0.43%-0.64% of the HVC neurons present at I year of age are labeled per day of [ 3H]-thymidine injection. 13H]-Thymidine labeling is a marker of cell birth, and during the first 4 months HVC neuron number increases, probably accounting for part of the HVC growth
observed. Thereafter, the number of HVC neurons remains constant, but neuronal birth persists. We infer from this that neuronal replacement starts as early as 4 months after hatching and perhaps before then. About half of the neurons born after posthatching day 10 grow an axon to RA to form the main efferent pathway exiting from HVC. HVC growth, neurogenesis, axogenesis, and the observed replacement of neurons happen during the period of juvenile vocal learning. However, the recruitment of neurons that are still present at I year shows no particular inflections corresponding to the various stages in song learning, and continues at essentially the same rate after the more stereotyped adult song has been acquired. We suggest that a combination of neurogenesis and neuronal replacement provides unique advantages for learning. Keywords: neurogenesis, neuronal replacement, songbirds, plasticity.
INTRODUCTION
sculpted. There may be several reasons for this phenomenon. An excess of neurons may have to be available to ensure that when some cells fail to make the right connections, there will still be others to constitute the circuits needed. Neurons may also be discarded to adjust the relation between parts of the central nervous system (CNS) and their target organs when the relative size of these components is independently programmed (Purves, 1988). The situation that we describe here, in a nucleus of the developing song system of birds, is different. In this case, net growth of neuronal numbers continues during a long period after hatching, while the behavior controlled by these cells is acquired.
The development of many vertebrate brain systems includes a stage during which there is an excess of neurons, dendrites, or synapses, which are subsequently pared down (reviewed in Jacobson, 1991; Cowan, Fawcett, O’Leary, and Stanfield, 1984; Oppenheim, 1981;Finley and Pallas, 1989). A surplus of raw materials seems to be provided, out of which the final configuration of circuits is Received January 28, 1992, accepted March 16, 1992 Journal of Neurobiology, Vol. 23, No. 4, pp. 396-406 (1992) 0 1992 John Wiley & Sons, Inc. CCC 0022-3034/92/040396-11$04.00
396
Late Development of Canary Song Nucleus
We will argue in the Discussion section that this form of ontogeny may be well suited for learned behaviors that are acquired late in development. The high vocal center (HVC) of songbirds plays an important role in the acquisition and production of learned song (Nottebohm, Stokes, and Leonard, 1976; McCasland, 1987; Simpson and Vicario, 1990). The HVC of songbirds grows markedly after hatching, as shown by Nissl staining (Nottebohm, Nottebohm, and Crane, 1986). This growth in volume is accompanied by an increase in cell number (Bottjer, Miesner, and Arnold, 1986; Bottjer, Glaessner, and Arnold, 1985; Nordeen, Marler, and Nordeen, 1989; Burek, Nordeen, and Nordeen, 1991; Kirn and DeVoogd, 1989; Nordeen and Nordeen, 1988; Alvarez-Buylla, Theelen, and Nottebohm, 1988). While these studies indicate that many neurons are added to HVC during juvenile and adult life, there is no systematic study of neuronal incorporation throughout juvenile development or of how this process relates to the growth of HVC. This information is important to the hypothesis on the role of postnatal neurogenesis in song learning (Nottebohm, 1985; AlvarezBuylla et al., 1988; Nordeen and Nordeen, 1988). Specifically,is neurogenesis of HVC cells restricted or predominantly represented during only a particular stage of song acquisition? Or, to put it differently, how many of the HVC neurons present at 1 year of age, when canaries are in stable adult song, were recruited at various early stages of song development? Previous studies determined the size of HVC using cresyl violet, a Nissl staining technique that can disguise the true size of HVC (Gahr, 1990). In this paper we use Fluoro-Gold (FG) backfills to determine the size of HVC, and [ 3H]-thymidine to document DNA synthesis and, by inference, cell birth. By combining these methods, we present a detailed morphometric description of the relation between neuronal birth dates, neuronal numbers, and the size of HVC during juvenile development. The HVC includes two types of projection neurons: those that project to area X of lobus parolfactorius and those that project to nucleus robustus archistriatalis (RA). HVC neurons that project to area X are born before hatching or soon thereafter; those that project to RA (HVC-RA neurons) are born mainly after hatching and continue to be produced in adulthood ( Alvarez-Buylla et d., 1988, 1990a). HVC-RA neurons form the main efferent output from HVC and constitute an important part of the motor circuit for song production (Alvarez-Buylla et al., 1990a; Kim, Alvarez-Buylla,
397
and Nottebohm, 1991; Nottebohm et al., 1976). However, the data on the time of birth of HVCRA neurons during juvenile development are, at present, fragmentary (Nordeen and Nordeen, 1988; Alvarez-Buylla et al., 1988). For this reason, our present analysis includes a more comprehensive investigation of the time of birth of HVC-RA neurons. We report here that the HVC of male canaries reaches adult volume at 4 months of age, well before sexual maturity. The birth of new HVC neurons, and, in particular, of new HVC-RA neurons, is fairly constant during the first 130 days, with no major peaks that correlate with particular phases of song learning. This rate of addition is comparable to that seen at 240 days (sexual maturity), when song is much more stable (Alvarez-Buylla et al., 1988) . Neuronal replacement is already present at 4 months, well before adult song has acquired its stable form. The combination of late neurogenesis and neuronal replacement during juvenile development and their persistence in adulthood may offer unique opportunities for learning. MATERIALS AND METHODS Male Waterschlager canaries kept under the natural photoperiod of New York state were bred during the spring and summer in individual cages and the hatching dates of young birds were noted. Birds hatched from mid-April to mid-August ( 1987-1990) and were separated from their parents into flight cages at 21 days of age. Birds treated with [ 3H]-thymidine were housed in individual cages after this age. Younger birds treated with [3H]thymidine were housed with their parents in their breeding cage.
Treatment with [ 3H]-Thymidine Nineteen canaries were used in this study. Birds received daily injections of [ 3H]-thymidine (6.7 Ci/mM, New England Nuclear) on days 1-5 ( n = 3), 6-10 ( n = 3), I 1-20 ( n = I ), 2 1-30 ( n = 3), 3 1-40 ( n = 2 ) , 41-50 ( n = 2 ) , 91-100 ( n = 2), and 121-130 ( n = 3 ) ofage. All birds were sacrificed during the following breeding season, when they were 10-13 months old (referred to as I-year-old). The dose of [ 3H]-thymidinefor young birds was based on their weight at first injection (2.5 pCi/g body weight). Canaries received the following amounts of [ 3H]-thymidine per injection: 10 pCi at 1-5 days, 25 pCi for birds 6-10 days old, and 50 pCi for birds oIder than 10 days. All birds received bilateral micro injections of retrograde tracer Fluoro-Gold into nucleus RA 4 days before sacrifice. This procedure has been described in detail before (Alvarez-Buylla et al., 1988; Kirn et al.,
398
Alvarez-Buyllu et al.
1991 ). Birds were killed by an overdose of anesthesia (Nembutal) and perfused intracardially with 0.9% saline followed by 3% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. Brains were post-fixed in the same solution for 1-3 days at 4°C. Sagittal sections (6-pm thick) were cut and mounted onto gelatinized glass slides after embedding in polyethylene glycol ( Alvarez-Buylla, Buskirk, and Nottebohm, 1987; Clayton and AlvarezBuylla, 1989). Sections were covered with NTB2 (Kodak) nuclear track emulsion and incubated at 4°C for 30 days. The autoradiography was developed in D19 (Kodak) for 3 min at 17°C. Sections were counterstained with fluorescent cresyl violet as described before (Alvarez-Buylla, Ling, and Kirn, 1990b), and coverslipped with Krystalon (Harleco).
Development of HVC Twenty-four canaries were used for this study: four (ages 34-38 days old) for the 1-month group: five (ages 56-66 days old) for the 2-month group: seven (ages 1 14-1 39 days) for the 4-month group; three (ages 308-3 10 days old) for the 10-month group, and four (ages 373-398 days old) for the 13-month group. The HVC was backfilled bilaterally with FG injected (40 nl) into nucleus RA. These birds were killed 4 days later, and their brains were sectioned as described above. The amount of FG injected into RA was larger in the younger birds ( 1 and 2 months old, 80 nl) to ensure that enough of this substance reached their relatively smaller RA or was picked up by axons from HVC that had not yet grown into RA (Konishi and Akutagawa, 1985). Sections were stained with fluorescent cresyl violet ( Alvarez-Buylla et al., 1990b) and analyzed.
the number of neurons counted, we calculated the number of neurons of each category per millimeter square. The nuclear diameter was measured for 50 cells of each type on each brain. From nuclear sizes and cells/mm2, we calculated cell densities ( cells/mm3) following the correction for cell splitting ( Weibel, 1979: Clark, Cynx, Alvarez-Buylla, O’Loughlin, and Nottebohm, 1990). The corrected densities were used to calculate total neuronal numbers. For purposes of documenting the birthdates of HVC neurons present in the 1-year-old birds, we used 10 evenly spaced sections for each HVC. We counted the number of [ 3H]-thymidine-labeled HVC neurons and [ 3H]-thymidine-labeled HVC neurons backfilled with FG (HVC-RA neuron) in each of these sections. Cells were identified under the 63X objective (lox oculars) and considered [ 3H]-thymidine labeled if their nuclei were overlaid by more than seven exposed silver grains (>20X background). In order to calculate the proportion of HVC neurons labeled with [)HI -thymidine, we also counted unlabeled HVC neurons (distinguishing between those that projected to RA and those that did not) in three sections per HVC. The position of all [3H]labeled HVC neurons was mapped to determine whether neurons born during any particular period were concentrated in specific subregions of HVC. Values obtained for left and right hemispheres were averaged for each brain. In cases where the FG injection missed RA in one side, only the side of the brain with a good backfill was used. No systematic right-left differences were observed. Data were analyzed with a one-way analysis of variance (ANOVA). Comparisons between pairs of data sets were analyzed with a Student’s t test; p values >0.05 were rejected as not significant.
Microscopic Analysis A computer-yoked microscope described before ( Alvarez-Buylla and Vicario, 1988) was used for all quantification described here. For volume calculations, the perimeter of HVC in one of every 16 sections was traced into the computer, and its area was calculated. The borders of the HVC were determined by the boundaries of fluorescence (neuropil and somata) produced by FG injections into RA (Fig. 3). The sum of areas was multiplied by 96 pm (interval between sections) to determine the volume. To estimate the total number of neurons in HVC at the different developmental ages studied, all neurons within the area defined by the FG backfill were counted in three evenly distributed sections for each HVC. Our counts distinguished between neurons with FG-backfilled somata (HVC-RA neurons) and those without FG (other neurons). HVC-RA were labeled by FG in their cytoplasm, and neurons not backfilled with FG were identified based on Nissl staining as described before (Goldman and Nottebohm, 1983). Only neurons with a clear neuronal morphology were included in our quantification. From the area of these HVC sections and
RESULTS Birth of Neurons in HVC
Figure 1 shows the rates at which HVC neurons present at 1 year of age were labeled with [3H]thymidine during the period between 1 and 130 days after hatching. About 1% of all neurons present in the 1-year-old HVC were labeled per day during the first 5 days after hatching. The actual number born may have been higher since the Sphase preceding mitosis is likely to be shorter than 24 hours (Korr, 1980); that is, during this 5-day period at least 5% of the adult HVC neurons were generated. This daily rate dropped to about half after day 5 and remained between 0.4% and 0.6% for the rest of the 130 days posthatching period (Fig. 1 ). The fraction of these new neurons that project to nucleus RA increased with age. Thus be-
Lute Development 0fCanur.v Song Nucleus
HVC Growth after Hatching
1.2 8
e
2 1 A
’
399
OTotal Neurons HVC-RA Neurons
0.8 0.6
Days
.-
Age at fH]-Thymidine Treatment
Figure 3 shows the change in volume of the HVC during posthatching development. The size of HVC increased dramatically between 1 and 4 months of age ( p < 0.005). At 4 months of age, HVC reached its 1-year-old size. The growth of HVC is illustrated in the photomicrographs of Fig. 4. HVC encompasses more of the caudal neostriatum at 4 months than it does at 1 month. Since the number of sections containing HVC increased and the area of HVC per section was larger, the growth occurred in both saggital and frontal planes. Bulging of the HVC into the overlying ventricle cannot account for the growth observed between 1 and 4
Figure 1 Proportion of neurons in the 10- to 13month-old HVC of male canaries labeled with injections of [ 3H]-thymidine at different ages after hatching. The total height of each histogram bar indicates the average percentage of HVC neurons labeled during that treatment period per day of [ 3H]-thymidine injection; the shaded part of each bar represents the fraction of the total neurons labeled during that period that projected to RA. Filled circles and open squares indicate the values for individual birds from which the means were derived.
tween 1 and 5 days only 19.6% of the new HVC neurons were HVC-RA neurons. By 6-10 days, this value was 33.8% and reached 40.1% between days 1 1 and 20. Between 2 1 and 30 days and thereafter more than half (52%-57%) of the new HVC neurons born became RA-projecting neurons. These rates and trends did not seem to be affected by whether the counting of HVC neurons and characterization of cells types were done at 10 or 13 months or at the ages in between. Since [ 3H]-thymidine is a reliable marker of cell division, our data indicate that many of the neurons in the 1-year-old HVC were born after hatching. The HVC-RA neurons were an important fraction of these new neurons. We found no particular peak or trough in the rate of labeling of HVCRA neurons during the developmental period studied (Fig. 1 ), with this caveat: our counts of labeled neurons at I year of age do not tell us about neurons that may have been born at that time and died before sampling time. New neurons born at each of the developmental ages sampled were found throughout the adult HVC (Fig. 2). We infer that the new neurons were interspersed amongst others born earlier and that this was accompanied by a restructuring of the spatial relations between cells within HVC.
Figure 2 Position of labeled neurons in the HVC of four canaries that received [3HJ-thymidine at the different ages indicated as days after hatching. Maps from three sagittal sections (48 pm between sections) were superimposed and the position of labeled neurons marked by circles; HVC-RA neurons are shown filled. Rostra1 is to the left and dorsal is up. The position of the lateral ventricle is indicated by a heavier line dorsal to each HVC.
400
Alvarez-Biiylla et al. 0.3
HVC Volume
Age (months)
Figure 3 HVC volume in mm3 in male canaries at different ages after hatching. Volumes were determined by backfills of FG injected into RA, and correspond to one hemisphere only. Error bars are standard deviations of the mean (ANOVA F = 12.17; p < 0.0001 ).
months. Since adult brain volume is achieved as early as 1 month after hatching (Nottebohm et al., 1986), we infer that HVC grows by encroaching on other adjacent tissue. Many new neurons are added to HVC during the first 4 months after hatching. Figure 5 shows that this addition results in a net increase in HVC neuronal numbers ( p < 0.05). After 4 months, the number of HVC neurons decreased slightly but this decrease was not significant. Since the density of neurons in HVC did not change significantly between 1 and 4 months (Fig. 5 ) , the new neurons plus the supporting vasculanzation and glia presumably account for the observed increase in HVC volume. The number of RA-projecting neurons was 1607 ( k 3 9 0 S.D.) at 1 month, 766 1 ( & 1250 S.D.) at 4 months, and 11,287 (21160 S.D.) at 13 months ( p < 0.02) (Fig. 2). Unlike the case for all neurons (Fig. 5 ) , the packing density of RA-projecting neurons increased between 1 and 13 months of age ( p < 0.04) (Fig. 4, and Fig. 6 ) . In addition, the HVC neuropil of birds that received FG injections into RA at a younger age was less fluorescent than that of older birds. We do not know if this is because of the lower packing densities of HVC-RA neurons in the younger birds, or because the processes of RA-projecting cells are less we11 in younger than in Older birds, or for both these reasons (Fig. 4). DISCUSSION
The present results show that the HVC of canaries triples its volume between 1 and 4 months of age.
Figure 4 Photomicrographs of sagittal sections through the HVC of birds 1, 2, 4, and 13 months old. HVC-RA neurons and neuropil in the HVC are fluorescently labeled by retrograde transport of FG injected into the ipsilateral RA. Notice how the number of backfilled axons forming discrete bundles increases with age. Photographs were taken of the sections showing the largest HVC area for each brain. Dorsal is up and caudal is right. Scale bar = 200 l m .
401
Late Development of Canary Song Nucleus 15
--0
T
Neuronal Density
2-
x
Figure 5 Density (top panel) and total number of neurons (bottom panel) in the HVC of birds at different ages after hatching. Error bars indicate standard deviations of the mean (ANOVA for top panel, F = 0.92; p < 0.5; bottom panel, F = 10.70; p < 0.0001 ).
This increase in volume is paralleled by an increase in neuronal number. Many new HVC neurons are born during this period including new RA-projecting neurons.
Posthatching Birth of HVC Neurons [ 3H]-Thymidine data show that HVC neurons present at 1 year of age are produced at a relatively constant rate during the period between 5 and 130 days after hatching and that during this period new neurons are intercalated amongst the older ones. This pattern of growth must lead to a constant rearrangement of the structure of this nucleus. Our count of labeled cells requires a comment. The labeling index observed (i.e., the percentage of labeled neurons per day of [ 3H]-thymidine treatment) cannot be equated to the number of neurons born per day. There are several reasons for this: 1 ) We do not know if [ 3H]-thymidine was metabolized at the same rate at all ages; 2 ) The S-phase preceding division in warm-blooded vertebrates usually lasts from 8 to 12 h (Korr, 1980). By giving one injection of [ 3H]-thymidine per day, the numhers Of that were labeled per is probably somewhere between one-third to one-half of the number that could have been labeled if [3H]thymidine had been available throughout the 24-h period. 3 ) Ventricular-zone stem cells divide at least once ( Alvarez-Buylla, Theelen, and Nottebohm, 1990c), and perhaps several times, before one of the daughter cells becomes a young migratory neuron. A labeled precursor cell may divide and give birth to a neuron several days after the last
10
HVC Volume HVC is small but clearly distinguishable at 1 month of age (Fig. 4). It achieves its 1-year old size by 4 months. These observations, using backfills from RA to define the boundaries of HVC, differ from those of an earlier study that used Nissl staining (Nottebohm et al., 1986). That study reported that HVC grew until the age of 7.5 months. Thus the present results confirm the posthatching growth of this nucleus, but show that HVC achieves full adult volume earlier than previously thought. Gahr ( 1990) has recently shown that HVC volume measured by Nissl staining can differ from that revealed by retrogradely transported substances. We agree with Gahr that when there is a disagreement between the two methods, the backfill method provides the more reliable definition of boundary. Nissl staining may reflect the metabolic state of cells (Peters, Palay, and Webster de F., 199 1 ) . Perhaps the metabolic status of neurons in the periphery of a nucleus can make this nucleus look larger or smaller (Gahr, 1990).
1
Density of HVC-RA Neurons
1
Total HVC-RA Neurons
I
Age (months) Figure 6 Density (top panel) and total number of RAprojecting neurons (bottom panel) in the HVC of birds at different ages after hatching. Error bars indicate standard deviations of the mean (ANOVA for top panel, F = 4 . 8 3 ;< ~ 0.009; bottom panel, F = 9 . 8 8 ; < ~ 0.0001).
402
Alvurez-Buyllu et al.
[ 3H]-thymidine treatment. For these reasons we have avoided indicating exact birth dates, and for the same reasons we have avoided using the phrase “number of neurons born per day.” Despite these caveats, our counts tell us about the relative numbers of neurons produced at different times and that survive to the 1-year mark. Had HVC-RA neurons present at 1 year of age been born only during a restricted period of the 130 days sampled, then our results should have shown this; [ 3H]-thymidine administration before or after this restricted period would have resulted in no labeling of these neurons. Thus, the 1-year old HVC is made up of neurons born throughout juvenile development.
Neuronal Replacement in Juveniles
At 4 months of age, HVC has reached the volume and neuronal density seen at 1 year. Since the neurons labeled between 120 and 130 days would not have appeared yet in the neuronal counts at 4 months, we suggest that those HVC neurons and others born after this time could not be part of a process of net neuronal addition, but are replacements for other HVC neurons born earlier. Thus the process of neuronal replacement is well under way by 4 months of age and perhaps even earlier, well before canaries reach sexual maturity and develop stable adult song. Neuronal replacement may start before HVC stops growing but our data does not address this point. Since neurogenesis continues at a comparable rate at 240 days (Alvarez-Buylla et al., 1988), we suggest that neuronal replacement is an ongoing process that affects all of the juvenile period. This inference offers a new outlook on the cell dynamics of HVC at the time song is first acquired. Pycnotic cells have been described in the HVC of juvenile zebra finches (Kim and DeVoogd, 1989) and, therefore, the replacement of HVC neurons might start before the occurrence of stable adult song in this species also. The number of [ 3H]-thymidine-labeled neurons that we observed at 1 year of age reflect both production and replacement. We do not know if neurons labeled at the various times during which we injected [3H]-thymidine were more or less likely to be present when sampling occurred at 1 year of age. Neurons born earlier might be of a kind that is not replaced, or they might have been replaced to a greater extent because the survival period was longer for these neurons than for neurons labeled later.
Birth of HVC-RA Projection Neurons
Many of the new neurons added to the juvenile HVC grow an axon that reaches the archistriatum and innervates RA. There is a sevenfold increase in the number of HVC-RA neurons between 1 and 13 months of age, and this is the most numerous type of neuron in the adult HVC. The total number of HVC-RA neurons continues to increase between lOand 13 monthsofage(p
