Neuropeptides in the retina: Evidence for differential topographical localization

Peptides, Vol. 2, pp. 153--157, 1981. Printed in the U.S.A.

Neuropeptides in the Retina: Evidence for Differential Topographical Localization E. F. E R I K S E N

A N D L.-I. L A R S S O N 1

hzstitltte o f Medical Biochemistry, University o f Aarhus DK-8000 Aarhus C, Denmark R e c e i v e d 22 O c t o b e r 1980 ERIKSEN, E. F. AND L.-I. LARSSON. Neuropeptides b~the retina:Evidencefor differential topographicalIocalizotion. PEPTIDES 2(2) 153--157, 1981.mThe retina has been examined for the presence of vasoactive intestinal polypeptide (VIP), substance P, somatostatin and gastrin/cholecystokinin immunoreactive structures by immunocytochemistry. The results indicate that all four peptides reside in neuron-like processes of the inner plexiform layer. Interestingly, the immunoreactive processes display distinctive distributional patterns in different strata of this layer, suggesting that they may serve distinct functions, possibly related to integration and modification of visual input. This supposition is strengthened by the observation that the immunoreactivity of the VIP cell bodies and processes varies extensively with the amount of light entering the eye. Vasoaetive intestinal polypeptide Inner plexiform layer Retina

Substance P Somatostatin lmmunoreactivity

N U M E R O U S secretory peptides have been localized to central and peripheral neurons, where they have been suggested to act as neurotransmitters or neuromodulators (for a recent review, see [15]). Studies on the synthesis, transport and physiological activities of neuropeptides in the nervous system are facilitated in regions possessing an ordered anatomical structure, making possible simultaneous examination of cell bodies and their processes. The retina constitutes such a specialized part o f the central nervous system, which easily can be physiologically manipulated. We have therefore examined this organ for the presence o f several neuropeptides, including vasoactive intestinal polypeptide (VIP), substance P (SP), somatostatin (SRIF) and gastrin/cholecystokinin. Moreover, preliminary data on the effects o f varying the light input to the eye on these peptides are reported. METHOD Adult male Wistar/BN rats were kept on an ordinary light schedule (12 hrs light and 12 hrs darkness; n=6), on constant illumination for 12 days (n=4) or in total darkness for 12 days (n=4). In addition, 2 rats received intravitreal injections of colchicine (50 /zg) 24 hrs before sacrifice. The rats were anaesthesized with diethyl ether and intracardially perfused with 20 ml saline followed by 60 ml 4 percent paraformaldehyde [13]. Simultaneously, fixative was also injected intraocularly. Eyes were halved, postfixed for 24 hrs in the formaldehyde solution, soaked overnight in 20 percent sucrose and frozen on cryostat chucks in melting F r e o n - 2 2 . 8 ~ m cryostat sections were exposed to antisera against VIP (c. 5603), substance P (K 22), somatostatin (R 213/3) and

Gast rin/cholecystokinin

gastrin/CCK (No. 4562). The antisera were applied at their optimal dilutions for 24 hrs at 4~ (5603; dilution 1:5120; K 22; 1:1000; R 213/3; 1:3200 and 4562; 1:50000), whereafter the site of antigefi antibody reaction was demonstrated by the peroxidase-antiperoxidase (PAP) technique of Sternberger [ 16]. The antisera have been characterized in detail elsewhere and shown not to cross-react with a number of related and unrelated peptides. Thus, the VIP antiserum (5603) used is fully specific for VIP and does not cross-react with the related peptides: secretin, glucagon or gastric inhibitory polypeptide [5,6]. Also the somatostatin (R 213/3) and substance P (K 22) antisera have been extensively analyzed previously [7,12] and do not cross-react with a number of structurally related peptides. The antiserum (4562) used for revealing CCK-like immunoreactivity recognizes the COOH-terminal tetrapeptide portion [8, 9, 10]. Since this region also occurs in gastrin the antibody does not differentiate between these two peptides, hence the designation: gastrin/CCK-like immunoreactivity. Controls consisted o f conventional staining controls [16] as well as of absorption controls including the pretreatment of the optimally diluted antisera with 100 ~g per ml of synthetic VIP, somatostatin, substance P, gastrin-17 o r 99 percent pure porcine CCK-33. All controls documented the specificity of the antisera for their respective antigens and excluded cross-reactivity with related and unrelated peptides. Stained sections were examined in the light microscope following counter-staining with haematoxylin (to reveal background details) or in a dark field microscope employing a Tiyoda condensor and monochromatic light of 546 nm.

ERIKSEN AND LARSSON

154

b FIG. 1. Sections of rat retina (normal light cycle) stained for vasoactive intestinal polypeptide (VIP) immunoreactivity (a) and somatostatin immunoreactivity (b), using the PAP procedure, and viewed in dark field. Note that whereas the VIP immunoreactive processes are distributed in two distinct strata of the inner plexiform layer (IPL) (arrows), the somatostatin immunoreactive processes distribute in only one single layer (arrow) (eL Fig. 2). Note also the presence of an intensely immunoreactive VIP cell body which sends out processes to the IPL. Unspecific light scatter is seen in the outer nuclear layer, x320

For studies on the influences of a variable light input to the eye, homogenous groups of rats were randomized and exposed to constant light or darkness or to the standard light scheme (cf. above). All rats were killed, fixed, processed, stained and analyzed in parallel to exclude methodological variation. RESULTS All four antisera revealed specifically stained structures in the retinae of all animals investigated. Staining was confined to numerous processes of the inner plexiform layer (ipl)

and, in the case of VIP, also occurred in cell bodies of the inner nuclear layer (ini). Absorption controls confirmed the specificity of the immunocytochemical staining in every case. In normal rats (kept on a schedule of 12 hrs light and 12 hrs darkness), VIP immunoreactive somas were detected in the inner aspect of the inl. These somas gave off processes which entered the ipl and either travelled at the zone of junction between the inl and the ipl or occupied a distinct stratum in the middle portion of the ipl (Figs. 1 and 2). Staining for somatostatin (Ab. R 213/3) failed to reveal immunoreactive somas even after intraocular injection of col-

RETINAL NEUROPEPTIDES

I

SP

<]GO

I SRIF

000

155

I

uw<~

o

VIP 00(2

`

0

I 00o~

0

CCK

I

~cou~

l~OVJ hu_ooP_o hO~o o~oo_Looo &oA^ O _ i o ~ oP~ooo,

,u

~ooo~~~ocYooo~

I

I"

I'""-~ I'--'~-'-t

I

I

I

I

I

I~

I

0 000dO0 I I

~-T~II

1

I

]

I

I

ipl

I O0 0 0

C m Y o o 001 I I

gel

FIG. 2. Schematic drawing illustrating the differential distribution of substance P (SP), somatostatin (SRIF), vasoactive intestinal polypeptide (VIP) and gastrin/cholecystokinin (CCK) immunoreactive processes in the rat retina. Note the striking differences in the distribution of the processes in the inner plexiform layer (ipl).

chicine. Instead, a dense collection of fibers was detected at the zone of junction between inl and ipl (Figs. 1 and 2). In addition, scarce somatostatin immunoreactive fibers were seen in the middle portion of the ipl. Similarly, staining for SP and CCK/gastrin failed to document cell bodies ir/either normal or colchicine-injected rats. Substance P immunoreactive fibers were more numerous than CCK/gastrin fibers, which were extremely rare. Both types of fibers occurred at the innermost aspect of the ipl, close to the ganglion cell layer. The spatial distribution of immunoreactive fibers in the ipl is schematized in Fig. 2, showing that both VIP and SRIF immunoreactive fibers occupy both outer and intermediate layers of the ipl, whereas the substance P and CCK/gastrin immunoreactive fibers occupy the inner, juxta-ganglionic part. Analysis of retinae from rats exposed to constant light or darkness for 12 days revealed striking differences in the VIP immunoreactive structures. Thus, constant light produced a marked increase in the immunoreactivity and apparent numbers of cell bodies and processes (Fig. 3). In constrast, following constant darkness, very few and faintly immunoreactire VIP cell bodies and processes were detected. Rats kept on the standard light schedule (12 hrs darkness and 12 hrs light) showed a pattern intermediate between these two extremes. It appeared that mainly the VIP cells were affected by variations in light input. Significant effects on the SP, CCK/gastrin and SRIF immunoreactive structures, as examined in sections stained in parallel with those above, could not be detected. DISCUSSION

Our results suggest that nerve-like processes containing V I P , substance P, somatostatin and gastrin/CCK immunoreactivity occupy distinct strata of the inner plexiform

layer of the rat retina. The localization of the VIP immunoreactive cell bodies and processes to the inner nuclear and plexiform layers, respectively, suggests that these cells may form part of the amacrine cell population. Previously, several peptides, including enkephalin, VIP, substance P , neurotensin, and glucagon and somatostatin, have been localized to amacrine cells of the avian retina [1, 2, 3, 11, 14]. In preliminary studies, we have been unable to detect enkephalin immunoreactive structures in the rat retina. This may suggest that species differences with respect to the peptide make-up of retinal cells may occur, although it should be emphasized that negative immunocytochemical data are essentially worthless. Thus, different enkephalin antisera react to different degrees with met- and leu-enkephalin and, moreover, certain antisera are capable to react only with pentapeptide enkephalins, whereas others also recognize Cor N-terminally elongated enkephalin congeners. Hence, generally speaking, absence of immunoreactivity can be due to the presence ofpeptide components not able to react with the antiserum employed. The cellular origins of the other types of immunoreactive processes (SRIF, SP and CCK/gastrin) of the inner plexiform layer have yet to be defined. The localization of these processes makes it possible that they may emanate from amacrine cell bodies of the inner nuclear layer. This, however, remains to be proven. Since previous work [14] has documented that retinal somatostatin concentrations are elevated after optic nerve transsection it is less likely that this neuropeptide is of extraocular origin~ Failure to detect the cell bodies from which the somatostatin, substance P and gastrin/CCK immunoreactive processes originate may relate to many factors, including low concentrations of the peptides in cell bodies or, alternatively, to storage of poorly immunoreactive precursors in the somas (cf. [4]).

ERIKSEN

156

a,a ,-dr,- &..,,

,...,,

tR

:? "

..

~:-X.i.Mi~ " ( .'z- " q /

''.

0

~..,,, :- 7"

q'e"

ei "a~,'o ~e,

1

~

~

9

/,,

. ~:

"4,,,

.. .~ _

.-

.,.~.,L.

a

" n

9.

,~

~

b

7

.o

r C

FIG. 3. Sections of retinae from rats exposed to constant light for 12 days (a), to a normal light schedule with 12 hrs light and 12 hrs darkness (b) or to constant darkness for 12 days (c). The sections were stained for VIP immunoreactivity using the PAP method. Note the differences in the number and intensity of both VIP immunoreactive cell bodies of the inner nuclear layer and processes of the inner plexiform layer during the different illumination conditions, x425.

AND LARSSON

R E T I N A L NEUROPEPTIDES

157

The functions of the processes detected are unknown. Although the antisera used are of high specificity [5-I0], cross-reactivity with yet uncharacterized peptides cannot be excluded. On the balance, however, all four types of peptides detected have previously been found in other types of neurons [15] and are known to affect neuronal firing and excitability. Moreover, in their recent paper, Rorstad et al. [14] have characterized extractable somatostatin from murine retinae and shown it to be indistinguishable from true somatostatin. The presence of neuroaetive peptides in amacrine-like processes of the ipl makes it tempting to suggest that the peptides participate in integrating and modulating visual signals. This hypothesis is, in the case of VIP, corroborated by the striking differences obtained by varying

the light input to the eye. If such integratory and/or modulatory influences can be proven, the data suggest that different peptides exert their actions at different levels of the inner plexiform layer.

ACKNOWLEDGEMENTS Grant support was from the Danish MRC and excellent technical assistance was provided by Mrs. Bente Blankenberg-Serup. Antisera used were generously donated by Drs. J. Fahrenkrug, G. Nilsson and J. F. Rehfeld. We also wish to thank Dr. J. Morley of ICI, Alderley Park for supplying synthetic human gastrin I and Prof. V. Mutt, Department of Biochemistry, Karolinska Institutet, Stockholm for supplying highly purified porcine CCK-33.

REFERENCES

I. Brecha, N., H. J. Karten and B. Davis. Localization of neuropeptides, including VIP and glucagon within the adult and developing avian retina. Soc. Neurosci. Abstr. 6: 346, 1980. 2. Brecha, N., H. J. Karten and C. Laverack. Enkephalincontaining amacrine cells in the avian retina: Immunohistocbemical localization. Proc. hath. Acad. ScL U.S.A. 76: 3010-3014, 1979. 3. Famiglietti, E. V., Jr., N. C. Brecha and H. J. Karten. Neural localization of substance P-like immunoreactivity in rabbit retina. Soc. Neurosci. Abstr. 6: 212, 1980. 4. Larsson, L.-I. Corticotropin and a-melanotropin in brain nerves: Immunocytochem~cal evidence for axonal transport and processing. In: Neural Peptides and Neuronal Communication. edited by E. Costa and M. Trabucchi. New York: Raven Press, 1980, pp. 101-108. 5. Larsson, L.-I., J. Fahrenkrug and O. B. Schaffalitzky de Muckadell. Vasoactive intestinal polypeptide occurs in nerves of the female genito-urinary tract. Sc&nce 197: 1374-1375, 1977. 6. Larsson, L.-I., J. Fahrenkrug, O. B. Schaffalitzky de Muckadell, F. Sundler, R. H~tkanson and J. F. Rehfeld. Localization of vasoactive intestinal polypeptide (VIP) to central and peripheral neurons. Proc. hath. Acad. Sci. U.S.A. 73: 3197-3200, 1976. 7. Larsson, L.-I., N. Goltermann, L. D. Magistris, J. F. Rehfeld and T. W. Schwartz. Somatostatin cell processes as pathways for paracrine secretion. Science 205: 1393-1394, 1979.

8. Larsson, L.-I. and J. F. Rehfeld. Evidence for a common evolutionary origin of gastrin and cholecystokinin. Nature 269: 335-338, 1977. 9. Larsson, L.-I. and J. F. Rehfeld. A peptide resembling COOHterminal tetrapeptide of gastrin from a new gastrointestinal endocrine cell type. Nature 277: 575-578, 1979. 10. Larsson, L.-I. and J. F. Rehfeld. Localization and molecular heterogeneity of cholecystokinin in the central and peripheral nervous system. Brain Res. 165: 201-218, 1979. 11. Lor6n, I., K. Tornqvist and J. Alumets. VIP-immunoreactive neurons in the retina of the rat. Cell Tissue Res. 210: 167-170, 1980. 12. Nilsson, G., L.-I. Larsson, R. H~tkanson, E. Brodin, B. Pernow and F. Sundler. Localization of substance P-like immunoreactivity in mouse gut. Histochemistry 43: 97-99, 1975. 13. Pease, D. C. Buffered formaldehyde as a killing agent and primary fixative for electron microscopy. Anat. Rec. 14Z: 342, 1962. 14. Rorstad, O. P., M. J. Brownstein and J. B. Martin. Immunoreactive and biologically active somatostatin-like material in rat retina. Proc. hath. Acad. Sci. U.S.A. 76: 3019--3023, 1979. 15. Snyder, S. H. Brain peptides as neurotransmitters. Science 209: 976--983, 1980. 16. Sternberger, L. A. hnmunocytochemistt3,, 2nd. edition. New York: Wiley, 1979.