Light and electron microscopic study of tyrosine hydroxylase-immunoreactive neurons within the developing rat arcuate nucleus

Brain Research. 439 (1988) 127-137 E~sevier

127

BRE 13205

Light and electron microscopic study of tyrosine hydroxylaseimmunoreactive neurons within the developing rat arcuate nucleus M. Piotte l'*, A. Beaudet 1"2and J.R. Brawer !"3 Departments of 1Anatomy and 2Neurologyand Neurosurgery. and 3The McGill Ccmrefor the Study of Reproduction. McGill University, Montreal, Que. (Canada) (Accepted 30 June 1987)

Key words: Dopaminergic: Hypothalamus: lmmunocytochemistry: Neonate: Dendrite: Tanycytic process

The topography, fine structure, and patterns of connections of tyrosine hydroxylase (TH)-immunoreactive tubero-infundibular dopaminergic (TIDA) neurons were examined by light and electron microscopic immunocytochemistry in the arcuate nucleus of 2-, 15-and 30-day-old female Wistar rats. In 2-day-old animals, TH-immunoreactive perikarya were mainly located in the ventrolateral portion of the arcuate nucleus. In 15 day-old rats numerous TH-positive cell bodies were still present ventrolateraily, but a cluster of labeled cells was also apparent in the mediodorsal segment of the nucleus. In the 30-day-old rats, most TH-immunoreactive neurons were concentrated mediodorsally, as seen in the adult. At the ultrastructural level, TH-immunoreactive somata exhibited, in all age groups, a large nucleus surrounded by a thin rim of cytoplasm containing mitochondria, Golgi apparatus, endoplasmic reticulum, multivesicular bodies and lysosomes. These labeled somata were synapticaily contacted by unlabeled axon terminals and often laid adjacent to either labeled or unlabeled dendrites. Similarly, in all age groups, labeled dendrites were synaptically contacted by unlabeled axon terminals and were often directly apposed to either labeled or unlabeled perikarya and dendrites, or to tanycytic processes. These results indicate that TIDA neurons establish extensive connections early in development, and that their pattern of intercellular relationships remains qualitatively unchanged from 2 days to adulthood. It is suggested that TIDA neurons may be already functional at birth, and could therefore, influence the maturation of other arcuate neuronal populations.

INTRODUCTION Tubero-infundibular dopaminergic ( T I D A ) neurons appear at an early stage of brain development. Histochemical studies, ba~ed on the use of the F a l c k Hillarp formaldehyde fluorescence technique, demonstrated the presence of dopaminergic perikarya within the hypothalamic arcuate nucleus of the fullterm rat fetus (20-22 days of gestation) s'25. Recently, application of the more sensitive 7 peroxidase-anti-peroxidase (PAP) immunocytochemica! me:hod of Sternberger 2s enabled detection of neuronal perikarya containing the catecholamine-synthesizing e~:zyme, tyrosine hydroxylase (TH), as early as on embryonic day 14.5 (ref. 26). Furthermore, the topography of the TH-immunoreactive neurons in the ar-

cuate nucleus of fetal rats closely resembled that of the adult by embryonic day 18 (ref. 27). Consistent with these obse~ations are those suggesting that neonatal T I D A neurons are biochemically and physiologically active. For example, these neurons can be depleted of neurotransmitter by reserpine and catecholamine synthesis inhibitors ~5a7 and can take up the amine precursor L-DOPA 16. In addition, the prolactin inhibitory function of the T I D A system appears to be operative at birth 18. The histological and apparent functional maturity of the neonatal T I D A system suggests that at least some of its connections are established early in development. Although information is currently available concerning the connectivity of dopaminergic neurons in the arcuate nucleus of the adult rat 2"3"!1'13a4'23"3°,

* Present Address: Department of Pharmacology and Therapeutics, McGill University, Montreal, Que., Canada. Correspondence: J.R. Brawer, Department of Anatomy, McGiU University, 3640 University Street, Montreal, Que., Canada H3A 2B2. 0006-8993/88/$03.50 © 1988 Elsevier Science Publishe~ B.V. (Biomedical Division)

128 nothing is known about the fine structural features and cellular relationships of these cells during development. In the present study, the fine structure and patterns of connections of TH-immunoreactive cell bodies and processes were therefore examined in the arcuate nucleus of 2-, 15- and 30-day-old Wistar female rats. MATERIALS AND MEI'HODS

Animals Two- end 15-day-old rats. Pregnant Wistar female rats were housed in individual cages in which they later gave birth. On days 2 and 15 postpartum, 3 female pups were killed by intracardiac whole body gravity perfusion. The perfusing solution contained a mixture of 0.5% glutaraldehyde, 4% paraforma~dehyde and 0.1% picric acid in 0.1 M Sorensen buffer. The amount of fixative delivered per animal was 50 ml for the 2-day-old rats and 100 ml for the 15-day-old rats. Immediately after fixation the brains were removed, and blocks including the entire hypothaiamus were postfixed for 1 h in the same fixative. The blocks were then put in a cold (4 °C) paraformaldehyde solution (4%) for 2-4 h. Thick (50 and 30/~m for the neonates and the 15-day-old rats respectively) coronal sections of the arcuate nucleus were cut on a Vibratome (Oxford), collected in phosphate buffer 0.1 M at room temperature, and immediately processed for immunocytochemistry. Thirty-day-old rats. Three 30-day-old Wistar females were perfused by intra-aortic arch perfusion of 300 ml of the fixative mixture described above, using a manostat varistaltic pump (80 ml/min). The brains were subsequently processed as those of 15-day-old animals, except that the hypothalamic blocks were not put in a cold paraformaldehyde solution after postfixation.

cording to a protocol modified from Pickei ~9. Briefly, the free floating sections were rinsed twice in Tris-saline (0.5 M) containing lysine (0.1 M), and sequentially incubated with: (1) normal goat serum (NGS) diluted 1:30 with Tris-saline (30 min); (2) the TH antibody diluted 1:750 with Tris-safine containing 1% NGS (overnight, at 4 °C); (3) goat anti-rabbit immunoglobulins (Miles) diluted 1:50 in Tris-saline containing 1% NGS (30 min); (4) PAP (Cedarlar~e) diluted 1:50 with Tris-saline containing 1% NGS (30 min). Between each incubation, the sections were rinsed twice in Tris-saline containing 1% NGS. They finally were incubated 10 min with CoC! 2 0.5%, rinsed twice in Tris-saline, and reacted for 6 min with 0.05% 3,3'-diaminobenzidine and 0.01% hydrogen peroxide in Tris buffer. Following immunocytochemical staining, every fourth section from the mid-arcuate nucleus region (Figs. lb,c, 2b,c and 3b,c) was postfixed for 1 h in 2% OsO4, block-stained with uranyl acetate, dehydrated in graded ethanols, and flat-embedded in Epon. These flat-embedded sections were then reincluded in Beem capsules and thin sections were cut from their surface, collected on Formvar-coated grids, stained with lead citrate for 2.5 min, a ' :! examined with a Siemens 101 electron microscope. ~ total of 18 thin sections from 2-day-old, 8 from 15-day-old, and 8 from 30-day-old rats (2-8 sections/animal) was systematically scanned and every TH-immunoreactive element encountered photographed at an original magnification of "<10.000. The remaining immunocytochemically stained sections were mounted on gelatin-coated glass slides, defat~ed, coverslipped, examined with a Wild ligt't microscope, and mapped at x 125 using a camera lucida. RESULTS

Irnmunocytochemical procedures

Topography

Antibodies against partially purified bovine adrenal TH were supplied by Dr. Tong Hyub Joh (Cornell University, New York, NY). The immunological characteristics of these rabbit-raised antibodies and their specificity towards catecholamine-containing neurons have been described in detail elsewhere 9'1°" 20,21,22. Sections of the arcuate nucleus were immunostained using the PAP method of Sternberger 2s, ac-

Two-day-old rats (neonates). At this age, the population of TH-immunoreactive nerve cell bodies was mainly located in the ventrolateral region of the arcuatc nucleus and extended laterally into the ventral portion of the lateral hypothalamus. Only a few lightly staining cells occurred in the mediodorsal area (Fig. la-d). Immunoreactive perikarya were also oc-

129

a

°

b

~

o.

,~*~r

~

..

C

/

__e

e"

~,~

, ,7

4m :1~ "q~

d

casionally located within the median eminence (Fig. lb). The labeled neurons were surrounded by an intricate network of straight, thick immunoreactive processes. In contrast, small immunopositive punc• " • "~3 tate elements, typical of labeled axonal vancosmes-, were scarce• Fifteen-day-old rats. In 15-day-old animals, a mediodorsa! cluster of labeled cells became apparent in addition to the labeled cells in the ventrolateral arcuate nucleus and those above the ventral border of the lateral hypothalamus (Fig. 2a-d). The TH-positive cells of the mediodorsal cluster generally stained more intensely than the cells of the ventrolateral arcuate region, or lateral hypothalamic area. In the most posterior region of the nucleus, the distinct ventrolateral and dorsomedial parcellation was lost as cells become more evenly distributed (Fig. 2d). Thirty-day-old rats. The distribution of TH-positive perikarya, processes and punctate elements within the anteroposterior extent of the arcuate nucleus was similar to that seen in the adult 5"6"23"3°. In the anterior portion of the nucleus (Fig. 3a), immunoreactive cells were distributed on either side of a slightly convex linear trajectory running from the mediodorsal to the ventrolateral arcuate areas. In the mid-arcuate region (Fig. 3b,c), numerous THpositive perikarya were present in the mediodorsal area of the nucleus, immediately adjacent to the third ventricle. Most of the remaining immunoreactive cells were scattered within the lateral region of the nucleus. These lightly staining cells appeared almost as a distinct ventrolateral cell cluster. Posteriorly (Fig. 3d), the labeled cells were distributed uniformly throughout the nucleus.

Electron microscopy Perikarya Two-day-old rats (neonates). Thirty TH-positive

Fig. 1. TH-positive nerve cell bodies and processes from representative coronal sections taken from the anterior (a), middle (b,c) and posterior (d) regions of a female neonatal arcuate nucleus. The intensely immunoreactive perikarya are filled in while the pale staining neurons are not. Bar = 100pm

perikarya were observed in our material. All of them were ovoid in shape. These somata were smal!, (5 x 8 pm in mean diameter) and contained a large, generally indented nucleus (Fig. 4). The majority of nuclear profiles exhibited a single nucleolus. The thin rim of cytoplasm was filled with a black, granular immunoprecipitate which tended to accumulate at the surface of organelles and inclusions. In some of the labeled perikarya cytoplasmic organelles were

130

,

o

o.~

a

a

e

9

: ;E /: ,',;

o

0 ~

• ~'

•

o

o

d, °

b.

° 0

o

,,

I

°

) ¢o

o~

"

~

oB

o b 0 ,o

~v"

b

b

d ;)

•

o ";'1 ~::.~"

o; qb °O

" "lb2 0 Q O

bes,

0o

O

I

Q

c

0V.,

¢.:lJl o.o

a o

•

oO

o~

6

•

a

o

~o

4-",

•

~v*

0

Q~

e o•

o

~ ~)

°,~

..

%m

o

J

d

\

•

•

(".: i" "\ / :-, !o o°') ( o.°:

~al, °

~•o

o

Fig. 2. TH-positive nerve cell bodies from representative coronal sections taken from the anterior (a), middle (b,c) and posterior (d) regions of a 15-day-old female arcuate nucleus. Bar = !00 #,m.

Fig. 3. TH-positive nerve cell bodies from representative coronal sections taken from the anterior (a), middle (b,c) and posterior (d) regions of a 30-day-old female arcuate nucleus. In this age group the ventrieular ependymal lining is fully differentiated (that is, no longer stratified) and is rer~resented as a selid black line. Bar = lO0/~m.

or,

"

."~.

....

~"

• "

~:

"~'~,

-'~" , ~."

,

..

;

....

~. p o

---

~

,

,

" ~- ~,

.

-, ; : 3 ~

,

'~".

.... . .

. ..,

,

~."

k--

"

I ~ " :..% ~ ..

- ;~

-

.-"-~'..,'"-

"

,

*~4 . ~,~.

~

-:'-'~-

,

".~-'".'

'-.

" ,, ~~ .

~

.

~~¢. ,;, . ~,..,~. "

t~

-

;

,

,~ .

-

:~.~

, ',t ~ .

'-"q

~y

'~'.~

-"

"

.;_.-.

"

~ .....

~

~. - - ' L r -.

'

'

~ . "

. . . .

.

.

r ,,~

.

"

.

~.

. . . .

:

. . . .

~"..

!~T!il~i

" ',

,,.

-~,~.

~

"

-"

"~.T-,;"'~

"

, -~, ..

.

-...

~ . , j ~

,~

"

.

"

.

.

"

-,~

!~.-~',,

" . . -

. t

"- •

:

~"

.

~-.

.

',~

"

,

.'.,~".-,-

~;.

,,,,

.~",~

...

~;,,;

~,

.

.

. •

•

-

~t

:-

",;." ,.

-

J

~o

~ ,

" ' ~ ~ , ~ ,,~

,

"

';" "

,

i.¢.

.....

_.

,,..

"

I:,:

-"

,.-,

::

4

Fig. 4. TH-positive nerve cell body from a neonatal arcuate nucleus (see text for description). The inset illustrates an asymmetrical axosomatic synapse involving an unlabeled axon terminal and a TH-immunoreactive perikaryon from a 15-day-old rat. Bars = 1 !~m and 0.5/~m (inset).

.~'::

~-~

::,:)!

,

•

~

-

~

~

e o"

~'.

~

~,~'

,

~J

.~

. .

..1

"%

; e~

!

~:

J:,

~,. % .

'°:'qA:" :

' :'

4

<*

133 obscured by the reaction product, but the majority of immunoreactive cell bodies exhibited mitochondria, a small and poorly developed Goigi apparatus, as well as a few muitivesicular bodies. In addition, short, sometimes distended, cisternae of endoplasmic reticulum were present and, occasionally, long individual cisternae extended for some distance parallel to the cell's contour. Lysosomes were rarely encountered. Neonatal TH-positive perikarya were synaptically contacted by numerous unlabeled axon terminals. Some were also directly apposed to labeled (in 4 instances) or to unlabeled (in 10 instances) dendrites (Fig. 5). At these sites of apposition, membranes were parallel to each other without any intervening glia, and vesicular aggregates and/or membrane differentiations were absent. Labeled neuronal perikarya were also occasionally contacted by tanycytic processes. Fifteen-day-old rats. Thirty-four TH-positive nerve cells were sampled in 15-day-old females. They were slightly larger than (6 x 10 ~m on average), but ultrastructuraily similar to, those in the neonates. In contrast to the neonatal perikarya, however, these cell bodies contained a well-developed Golgi apparatus. The cisternae making up each Golgi stack were larger, and the stacks more numerous than in the neonate. Labeled perikarva were frequently seen in synaptic contact with unlabeled axon terminals (Fig. 4, inset). Appositions similar to those described in the neonate were also observed: some labeled perikarya were adjacent to labeled (in 5 instances) or to unlabeled (in 17 instances) dendrites~ without any obvious membrane specialization at the site of apposition. Thirty-day-old rats. In 30-day-old female rats, 20 immunoreactive perikaryal profiles were examined. They averaged 7/~m in short diameter and 9/~m in long diameter. Their ultrastructural features were similar to those of TH-positive neurons in 15-day-old

rats. As in neonate and 15-day-old animals, these nerve cell bodies were synaptically contacted by unlabeled axon terminals. A few were directly apposed to labeled (in 4 instances) or unlabeled (in 4 instances) dendrites. Finally, some of the perikarya were ensheathed by glial and tanycytic processes. Dendrites Numerous TH-immunoreactive dendritic profiles, morphologically similar to those described in adults 23, were visible in all age groups. These dendritic profiles showed, on average, the same crosssectional diameter in all age groups. Examination of 192 of these profiles in the neonate, 353 in the 15day-old, and 401 in the 30-day-old, revealed that all types of relationships established by adult TH-positive dendrites were present at each age studied. These included direct appositions, without synaptic specializations, to unlabeled perikarya, dendrites (Fig. 6) and tanycytic processes (Fig. 7), and synaptic contacts with unlabeled axon terminals (Fig. 8). A few of the labeled dendrites also directly abutted other labeled dendrites (Fig. 9) and perikarya (Fig. 10) without any obvious membrane differentiation and/or vesicular clustering at the site of apposition. Labeled axon termblals Although they were relatively scarce in all age groups~ immunoreactive axon terminals were most frequently encountered in the neonates. Totals of 48, 11, 19 and 27 such terminals were observed in the arcuate nucleus of 2-, 15- and 30-day-old, and normally cycling adult rats respectively. These labeled terminals showed no significant change in cross-sectional diameter between day 2 and ~du!thood. In all age groups, they formed synaptic contacts with unlabeled dendrites (Fig. 11). Only in neonates, however, were they seen in synaptic relationship with unlabeled somat:.~ or TH-positive dendrites.

Fig. 5. Dendrosomatic apposition between an unlabeled dendrite (arrow) and a labeled perikaryon within the neonatal arcuate nucleus. Bar = 0.5/tin. Fig. 6. One TH-positive dendrite in direct apposition to two unlabeled dendrites within the arcuate nucleus of a female neonate. There is no membrane specialization nor vesicular cluster at the points of contact. An unlabeled axon terminal, which makes an asymmetrical synapse on the labeled dendrite, is also present. Bar = 0.3.urn, Fig. 7. A neonatal dendritic profile in direct apposition (arrow) to an extension of a tanycytic process (tan). Note the characteristic lipid droplet within the tanycyticprocess of origin (upper portion of the figure). Bar = 0.3 ~tm.

•

0

•

•

•

-..

•

.,

.

..

'

..

.

•-

.z~

1

"-":-~'~i~i

-~',~

-

~.~

,.:?" ,,~k.v~,l

.... ....,~:::,~.:~,_., Y~

.,:,;

~" '

..,

.... ~

~,~,.

.

.

& .... ~..

, .~:~..~.,,

.~.~,

,¢ ...., :.-

L

•

•

_

"""~

,~,

J,~" ~'>' ~" " ~

," •

"

~:.

~

" '"f

~

~

135

Fig. 11. TH-immunoreactive axon terminal forming a symmetrical synapse with an unlabeled dendrite from a 30-day-old rat. Bar = 0.5~um.

DISCUSSION

Topography The presence within the neonatal arcuate nucleus of mature-looking TH-positive neurons in consistent with the results of Specht et al. z7 who observed such neurons as early as embryonic day 18. These investigators, however, also reported that the distributional pattern of TH-positive perikarya in fetal arcuate nu-

cleus (embryonic day 21) was similar to that in the adult. In contrast, we have shown a difference between the neonatal and the adult topography. In the neonate most TH-posit~ve cells are located in the ventrolateral region of the nucleus while in the adult these cells occur predominantly in the mediodorsal region. This discrepancy may be explained by the fact that Specht et al. 27 described the location of THimmunoreactive cells within the whole brain and, consequently, did not study the arcuate nucleus in detail. The present results indicate that the characteristic adult distribution is not present until at least the third postnatal week. Indeed, in 15-day-old animals, the distribution of T I D A neurons is intermediate between the neonatal pattern and that of the adult. Daikoku et al. 6 also observed the adult distributional pattern of TH-immunoreactive neurons relatively late in postnatal development (20th postnatal day). Both our results and those of Daikoku et al. 6 suggest that cell migration may continue well into postnatal life. The mediolateral and dorsoventral expansion of the arcuate nucleus during development may also contribute to the final adult cell distribution. The observed range of staining intensity among the TH-immunoreactive perikarya of all age groups probably reflects different T H expression by different T I D A neurone. In all age groups examined, TH-immunoreactive neurons were generally located along the slightly curved mediolateral trajectory followed by tanycytic processes. Furthermore, in the electron microscope, TH-immunoreactive perikarya and dendrites were found to directly abut tanycytic processes. It is therefore tempting to speculate that immature TH-positive neurons migrate along tanycytic processes during development. The concept of glial processes gaiding migrating cells during development was proposed by Rakic 24 following his observation, within

Fig. 8. An unlabeled axon terminal svnapses, on a labeled dendrite within the neonatal arcuate nucleus. Both vesicular cluster and membrane specialization are clearly seen at the svnaptic junction. "lhe labeled dendrite is also in direct apposition to another immunoreactive dendrite and. even though no vesicularcluster is seen at the point of contact (arrow), a membrane specialization is present. Bar = 0.3 gtm. Fig. 9. Dendrodendritic apposition involving two labeled dendrites within the arcuate nucleus of a 15-day-old rat. There is no ve,Acular cluster nor membrane specialization at the point of contact. Bar = 0.5 ~,m. Fig. 10. Dendrosomatic apposition between a labeled dendrite and a labeled perikaryon within the arcuate nucleus of a 15-day-old animal. The TH-positive dendritic profile is also synaptically contacted by an unlabeled axon terminal. Bar = 0.3 jtm.

136 the monkey cerebellar cortex, of a close association between migrating granule cells and b~ghly oriented g!ial fibers. It is also possible, however, that the similarity of orientation between the course of the tanycytic processes i and the distribution of TH-positive neurons is a fortuitous event. Ultrastructural organization The ultrastructural similarity between immature and adult T I D A perikarya and dendrites, and the presence in the neonate of the same types of cellular relationships as those encountered in the adult, suggest that neonatal TIDA neurons are synthesizing and releasing their neurotransmitter at birth. These findings conform very well to biochemical and physiological evidence indicating an early onset of T I D A function 8'17"18. The argument has been advanced that, in the adult, dendritic release of dopamine may be a major mechanism of communication in the TIDA system, as it is in the substantia nigra 23, and that dendrosomatic and dendrodendritic appositions involving TH-immunoreactive neurons may represent functional contacts despite the absence of any synaptic specialization 23'29. If such is the case, the neonatal dendrosomatic and dendrodendritic appositions between dopaminergic neurons may form part of the connections necessary for the self-inhibitory activity within the TIDA system 12. Thus, SOlne of these self-inhibitory connections would be established and functional at birth. Similarly, assuming

that appositions between TH-immunoreactive dendrites and non-immunoreactive dendritic and perikaryal profiles are functionally significant, they might represent a means for T I D A neurons to influence other developing non-dopaminergic neuronal populations within the arcuate nucleus 23.

REFERENCES 1 Airman, J. and Bayer, S.A., Development of the diencephalon in the rat. III. Ontogeny of the specialized ventricular linings of the hypothalamic third ventricle, J. Comp. Neurol., 182 (1978) 995-1016. 2 Bosler, O. and Beaudet, A., Relations ultrastructurales entre syst~mes monoaminergiques et peptidergiques dans I'hypothalamus. Approche radioautographique et immunocytochimique coupl6e dans le noyau arque et le noyau suprachiasmatique du rat, Ann. Endocrinol. (Paris), 46 (1985) 19-26.

In all age groups, TH-immunoreactive axon terminals were found to establish synaptic contacts with unlabeled dendrites, as in the adult t3'22'29. In the neonate, a few TH-positive axons were also seen in contact with unlabeled perikarya. The fact that these were not observed in 15- or 30-day-old animals may reflect a reorganization of the dopaminergic i~nervation due to progressive dendritic development of non-dopaminergic target neurons. Although axosomatic and axodendritic synapses involving two labeled elements have been described in the adult arcuate nucleus t4, they do not occur in our material at any age, including adulthood 2:a. This discrepancy could be explained by the fact that our antibody may not be sensitive enough to detect the majority of dopaminergic axon terminals. 'Thus, many of the unlabeled axon terminals observed in contact with TH-immunoreactive neuronal elements may in fact be dopaminergic. In summary, the TH-immunoreactive neurons of the arcuate nucleus make extensive contacts with one another very early in postnatal development. These intrinsic connections, which are potentially functional at birth, are identical to those observed in the adult. This raises the possibility that the early maturing TIDA system may influence the development of other arcuate neuronal systems. ACKNOWLEDGEMENTS The authors sincerely thank Mr. John Bertley, Mrs. Jeannie Wong-Mui, and Ms. Patricia Hales for their skillful ~technical assistance. This study was supported by Grants MA-7366 and PG-19, and by a Scientist Award to A.B. from the Medical Research Council of Canada.

3 Bosler, O., Joh, T.H. and Beaudet, A., Ultrastructural re-. lationships between serotonin and dopamine neurons in the rat arcuate nucleus and medial zona "ncerta: a combined radioautographic and immunocytochemical study, Neurosci. Lett., 48 (1984) 279-285. 4 Brawer, J., Bertley, J. and Beaudet, A., Testosterone inhibition of tyrosine hydroxylase expression in the hypothalamic arcuate nucleus, Neurosci. Lett., 67 (1986) 313-318. 5 Chan-Palay, V., Z~iborszky, L., K6hler, C., Goldstein, M. and Palay, S.L., Distribution of tyrosine-hydroxylase-immunoreactive neurons in the hypothalamus of rats, J. Comp. Neurol., 227 (1984) 467-496.

137 6 Daikoku, S., Kawano, H., Okamura, Y., Tokuzen, M. and Nagatsu, I., Ontogenesis of immunoreactive tyrosine hydroxylase-containing neurons in the rat hypothalamus, Dev. Brain Res., 28 (1%6) 85-98. 7 H6kfelt, T., Johansson, O. and Goldstein, M., Central catecholamine neurons are revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Part 1, Classical Transmitters in the CNS, Elsevier, New York, 1984, pp. 217-248. 8 Hyypp/i, M., A histochemical study of the primary catecholamines in the hypothalamic neurons of the rat in relation to the ontogenic and sexual differentiation, Z. Zeilforsch., 98 (1969) 550-56~,. 9 Joh, T.H., Geghman, C. and Reis, D.J., Immunochemical demonstra,,en of increased tyrosine hydroxylase protein in sympathetic ganglia and adrenal medulla elicited by reserpine, Proc. Natl. Acad. Sci. USA, 70 (1973) 2767-2771. 10 Joh, T.H. and Ross, M.E., Preparation of catecholaminesynthesizing enzymes as immunogens for immunoh]stochemistry. In A.C. Cuello (Ed), Immunohistochemistry, IBRO Handbook Series: Methods in the Neurosciences, Vo!. 3, Wiley, Chichester, 1983, pp. 121-138. 11 Kiss, J. and Halfisz, B., Synaptic connections between serotoninergic axon terminals and tyrosine hydroxylase-immunoreactive neurons in the arcuate nucleus of the rat hypothalamus. A combination of electron microscopic autoradiography and immunocytochemistry, Brain Research, 364 (1986) 284-294. 12 Kizer, J.S., Hnmm, S., Nicholson, G., Greeley, G. and Youngblood, W., The effect of castration, thyroidectomy and haloperidoi upon the turnover rates of dopamine and norepinephrine and the kinetic properties of tyrosine hydroxylase in discrete hypothalamic nuclei of the male rat, Brain Research, 146 (1978) 95-107. 13 L~r~inth, C., Sakamoto, H., Maclusky, N.J., Shanabrough, M. and Naftolin, F., Application of avidin-ferritin and peroxidase as contrasting electron-dense markers for simultaneous electron microscopic immunocytochemical labelling of glutamic acid decarboxylase and tyrosine hydroxylase in the rat arcuate nucleus, Histochemistry. 82 (1985) 165-168. 14 L6rfinth, C., Sakamoto, H., Maclusky, N.J., Shanabrough, M. and Naftolin, F., Intrinsic tyrosine hydroxylase (TH) immunoreactive axons synapse with TH immunopositive neurons in the rat arcuate nucleus, Brain Research, 331 (1985) 371-375. 15 Loizou, L.A., The development of monoamine-containing neurones in the brain of the albino rat, J. Anat. (London), 104 (1969) 588. 16 Loizou, L.A., The postnatal development of monoaminecontaining structures in the hypothalamo-hypophyseal system of the albino rat, Z. Zellforsch., 114 (1971) 234-252. 17 Loizou, L.A., The postnatal ontogeny of monoamine-con-

taining neurones in the central nervous system of the albino rat, Brain Research, 40 (1972) 395-418. 18 Ojeda, S.R. and McCann, S.M., Deve!cpment of dopaminergic and estrogenic control of prolactin release in the female rat, Endocrinology, 95 (1974) 1499-1505. 19 Pickel, V.M., Immunocytochemical methods. In L. Heimer and M.J. Robards (Eds.), Neuroanatomical Tract-Tracing Methods, Plenum, New York, 1981, pp. 483-509. 20 Pickel, V.M., Joh. T.H., Field, P.M., Becker, C.G. and Reis, D.J., Cellular localization of tyrosine hydr6xylase by immunohistochemistry, J. Histochem. Cytochem., 23 (1975) 1-12. 21 Pickel, V.M., Joh, T.H. and Reis, D.J., Monoamine-synthesizing enzymes in central dopaminergic, noradrenergic and serotonergic neurons, lmmunocytochemical localization by light and electron microscopy, J. Histochem. Cytochem., 24 (1976) 792-806. 22 Pickel, V.M., Joh, T.H. and Reis, D.J., Ultrastructural localization of tyrosine hydroxylase in noradrenergic neurons of brain, Proc. NatL Acad. Sci. USA, 72 (1975) 659-663. 23 Piotte~ M., Beaudet, A., .~oh, T.H. and Brawer, J.R,, The fine structural organization of tyrosine hydroxylase immunoreactive neurons in rat arcuate nucleus, J. Comp. Neurol., 239 (1985) 44-53. 24 Rakic, P., Neuron-gila relationship during granule cell migration in developing cerebellar cortex. A Golgi and electron microscopic study in Macacus rhesus, l. Comp. Neurol., 141 (1971) 283-312. 25 Seiger, A. and OIson, L., Late prenatal ontogeny of central monoamine neurons in the rat: fluorescence histochemical observations, Z. Anat. Entwickl.-Gesch., 140 (1973) 281-318. 26 Specht, L.A., Pickel, V.M., Joh, T.H. and Reis, D.J., Light-microscopic immunocytochemical localization of tyrosine hydroxylase in prenatal rat brain. I. Early ontogeny, J. Comp. Neurol., 199 (1981) 233-253. 27 Specht, L.A., Pickel, V.M., Joh, T.H. and Reis, D.J., Light-microscopic immunocytochemical localization of tyrosine hydroxylase in prenatal rat brain. II. Late ontogeny, J. Comp. Neurol., 199 (1981) 255-276. 28 Sternberger, L.A., The unlabeled antibody peroxidase-antiperoxidase (PAP) method. In S. Cohen and T. McCluskey lmmunocvtochemistrv, Wiley, New York, 1979, pp. 104-169. 29 Thind, K.K. and Goldsmith, P.C., Ultrastructural analysis of synapses involving tyrosine hydroxylase-containing neurons in the ventral periventricular hypothalamus of the macaque, Brain Research, 366 (1986) 37-52. 30 Van den Pol, A.N., Herbst, R.S. and PowelL J.F., Tyrosine hydroxylase-immunoreactiw-~ neurons of ihe hypothalamus: a light and electron microscopic study, Neuroscience, 13 (1984) 1117-1156.