THE JOURNAL OF COMPARATIVE NEUROLOGY 294399-417 (1990)

Development of Catecholaminergic Projections to the Spinal Cord in the North American Opossum, Didelphis virginianu RONDA R. PINDZOLA, RAYMOND H. HO, AND GEORGE F. MARTIN Department of Anatomy and Neuroscience Program, The Ohio State University College of Medicine, Columbus, Ohio 43210

ABSTRACT The intent of our study was to determine when catecholaminergic axons grow into each of their adult targets in the spinal cord of the North American opossum (Didelphis uirginiana) and to identify the origin of catecholaminergic axons in the lumbosacral cord at different stages of development. Tyrosine hydroxylase-like immunoreactive axons, presumed to be catecholaminergic, were demonstrated at different stages of development by the indirect antibody peroxidaseantiperoxidase technique of Sternberger. The neurons giving rise to such axons in the lumbosacral cord were identified by using the retrograde transport of Fast Blue and immunofluorescence for tyrosine hydroxylase-like immunoreactive neurons, At birth, 12-13 days after conception, tyrosine hydroxylase-like immunoreactive axons are present in the marginal zone throughout the length of the spinal cord. Such axons are particularly numerous in the dorsolateral marginal zone, the region containing most of them in adult animals. By postnatal day 3, a few immunoreactive axons are present in the intermediate (mantle) zone of the spinal cord; and by postnatal day 8, they are most concentrated in the presumptive intermediolateral cell column. Laminae I and I1 of the dorsal horn are not innervated by such axons until approximately postnatal day 15. By postnatal day 44, the distribution of tyrosine hydroxylase-like immunoreactive axons in the spinal cord resembles that in adult animals, although some areas may be hyperinnervated. At birth, tyrosine hgdroxylase-like immunoreactive cell bodies are present in all of the brainstem areas providing catecholaminergic projections to the spinal cord in adult animals (Pindzola et al.: Bruin Behac. Evol. 32:281-292, '88); and by at least postnatal day 5, lumbosacral injections of Fast Blue retrogradely label tyrosine hydroxylase-like immunoreactive neurons in all such areas. Retrogradely labeled immunoreactive neurons were also found in areas that do not contain them in adult animals. Such areas include the dorsal part of the nucleus coeruleus and certain areas of the reticular formation. During development, spinally projecting tyrosine hydroxylase-like immunoreactive neurons are numerous medial to the nucleus ventralis lemnisci lateralis (the paralemniscal region), whereas only a few are present in the same location in adult animals. Our results suggest that catecholaminergic axons grow into the spinal cord prenatally, that they innervate their adult targets postnatally and over an extended time period, and that during some stages of development they originate from areas that do not supply them in the adult animal. Key words: tyrosine hydroxylase, immunofluorescence, double labeled neurons, Fast Blue, P A P immunohistochemistry

Cat,echolamines are present in the spinal cord of rats very early in development (Loizou, '72; Olson and Seiger, '72; Seiger and Olson, '73; Commissiong, '83a.b; Aramant et al., '86),and they may influence motility in developing (Commissiong, '83a,b) as well as adult animals (Fossberg and Grillner, '73; Grillner, '76: Westlund and Coulter. '80). It is possible, therefore, that catecholaminergic (CA) axons influthe Of spontaneous limb but that be to test in most placental mammals because limb movements first occur prenatally. It 0 1990

WILEY-LISS, INC.

should be testable in North American opossums, however, because they are born in a premature state, 12-13 days after conception, and hindlimb motility does not develop until 7-10 days later.

Accepted December 21,1989. Ronda R. Pindzola's current address is Department of Neuroscience, School of Medicine, Case Western Reserve University, 2119 Abington Rd., Cleveland, OH 44106.

R.R. PINDZOLA ET AL.

400

Abbreviations bc Cb cc CeS Coe CI Fac fx GLD GMc GrP Hg

HYP ILC I2 LRCm mth

MZ 01

0s

PH

PHD

brachium conjunctivum cerebellum central canal nucleus centralis superior nucleus coeruleus colliculus inferior nucleus nervi facialis columna fornicis nucleus corporis geniculati lateralis dorsalis nucleus corporis geniculati medialis: pars centralis griseum pontis nucleus nervi hypoglossi area hypothalamica posterior intermediolateral cell column intermediate zone nucleus lateralis reticularis caudalis: pars magnocellularis fasciculus mamillothalamicus marginal zone nucleus olivaris inferior nucleus olivaris superior nucleus paraventricularis hypothalami

In order to assess the influence of CA axons on the development of hindlimb motility, it is important to know their developmental history. We have shown previously, with the Falck-Hillarp technique, that CA axons are present in the spinal cord a t birth in opossums and that they innervate their adult targets postnatally over an extended period of time (Martin et al., '78; Humbertson and Mart.in, '79). It is difficult to evaluate developmental material processed by the Falck-Hillarp technique, however, because the fluorescence fades rapidly. In the present study, we examined the development of CA axons in the spinal cord by using the indirect antibody peroxidase-antiperoxidase(PAP) technique of Sternberger et al. ('70) for tyrosine hydroxylase (TH). Since TH is one of the enzymes used in catecholamine biosynthesis. we have assumed that TH-like immunoreactive (IR) axons are catecholaminergic. Although the lumbosacral cord is of particular interest because of its control over hindlimb motility, we have examined the growth of TH-IR axons into the entire cord. In addition, the origin of TH-IR axons reaching lumbosacral levels of the cord was studied by means of the retrograde transport of Fast Blue (FB) in combination with irnmunofluorescence to label TH-IR neurons. The distribution of TH-IR axons in the spinal cord and the origins of those which project to the cervical cord has been reported previously for adult opossums (Pindzola e t al., '88).

MATERIALS AND METHODS Developing opossums Pouch-young opossums were obtained from adults purchased from licensed collectors in Florida and from animals bred a t The Ohio State University. The age of the pouchyoung from captured opossums was estimated by comparing their snout-rump length (SRL) and external features with those of animals a t known ages from timed litters (Cutts et al., '78, and data from our collection).

Antisera The tyrosine hydroxylase antibody was obtained from Eugene Technical (Allendale, New Jersey). Eugene Technical has characterized the antiserum by immunohistochemical procedures and by other immunological techniques. We

nucleus paraventricularis hypothalami dorsalis paralemniscal area tractus pyramidalis nucleus retrofacialis fasciculus retroflexus nucleus reticularis gigantocellularis nucleus reticularis gigantocellularis: pars ventralis nucleus reticularis l a t e r a h nucleus reticularis pontis nucleus reticularis pontis: pars ventralis nucleus reticularis medullae oblongatae ventralis trigeminal ganglion nucleus motorius nervi trigemini tractus opticus tractus spinalis nervi trigemini nucleus tractus spinalis nervi trigemini: pars caudalis nucleus corporis trapezoidei nucleus ventralis lemnisci lateralis: pars dorsolateralis nucleus ventralis lemnisci lateralis: pars ventromedialis ventrolateral medulla nuclei ventralis thalami basalis

pVLL PY* RFc rfl

RGC RGcv

RL RP RPV RV TriG TrMo tro trs TrSc

Tz VLLd VLLv VLM

xVB

have used the antibody on the rat spinal cord and brainstem and have produced immunostaining similar to that reported by others (Pickel et al., '75; Chan-Palay et al., '84; Hokfelt et al., '84). Controls for specificity of the antibody are described elsewhere (Pindzola et al., '88). Although immunohistochemical localization of T H indicates the probable location of catecholamine synthesis, it does not show the end products of synthesis. Even so, we presume that TH-IR axons and neurons are catecholaminergic. Our studies only reveal substance-like immunoreactivity since antibodies may bind to their respective antigen when it is part of a larger molecule or to antigens with identical or similar molecular sequences. For PAP immunohistochemistry the antibody to TH was diluted 1:5,000. The diluent consisted of borate buffered saline (BBS, pH 8.2) containing 0.5% bovine serum albumin (BSA) and 0.3% Triton X-100. The sheep anti-rabbit IgG antiserum (SAR) was diluted 1:600 in BBS, BSA, and Triton X-100. The rabbit peroxidase-antiperoxidase complex was diluted 1:1,000 in BBS and BSA. For immunofluorescence the antibody to TH was diluted 1:4,000 in phosphate buffered saline (PBS; pH 7.2) with 0.370 Triton X-100. The fluorescein-isothiocyanate (F1TC)-conjugated goat antirabbit IgG (GAR) was diluted 1:700 in the same diluent. We did not inhibit nonspecific activity or endogenous peroxidase in order to maximize immunolabeling.

The developmentof TH-IR axons and neurons The pouch-young animals ranged from postnatal day (PD) 1 to estimated PD105 (Table 1). They were removed from the pouch, anesthetized by metofane (2,2-dichloro-l,ldifluoroethyl methyl ether) inhalation, and sacrificed by TABLE 1. Animals Studied by PAP Immunohistochemjstry' PD

SRL N PD

SRL N PD

SRL N

3 23

24

45 3

46

4748

2

2

6 24 2 26 50 2

39

44

54

60

75

27-29 4 26 58 1 90

69

76

90

99

1

2

2

1

130 2

200 1

1 15 6 21 44 3 38 67

2 17 2

1

22

3 18

4 19 2

8 25 2 27 54 2

%11

1P15 3435 2 32 31 fi0 59 1 1 105

18

19

20

39 2 33

40 2

43

64

65 1

1

35

1 36 67

3

2%

1

Age in postnatal days (PD), snout-rumplength (SRL),and number D f m h a k at each age (N).

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD transcardiac perfusion with saline (0.9% NaCl) followed by Zamboni's fixative (Stephanini et al., '67) with the aid of a Cole-Palmer minipump. A few animals were perfused with 3.5 % paraformaldehyde. Animals younger than PD7 (under 24 mm SRL) were decapitated and fixed by immersion in Zamboni's fixative for 24 hours. The spinal cord and brain were removed promptly and placed in fixative at 4OC (4-6 hours). After fixation the tissue was stored overnight in Sorenson's phosphate buffer (pH 7.2) containing 30% sucrose. Frozen sections (80 pm) from representative spinal levels, including the lumbosacral cord and from the entire brain, were cut in the coronal or sagittal plane and collected in ice cold BBS. The tissue sections were processed for T H according to the PAP technique of Sternberger et al. ('70) as described previously (Pindzola et al., '88).

TH-IR neurons that project to the lumbosacral cord at different stages of development Surgery was performed on the developing opossums while they remained in the pouch. The mother was anesthetized with intramuscular injections of ketamine (1.2 ml, 100 mg/ml) and metofane inhalation, which makes the pouch sphincter relax, exposing the litter. Pups a t representative ages (Table 2) were anesthetized by hypothermia. The lumbosacral spinal cord was exposed by laminectomy and pressure injections of F B (0.2-1.5 pl) in 5 % aqueous suspensions (sonicated for 5-10 minutes) were made by using a micropipette (50 pm tip) attached to a 1 p1 Hamilton syringe. The injections were intentionally large and included the lateral funiculus, where TH-IR axons are most numerous in adult animals (Pindzola et al., '86, '88). After delivery of the dye, the exposed cord was covered with gelfoam (older animals only) and the skin sutured. The animals were allowed to survive for 72-96 hours in the pouch before being anesthetized by metofane and perfused transcardially with saline followed by 4 % cold buffered paraformaldehyde. Pups younger than PD7 (less than 24 mm SRL) were perfused with a micropipette attached to a 26 gauge syringe needle. The brains and spinal cords were removed and immersed in the same fixative containing 30% sucrose for 6-18 hours. After postfixation, frozen sections of the brain and spinal cord were cut a t 30-40 pm into PBS and processed for TH-IR according to the indirect immunofluorescent technique of Sawchenko and Swanson ('81).Details of the technique are described elsewhere (Pindzola et al., '88).

Data analysis The PAP preparations were examined and photographed with brightfield and Nomarski optics attached to a Leitz (Orthoplan) microscope. The sections processed for immunofluorescence were viewed in darkfield to determine nuclear boundaries and epi-illumination to observe fluorescence. The locations of FB labeled neurons, TH-IR neurons, TABLE 2. Animals Subjected to Injections of FB and Indirect Immunofluarescence'

PD 5

6

12

15

20

24

35

41

43-44

24 2

26 2

30 1

35 2

42 2

47 2

63 2

72 2

7F-76 2

53

~

SRL N

57 ~

88 2

95 1

'Ape in postnatal days (PD).snoutrump length (SRL),and number of animals at each nge (N).

40 1

and TH-IR neurons that contained FB were plotted on drawings by using an X-Y plotter interfaced by position transducers with the microscope stage (Leitz Orthoplan). Filter cube A (excitation wavelength = 340-380 nm) of the Ploem illumination system was used to visualize FB, and filter cube I2 (excitation wavelength = 450-490 nm) was used to visualize fluorescein. We presume that the double labeled neurons contribute to the TH-IR axons found in the spinal cord by using PAP immunohistochemistry. The extent of the injection in each case was outlined on drawings by using the plotter. The terminology for nuclei in the opossum's brain was taken from Oswaldo-Cruz and RochaMiranda ('68).

RESULTS The development of TH-IR axons in the spinal cord TH-IR axons were examined in sections of representative spinal levels from P D l to PD105 (Table 1).When possible, Rexed's laminae (Rexed, '54) were used to describe their locations. Such laminae were not distinct, however, so we have often referred to them as presumptive laminae based primarily on relative positions. In PD1 (15 mm SRL) animals, TH-IR axons are found throughout the length of the cord where they are most numerous in the dorsolateral marginal zone (arrows, Figs. 1, 2A,B). Only an occasional TH-IR axon is found in the intermediate (mantle) zone. At PD3 (18 mm SRL), TH-IR axons appear more numerous in the developing lateral funiculus, and more of them have reached sacral levels. Many of the immunostained axons in the lateral funiculus are close to the intermediate zone, and some of them extend into the presumptive intermediolateral cell column (ILC). By PD3 a few TH-IR axons can also be found in the ventral horn and near the central canal at cervical and lumbar levels. By PD8 (24-25 mm SRL), TH-IR fibers are demonstrable in the dorsal and ventral funiculi, and a few have grown into the dorsal and ventral horns at most spinal levels (Fig. 1). Most of the immunostained axons in the dorsal horn are located in presumptive laminae IV-V. In coronal sections, TH-IR axons in the grey matter are cut primarily in cross section, but axonal segments exhibiting varicosities can be seen. Immunoreactive axons are present in all subdivisions of the presumptive ILC at thoracic levels (Figs. 1,2C,D). By PD15 (34 mm SRL) an occasional immunostained axon is found in presumptive lamina I and I1 at cervical, thoracic, and lumbar levels. By PD23 (46 mm SRT,) such axons are more numerous in both laminae and they are present at all spinal levels (Fig. 3). The density of TH-IR axons in the grey matter increases a t all levels with age, but particularly within the dorsal horn and presumptive ILC. By PD36 (65 mm SRL), TH-IR axons in the ILC are distributed as in the adult animal, and a few immunoreactive fibers in the ventral horn surround presumptive motoneurons in lamina IX. This apparent relationship with motoneurons persists into adulthood. By PD39 (69 mm SRL), TH-IR axons are more numerous in the lateral funiculus than a t early ages, and their density within the ILC and lamina IX is greater than in adults. By PD44 (76 mm SRL), the distribution of TH-IR axons (Figs. 3,4) resembles that described for adult animals (Pindzola et al., '88). lmmunoreactive axons are now fairly numerous in presumptive laminae I and I1 a t all levels (Figs. 3, 4A) and

R.R. PINDZOLA ET AL.

402

PDI

PD8 250pm

Cervical

Thoracic

Lumbar

Fig. 1. Distribution of TH-IR axons in the spinal cord of P D l (15 mm SRL) and PD8 (25 mm SRL) opossums. The arrows point to the TH-IR s o n s in the marginal zone. The presumptive intermediolateral cell column is indicated (ILC) in the thoracic cord of the PD8 case.

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD

Fig. 2. Low (A) and high (B) magnification photomicrographs of TH-IR axons (large arrows) in the dorsolateral marginal zone of the lumbar spinal cord of a PD1 opossum (15 mm SRL). B shows the area outlined in A. Low (C)and high (D) magnification photomicrographs of TH-IR axons in the dorsolateral marginal zone (arrows) and in the edge of the presumptive intermedinlateral cell column (ILC) at PD8. D was

403

taken from the area indicated in C. The central canal (cc),intermediate zone (IZ), and the marginal zone (MZ) are indicated. In A and C, the border between the marginal zone and the intermediate zone is indicated by small arrows. Differential interference contrast optics was used for all photomicrographs. Nucleated red blood cells are stained because of endogenous peroxidase activity.

R.R. PINDZOLA ET AL.

404

PD23

PD44 500pm

Cervica

Thoracic

Lumbar

Sacral

Fig. 3. Distribution of TH-IR axons in the spinal cords of PD23 (46 mm) and PD44 (76 mm SRL) opossums.

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD

Fig. 4. Photomicrographs of TH-IR axons in laminae I and I1 (A) and I11 and IV (B) of the sacral cord, the intermediolateral cell column of the thoracic cord (C), and lamina JX (D) of the lumbar cord from an

405

animal sacrificed at PD44. Differential interference contrast optics were used for all photomicrographs.

Fig. 5. TH-IR neurons in the dorsal horn a t PD44 (A),adjacent to the central canal (cc) a t PD32 (B),and in lamina IX of the ventral horn a t PD36 (C). Laminae I and I1 are indicated in A. Differential interference contrast optics were used for all photomicrographs.

R.R. PINDZOLA ET AL.

406

PDI

500

i

i

.

.*

I

C

Fig. 6. Distribution of TH-IR neurons (dots) in three sections of the brainstem of a P D l opossum demonstrated by PAP immunohistochemistry.

their density within presumptive laminae 111-IV (Figs. 3, 4B), VII, and VIII; the ILC (Figs. 3, 4C); and lamina IX (Figs. 3,4D) is greater than in adult opossums (Pindzola et al., '88). By PD90 (200 mm SRL), the density of TH-IR axons in these regions appears comparable to that of adult animals.

TH-IR perikarya in the developing spinal cord Tyrosine hydroxylase-like immunoreactive neurons are found in the spinal cord from PD1 to PD90. A t PD1 they are sparse and found mainly in the ventral horn at cervical

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD

407

Fig. 7. Photomicrographs of TH-IR neurons (PDI) in the presumptive nucleus paraventricularis hypothalami (PH; A,B), the presumptive nucleus coeruleus (Coe; C,D), and the trigeminal ganglion (TriG, C).

The areas outlined in A and C are shown in B and D, respectively. Differential interference contrast optics were used for all photomicrographs.

levels. Immunoreactive neurons increase with age, however, and become particularly numerous between PD32 and PD90 (Fig. 5). From P D l to PD38, TH-IR perikarya are encountered most readily a t cervical and thoracic levels, but their number increases a t lumbosacral levels with age. Between PD32 (59 mm SRL) and PD7.5 (130 mm SRL), TH-IR

neurons can be found in the dorsal horn (presumptive laminac I, 11, and V), around the central canal in presumptive lamina X (especially ventral to it where they are most numerous), in the marginal zone ventral to the central canal, and in the lateral and medial part of the ventral horn (presumptive laminae VII-IX). Many of the immunoreac-

R.R. PINDZOLA ET AL.

408

tive neurons around the central canal are located in the ependymal zone, and some have processes which extend into the canal itself (Fig. 5B). After PD90, the number of TH-IR neurons decreases but in adult animals an occasional one can still be found around the central canal a t cervical levels.

TH-IR perikarya in the developing brainstem At birth, TH-IR neurons are present in hypothalamic and brainstem areas presumed by position to be the nucleus paraventricularis hypothalami (Figs. 6A, 7A,B), the area hypothalamica lateralis, the nucleus coeruleus (Figs. 6B, 7C,D), the nucleus reticularis pontis (Fig. 6B), the ventrolateral medulla (Fig. 6C), the hypoglossal nucleus (Fig. 6C), and possibly the nucleus of the tractus solitarius (Fig. 6C). Those within the nucleus reticularis pontis, the ventrolateral medulla, and the solitary complex show relatively little immunostaining. TH-IR neurons are also seen in the ventral mesencephalon, where they appear to be migrating into the ventral tegmental area and substantia nigra, and within the trigeminal ganglion (Figs. 6B, 7C). By PD8, TH-IR neurons are more obvious in the nucleus of the tractus solitarius and by PD19 they can be detected in the nucleus periventricularis hypothalami. Immunostained neurons are seen in the nucleus reticularis gigantocellularis from PD4 to PD54 and medial to the nucleus ventralis lemnisci lateralis, the paralemniscal area, from PD8 to PD90. Immunostained neurons in the hypoglossal nucleus are not seen after PD60. Tyrosine hydroxylase-like immunoreactive neurons are present in the trigeminal ganglion until a t least PD22, but they might be present even later since we did not process that area at older ages.

The origins of TH-IR projections to the lumbosacral cord at different stages of development The locations of TH-IR neurons which project to the lumbosacral cord were studied at selected ages (Table 2) using the retrograde transport of FB and irnmunofluorescence. Three populations of neurons were present: 1) those containing only retrogradely transported FB, 2) those demonstrating only the FITC used to identify TH, and 3) double labeled neurons which contained both F B and FITC. The youngest successful cases were injected at PD5, and in both of them, F B spread bilaterally over several segments of the cord. Many double labeled neurons were found in areas presumed to be the nucleus paraventricularis hypothalami (Fig. 8A) and the area hypothalamica lateralis; the paralemniscal region of the nucleus reticularis pontis (Fig. 8R); the presumptive nucleus coeruleus (Fig. 8C,D), both dorsally and ventrally; non-paralemniscal areas of the nucleus reticularis pontis (Fig. 8D): and the rostral part of the ventrolateral medulla (Fig. 8E). The latter area appears to include the nucleus reticularis lateralis. Ventral to the presumptive locus coeruleus, TH-IR neurons were also labeled within the ependymal zone. Comparable results were obtained by injections a t PD6,12, and 15. In the cases injected at PD20, F B still spread to both sides of the cord, but it was densest on the side of the injection. The results from one case are plotted (Fig. 9) and representative photomicrographs are supplied in Figures 10 and 11. Tyrosine hydroxylase-like immunoreactive neurons were labeled in all of the areas described above, e.g., the presump-

tive nucleus paraventricularis hypothalami dorsalis (Figs. 9A, 10A,B), the paralemniscal area (Figs. 9B, 10C,D), the presumptive locus coeruleus (Figs. 9C,D, 11A,B), and the ventrolateral medulla (Figs. 9F,G, llC,D); but a few were also labeled in the reticular formation near the nucleus n. facialis and the nucleus reticularis gigantocellularis (Fig. 9F). After injections a t PD35, double labeled neurons could also be identified adjacent to the nucleus retrofacialis. When the injections were made between PD20 and PD44, double labeled neurons were more numerous in the rostral part of the ventrolateral medulla than a t earlier ages, and a t PD24 they were present more caudally, in the nucleus lateralis reticularis caudalis. It should be noted that the injection in the latter cases spread to the sacral cord. In addition to the labeling noted above, a few double labeled neurons were present dorsal to the superior olivary complex in animals injected a t PD44. By PD53-57, double labeled neurons in the paralemniscal area had decreased in number (Fig. 12B). Tyrosine hydroxylase-like immunoreactive neurons were still labeled in the dorsal part of the locus coeruleus (Fig. 12C,D) and in the reticular formation near the nuclei n. facialis and retrofacialis (Fig. lZF,G). The injections in the PD53 and PD57 animals were relatively small and more confined to the ipsilateral side than in previous cases. In adult animals, double labeled neurons were sparse in the paralemniscal region (Fig. 13C) and absent in the dorsalmost part of the rostral nucleus coeruleus (Fig. 13D). In the case plotted in Figure 13, the injection was intentionally large, covering all ofthe white matter and most of the gray matter bilaterally. Only part of contralateral ventral horn was spared. The rostral to caudal extent of the injection was not as great as in developing animals, however. Because our injections were large, we cannot make conclusions on the laterality of projections from the areas labeled. Comparison of the results obtained in developing and adult animals suggests that exuberant projections which are pruned prior to maturity existed during some stages of development.

DISCUSSION Technical considerations Antibodies to T H have been reported to label preferentially dopaminergic axons in several areas of the brain (Hokfelt et al., '77; Lewis et al., '87). It appeared that most CA axons were demonstrated in our experiments, however, since the distribution of immunolabeled axons was similar to that seen after processing by the Falck-Hillarp technique (Martin et al., '78). Of course, absence of immunoreactivity does not necessarily indicate absence of the enzyme since i t may be present at undetected levels. Unfortunately, use of antibodies to dopamine-beta-hydroxylase and phenylethanolamine N-methyltransferase (PNMT) did not result in specific immunostaining in the opossum, although a parallel study in the rat produced the expected results. It is possible that the molecular characteristics of the corresponding enzymes in opossums are different from those of rats. It appears that available antibodies to PNMT also failed to immunostain tissue in rabbits (Blessing et al., '86). In order to label as many spinally projecting neurons as possible, we made large pressure injections of FB. In the smaller animals the injections were large relative to those in older cases and they usually included more than one spinal segment. Since the injections were not comparable, the

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD

Fig. 8. Distribution of retrogradely labeled neurons (dots), ‘I’H-IK neurons (crosses), and neurons that contain F B and are TH-IR (stars) in an animal injected with F B a t PD5. For this figure and Figures 9,12, and 13, the sections are arranged from rostra1 (A) to caudal (E).

409

410

R.R. PINDZOLA ET AL.

Fig. 9. Distribution of FB (dots), TH-IR (crosses), and double labeled (stars) neurons in an animal injected at PD20.

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD

Fig. 10. Pairs of fluorescent photomicrographs from an animal injected a t PD20. Double labeled neurons (arrows) are shown in the presumptive nucleus paraventricularis hypothalami dorsalis (PHD; A,B) and t h e paralemniscal region of the reticularis pontis (pVLL;

411

C,D). T h e photomicrographs on the left were taken a t an excitation wavelength of 360 nm to demonstrate FB, whereas those on the right show the identical fields and plane of focus illuminated a t 490 nm to visualize FITC.

412

Fig. 11. Pairs of fluorescent photomicrographs from an animal injected at PD20. Double labeled neurons (arrows) are shown in the nucleus coeruleus (Coe, A,B) and the ventrolateral medulla (VLM,

R.R. PINDZOLA ET AL.

C,D). The photomicrographs on the left were taken a t an excitation wavelength of 360 nm for FB, whereas those on the right show the same field photographed at 490 nm for FITC.

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD

Fig. 12. Distribution of FB (dots), TH-IR (crosses), and double labeled (stars) neurons in a n animal injected with FB a t PD57.

413

R.R. PINDZOLA ET AL.

414

Fig. 13. Distribution of True Blue (dots),TH-IR (crosses), and double labeled (stars) neurons in an adult animal injected with True Blue.

apparent number of labeled neurons at different ages may not be meaningful. Some of the retrogradely labeled neurons that were not immunostained may contain low concentrations of TH, although we processed the tissue free-floating with rela-

tively concentrated antibody to minimize that possibility. It is also possible that TH may be present earlier than catecholamines (Specht et al., 'Ha). It should be noted, however, that neurons showing catecholamine-like fluorescence can be demonstrated in Falck-Hillarp material at

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD birth (Martin et al., '78); and they are located in many of the areas which contained immunolabeled neurons in the present study.

Innervation of the developing spinal cord by TH-IR axons Tyrosine hydroxylase-like immunoreactive axons are present in the marginal zone of the spinal cord at birth and they grow into both the ventral and dorsal horns by PD8. At PD8, TH-IR axons are most numerous within the presumptive ILC, suggesting that catecholamines may influence autonomic function early in development. It is not until PD44, however, that the pattern of TH-IR innervation in the dorsal horn resembles that in adult opossums. If sequences in axonal ingrowth reflect sequences in functional development, it is possible that CA modulation of pain processing occurs later than that of somatic motor and autonomic functions. Our results also suggest that TH-IR axons grow into the grey matter following rough rostra1 to caudal, as well as a ventral to dorsal gradients, similar to those found by Singer et al. ('80) for norepinephrinergic (NE) axons as well as Humbertson et al. ('82) and Bregman ('87) for 5-HT axons. Since our study was limited to light microscopy we cannot confirm the presence of synapses at any age. Assuming, however, that TH-IR axons form synapses when they grow into the grey matter, or soon thereafter, we suggest that they hyperinnervate the opossum's spinal cord between PD39 (69 mm SRL) and 90 (200 mm SRL). Commissiong ('83a) suggested that NE hyperinnervation and subsequent retraction occurs in the rat's cord and Bernstein-Goral and Bohn ('88) reported evidence for hyperinnervation of spinal sympathetic nuclei by epinephrinergic axons in the same species. It is possible, however, that what appears to be increased innervation density in the developing animal merely reflects a smaller volume of neuropil. As the animal matures, the neuropil enlarges due to dendritic and axonal growth, myelination, and other factors.

TH-IR perikarya in the developing brainstem and spinal cord At birth, TH-IR neurons are present in most of the brainstem areas which contain them in adult animals. It was difficult to identify some areas of the metencephalon with certainty, however, because of their immaturity and the presence of the pontine flexure. Between PD8 and PD90, TH-IR neurons were more numerous in the paralemniscal area of the lateral pons than in adult animals. In general, the distribution of TH-IR perikarya was similar to that of presumed CA perikarya demonstrated by fluorescent, histochemistry in the fetal rat (Olson and Seiger, '72), mouse (Golden, '73), rabbit (Tennyson et al., '72), and human (Olson et al., '73) and that demonstrated by immunohistochemistry for TH-IR in the prenatal rat (Specht et al., '81a,b). The transient existence of TH-like immunoreactivity in the nucleus n. hypoglossi and trigeminal ganglia may reflect cell death or loss of the ability to synthesize sufficient amounts of T H to be demonstrated by our protocol during later stages of development. Since dorsal root ganglia were not processed, we do not know whether they contain TH. Tyrosine hydroxylase-like irnmunoreactive perikarya are found in the spinal cord of developing opossums more frequently than in adult animals. In chicks, TH-IR perikarya

415

are also located around the central canal and, infrequently, within the dorsal horn (Wallace et al., '87). Dorsally positioned TH-IR neurons have also been seen in higher vertebrates (Singhaniyom et al., '83; Diet1 et al., '85). In chicks and opossums, the TH-IR neurons ventral to the central canal are comparable in position to the dopamine-containing cells described in lower vertebrates (Sims, '77, '86; Ochi et al., '79; Parent and Northcutt, '82; Wolters et al., '84). Sims ('86) found CA neurons in the ventral horn of the salamander cord which were only detected after transection of the spinal cord. Commissiong ('8%) found CA cells in the spinal cord of the rat at fetal day 18. The developmental and/or functional significance of such neurons is not known.

The origin of TH-IR projections to the spinal cord at different stages of development The present study is the first to identify the origin of CA axons within the developing lumbosacral spinal cord in any species. In newborn opossums, TH-IR cell bodies are found in all of the brainstem areas providing TH-IR projections to the spinal cord in adult animals (Pindzola et al., '86, '88) and by PD3 lumbar injections of FB label neurons in all such areas (unpublished results). I t is unfortunate that we were unable to obtain good results in the experiments which combined retrograde tracing with immunofluorescence prior to PD5, but it was difficult to maintain tissue integrity during processing and the FB faded rapidly. By PD5, however, lumbar injections of FB labeled TH-IR neurons within areas presumed to be the nucleus paraventricularis hypothalami, the nucleus coeruleus, and the ventrolateral medulla. These areas also project to the lumbar spinal cord in adult opossums. During development, the spinal cord receives TH-IR innervation from brainstem areas which provide little or no comparable innervation in adult animals. Such areas include the dorsal extreme of the nucleus coeruleus, the paralemniscal region, an area near the nucleus n. facialis, the reticular formation around the nucleus retrofacialis, the nucleus reticularis gigantocellularis and an area dorsal to the superior olive. It is possible that TH-IR neurons in such areas die or withdraw their spinal axon, but i t is equally possible that they simply change their phenotype. Cabana and Martin ('84) reported previously that the dorsal part of the locus coeruleus provides transient projections to the spinal cord and their observation has been supported in the rat (Chen and Stanfield, '87). In the rat, such projections are apparently lost by the elimination of spinal collaterals from axons which retain projections to other areas.

Possible correlation of presumptive CA innervation with the development of hindlimb motility There is evidence that descending projections to the spinal cord affect the development of motor activity. Supraspinal axons first reach lumbar levels on embryonic day (E) 5 in the chick (Okado and Oppenheim, '85), but changes in leg motility do not occur until El0 following transection of the spinal cord on E2 (Oppenheim, '75). Some of the supraspinal axons which infiuence embryonic movement in the chick may be catecholaminergic, since Singer et al. ('80) found that norepinephrinergic axons, detected by uptake studies, extend throughout the spinal cord by 10 days of incubation. During the next few days these axons can

R.R. PINDZOLA ET AL.

416 synthesize and store NE and this becomes well established by incubation day 17. Refinement of embryonic motility occurs on incubation days 16-17, and it depends on inhibitory supraspinal influences (Oppenheim, '75; Sastry and Sinclair, '76). A possible role for catecholamines on the development of spontaneous limb movements in the rat is suggested by the work of Commissiong ('83b). In opossums, brainstem axons are present within the marginal zone of the lumbosacral cord before spontaneous hindlimb movements begin (PD7-10; Martin et al., '78), and our results suggest that some of them are catecholaminergic. Since dendrites of spinal neurons extend into the marginal zone at early stages of development, it is possible that CA axons influence pattern generators for limb movement even before they reach the intermediate zone. In any case, a few CA axons have grown into presumptive laminae V-VII of the lumbar cord by the time spontaneous hindlimb movements begin, and when hindlimb movements can be altered by cutting brainstem projections to the lumhosacral cord (Martin et al., '78), CA axons have grown into most of the areas they occupy in the adult animal. Although our results establish a temporal correlation between growth of CA axons into the lumbosacral cord and the development of hindlimb motility, they do not prove that catecholamines influence motility. Proof awaits manipulation of spinal catecholamines and observing its effect on behavior.

SUMMARY AND CONCLUSIONS Our results suggest that in opossums 1) TH-IR axons begin to grow into the spinal cord prenatally, but innervate their adult targets postnatally and over an extended period of time; 2) during most, if not all, stages of development TH-IR axons originate from areas which contribute to such axons in the adult animal, and 3) during some stages of development TH-IR axons grow into the lumhosacral cord from areas which do not give rise to comparable axons in adult animals.

Chan-Palay, V., L. Zahorszky, C. Kohler, M. Goldstein, and S.L. Palay (1984) Distribution of tyrosine-hydroxylase immunoreactive neurons in the hypothalamus of rats. J. Comp. Neurol. 227:467-496. Chen, K.S., and B.B. Stanfield (1987) Evidence that selective collateral elimination during postnatal development results in a restriction in the distribution of locus coeruleus neurons which project to the spinal cord in rats. Brain Res. 420:154-158. Commissiong, J.W. (19838) Development of catecholaminergic nerves in the spinal cord of the rat. Brain Res. 264:197-208. Commissiong, J.W. (1983b) The development of catecholaminergic nerves in the spinal cord of the rat. 11. Regional development. Dev. Brain Res. 11:75-92.

Cutts, J.H., W.J. Krause, and C.R. Leeson (1978) Generalobservations on the growth and development of the pouch-young opossum (Didelphis oirginzana). Biol. Neonate 33264-272. Dietl, M., M. Arluison, P. Mouchet, C. Feuerstein, M. Manier, and 3. Thibault (1985) Immunobistochemical demonstration of catecholaminergic cell bodies in the spinal cord of the rat. Histochemistry 82:385-389. Fossberg, H., and S. Grillner (1973) The locomotion of the acute spinal cat injected with Clonidine, i.v. Brain Res. 50:184-186. Golden, G.S. (1973) Prenatal development of the biogenic amine systems of the mouse brain. Dev. Biol. 33.300-311. Grillner, S. (1976) Some aspects on the descending control of the spinal circuits generating locomotor movements. In R.M. Herman, S. Grillner, P.S.G. Stein, and D.G. Stuart (eds): Neural Control of Locomotion. New York Plenum Press, pp. 351-375. Hokfelt, T., 0. Johansson, K. Fuxe, M. Goldstein, and D. Park (1977) Immunohistochemical studies on the localization and distribution of monoamine neuron systems in the rat brain. 11. Tyrosine hydroxylase in the telencephalon. Med. Biol. 5 5 . 2 4 0 . Hokfelt, T., 0. Johansson, and M. Goldstein (1984) Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bjorklund and T. Hokfelt (eds): Handbook of Chemical Neuroanatomy. Amsterdam: Elsevier Science Publishers, Val. 2, pp. 157-276. Humhertson, A.O., Jr., T. Cabana, F.J. DiTirro, R.H. Ho. and G.F. Martin (1982) Development of raphe-spinal connections in the North American opossum. Brain Res. Bull. 9:627-633. Humhertson, A.O., Jr., and G.F. Martin (1979)The development of monoaminergic brainstem-spinal systems in the North American opossum. Anat. Embryol. (Berl.) I56:301-318. Lewis, D.A., M.J. Campbell, S.L. Foote, M. Goldstein, and J.H. Morrison (1987) The distribution of tyrosine hydroxylase-immunoreactive fibers in primate neocortex is widespread but regionally specific. J. Neurosci. 7(1):279-290.

Loizou, L.A. (1972) The postnatal ontogeny of monoamine-containing neurons in the central nervous system of the albino rat. Brain Res. 40.395-

ACKNOWLEDGMENTS The authors acknowledge Mr. John M. Pindzola for assistance with -the graphics and Mr. Karl Ruben for help with photographic processing. This study was supported by NS-25095.

LITERATURE CITED Aramant, R.B., L.T. Giron, Jr., and M.G. Ziegler (1986) Postnatal development of dopamine-&hydroxylase-immunoreactive fibers of the spinal cord of the rat. Dev. Brain Res. 25:161-171. Bernstein-Goral, H., and M.C. Bohn (1988) Ontogeny of adrenergic fibers in rat spinal cord in relationship to adrenal preganglionic neurons. J. Neurosci. Res. 21:333-351. Blessing, W.W., P.R.C. Howe, T.H. Joh, J.R. Oliver, and J.O. Willoughby (1986) Distribution of tyrosine hydroxylase and neuropeptide Y-like immunoreactive neurons in rabbit medulla oblongata, with attention to colocalization studies, presumptive adrenaline-synthesizing perikarya, and vagal preganglionic cells. J. Comp. Neurol. 248.285-300. Bregman, B.S. (1987) Development of serotonin immunoreactivity in the rat spinal cord and its plasticity after neonatal spinal cord lesions. Dev. Brain Res. 34 245-263. Cabana, T., and G.F. Martin (1984) Developmental sequences in the origin of descending spinal pathways. Studies using retrograde transport techniques in the North American Opossum (Didelphis uirginiana). Dev. Brain Res. 15.247-263.

418.

Martin, G.F., J.K. Beals, J.L. Culberson, R. Dom, G. Goode, and A.O. Humhertson, Jr. (1978) Observations on the development of hrainstemspinal systems in the North American opossum. J. Comp. Neurol. 181:271-290.

Ochi, J., T. Yamamoto, and Y. Hosoya (1979) Comparative study of the monoamine neuron system in the spinal cord of the lamprey and hagfish. Arch. Histol. Jpn. 42:327-336. Okado, N., and R.W. Oppenheim (1985) The onset and development of descending pathways to the spinal cord in the chick embryo. J. Comp. Neurol. 232:143-161. Olson, L., and A. Seiger (1972) Early prenatal ontogeny of central monoamine neurons in the rat: Fluorescence histochemical observations, Z. Anat. Entwickl. Gesch. 137:301-316. Olson, L., O.L. Boreus, and A. Sieger (1973) Histochemical demonstration and mapping of 5-hydroxytryptamine- and catecholamine-containing neuron systems in the human fetal brain. Z. Anat. Entwickl. Gesch. 139:259-282.

Oppenheim, R.W. (1975) The role of supraspinal input in embryonic motility: A re-examination in the chick. J. Comp. Neurol. 160:37-50. Oswaldo-Cruz, E., and C.E. Rocha-Miranda (1968) The Brain of the Opossum (Didelphis marsupialis). Rio de Janeiro: Instituto de Biofiuica, Universidade Federal de Rio de Janeiro. Parent, A., and R.G. Northcutt (1982) The monoamine-containing neurons in the brain of the garfish, Lepisosteus osseu.7. Brain Res. Bull. 9:189-204. Pickel, V.M., T.H. Joh, P.M. Field, C.G. Becker, and D.J. Reis (1975) Cellular localization of tyrosine hydroxylase by immunohistochemistry. d. Histochem. Cytochem. 23(1):1-12.

DEVELOPMENT OF TH-IR PROJECTIONS TO SPINAL CORD Pindzola, R.R., G.F. Martin, and R.H. Ho (1986) Catecholaminergic innervation of the spinal cord in the North American opossum, Didelphis uirgrniana Proc. Soc. Neurosci. 12(2):1547 (Abstract). Pindzola, R.R., R.H. Ho, and G.F. Martin (1988) Catecholaminergic innervation of the spinal cord in the North American opossum, Didelphis uirginiana. Brain Behav. Evol. 32281-292. Rexed, B. (1954) A cytoarchitectanic atlas of the spinal cord in the cat. J. Comp. Neurol. 100:297-379. Sastry, B.S.R., and J.G. Sinclair (1976) Tonic inhibitory influence of a supraspinal monoaminergic system on recurrent inhibition of an extensor monosynaptic reflex. Brain Res. 117:69-76. Sawchenko, P.E., and L.W. Swanson (1981) A method for tracing biochemically defined pathways in the central nervous system using combined fluorescence retrograde transport and immunohistochemical techniques. Brain Res. 210:31-51. Seiger, A,, and L. Olson (1973) Late prenatal ontogeny of central monoamine neurons in the rat: Fluorescence histochemical observations. Z. Anat. Entwickl. Gesch. 140:281-318. Sims, T.J. (1977) The development of monoamine-containing neurons in the brain and spinal cord of the salamander (Ambystoma mexiconum). J. Comp. Neurol. 173:319-336. Sims, T.J. (1986) Identification of a second type of catecholaminergic neuron in the spinal cord of the axolotl salamander. Exp. Neurol. 93.42b433. J.T. Coyle, N.Vernon, C.H. Kallman, and D.L. Price (1980) The Singer, H.S., development of catecholamine innervation in chick spinal cord. Brain Res. 191:417-428. Singhaniyom, W., N.G.M. Wreford, and F.-H. Guldner (1983) Asymmetric distribution of catecholamine-containing neuronal perikarya in the upper cervical spinal cord of rat. Neurosci. Lett. 41:Yl-97.

417

Specht, L.A., V.M. Pickel, T.H. Joh, and D.J. Reis (1981a) Light-microscopic immunocytochemical localization of tyrosine hydroxylase in prenatal rat brain. I. Early ontogeny. J. Comp. Neurol. 199233-253. Specht, L.A., V.M. Pickel, T.H. Job, and D.J. Reis (1981b) Light-microscopic immunocytochemical localization of tyrosine hydroxylase in prenatal rat brain. 11. Late ontogeny. J. Comp. Neurol. 199255-276. Stephanini, M., C.D. Martinio, and L. Zamboni (1967) Fixation of ejaculated spermatozoa for electron microscopy. Nature (London) 216:173-174. Sternherger, L.A., P.H. Hardy, J.J. Cuculus, and H.G. Meyer (1970) T h e unlabelled antibody-enzyme method of immunohistochemistry. Preparation and properties of soluable antigen-antibody complex (horseradish peroxidase-antihorseradish peroxidase) and its use in identification of spirochetes. J . Histochem. Cytochem. 18:315-333. Tennyson, V.M., R.E. Barrett, G. Cohen: L. Cote, R. Heikkela, and C. blytilineou (1972) T h e developing neostriatum of the rabbit: Correlation of fluorescence histochemistry, electron microscopy, endogenous dopadopamine uptake. Brain Res. 46:251-285. mine levels and [3H] Wallace, J.A., R.M. Mondragon, P.C. Allgood, T.J. Hoffman, and R.R. Maez (1987) Two populations of tyrosine hydroxylase-positive cells occur in the spinal cord of the chick embryo and hatchling. Neurosci. Lett. 83253258. Westlund, K.N., and J.D. Coulter (1980) Descending projections of the locus coeruleus and subcoeruleus/medial parahrachial nuclei in monkey: Aximmunocytochemonal transport studies and dopamine-beta-hydroxylase istry. Brain Res. Rev. 2235-264. Wolters, J.G., H.J. Ten Donkelaar, and A.A.J. Verhofstad (1984) Distribution of catecholamines in the brainstem and spinal cord of the lizard Varanus exanthematicus: An immrinohistochemical study based on the use of antibodies to tyrosine hydroxylase. Neuroscience 13(2):469-493.