Neuropeptide Y-like Immunoreactivity Localizes to Preganglionic Axon Terminals in the Rhesus Monkey Ciliary Ganglion

Transcription

1 Neuropeptide Y-like Immunoreactivity Localizes to Preganglionic Axon Terminals in the Rhesus Monkey Ciliary Ganglion Patricia A. Grimes, 1 Brigitte Koeberlein, 1 Margarete Tigges, 2 and Richard A. Stone 1 PURPOSE. TO characterize neuropeptide distribution in the ciliary ganglion of rhesus monkeys (Macaca mulatto). METHODS. Cryostat tissue sections of fixed rhesus monkey ciliary, pterygopalatine, superior cervical, and trigeminal ganglia were incubated with antisera to neuropeptide Y (NPY), calcitonin generelated peptide (CGRP), substance P (SP), vasoactive intestinal peptide (VIP), tyrosine hydroxylase (TH), and dopamine-/3-hydroxylase (DBH). Antibody binding was visualized by indirect immunofluorescence. RESULTS. NPY-like immunoreactive (LI) nerve terminals surrounded 80% of ciliary ganglion cells, but ciliary ganglion cell somata were unstained. NPY-LI cells were present in the superior cervical ganglion, in which almost all cells were TH- and DBH-LI, and in the pterygopalatine ganglion, in which almost all cells were VIP-LI. Because neither TH, DBH, nor VIP immunoreactivity was detected in nerves contacting ciliary ganglion cells, the NPY-LI input to ciliary neurons does not likely derive from the autonomic ganglia. The trigeminal ganglion, another potential source, had no NPY-LI neurons. CGRP- and SP-LI axons from the nasociliary nerve traversed the ciliary ganglion; a small number of varicose axons were distributed among ganglion cells and rarely surrounded cell somata. Most ciliary ganglion cells were TH-LI, but only a few were DBH-LI. CONCLUSIONS. Based on these patterns of peptide immunoreactivities, the NPY-LI nerve fibers investing ciliary ganglion cells in the rhesus monkey are most likely preganglionic axon terminals of mesencephalic parasympathetic neurons. Although the origin and function of these NPY-LI nerves remains to be established, the present finding adds to the remarkable diversity of neuropeptide immunoreactivity so far identified in preganglionic and postganglionic cells of the ciliary ganglion in different species of birds and mammals, including primates. (Invest Ophthalmol Vis Sci. 1998;39: ) The ciliary ganglion provides parasympathetic innervation to the intrinsic ocular muscles responsible for pupil movement and accommodation. Immunohistochemical studies have localized several neuropeptides to subsets of ciliary neurons in mammals and birds that possibly reflect functional diversity in the neuronal population. In the rat, the ciliary ganglion contains subgroups of cells immunoreactive to neuropeptide Y (NPY), calcitonin gene-related peptide (CGRP), enkephalin, vasoactive intestinal peptide (VIP), and pituitary adenylate cyclase-activating polypeptide, a seemingly large diversity of neuronal subtypes given the animal's rudimentary accommodative capacity; no peptides have been identified in preganglionic nerves supplying the rat ciliary gangli- From the 'Department of Ophthalmology and Scheie Eye Institute, University of Pennsylvania, Philadelphia; and the 2 Yerkes Regional Primate Center, Atlanta, Georgia. Presented in part at the 1994 annual meeting of the Association for Research in Vision and Ophthalmology, Sarasota, Florida. Supported by research grants EYO5454 and EY09737 from the National Institutes of Health, Bethesda, Maryland and by an unrestricted grant from Research to Prevent Blindness. RAS is a Research to Prevent Blindness Senior Scientific Investigator. Submitted for publication May 7, 1997; revised September 19, 1997; accepted November 3, Proprietary interest category: N. Reprint requests: Richard A. Stone, D-603 Richards Building, University of Pennsylvania, Philadelphia, PA on. 1 In the cat, NPY- and galanin-like immunoreactive (LI) cells comprise approximately 25% each of the ciliary ganglion population, with only 5% of cells immunoreactive to both peptides; the NPY-LI cells are preferentially contacted by somatostatin- and CGRP-LI preganglionic nerve terminals. 5 " 7 The monkey has a much greater accommodative range than either the rat or the cat; most ciliary ganglion cells supply the ciliary muscle, and few target the iris muscles. 8 In the cynomolgus monkey (Macaca fascicularis), NPY, VIP, and substance P (SP) immunoreactivities localize to nerve fibers between and around ciliary ganglion cells; a portion of ganglion cell somata are immunoreactive for VIP but not for the other peptides. 9 In another study of the same monkey species, faint to moderate SP immunoreactivity was detected in 60% of cell somata. 10 Some cholinergic ciliary ganglion neurons of monkeys 1112 show tyrosine hydroxylase (TH) immunoreactivity, a property also noted in rats, cats, and dogs."" 13 In the only immunohistochemical study of the human ciliary ganglion, both SP and CGRP immunoreactivity localized in varicose axons contacting ciliary neurons, and approximately one fourth of ciliary perikarya showed TH immunoreactivity. 14 The ciliary ganglion of birds contains a distinct population of neurons innervating the choroid in addition to those innervating iris and ciliary muscle, and it is therefore more complex than the mammalian ganglion. In the pigeon, immunoreactivities to SP, enkephalin, and VIP colocalize in preganglionic Investigative Ophthalmology & Visual Science, February 1998, Vol. 39, No. 2 Copyright Association for Research in Vision and Ophthalmology 227

2 228 Grimes et al. IOVS, February 1998, Vol. 39, No. 2 terminals contacting many ciliary ganglion cells of both types. l5 ~ 17 In the chicken ciliary ganglion, however, only SP-LI preganglionic terminals have been identified, and these selectively target a group of somatostatin-li ganglion cells that innervate the choroid.' 8 In the present study, the ciliary ganglion of rhesus monkeys was examined for immunoreactivity to several neuropeptides and neurotransmitter-related enzymes. The most notable finding is that NPY-LI preganglionic nerve processes surround most ciliary neurons in this primate. NPY immunoreactivity has not been identified previously in preganglionic terminals of either the ciliary ganglion or other parasympathetic ganglia in any species. METHODS Ciliary, pterygopalatine, superior cervical, and trigeminal ganglia were obtained from rhesus monkeys (Macaca mulatto) killed for other purposes at the Yerkes Primate Center (Atlanta, GA). Nine monkeys, aged 9 to 11 months, had either unilateral surgical aphakia (five animals) or occlusion of the eye (four animals) from shortly after birth; one mature 18-year-old monkey was not part of any ocular study. This investigation adhered to the tenets of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. While under deep general anesthesia, all animals were perfused transcardially with 0.9% NaCl solution followed by 4% paraformaldehyde in 0.1 M phosphate buffer; the eyes and brains, used by other investigators for unrelated experiments, were unavailable for study. The ganglia, removed approximately 16 to 20 hours after perfusion, were immersed either in 4% paraformaldehyde solution for 24 hours or in Zamboni's solution 19 for 24 or 48 hours and then were transferred to 0.1 M phosphate buffer containing 30% sucrose for 24 hours. Cryostat tissue sections, 16- to 20-/xm thick, were cut and thaw mounted on gelatincoated slides. Tissue sections were incubated with one of three different rabbit polyclonal antisera raised against porcine NPY (Incstar, Stillwater, MN; Peninsula Laboratories, Belmont, CA; Affiniti Research Products, Mamhead, Exeter, UK) and diluted 1:500 in 0.05 M phosphate-buffered saline containing 0.3% Triton X-100. All three NPY antisera showed the same quality and distribution of specific immunofluorescence. The NPY antiserum supplied by Incstar was used for the photomicrographs provided here. Additional rabbit polyclonal antisera against the following antigens were also used at the indicated dilutions: CGRP (Cambridge Research Biochemicals, Wilmington, DE), 1:500; SP (mcstar), 1:500; VIP (Incstar), 1:500; TH (Eugene Tech International, Ramsey, NJ), 1:400; and dopamine-j8-hydroxylase (DBH; Eugene Tech), 1:200. After incubation with primary antisera for 18 hours at room temperature, the tissue sections were incubated for 1 hour with biotinylated goat anti-rabbit Immunoglobulin G (Zymed Laboratories, San Francisco, CA) and then for 1 hour with streptavidin-texas red (Amersham, Arlington Heights, IL). The immunostaining specificity was established by omitting the primary antiserum from die incubation procedure and, for NPY, by preadsorbing the diluted primary antiserum with 1 ixm porcine NPY (Peninsula Laboratories). The proportion of ciliary ganglion cells surrounded by NPY-LI cell processes was estimated from counts of all somata receiving labeled and unlabeled inputs in a series of nonoverlapping photographs of a single section from each ganglion. Approximately 250 cells were counted in each section. Except where specifically noted, the observations described are based on examinations of ganglia ipsilateral to the control phakic or nonoccluded eyes. RESULTS The ciliary ganglia in this study showed the same anatomic features and neural connections described previously for rhesus monkeys. 20 Preganglionic parasympathetic nerves enter the ganglion at its attachment to the oculomotor nerve at the origin of the branch to the inferior oblique muscle. The ganglion also is joined by two or three fine branches of the nasociliary nerve carrying both sensory and sympathetic axons that are thought to traverse the ganglion without synapse and to exit with postganglionic parasympathetic axons in the multiple ciliary nerves supplying the eye. Many fine nerves (rami orbitales) originating from the pterygopalatine ganglion lie in proximity to the ciliary ganglion 21 and could join the ganglion. In all ciliary ganglia examined, the majority of cell somata were surrounded by a dense array of NPY-LI nerve processes (Fig. 1). Tangential sections passing through the perisomatic region showed that the NPY-LI processes consisted of fine fibers often terminating in bulbous expansions. Some smooth nerve fibers coursing among the ganglion cells also were positively stained. All ganglion cell somata were NPY negative. The proportion of ciliary neurons surrounded by NPY-LI nerve processes averaged 80% (range, 55%-96%) in ganglia innervating the untreated eyes of the nine young monkeys; counts from three ganglia innervating aphakic eyes gave the same value (average, 80%; range, 66%- 89%). In the two ganglia from one untreated mature monkey, the proportions were 66% and 76%. NPY immunoreactivity was not detected in cross-sections of the inferior ramus of the oculomotor nerve or in longitudinal sections of the parasympathetic root supplying the ganglion. The localization of NPY immunoreactivity to nerve processes surrounding ciliary ganglion cells contrasts notably with its localization in the other ganglia supplying the rhesus monkey eye. We found NPY immunoreactivity in the postganglionic neurons and not in preganglionic processes of both the sympathetic superior cervical ganglion and the parasympathetic pterygopalatine ganglion. These observations are fully consistent with other reports. 9 ' 22 In the superior cervical ganglion, approximately half of the cell somata were NPY-LI; almost all cells were TH positive (Figs. 2A, 2B) and DBH positive (data not shown). In the pterygopalatine ganglion, approximately half of the cells were NPY-LI and almost all cells were VIP-LI (Figs. 2C, 2D). Neuropeptide Y immunoreactivity was not detected in sensory neurons of the trigeminal ganglion, but CGRP-LI and SP-LI cells were present (data not shown). In the ciliary ganglion, the NPY-LI nerve processes surrounding ganglion cells were not immunoreactive for TH or DBH, the adrenergic marker enzymes that colocalize with NPY in postganglionic sympathetic neurons (Fig. 3). However, almost all the cholinergic ciliary ganglion cells (>90%) displayed TH immunoreactivity (Fig. 3A), and rare ciliary ganglion cells (<1%) were immunoreactive for DBH (Figs. 3B, 3Q. Small bundles of NPY-, TH-, or DBH-LI nerve fibers, presumably postganglionic sympathetic axons, were present in nasociliary nerve branches entering the ganglion.

3 IOVS, February 1998, Vol. 39, No. 2 Neuropeptide Y-Like Immunoreactivity in Monkey Ciliary Ganglion 229 FIGURE 1. Neuropeptide Y (NPY) immunoreactivity in the ciliary ganglion. (A) Most ganglion cell somata are surrounded by NPY-like Lmmunoreactive (LI) nerve processes. Where the perisomatic region is sectioned tangentiaily, a complex array of stainedfibersand bulbous profiles is visible {arrow). Small granules in the ganglion cell nuclei are nonspecifically stained. (B) A small proportion of ciliary neurons scattered throughout the ganglion is not invested with NPY-LI nerve processes (arrows). Some thick immunoreactive nerve fibers (arrowheads) course within die ganglion. (C) In sections incubated with preabsorbed antiserum, staining of nerve processes is absent, but nonspecific staining of granular structures in the cell nuclei persists (arrows). Magnification bar, 50 /nm. VIP immunoreactivity, which coexists with NPY immunoreactivity in pterygopalatine ganglion cells, was not detected in any neurons of the ciliary ganglia (data not shown). VIP-LI varicose nerve fibers, always seen around adjacent blood vessels, provided evidence of effective immunohistochernical staining. CGRP and SP immunoreactivity, identified in sensory neurons of the trigeminal ganglion, localized to axons of nasociliary nerve branches joining the ciliary ganglion and in bundles of axons traversing the ganglion (Fig % 4). In addition, a small number of isolated varicose nerve fibers immunoreactive to either peptide were diffusely distributed throughout the ganglion; rarely, these varicose fibers partially or completely surrounded a ciliary ganglion cell (Fig. 4). No ciliary ganglion cell somata stained positively for either CGRP or SP. DISCUSSION The present results demonstrate dense networks of NPY-LI nerve fibers surrounding approximately 80% of rhesus monkey ciliary ganglion cells. Unilateral aphakia or occlusion of the eye in the young monkeys studied did not cause a difference in the distribution of NPY immunoreactivity between ipsilateral and contralateral ciliary ganglia. Demonstration of the same immunohistochemical staining pattern in the ciliary ganglia of an untreated monkey indicates that the unilateral ocular procedures in the experimental monkeys did not induce expression of NPY-like immunoreactivity in both ganglia. The distribution of NPY-LI nerve fibers, clustered densely around most ciliary ganglion cell somata, suggests they are terminals of preganglionic parasympathetic neurons located in the mesencephalon. NPY-LI axons were not detected in the oculomotor nerve supplying the ganglion, but it is possible that the amount of peptide in preterminal axons is below the level detectable with immunohistochemical analysis. In support of a mesencephalic origin, the NPY-positive nerves investing ciliary ganglion cells clearly differ from NPY-positive nerves that join or potentially join the ganglion through connections from the superior cervical, pterygopalatine, or trigeminal ganglia. Superior cervical ganglion cells immunoreactive for NPY are also immunoreactive for TH and DBH, enzymes present in all postganglionic sympathetic neurons. Although bundles of sympathetic axons joining the ciliary ganglion are NPY-, TH-, and DBH-LI and many ciliary ganglion cell somata are TH-positive, the NPY-LI nerves surrounding ciliary ganglion cells are TH and DBH negative. Many pterygopalatine neurons also are NPY-LI, but almost all are immunoreactive for VIP, indicating extensive, if not complete, coexistence of the two inimunoreactivities in these pterygopalatine neurons. Because we found no VIP immunoreactivity in the ciliary ganglia, a major projection of NPY-positive nerves from the pterygopalatine ganglion to ciliary ganglion cells seems unlikely. No NPY immunoreactivity

4 230 Grimes et al. IOVS, February 1998, Vol. 39, No. 2 FIGURE 2. Immunoreactivity to neuropeptide Y (NPY) and other markers in the superior cervical ganglion (SCG) and pterygopalatine ganglion (PPG). (A) In the SCG, NPY-like immunoreactive (LI) granules are present in the perikaryal cytoplasm of approximately 50% of the sympathetic ganglion cells, and (B) almost all cells are tyrosine hydroxylase (TH) positive. (C) NPY-LI granules are present in approximately half of the parasympathetic ganglion cells of the PPG, whereas (D) approximately 90% of the cells demonstrate varying levels of vasoactive intestinal peptide (VIP) immunoreactivity. Magnification bar, 50 /xm. FIGURE 3. Tyrosine hydroxylase (TH) and dopamine-js-hydroxylase (DBH) immunoreactivity in the ciliary ganglion, (A) Nearly all ciliary ganglion cells are TH positive with staining varying from bright to dim. Rare, dim cells were considered negative. No staining is evident in the perisomatic nerve processes, although many TH-positive smooth nerve fibers course between the cell somata, (B, C) DBH immunoreactivity is detected only in a few scattered ganglion cells. Magnification bar, 50 fun.

5 IOVS, February 1998, Vol. 39, No. 2 Neuropeptide Y-Like Immunoreactivity in Monkey Ciliary Ganglion 231 FIGURE 4. Calcitonin gene-related peptide (CGRP) and substance P (SP) immunoreactivity in the ciliary ganglion. (A) CGRP-like immunoreactive (LI) nervefibersare prominent in a nasociliary nerve branch entering the ganglion (arrowheads). A few bright varicose fibers course between ganglion cells. Some immunoreactive processes cluster around one small ganglion cell (arrow). (B) SP-LI varicose nerve fibers cluster irregularly around two adjacent cell somata (arrows). Some axons of moderate staining intensity are visible in a small nasociliary nerve branch entering the ganglion (arrowheads). Magnification bar, 50 ;nm. was identified in sensory neurons of the trigeminal ganglion in these rhesus monkeys, a rinding consistent with observations of others in cynomolgus monkeys 9 and in the Japanese macaque 22 and excluding a sensory origin for the NPY-positive terminals in the ciliary ganglion. A major NPY-LI preganglionic input has not been identified in previous inimunohistochemical studies of the ciliary ganglion of any species, including that of cynomolgus monkey 9 and humans. 14 In cynomolgus monkeys, however, van der Werf 9 observed NPY-LI nerve fibers surrounding some ciliary ganglion cells and assumed that they originated from the pterygopalatine ganglion; an alternative origin from preganglionic mesencephalic neurons seems equally possible based on the present results in rhesus monkey. NPY-LI nerve fibers in the human ciliary ganglion are extremely rare and are not in contact with ciliary neurons. 14 Besides NPY immunoreactivity, other differences are evident between primates in the immunohistochemistry of the ciliary ganglion with respect to both neural somata and nerve fiber staining. Although one study 10 reported SP immunoreactivity in approximately 60% and 40% of ciliary neurons of the cynomolgus monkey and the cat, respectively, this localization had not been detected in either animal by other investigators. 6 ' 9 Approximately 30% of ciliary ganglion cell somata have been reported to be V1P-LI in the cynomolgus monkey. 9 In comparison, cell somata in the ciliary ganglion of rhesus monkeys and humans 14 are negative for all neuropeptides examined, including SP, VIP, CGRP, and NPY. One consistent characteristic of primates is the presence of many TH-positive-DBH-negative ciliary ganglion cells. In rhesus, cynomolgus,' 2 and Japanese macaque 1 ' monkeys, these cells comprise 70% to 90% of the ganglion cell population; in the human ciliary ganglion, 25% of cell somata are TH positive and DBH negative. 14 The function of TH in cholinergic ciliary neurons of primates and of other mammals remains unknown. Differences also occur in the nerve fiber localization of neuropeptides within the ciliary ganglion. Although bundles of sensory CGRP- and SP-LI nerves from the nasociliary nerve enter and traverse the rhesus monkey ganglion, we rarely found CGRP- or SP-positive varicose axons surrounding ciliary neurons, hi cynomolgus monkeys, SP- but not CGRP-LI varicose nerve fibers invest a small proportion of ciliary ganglion cells. 910 In the human ciliary ganglion, 18% of cells are contacted by SP-LI varicose axons and 12% by CGRP-LI axons, almost all of which were also SP positive. 14 Some of die sensory axons passing through the ciliary ganglion could give rise to the fine varicose nerves that contact the ganglion cells in these species; alternatively, the CGRP- and SP-LI perisomatic nerve fibers may originate from another source. Inimunohistochemical studies of neuropeptide distribution in the primate ciliary ganglion are still limited in number, and the discrepancies in results may be attributed in part to technical problems. The neuropeptides identified in neuronal elements of the ciliary ganglion in lower mammals, however, vary so markedly from one species to another that the observed diversity among primates may be valid. Although the distribution of specific neuropeptides in the preganglionic input to the ciliary ganglion of birds has been well demonstrated in both the ganglion and the Edinger-Westphal nucleus, 23 neuropeptides in the mesencephalic neurons projecting to the ciliary ganglion of primates or other mammals have not been described. Localization of candidate neuropeptides, including NPY, in these neurons may clarify the functional organization of the autonomic pathway regulating pupillary movement and accommodation. References 1. Kuwayama Y, Grimes PA, Ponte B, Stone RA. Autonomic neurons supplying the rat eye and the intraorbital distribution of vasoactive intestinal polypeptide (VIP>like immunoreactivity. Exp Eye ties. 1987;44:

6 232 Grimes et al. IOVS, February 1998, Vol. 39, No Leblanc GG, Trimmer BA, Landis SC. Neuropeptide Y-like immunoreactivity in rat cranial parasympathetic neurons: coexistence with vasoactive intestinal peptide and choline acetyltransferase. Proc Natl Acad Sci USA. 1987;84: Stone RA, McGlinn AM, Kuwayama Y, Grimes PA. Peptide immunoreactivity of the ciliary ganglion and its accessory cells in the rat. Brain Res. 1988;475: Elsas T, Uddman R, Mulder H, Sundler F. Pituitary adenylate cyclase activating polypeptide and nitric oxide synthase are expressed in the rat ciliary ganglion. BrJ Ophthalmol. 1997;81: Grimes PA, McGlinn AM, Koeberlein B, Stone RA. Galanin immunoreactivity in autonomic innervation of the cat eye. / Comp Neurol. 1994;348: Grimes PA, McGlinn AM, Stone RA. An immunohistochemically distinct population of cat ciliary ganglion cells. Brain Res. 1990; 535: Kondo H, Katayama Y, Yui R. On the occurrence and physiological effect of somatostatin in the ciliary ganglion of cats. Brain Res. 1982;247: Warwick R. The ocular parasympathetic nerve supply and its mesencephalic sources. J Anat. 1954;88: van der Werf F. Autonomic and sensor)' innervation of some orbital structures in the primate. Universiteit van Amsterdam Thesis. 10. Zhang YL, Tan CK, Wong WC. Localization of substance P-like immunoreactivity in the ciliary ganglia of monkey (Macaca fascicularis) and cat: a light- and electron-microscopic study. Cell Tissue Res. 1994;276:l Uemura Y, Sugimoto T, Nomura S, Nagatsu I, Mizuno N. Tyrosine hydroxylase-like immunoreactivity and catecholamine fluorescence in ciliary ganglion neurons. Brain Res. 1987;4l6: Tan CK, Zhang YL, Wong WC. A light- and electron microscopic study of tyrosine hydroxylase-like immunoreactivity in the ciliary ganglia of monkey (Macaca fascicularis) and cat. Histol Histopathol. 1995;10: Landis SC, Jackson PC, Fredieu JR, Thibault J. Catecholaminergic properties of cholinergic neurons and synapses in adult rat ciliary ganglion. / Neurosci. 1987;7: Kirch W, Neuhuber W, Tamm ET. Immunohistochemical localization of neuropeptides in the human ciliary ganglion. Brain Res. 1995;681: Erichsen JT, Reiner A, Karten H. Co-occurrence of substance P-like and leu-enkephalin-like immunoreactivities in neurones and fibers of avian nervous system. Nature. 1982;295: Erichsen JT, Karten H, Eldred W, Brecha N. Localization of substance P-like and enkephalin-like immunoreactivity within preganglionic terminals of the avian ciliary ganglion: light and electron microscopy. / Neurosci. 1982;2: Reiner A. A VIP-like peptide co-occurs with substance P and enkephalin in cholinergic preganglionic terminals of the avian ciliary ganglion. Neurosci Lett. 1987;78: Epstein M, Davis J, Gellman L, Lamb J, Dahl J. Cholinergic neurons of the chicken ciliary ganglion contain somatostatin. Neurosdence. 1988;25: Zamboni L, De Martino C. Buffered picric-acid formaldehyde; a new, rapid fixative for electron microscopy. / Cell Biol. 1967;35: 148A. 20. Grimes PA, von Sallmann L. Comparative anatomy of the ciliary nerves. Arch Ophthalmol. 1960;64: Ruskell GL. An ocular parasympathetic nerve pathway of facial nerve origin and its influence on intraocular pressure. Exp Eye Res. 197O;1O:319-33O. 22. Matsumoto Y, Tanabe T, Ueda S, Kawata M. Immunohistochemical and enzymehistochemical studies of peptidergic, aminergic and cholinergic innervation of the lacrimal gland of the monkey (Macaca fuscatd). J Auton Nerv Syst. 1992;37:207-2l Reiner A, Erichsen JT, Cabot JB, Evinger C, Fitzgerald M, Karten HJ. Neurotransmitter organization of the nucleus of Edinger-Westphal and its projection to the avian ciliary ganglion. Vis Neurosci. 1991;6: