- Email: [email protected]
Differential ontogeny of three putative catecholamine cell types in the postnatal rat retina
DevelopmentaIBrain Research, 22 (1985) 187-196 Elsevier
187
BRD 50270
Differential Ontogeny of Three Putative Catecholamine Cell Types in the Postnatal Rat Retina GEORGE A. FOSTER l, MARIANNE SCHULTZBERG 1, MENEK GOLDSTEIN2 and TOMAS HOKFELT 1 1Department of Histology, Karolinska Institute, Stockholm (Sweden) and 2Department of Psychiatry, New York University Medical Center, New York, NY (U.S.A.) (Accepted March 12th, 1985) Key words: retina - - postnatal ontogeny - - phenylethanolamine N-methyltransferase - tyrosine hydroxylase --immunofluorescence
The development of tyrosine hydroxylase (TH) and phenylethanolamine N-methyltransferase (PNMT) in the rat retina was investigated using the indirect immunofluorescence technique. Two types of TH-positive cells were found. The first appeared at postnatal day 2, in the vitreal half of the inner nuclear layer (INL). Single fibres from these neurones, bifurcating in and innervating layer 3 of the inner plexiform layer (IPL) were seen at day 4. This first type of TH-positive cell was most numerous at day 15, but thereafter disappeared before adulthood. At day 5, a more intensely staining TH-immunoreactive neurone became visible, occupying a more proximal part of the INL, and projecting multiple fibres to layer 1 of the IPL. In contrast, PNMT-positive cells, in the vitreal half of the INL and in the ganglion cell layer (GCL), sending single varicose axons to layer 3 of the IPL, were first apparent only at day 10, achieving a disposition similar to that of the adult by days 15-16 postnatal. Analysis of adjacent sections stained with antibodies to TH and PNMT revealed that neither type of TH-positive neurone also contained PNMT-Iike immunoreactivity. It is concluded that although both of the rate-limiting enzymes of the catecholamine synthetic pathway are present in the developing rat retina, they occur in 3 mutually exclusive populations of neurones. INTRODUCTION Various aspects of the postnatal development of retinal dopamine neurones have been analyzed in several studies3,16,20,24,28,29. Using immunocytochemical techniques, the ontogeny of tyrosine hydroxylase (TH), the enzyme converting tyrosine to L-dihydroxyphenylalanine ( L - D O P A ) in the postnatal rat retina has recently been reported 26. Two sorts of TH-positive cells are visible in the maturing retina. The first type is seen initially at postnatal days 3-426 , but declines after days 15-21, such that in the adult retina it is found only extremely rarely2,6,7,14,21. These cells are weakly immunoreactive to T H antisera, are pearshaped, and inhabit the vitreal part of the inner nuclear layer (INL) (nomenclature according to ref. 30). Processes from the cells bifurcate and ramify in predominantly layer 3 of the inner plexiform layer (IPL). The second TH-positive cell type is larger and more strongly immunoreactive. These cells are found
in the innermost layer of the INL, while their processes ramify mainly in layer 1 of the IPL. In contrast to the weakly staining TH-positive retinal neurones, the latter cells, which are thought to be dopaminergic2,5,14,21, are first apparent only at postnatal days 6-7. They increase in number until day 20, but exhibit no subsequent decline. Phenylethanolamine N-methyltransferase (PNMT) is the enzyme which converts noradrenaline to adrenaline. (PNMT)-immunoreactive cells and fibres with a morphology and disposition similar to that of the weakly-staining neonatal TH-positive neurones, have been described in the adult retinaT, 13. Of particular interest was the observation that the PNMT-positive neurones failed to display detectable amounts of TH-like immunoreactivity (TH-LI) 7, prompting the hypothesis that these were indeed the same as the neonatal pear-shaped cell type, but that their production of T H was curtailed during maturation of the retina. Accordingly, in an effort to discover if the two
Correspondence: G.A. Foster, Department of Histology, Karolinska Institute, P.O. Box 60400, S-104 01 Stockholm, Sweden. 0165-3806/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
188 substances were ever co-localized, the ontogeny of PNMT- and TH-LI in the rat retina was studied at a time when it was known that the weakly staining, pear-shaped TH-positive neurones were present, namely during neonatal life.
conjunction with objective lenses of 2.5 or 4 x magnification was used 23. Afterwards, some sections were stained with cresyl violet to confirm the anatomical landmarks of the retina, RESULTS
MATERIALS AND METHODS
Postnatal days 2-9 Antibodies to bovine and rat adrenal PNMT TM were raised in rabbits10,11, and their specificity tested by immunoelectrophoresis. T H was isolated from a rat phaeochromocytoma cell line, and the antibodies raised in rabbits, as previously described2L The specificity of the TH-antiserum has also been validated 22. In the present experiments, normal rabbit serum was used in place of the primary antibody for the purpose of a control. Pregnant Sprague-Dawley rats (Anticimex, Stockholm, Sweden) at various stages of gestation were anaesthetized, the embryos removed, and the retinas dissected under the surface of an ice-cold solution of 4% paraformaldehyde containing 0.2% pictic acid 33. Rats at various postnatal stages were sacrificed, and the retinas dissected in the same way. After immersion in the same fixative for 2 h at 4 °C, the retinas were rinsed for at least 24 h in several changes of 10% sucrose in 0.1 M phosphate buffer (pH 7.4) at 4 °C. Serial cryostat sections at 14 ktm thickness were cut through the whole retina. The sections were then processed according to the indirect immunofluorescence technique 4. In brief, sections were rinsed for 15 min in 0.01 M phosphate buffer (pH 7.4) containing 0.9% sodium chloride (PBS), and then incubated in a humid atmosphere with PNMT or TH antiserum (both diluted 1:400) at 4 °C for 24 h. After rinsing in PBS at room temperature, fluorescein isothiocyanate (FITC) conjugated to swine anti-rabbit antibodies (Dakopatts, Copenhagen, Denmark) (diluted 1:10), was applied to the sections at 37 °C for 30 min. Finally, after rinsing in PBS, the sections were mounted in a mixture of glycerol and PBS (3:1) containing 0.2% p-phenylenediamine ~9. The sections were examined in a Zeiss fluorescence microscope, fitted with an oil dark-field condenser, a KP500 excitation filter and an LP520 stop filter. Scopix RP1 black and white film (Gevaert, Belgium) was used for photography. For some larger micrographs, a motor driven scanning condenser in
No TH- or PNMT-LI was discovered in the rat retina between day 16 of gestation and postnatal day 1. On day 2, however, TH-immunoreactive cells were observed for the first time (Fig. 1A). They were 7-8 /zm in diameter, pear-shaped, and occupied a position in the medial part of the primordium of the INL (Fig. 1B). At this stage, it was also possible to find very short TH-positive processes emanating from these cells, but no varicosities were apparent along their length. The outer nuclear layer (ONL) and the photoreceptor layer (PRL) became distinguishable at day 4, whereas it was not until 5 days later that the outer plexiform layer (OPL) could be delineated. As the age of the animal increased, progressively more neurones exhibited TH-LI. By days 4 and 5, for example, many more of the pear-shaped cells were visible, occupying a position in the middle of the INL, and sending out single processes which traversed orthogonally to the IPL where they then bifurcated (Fig. 1C, D, F - H ) . A very thin band of fibres, presumably from these same pear-shaped cells, was apparent in the middle layer of the IPL. Also at day 5, a second, more strongly fluorescent type of TH-positire neurone became visible (Fig. 1C, D). The perikarya of these cells were similar in diameter to the pear-shaped cells, and also contained a large, welldefined nucleus, but they were found only in the most proximal layers of the INL. By day 7 their size had increased to about 12-15 ~tm in diameter. No long processes from these strongly TH-immunoreactive neurones were noted between days 5 and 7 postnatal, although some short axon fibres could occasionally be seen at the INL/IPL border (Fig. 1C, D). In layer 3 of the IPL a dense band of varicose axons, thought to arise from the weakly TH-positive neurones of the medial INL, was now prominent, but no fibres were found elsewhere. On day 8 postnatal, a few varicose fibres in layer 1 of the IPL appeared, and were considered to be projections of the strongly TH-immunoreactive cells. The extent of these fibres increased
189
Fig. 1. Immunofluorescence photomicrographs of the ontogeny of TH-LI in the rat retina. A: the first TH-positive cells, at day 2 postnatal, situated in the middle of the primordial INL. B: a higher magnification of the same cell. C: at postnatal day 5, many more qq-Ipositive cells are visible, and they can be divided into two g r o u p s - either weakly (short arrows) or strongly staining (long arrows). D: higher power micrograph of C. E: a strongly staining TH-immunoreactive cell at day 9 postnatal, showing a long varicose axon travelling along the border of the IPL and INL. The weakly staining cells, in contrast, send out a sclerally oriented process, which bifurcates and ramifies in the middle layer of the IPL (F, higher magnification seen in G and H, all at day 5 postnatal). The numbers in G and H refer to the layers of the IPL. Other abbreviations are as in the text. Scale bar indicates 50pm for A, C, F and 20pro for B, D, E, G and H. considerably on day 9, such that individual beaded axons could then be traced in the section for several hundred microns (Fig. 1E). No PNMT-LI was found in the retina between day 16 of gestation and day 9 postnatal.
Postnatal days 10-15 At day 10 postnatal, TH-positive fibres in layer 3 of the IPL were by now very dense, and strongly immunoreactive. The fibres of the large TH-positive neurones of the INL branched m a n y times (Fig. 2A),
190
Fig. 2. Immunofluorescence photomicrographs of TH-LI in the developing rat retina. A: a section taken tangentially to the retinal layering, at day 10 postnatal and shows several strongly staining TH-positive cells with multiply dividing varicose fibres. B and C: sections taken approximately radially to the retinal layering at 13 days postnatal, showing strongly staining TH-immunoreactive cells which appear to innervate both layers 1 and 3 of the IPL. Abbreviations are as in the text. Scale bar indicates 50 um in A, B and 20/tm in C. and it was thought that some cells might send processes to more than one layer of the IPL. Two examples are shown in Fig. 2B, C, at day 13, the age when this was first apparent. Neurones containing PNMT-LI were observed for the first time in the retina at day 10 after birth (Fig. 3A). The immunoreactive somata were about 8/~m in diameter, and were found in the proximal half of the INL. Emerging from the cell body could be seen a single vitreally directed process (Fig. 3A, C) which bifurcated in layer 3 of the 1PL (Fig. 3 C - E ) , and which then formed a network of fibres. At later postnatal stages, the intensity of the immunostaining of the cell bodies and fibres increased. In addition, the meshwork of PNMT-positive fibres became denser over this period, resembling the adult situation by day 14 (Fig. 5A, B). Somata immunoreactive to
PNMT antisera were also visible in the ganglion cell layer (GCL) at day 10 (Fig. 3A, C, F). These cell bodies were similar in shape and size to those in the INL, and were considered, as in the adult 7, to be displaced amacrine cells. Single processes from these cells were visible (Figs. 3F, 4A), coursing obliquely to and ramifying in the medial layer of the 1PL (Fig. 4A). At day 13, a sparse plexus of fibres originating from the PNMT-immunoreactive cells in the INL became apparent in layer 1 of the IPL (Fig. 4B). These fibres grew progressively denser at days 14 and 15, when a situation comparable to that in the adult was found. Sections adjacent to those used for PNMT immunocytochemistry were stained for TH-LI. Areas of the retina that contained the PNMT-immunoreactive cells of the medial INL occasionally also exhibited
191
I
Fig. 3. Immunofluorescence photomicrographs of PNMT~ (A, C-F) and TH-LI (B) in the rat retina at day 10 postnatal. Although many PNMT-positive neurones are visible in the vitreal half of the INL and in the GCL (A), an adjacent section (B) stained for TH showed that no TH-positive cells were detectable in this region of the retina. C: higher magnification of the PNMT-containing neurones, in the INL (arrows) and the GCL (arrowheads), and the plexus of fibres in layer 3 of the IPL. Although PNMT-immunoreactive cells exhibited short processes (F), it was only those PNMT-positive neurones of the INL which displayed bifurcating axons in the IPL (D, E) at this age. Abbreviations are as in the text. Scale bar indicates 65 ~m in A, B and 20/~m in C-F.
the more proximally located large, strongly-staining, TH-positive cells of the first layer of the INL (cf. Fig. 5A and B). However, the weakly T H - i m m u n o r e a c tive perikarya found in the same layer of the INL as the PNMT-positive cell bodies were never found in
the same gross regions of the retina (cf. Figs. 5 A - D and 6 A - D ) .
Postnatal days 16-20 By day 16, the disposition of P N M T - L I was very
192
Fig. 4. Immunofluorescence photomicrographs of PNMT-LI in the retina of the 13-day-old rat. A: processes from the PNMT-positive displaced amacrine cells in the GCL (arrows) reach as far as layer 3 of the IPL, where they contribute to a, by now, dense fibre plexus. B: PNMT-immunoreactive neurone in the INL which displays a bifurcating axon in layer 1 of the IPL, where a sparse network of fibres is also apparent. The density of fibres in layer 3 of the IPL is similar to that of the adult retina. Abbreviations are as in the text. Scale bar indicates 20 ~m.
similar to that f o u n d in the m a t u r e retina. M a n y positive n e u r o n e s in the I N L and the G C L w e r e apparent, and a d e n s e b a n d of fibres in layer 3, with a sparser n e t w o r k of fibres in layer 1 of the I P L c o u l d also be seen. A t days 16 and 18 the T H - L I also bec a m e m o r e closely aligned to that of the adult: the weaker-staining
cells b e c a m e
p r o g r e s s i v e l y rarer,
while the m e s h w o r k of fibres in layer 3 of the I P L decreased (Fig. 7). In contrast, the strongly-staining T H - i m m u n o r e a c t i v e n e u r o n e s i n c r e a s e d in n u m b e r , and their p r o c e s s e s in layer 1 of the I P L w e r e similarly e n h a n c e d . A t no t i m e d u r i n g the p e r i o d s t u d i e d did any of the P N M T - p o s i t i v e n e u r o n e s also display detectable levels of T H - L I . A s c h e m a t i c s u m m a r y of
Fig. 5. Immunofluorescence photomicrographs of adjacent sections (.A, B and C, D) stained with PNMT (A, C) or TH (B, D) antisera at day 14 postnatal. Many PNMT-positive cells are visible in A, but only a single large TH-immunoreactive neurone is apparent in the adjacent section (B). Similarly, although many large, strongly staining and small, weakly staining TH-positive neurones exist in D, no cells containing PNMT-LI are manifest in the adjacent section. Note also the virtual absence of TH-LI from layer 3 of the IPL near the PNMT cells, and of PNMT-LI from layer 3 near the weakly staining TH-positive neurones. Abbreviations are as in the text. Scale bar indicates 50 ~m. The magnification was the same for all micrographs.
193
~ii ~
Fig. 6. Low-power immunofluorescence photomicrographs of adjacent sections (A, B and C, D) of 14-day-old rat retina, stained for PNMT- (A, C) and TH-LI (B, D). The micrographs show sections at two levels in the same retina. A long single-headed arrow in D indicates a weakly staining TH-neurone. In the regions where these cells are seen (between the short arrows), no PNMT-containing neurones are demonstrable. However, the large, strongly staining TH-positive cells (long double-headed arrows in D) of the first layer of the IPL, occupy the same zones of the retina as both the more distally stratified PNMT-immunoreactive neurones and the weaker staining TH-neurones. C, choroid; CP, ciliary process. Scale bar indicates 250,um. Magnification was the same for all micrographs.
194 PNMT
TH
2d C)=~L OPI
.
INL
.
.
.
.
.
.
.
.
.
.
.
_ Y
.
.
tPt
GCL .
.
.
.
.
.
.
PRt ONL OPt .
IPL Get .
.
.
.
.
.
.
• •
.
.
• -
.
•, .
.
.
.
.
." . . . . . . . . . .
.
.
.
.
.
.
.
.
• ' .
.
.
.
.
.
.
lOd ONL OPk*
Z
--
-"
,,L2. . . . . .
:
--
-"
2
....... i...... :
fiCL
15d ONL 0PL:
:
:
:
"-
"-
"-
:
=
:
:
:
:
:
:
:
Adult ONL OPL[
[
[
--
_"
_ . . . _ . . . : . , . . , . ,
GCL .
.
.
.
.
_-
~.
,.~., .
.
"_
Z
~.
Z
-"
Z
Z
Z
Z
,_..~.., .
.
.
.
.
.
.
.
.
Fig. 7. Schematic diagram of the ontogeny of TH- and PNMTLI neurones of the postnatal rat retina. The fibres are symbolized by dots in the IPL. No attempt was made to express different levels of immunoreactivity. Abbreviations are as in the text. Scale bar indicates 50 #m.
the ontogeny of PNMT- and T H - L I in the rat retina is shown in Fig. 7. DISCUSSION
The development of TH- and PNMT-LI in the maturing rat retina has been described. Weak TH-immunoreactive cells in the second layer of the INL were first demonstrable at postnatal day 2, whereas the stronger-staining neurones were not apparent until day 5. This is rather earlier than that found by other workers2S, and may result from the use of pictic acid in our fixation medium; TH-LI was undetectable in the small pear-shaped cells at day 2 in retinas
which had been fixed in 4% paraformaldehyde alone. The ontogeny of the stronger-staining THpositive cells and of the PNMT-immunoreactive cells was, however, the same with both types of fixative. At later postnatal stages the weakly-fluorescent neurones began to disappear, such that by day 20 they were extremely rare. Fibres in layer 3 of the IPL were visible at day 5, were of maximal density at day 15, but had diminished greatly in number by day 20. Conversely, the TH-positive processes in layer 1 of the IPL, which were sparse at day 8, had increased to reach mature levels by day 20. PNMT-Iike immunoreactivity was first found in neurones of the developing rat retina at postnatal day 10. These cells were located in the proximal half of the INL, and sent out processes which bifurcated, and then ramified in layer 3 of the IPL. By day 14 the density of the fibres in the medial layer closely resembled that found in the adult, whereas fibres in layer 1 were maximal at about day 15. Despite the known dependence of adrenaline concentrations 12 and of TH activity on photic stimulation 17, neither the number nor the intensity of staining of the TH- and PNMT-immunoreactive neurones exhibited any apparent fluctuations at the time of opening of the eyes, namely 16 days. It could be that the levels of the enzymes are not changed during photic stimulation, but, instead, that the degree of their activation by cofactors is responsible for the observed biochemical increases. Alternatively, an inherent diurnal rhythm may control PNMT and TH induction prior to eyeopening, which is subsequently replaced or combined with photic control of enzyme levels. Experiments are in progress in the adult retina to determine if either of these hypotheses is correct. It is apparent that in contrast to the TH-positive neurones, those containing PNMT-LI are morphologically well-developed before the detection of the antigen. Even on day 10, when the cell bodies were first visible, large numbers of PNMT-immunoreactive fibres were also seen. This is reminiscent of the situation in the developing rat medulla oblongata, where at day 13 of gestation PNMT-positive cell bodies and their extensive ascending and descending projections become apparent almost simultaneouslyS. Given that neurite extension in vitro a and in vivo 32 is about 40/tm/h, it would seem that the morphological development of the PNMT-containing
195 neurones precedes development of their transmitter synthetic enzyme. In contrast, T H - L I appears in processes of the neurones of the I N L over several days, suggesting that its induction occurs prior to the morphological maturation of the neurone. Of particular interest was the discovery that none of the PNMT-immunoreactive cell bodies contained detectable amounts of TH-LI. Although the possibility cannot be excluded that the levels of T H - L I may have been below the limits of sensitivity of the immunocytochemical technique, our experience is that the use of the TH antibody represents the most potent method of visualizing catecholaminergic neurones. An inability to detect T H - L I would therefore suggest that less than significant amounts are present. The observation of PNMT-positive, TH-negative neurones in the retina was all the more surprising in view of the similar stratification within the I N L of the PNMT-positive cells and the weakly staining TH-immunoreactive somata during early postnatal life. However, the two types of neurones occupied almost mutually exclusive zones of the retina. It would ap2 pear, therefore, that despite their commonalities in terms of morphology and layering, the two types of cell are not the same. Although the weak TH-positive cells either die or lose their ability to synthesize TH during development, those containing P N M T - L I alone continue to be manifest, even in the adult retina 7. Although the normal control studies were carried Out 1°,11.22, the possibility that our P N M T and T H antisera are cross-reacting with other substances cannot be wholly eliminated. However, in view of the similar results obtained with completely different antibodies this hypothesis seems untenable. For a fuller discussion see ref. 15. It is not clear what function the PNMT-positive, TH-negative cells in the retina possess. Neurones with PNMT-LI, but without detectable T H - L I have
REFERENCES 1 Bray, D., Branching patterns of individual sympathetic neurones in culture, J. Cell Biol., 56 (1973) 702-712. 2 Brecha, N., Retinal neurotransminers: histochemical and biochemical studies. In P.C. Emson (Ed.), Chemical Neuroanatomy, Raven Press, New York, 1983, pp. 85-129. 3 Cohen, J. and Neff, N.H., Retinal amacrine cell system tyrosine hydroxylase: the development of responsiveness to
also been described in the adult rat dorsal hypothalamus 3~ and posterior arcuate nucleus 7. It seems unlikely that nerve cells which are unable to mediate neurotransmission because of the absence of their endogenous transmitter would maintain projections or survive in the adult central nervous system. Three obvious explanations exist: the first is that an alternative synthetic pathway for adrenaline synthesis, not via TH, is in operation. In support of this, it is known that the retina does indeed contain adrenaline12, 27, and that this is unlikely to be derived from the circulation. Second, the P N M T may be catalyzing the methylation of some substance other than noradrenaline, perhaps unrelated to a transmitter role. This implies that adrenaline is not the final reaction product. Lastly, these neurones may utilize an alternative transmitter which is enabling them to function normally; the PNMT-LI may then be merely the result of transcription of a vestigial gene. A third putative catecholaminergic cell type has been recently described in the adult retina 9. This is also found in the proximal zone of the INL, but is only rendered visible in Falck-Hillarp histofluorescence by loading with exogenous noradrenaline, but not with its precursors dopamine or L - D O P A . These cells were therefore considered to possess the capacity for noradrenaline uptake, but not for its synthesis. It remains to be determined whether they are related to either the weakly-staining TH-cells seen during neonatal life, or to the PNMT-positive, TH-negative cells found both in the neonate and the mature animal 7. ACKNOWLEDGEMENTS This work was supported by grants from the Swedish Medical Research Council (04X-2887) and Knut och Alice Wallenbergs Stiftelse. G . A . F . is a N A T O Overseas Research Fellow.
light and neuroleptic drugs, Dev. Brain Res., 255 (1982) 1960-1963. 4 Coons, A.H., Fluorescent antibody methods. In J.F. Danielli (Ed.), General Cytochemical Methods, Academic Press, New York, 1958, pp. 399-422. 5 Dowling, J.E. and Ehinger, B., Synaptic organization of the dopaminergic neurons in the rabbit retina, J. Comp. Neurol., 180 (1978) 203-220. 6 Ehinger, B., Adrenergie nerves to the eye and related
196
7
8
9
10
11
12
13
14
15
16
17
18
structures in man and in the cynomolgus monkey (Macaca irus), Invest. Ophthalmol., 5 (1966) 42-52. Foster, G.A., H6kfelt, T., Coyle, J.T. and Goldstein, M., lmmunohistochemicat evidence for phenylethanolamine N-methyltransferase-positive, tyrosine hydroxylase-negative neurones in the retina and the posterior hypothalamus of the rat, Brain Res., in press. Foster, G.A., Schultzberg, M., Goldstein, M. and H6kfelt, T., Ontogeny of phenylethanolamine N-methyltransferaseand tyrosine hydroxylase-like immunoreactivity in presumptive adrenaline neurones of the foetal rat central nervous system, J. Comp. Neurol., in press. Fukuda, M., Ishimoto, I., Kuwayama, Y., Shimizu, Y., Shiosaka, S., Takagi, H. and Tohyama, M., Monoamine accumulating neuron system in the rat retina with special reference to noradrenaline accumulating neurons, Exp. Eye Res., 34 (1982) 487-491. Goldstein, M., Fuxe, K. and H6kfelt, T., Characterization and tissue localization of catecholamine synthesizing enzymes, Pharrnacol. Rev., 24 (1972) 293-309. Goldstein, M., Lew, J.Y., Matsumoto, Y., H6kfelt, T. and Fuxe, K., Localization and function of PNMT in the central nervous system. In M,A. Lipton, A. DiMascio and K.F. Killam (Eds.), Psychopharmacology, Raven Press, New York, 1978, pp. 261-269. Hadjiconstantinou, M,, Cohen, J. and Neff, N.H., Epinephrine: a potential neurotransmitter in retina, J. Neurochem., 41 (1983) 1440-1444. Hadjiconstantinou, M,, Mariani, A.P., Panula, P., Joh, T.H. and Neff, N.H., Immunohistochemical evidence for epinephrine-containing retinal amacrine cells, Neuroscience, 13 (1984) 547-551. Hfiggendal, J. and Malmfors, T., Identification and cellular localization of catecholamines in the retina and choroid of the rabbit, Acta Physiol. Scand., 64 (1965) 58-66. H6kfelt, T., Johansson, O. and Goldstein, M., Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS, Part 1, Elsevier, Amsterdam, 1984, pp. 157-276. Kato, S., Nakamara, T. and Negishi, K., Postnatal development of dopaminergic cells in the rat retina, J. Comp. Neurol., 191 (1980) 227-236. Iuvone, P.M., Galli, C.L., Garrison-Gund, C.K. and Neff, N.H., Light stimulates tyrosine hydroxylase activity and dopamine synthesis in retinal amacrine cells, Science, 202 (1978) 901-902. Joh, T.H. and Goldstein, M., Isolation and characterization of multiple forms of phenylethanolame N-methyltransferase, Mol. Pharmacol., 9 (1973) 117-129.
19 Johnson, D.G. and De C. Nogueira Araujo, G.M.. A sin> ple method of reducing the fading of immunofluorescencc during microscopy, J. lmmunol. Meth., 43 (1981) 349. 20 Lain, D.M.K., Fung, S.C. and Kong, Y.C., Postnatal development of dopaminergic neurons in the rabbit retina, J. Neurosci.. 1 (1981) 1117-1132. 21 Malmfors, T., Evidence of adrenergic neurons with synaptic terminals in the retinas of rats demonstrated with fluorescence and electron microscopy, Acta Physiol. Scand., 58 (1963) 99-100. 22 Markey, K.A., Kondo, S., Shenkman, i. and Goldstein, M., Purification and characterization of tyrosine hydroxylase from a clonal phaeochromocytoma cell line, Mol. Pharmacol., 17 (1980) 79-85. 23 M~rtensson, R. and BjOrklund, A., Low power photography in the fluorescence microscope using an anatomical dark-field condenser scanner. In A. Bj6rklund and T, H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS, Part 1. Elsevier, Amsterdam, 1984. 24 Morgan, W.W. and Kamp, C.W., Postnatal development of the light response of the dopaminergic neurons in the rat retina, J. Neurochem., 39 (1982) 283-285. 25 Nguyen-Legros, J., Berger, B., Vigny, A. and Alvarez, C.. Tyrosine-hydroxylase-like immunoreactive interplexiform cells in the rat retina, Neurosci. Lett., 27 (1981) 255-259. 26 Nguyen-Legros, J., Vigny, A. and Gay, M., Post-natal development of TH-like immunoreactivity in the rat retina, Exp. EyeRes., 37 (1983) 23-32. 27 Osborne, N.N. and Nesselhut, T., Adrenaline: occurrence in the bovine retina, Neurosci. Lett., 39 (1983) 33-36. 28 Parkinson, D. and Rando, R.R., Ontogenesis of dopaminergic neurons in the post-natal rabbit retina: pre- and post-synaptic elements, Dev. Brain Res., 13 (1984) 207-217. 29 Patten, M., Parkinson, D., Wyse, P. and Spira, A., Ontogenesis of dopaminergic neurons in the fetal and postnatal guinea pig retina, Invest. Ophthalmol. Vis. Sci., Suppl., 22 (1982) 114. 30 RamOn y Cajal, S., La rdtine des vert6bres, La Cellule, 9 (1893) 17-257. 31 Ross, C.A., Ruggiero, D.A., Meeley, M.P., Park, D.H., Job, T.H. and Reis, D.M., A new group of neurons in hypothalamus containing phenylethanolamine N-methyltransferase (PNMT) but not tyrosine hydroxylase, Brain Res., 306 (1984) 349-353. 32 Speidel, C.C., Adjustments of nerve endings, Harvey Lect., 36 (1941) 126-158. 33 Zamboni, L. and De Martino, C., Buffered picric-acid formaldehyde: a new rapid fixative for electron-microscopy, J. Cell Biol., 35 (1967) 148A.
187
BRD 50270
Differential Ontogeny of Three Putative Catecholamine Cell Types in the Postnatal Rat Retina GEORGE A. FOSTER l, MARIANNE SCHULTZBERG 1, MENEK GOLDSTEIN2 and TOMAS HOKFELT 1 1Department of Histology, Karolinska Institute, Stockholm (Sweden) and 2Department of Psychiatry, New York University Medical Center, New York, NY (U.S.A.) (Accepted March 12th, 1985) Key words: retina - - postnatal ontogeny - - phenylethanolamine N-methyltransferase - tyrosine hydroxylase --immunofluorescence
The development of tyrosine hydroxylase (TH) and phenylethanolamine N-methyltransferase (PNMT) in the rat retina was investigated using the indirect immunofluorescence technique. Two types of TH-positive cells were found. The first appeared at postnatal day 2, in the vitreal half of the inner nuclear layer (INL). Single fibres from these neurones, bifurcating in and innervating layer 3 of the inner plexiform layer (IPL) were seen at day 4. This first type of TH-positive cell was most numerous at day 15, but thereafter disappeared before adulthood. At day 5, a more intensely staining TH-immunoreactive neurone became visible, occupying a more proximal part of the INL, and projecting multiple fibres to layer 1 of the IPL. In contrast, PNMT-positive cells, in the vitreal half of the INL and in the ganglion cell layer (GCL), sending single varicose axons to layer 3 of the IPL, were first apparent only at day 10, achieving a disposition similar to that of the adult by days 15-16 postnatal. Analysis of adjacent sections stained with antibodies to TH and PNMT revealed that neither type of TH-positive neurone also contained PNMT-Iike immunoreactivity. It is concluded that although both of the rate-limiting enzymes of the catecholamine synthetic pathway are present in the developing rat retina, they occur in 3 mutually exclusive populations of neurones. INTRODUCTION Various aspects of the postnatal development of retinal dopamine neurones have been analyzed in several studies3,16,20,24,28,29. Using immunocytochemical techniques, the ontogeny of tyrosine hydroxylase (TH), the enzyme converting tyrosine to L-dihydroxyphenylalanine ( L - D O P A ) in the postnatal rat retina has recently been reported 26. Two sorts of TH-positive cells are visible in the maturing retina. The first type is seen initially at postnatal days 3-426 , but declines after days 15-21, such that in the adult retina it is found only extremely rarely2,6,7,14,21. These cells are weakly immunoreactive to T H antisera, are pearshaped, and inhabit the vitreal part of the inner nuclear layer (INL) (nomenclature according to ref. 30). Processes from the cells bifurcate and ramify in predominantly layer 3 of the inner plexiform layer (IPL). The second TH-positive cell type is larger and more strongly immunoreactive. These cells are found
in the innermost layer of the INL, while their processes ramify mainly in layer 1 of the IPL. In contrast to the weakly staining TH-positive retinal neurones, the latter cells, which are thought to be dopaminergic2,5,14,21, are first apparent only at postnatal days 6-7. They increase in number until day 20, but exhibit no subsequent decline. Phenylethanolamine N-methyltransferase (PNMT) is the enzyme which converts noradrenaline to adrenaline. (PNMT)-immunoreactive cells and fibres with a morphology and disposition similar to that of the weakly-staining neonatal TH-positive neurones, have been described in the adult retinaT, 13. Of particular interest was the observation that the PNMT-positive neurones failed to display detectable amounts of TH-like immunoreactivity (TH-LI) 7, prompting the hypothesis that these were indeed the same as the neonatal pear-shaped cell type, but that their production of T H was curtailed during maturation of the retina. Accordingly, in an effort to discover if the two
Correspondence: G.A. Foster, Department of Histology, Karolinska Institute, P.O. Box 60400, S-104 01 Stockholm, Sweden. 0165-3806/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
188 substances were ever co-localized, the ontogeny of PNMT- and TH-LI in the rat retina was studied at a time when it was known that the weakly staining, pear-shaped TH-positive neurones were present, namely during neonatal life.
conjunction with objective lenses of 2.5 or 4 x magnification was used 23. Afterwards, some sections were stained with cresyl violet to confirm the anatomical landmarks of the retina, RESULTS
MATERIALS AND METHODS
Postnatal days 2-9 Antibodies to bovine and rat adrenal PNMT TM were raised in rabbits10,11, and their specificity tested by immunoelectrophoresis. T H was isolated from a rat phaeochromocytoma cell line, and the antibodies raised in rabbits, as previously described2L The specificity of the TH-antiserum has also been validated 22. In the present experiments, normal rabbit serum was used in place of the primary antibody for the purpose of a control. Pregnant Sprague-Dawley rats (Anticimex, Stockholm, Sweden) at various stages of gestation were anaesthetized, the embryos removed, and the retinas dissected under the surface of an ice-cold solution of 4% paraformaldehyde containing 0.2% pictic acid 33. Rats at various postnatal stages were sacrificed, and the retinas dissected in the same way. After immersion in the same fixative for 2 h at 4 °C, the retinas were rinsed for at least 24 h in several changes of 10% sucrose in 0.1 M phosphate buffer (pH 7.4) at 4 °C. Serial cryostat sections at 14 ktm thickness were cut through the whole retina. The sections were then processed according to the indirect immunofluorescence technique 4. In brief, sections were rinsed for 15 min in 0.01 M phosphate buffer (pH 7.4) containing 0.9% sodium chloride (PBS), and then incubated in a humid atmosphere with PNMT or TH antiserum (both diluted 1:400) at 4 °C for 24 h. After rinsing in PBS at room temperature, fluorescein isothiocyanate (FITC) conjugated to swine anti-rabbit antibodies (Dakopatts, Copenhagen, Denmark) (diluted 1:10), was applied to the sections at 37 °C for 30 min. Finally, after rinsing in PBS, the sections were mounted in a mixture of glycerol and PBS (3:1) containing 0.2% p-phenylenediamine ~9. The sections were examined in a Zeiss fluorescence microscope, fitted with an oil dark-field condenser, a KP500 excitation filter and an LP520 stop filter. Scopix RP1 black and white film (Gevaert, Belgium) was used for photography. For some larger micrographs, a motor driven scanning condenser in
No TH- or PNMT-LI was discovered in the rat retina between day 16 of gestation and postnatal day 1. On day 2, however, TH-immunoreactive cells were observed for the first time (Fig. 1A). They were 7-8 /zm in diameter, pear-shaped, and occupied a position in the medial part of the primordium of the INL (Fig. 1B). At this stage, it was also possible to find very short TH-positive processes emanating from these cells, but no varicosities were apparent along their length. The outer nuclear layer (ONL) and the photoreceptor layer (PRL) became distinguishable at day 4, whereas it was not until 5 days later that the outer plexiform layer (OPL) could be delineated. As the age of the animal increased, progressively more neurones exhibited TH-LI. By days 4 and 5, for example, many more of the pear-shaped cells were visible, occupying a position in the middle of the INL, and sending out single processes which traversed orthogonally to the IPL where they then bifurcated (Fig. 1C, D, F - H ) . A very thin band of fibres, presumably from these same pear-shaped cells, was apparent in the middle layer of the IPL. Also at day 5, a second, more strongly fluorescent type of TH-positire neurone became visible (Fig. 1C, D). The perikarya of these cells were similar in diameter to the pear-shaped cells, and also contained a large, welldefined nucleus, but they were found only in the most proximal layers of the INL. By day 7 their size had increased to about 12-15 ~tm in diameter. No long processes from these strongly TH-immunoreactive neurones were noted between days 5 and 7 postnatal, although some short axon fibres could occasionally be seen at the INL/IPL border (Fig. 1C, D). In layer 3 of the IPL a dense band of varicose axons, thought to arise from the weakly TH-positive neurones of the medial INL, was now prominent, but no fibres were found elsewhere. On day 8 postnatal, a few varicose fibres in layer 1 of the IPL appeared, and were considered to be projections of the strongly TH-immunoreactive cells. The extent of these fibres increased
189
Fig. 1. Immunofluorescence photomicrographs of the ontogeny of TH-LI in the rat retina. A: the first TH-positive cells, at day 2 postnatal, situated in the middle of the primordial INL. B: a higher magnification of the same cell. C: at postnatal day 5, many more qq-Ipositive cells are visible, and they can be divided into two g r o u p s - either weakly (short arrows) or strongly staining (long arrows). D: higher power micrograph of C. E: a strongly staining TH-immunoreactive cell at day 9 postnatal, showing a long varicose axon travelling along the border of the IPL and INL. The weakly staining cells, in contrast, send out a sclerally oriented process, which bifurcates and ramifies in the middle layer of the IPL (F, higher magnification seen in G and H, all at day 5 postnatal). The numbers in G and H refer to the layers of the IPL. Other abbreviations are as in the text. Scale bar indicates 50pm for A, C, F and 20pro for B, D, E, G and H. considerably on day 9, such that individual beaded axons could then be traced in the section for several hundred microns (Fig. 1E). No PNMT-LI was found in the retina between day 16 of gestation and day 9 postnatal.
Postnatal days 10-15 At day 10 postnatal, TH-positive fibres in layer 3 of the IPL were by now very dense, and strongly immunoreactive. The fibres of the large TH-positive neurones of the INL branched m a n y times (Fig. 2A),
190
Fig. 2. Immunofluorescence photomicrographs of TH-LI in the developing rat retina. A: a section taken tangentially to the retinal layering, at day 10 postnatal and shows several strongly staining TH-positive cells with multiply dividing varicose fibres. B and C: sections taken approximately radially to the retinal layering at 13 days postnatal, showing strongly staining TH-immunoreactive cells which appear to innervate both layers 1 and 3 of the IPL. Abbreviations are as in the text. Scale bar indicates 50 um in A, B and 20/tm in C. and it was thought that some cells might send processes to more than one layer of the IPL. Two examples are shown in Fig. 2B, C, at day 13, the age when this was first apparent. Neurones containing PNMT-LI were observed for the first time in the retina at day 10 after birth (Fig. 3A). The immunoreactive somata were about 8/~m in diameter, and were found in the proximal half of the INL. Emerging from the cell body could be seen a single vitreally directed process (Fig. 3A, C) which bifurcated in layer 3 of the 1PL (Fig. 3 C - E ) , and which then formed a network of fibres. At later postnatal stages, the intensity of the immunostaining of the cell bodies and fibres increased. In addition, the meshwork of PNMT-positive fibres became denser over this period, resembling the adult situation by day 14 (Fig. 5A, B). Somata immunoreactive to
PNMT antisera were also visible in the ganglion cell layer (GCL) at day 10 (Fig. 3A, C, F). These cell bodies were similar in shape and size to those in the INL, and were considered, as in the adult 7, to be displaced amacrine cells. Single processes from these cells were visible (Figs. 3F, 4A), coursing obliquely to and ramifying in the medial layer of the 1PL (Fig. 4A). At day 13, a sparse plexus of fibres originating from the PNMT-immunoreactive cells in the INL became apparent in layer 1 of the IPL (Fig. 4B). These fibres grew progressively denser at days 14 and 15, when a situation comparable to that in the adult was found. Sections adjacent to those used for PNMT immunocytochemistry were stained for TH-LI. Areas of the retina that contained the PNMT-immunoreactive cells of the medial INL occasionally also exhibited
191
I
Fig. 3. Immunofluorescence photomicrographs of PNMT~ (A, C-F) and TH-LI (B) in the rat retina at day 10 postnatal. Although many PNMT-positive neurones are visible in the vitreal half of the INL and in the GCL (A), an adjacent section (B) stained for TH showed that no TH-positive cells were detectable in this region of the retina. C: higher magnification of the PNMT-containing neurones, in the INL (arrows) and the GCL (arrowheads), and the plexus of fibres in layer 3 of the IPL. Although PNMT-immunoreactive cells exhibited short processes (F), it was only those PNMT-positive neurones of the INL which displayed bifurcating axons in the IPL (D, E) at this age. Abbreviations are as in the text. Scale bar indicates 65 ~m in A, B and 20/~m in C-F.
the more proximally located large, strongly-staining, TH-positive cells of the first layer of the INL (cf. Fig. 5A and B). However, the weakly T H - i m m u n o r e a c tive perikarya found in the same layer of the INL as the PNMT-positive cell bodies were never found in
the same gross regions of the retina (cf. Figs. 5 A - D and 6 A - D ) .
Postnatal days 16-20 By day 16, the disposition of P N M T - L I was very
192
Fig. 4. Immunofluorescence photomicrographs of PNMT-LI in the retina of the 13-day-old rat. A: processes from the PNMT-positive displaced amacrine cells in the GCL (arrows) reach as far as layer 3 of the IPL, where they contribute to a, by now, dense fibre plexus. B: PNMT-immunoreactive neurone in the INL which displays a bifurcating axon in layer 1 of the IPL, where a sparse network of fibres is also apparent. The density of fibres in layer 3 of the IPL is similar to that of the adult retina. Abbreviations are as in the text. Scale bar indicates 20 ~m.
similar to that f o u n d in the m a t u r e retina. M a n y positive n e u r o n e s in the I N L and the G C L w e r e apparent, and a d e n s e b a n d of fibres in layer 3, with a sparser n e t w o r k of fibres in layer 1 of the I P L c o u l d also be seen. A t days 16 and 18 the T H - L I also bec a m e m o r e closely aligned to that of the adult: the weaker-staining
cells b e c a m e
p r o g r e s s i v e l y rarer,
while the m e s h w o r k of fibres in layer 3 of the I P L decreased (Fig. 7). In contrast, the strongly-staining T H - i m m u n o r e a c t i v e n e u r o n e s i n c r e a s e d in n u m b e r , and their p r o c e s s e s in layer 1 of the I P L w e r e similarly e n h a n c e d . A t no t i m e d u r i n g the p e r i o d s t u d i e d did any of the P N M T - p o s i t i v e n e u r o n e s also display detectable levels of T H - L I . A s c h e m a t i c s u m m a r y of
Fig. 5. Immunofluorescence photomicrographs of adjacent sections (.A, B and C, D) stained with PNMT (A, C) or TH (B, D) antisera at day 14 postnatal. Many PNMT-positive cells are visible in A, but only a single large TH-immunoreactive neurone is apparent in the adjacent section (B). Similarly, although many large, strongly staining and small, weakly staining TH-positive neurones exist in D, no cells containing PNMT-LI are manifest in the adjacent section. Note also the virtual absence of TH-LI from layer 3 of the IPL near the PNMT cells, and of PNMT-LI from layer 3 near the weakly staining TH-positive neurones. Abbreviations are as in the text. Scale bar indicates 50 ~m. The magnification was the same for all micrographs.
193
~ii ~
Fig. 6. Low-power immunofluorescence photomicrographs of adjacent sections (A, B and C, D) of 14-day-old rat retina, stained for PNMT- (A, C) and TH-LI (B, D). The micrographs show sections at two levels in the same retina. A long single-headed arrow in D indicates a weakly staining TH-neurone. In the regions where these cells are seen (between the short arrows), no PNMT-containing neurones are demonstrable. However, the large, strongly staining TH-positive cells (long double-headed arrows in D) of the first layer of the IPL, occupy the same zones of the retina as both the more distally stratified PNMT-immunoreactive neurones and the weaker staining TH-neurones. C, choroid; CP, ciliary process. Scale bar indicates 250,um. Magnification was the same for all micrographs.
194 PNMT
TH
2d C)=~L OPI
.
INL
.
.
.
.
.
.
.
.
.
.
.
_ Y
.
.
tPt
GCL .
.
.
.
.
.
.
PRt ONL OPt .
IPL Get .
.
.
.
.
.
.
• •
.
.
• -
.
•, .
.
.
.
.
." . . . . . . . . . .
.
.
.
.
.
.
.
.
• ' .
.
.
.
.
.
.
lOd ONL OPk*
Z
--
-"
,,L2. . . . . .
:
--
-"
2
....... i...... :
fiCL
15d ONL 0PL:
:
:
:
"-
"-
"-
:
=
:
:
:
:
:
:
:
Adult ONL OPL[
[
[
--
_"
_ . . . _ . . . : . , . . , . ,
GCL .
.
.
.
.
_-
~.
,.~., .
.
"_
Z
~.
Z
-"
Z
Z
Z
Z
,_..~.., .
.
.
.
.
.
.
.
.
Fig. 7. Schematic diagram of the ontogeny of TH- and PNMTLI neurones of the postnatal rat retina. The fibres are symbolized by dots in the IPL. No attempt was made to express different levels of immunoreactivity. Abbreviations are as in the text. Scale bar indicates 50 #m.
the ontogeny of PNMT- and T H - L I in the rat retina is shown in Fig. 7. DISCUSSION
The development of TH- and PNMT-LI in the maturing rat retina has been described. Weak TH-immunoreactive cells in the second layer of the INL were first demonstrable at postnatal day 2, whereas the stronger-staining neurones were not apparent until day 5. This is rather earlier than that found by other workers2S, and may result from the use of pictic acid in our fixation medium; TH-LI was undetectable in the small pear-shaped cells at day 2 in retinas
which had been fixed in 4% paraformaldehyde alone. The ontogeny of the stronger-staining THpositive cells and of the PNMT-immunoreactive cells was, however, the same with both types of fixative. At later postnatal stages the weakly-fluorescent neurones began to disappear, such that by day 20 they were extremely rare. Fibres in layer 3 of the IPL were visible at day 5, were of maximal density at day 15, but had diminished greatly in number by day 20. Conversely, the TH-positive processes in layer 1 of the IPL, which were sparse at day 8, had increased to reach mature levels by day 20. PNMT-Iike immunoreactivity was first found in neurones of the developing rat retina at postnatal day 10. These cells were located in the proximal half of the INL, and sent out processes which bifurcated, and then ramified in layer 3 of the IPL. By day 14 the density of the fibres in the medial layer closely resembled that found in the adult, whereas fibres in layer 1 were maximal at about day 15. Despite the known dependence of adrenaline concentrations 12 and of TH activity on photic stimulation 17, neither the number nor the intensity of staining of the TH- and PNMT-immunoreactive neurones exhibited any apparent fluctuations at the time of opening of the eyes, namely 16 days. It could be that the levels of the enzymes are not changed during photic stimulation, but, instead, that the degree of their activation by cofactors is responsible for the observed biochemical increases. Alternatively, an inherent diurnal rhythm may control PNMT and TH induction prior to eyeopening, which is subsequently replaced or combined with photic control of enzyme levels. Experiments are in progress in the adult retina to determine if either of these hypotheses is correct. It is apparent that in contrast to the TH-positive neurones, those containing PNMT-LI are morphologically well-developed before the detection of the antigen. Even on day 10, when the cell bodies were first visible, large numbers of PNMT-immunoreactive fibres were also seen. This is reminiscent of the situation in the developing rat medulla oblongata, where at day 13 of gestation PNMT-positive cell bodies and their extensive ascending and descending projections become apparent almost simultaneouslyS. Given that neurite extension in vitro a and in vivo 32 is about 40/tm/h, it would seem that the morphological development of the PNMT-containing
195 neurones precedes development of their transmitter synthetic enzyme. In contrast, T H - L I appears in processes of the neurones of the I N L over several days, suggesting that its induction occurs prior to the morphological maturation of the neurone. Of particular interest was the discovery that none of the PNMT-immunoreactive cell bodies contained detectable amounts of TH-LI. Although the possibility cannot be excluded that the levels of T H - L I may have been below the limits of sensitivity of the immunocytochemical technique, our experience is that the use of the TH antibody represents the most potent method of visualizing catecholaminergic neurones. An inability to detect T H - L I would therefore suggest that less than significant amounts are present. The observation of PNMT-positive, TH-negative neurones in the retina was all the more surprising in view of the similar stratification within the I N L of the PNMT-positive cells and the weakly staining TH-immunoreactive somata during early postnatal life. However, the two types of neurones occupied almost mutually exclusive zones of the retina. It would ap2 pear, therefore, that despite their commonalities in terms of morphology and layering, the two types of cell are not the same. Although the weak TH-positive cells either die or lose their ability to synthesize TH during development, those containing P N M T - L I alone continue to be manifest, even in the adult retina 7. Although the normal control studies were carried Out 1°,11.22, the possibility that our P N M T and T H antisera are cross-reacting with other substances cannot be wholly eliminated. However, in view of the similar results obtained with completely different antibodies this hypothesis seems untenable. For a fuller discussion see ref. 15. It is not clear what function the PNMT-positive, TH-negative cells in the retina possess. Neurones with PNMT-LI, but without detectable T H - L I have
REFERENCES 1 Bray, D., Branching patterns of individual sympathetic neurones in culture, J. Cell Biol., 56 (1973) 702-712. 2 Brecha, N., Retinal neurotransminers: histochemical and biochemical studies. In P.C. Emson (Ed.), Chemical Neuroanatomy, Raven Press, New York, 1983, pp. 85-129. 3 Cohen, J. and Neff, N.H., Retinal amacrine cell system tyrosine hydroxylase: the development of responsiveness to
also been described in the adult rat dorsal hypothalamus 3~ and posterior arcuate nucleus 7. It seems unlikely that nerve cells which are unable to mediate neurotransmission because of the absence of their endogenous transmitter would maintain projections or survive in the adult central nervous system. Three obvious explanations exist: the first is that an alternative synthetic pathway for adrenaline synthesis, not via TH, is in operation. In support of this, it is known that the retina does indeed contain adrenaline12, 27, and that this is unlikely to be derived from the circulation. Second, the P N M T may be catalyzing the methylation of some substance other than noradrenaline, perhaps unrelated to a transmitter role. This implies that adrenaline is not the final reaction product. Lastly, these neurones may utilize an alternative transmitter which is enabling them to function normally; the PNMT-LI may then be merely the result of transcription of a vestigial gene. A third putative catecholaminergic cell type has been recently described in the adult retina 9. This is also found in the proximal zone of the INL, but is only rendered visible in Falck-Hillarp histofluorescence by loading with exogenous noradrenaline, but not with its precursors dopamine or L - D O P A . These cells were therefore considered to possess the capacity for noradrenaline uptake, but not for its synthesis. It remains to be determined whether they are related to either the weakly-staining TH-cells seen during neonatal life, or to the PNMT-positive, TH-negative cells found both in the neonate and the mature animal 7. ACKNOWLEDGEMENTS This work was supported by grants from the Swedish Medical Research Council (04X-2887) and Knut och Alice Wallenbergs Stiftelse. G . A . F . is a N A T O Overseas Research Fellow.
light and neuroleptic drugs, Dev. Brain Res., 255 (1982) 1960-1963. 4 Coons, A.H., Fluorescent antibody methods. In J.F. Danielli (Ed.), General Cytochemical Methods, Academic Press, New York, 1958, pp. 399-422. 5 Dowling, J.E. and Ehinger, B., Synaptic organization of the dopaminergic neurons in the rabbit retina, J. Comp. Neurol., 180 (1978) 203-220. 6 Ehinger, B., Adrenergie nerves to the eye and related
196
7
8
9
10
11
12
13
14
15
16
17
18
structures in man and in the cynomolgus monkey (Macaca irus), Invest. Ophthalmol., 5 (1966) 42-52. Foster, G.A., H6kfelt, T., Coyle, J.T. and Goldstein, M., lmmunohistochemicat evidence for phenylethanolamine N-methyltransferase-positive, tyrosine hydroxylase-negative neurones in the retina and the posterior hypothalamus of the rat, Brain Res., in press. Foster, G.A., Schultzberg, M., Goldstein, M. and H6kfelt, T., Ontogeny of phenylethanolamine N-methyltransferaseand tyrosine hydroxylase-like immunoreactivity in presumptive adrenaline neurones of the foetal rat central nervous system, J. Comp. Neurol., in press. Fukuda, M., Ishimoto, I., Kuwayama, Y., Shimizu, Y., Shiosaka, S., Takagi, H. and Tohyama, M., Monoamine accumulating neuron system in the rat retina with special reference to noradrenaline accumulating neurons, Exp. Eye Res., 34 (1982) 487-491. Goldstein, M., Fuxe, K. and H6kfelt, T., Characterization and tissue localization of catecholamine synthesizing enzymes, Pharrnacol. Rev., 24 (1972) 293-309. Goldstein, M., Lew, J.Y., Matsumoto, Y., H6kfelt, T. and Fuxe, K., Localization and function of PNMT in the central nervous system. In M,A. Lipton, A. DiMascio and K.F. Killam (Eds.), Psychopharmacology, Raven Press, New York, 1978, pp. 261-269. Hadjiconstantinou, M,, Cohen, J. and Neff, N.H., Epinephrine: a potential neurotransmitter in retina, J. Neurochem., 41 (1983) 1440-1444. Hadjiconstantinou, M,, Mariani, A.P., Panula, P., Joh, T.H. and Neff, N.H., Immunohistochemical evidence for epinephrine-containing retinal amacrine cells, Neuroscience, 13 (1984) 547-551. Hfiggendal, J. and Malmfors, T., Identification and cellular localization of catecholamines in the retina and choroid of the rabbit, Acta Physiol. Scand., 64 (1965) 58-66. H6kfelt, T., Johansson, O. and Goldstein, M., Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS, Part 1, Elsevier, Amsterdam, 1984, pp. 157-276. Kato, S., Nakamara, T. and Negishi, K., Postnatal development of dopaminergic cells in the rat retina, J. Comp. Neurol., 191 (1980) 227-236. Iuvone, P.M., Galli, C.L., Garrison-Gund, C.K. and Neff, N.H., Light stimulates tyrosine hydroxylase activity and dopamine synthesis in retinal amacrine cells, Science, 202 (1978) 901-902. Joh, T.H. and Goldstein, M., Isolation and characterization of multiple forms of phenylethanolame N-methyltransferase, Mol. Pharmacol., 9 (1973) 117-129.
19 Johnson, D.G. and De C. Nogueira Araujo, G.M.. A sin> ple method of reducing the fading of immunofluorescencc during microscopy, J. lmmunol. Meth., 43 (1981) 349. 20 Lain, D.M.K., Fung, S.C. and Kong, Y.C., Postnatal development of dopaminergic neurons in the rabbit retina, J. Neurosci.. 1 (1981) 1117-1132. 21 Malmfors, T., Evidence of adrenergic neurons with synaptic terminals in the retinas of rats demonstrated with fluorescence and electron microscopy, Acta Physiol. Scand., 58 (1963) 99-100. 22 Markey, K.A., Kondo, S., Shenkman, i. and Goldstein, M., Purification and characterization of tyrosine hydroxylase from a clonal phaeochromocytoma cell line, Mol. Pharmacol., 17 (1980) 79-85. 23 M~rtensson, R. and BjOrklund, A., Low power photography in the fluorescence microscope using an anatomical dark-field condenser scanner. In A. Bj6rklund and T, H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS, Part 1. Elsevier, Amsterdam, 1984. 24 Morgan, W.W. and Kamp, C.W., Postnatal development of the light response of the dopaminergic neurons in the rat retina, J. Neurochem., 39 (1982) 283-285. 25 Nguyen-Legros, J., Berger, B., Vigny, A. and Alvarez, C.. Tyrosine-hydroxylase-like immunoreactive interplexiform cells in the rat retina, Neurosci. Lett., 27 (1981) 255-259. 26 Nguyen-Legros, J., Vigny, A. and Gay, M., Post-natal development of TH-like immunoreactivity in the rat retina, Exp. EyeRes., 37 (1983) 23-32. 27 Osborne, N.N. and Nesselhut, T., Adrenaline: occurrence in the bovine retina, Neurosci. Lett., 39 (1983) 33-36. 28 Parkinson, D. and Rando, R.R., Ontogenesis of dopaminergic neurons in the post-natal rabbit retina: pre- and post-synaptic elements, Dev. Brain Res., 13 (1984) 207-217. 29 Patten, M., Parkinson, D., Wyse, P. and Spira, A., Ontogenesis of dopaminergic neurons in the fetal and postnatal guinea pig retina, Invest. Ophthalmol. Vis. Sci., Suppl., 22 (1982) 114. 30 RamOn y Cajal, S., La rdtine des vert6bres, La Cellule, 9 (1893) 17-257. 31 Ross, C.A., Ruggiero, D.A., Meeley, M.P., Park, D.H., Job, T.H. and Reis, D.M., A new group of neurons in hypothalamus containing phenylethanolamine N-methyltransferase (PNMT) but not tyrosine hydroxylase, Brain Res., 306 (1984) 349-353. 32 Speidel, C.C., Adjustments of nerve endings, Harvey Lect., 36 (1941) 126-158. 33 Zamboni, L. and De Martino, C., Buffered picric-acid formaldehyde: a new rapid fixative for electron-microscopy, J. Cell Biol., 35 (1967) 148A.