Immunohistochemical Classification and Functional Morphology Of

ARTICLES Immunohistochemical Classiﬁcation and Functional Morphology of Human Choroidal Ganglion Cells

Christian Albrecht May,1 Winfried Neuhuber,2 and Elke Lu¨tjen-Drecoll1

PURPOSE. To characterize human choroidal ganglion cells Approximately, 2000 of these neurons are located in the inner (CGCs) further, regarding their immunohistochemical and ul- portion of the ciliary muscle and presumably are involved in trastructural appearance and their pre- and postsynaptic con- fine regulation of the accommodation.3 In addition, approxi- nections. mately 2000 neurons are present in the choroid. Most of the METHODS. Choroidal wholemounts and serial sections of human postganglionic nerve fibers of these choroidal ganglion cells (CGCs) join the perivascular nerve fiber plexus that supports donor eyes were stained with antibodies against neuronal ni- 1,2 tric oxide synthase (nNOS), vasoactive intestinal peptide (VIP), the vasodilative innervation of the choroidal vasculature. A tyrosine hydroxylase (TH), vesicular monoaminergic trans- small number of these CGCs is located within the ciliary porter (VMAT)-2, vesicular acetylcholine transporter (VAChT), nerves. The postganglionic nerve fibers of this group of neurons do not enter the choroid but join the nerve fiber plexus neuropeptide Y (NPY), substance P (SP), calcitonin gene-re- 4 lated peptide (CGRP), calretinin, galanin, synaptophysin, and that innervates the outflow system. ␣-smooth muscle actin. Ultrathin sections of glutaraldehyde- When different mammalian eyes were compared, substan- fixed eyes were studied with an electron microscope. tial numbers of CGCs in the posterior choroid were found only in higher primates and human eyes containing a fovea centra- RESULTS. All CGCs stained for nNOS, most for VIP, approxi- lis.5 CGCs are also present in a number of birds with a special- mately 45% for calretinin, and only single neurons for NPY and ized accommodative system and one or more foveae.6,7 Further galanin. Ultrastructurally, the CGCs showed an incomplete characterization of the CGCs in the duck revealed that in glial sheath and, in places, showed close contact to surround- addition to NO and VIP, the CGCs express galanin. The post- ing collagen fibrils. The CGCs were in close contact with ganglionic fibers of these cells innervate not only the choroidal numerous boutons staining for the different neuronal markers vasculature, but also the numerous non–vascular smooth mus- including synaptophysin, nNOS, VIP, NPY, TH, VMAT-2, cle cells (NVSMCs) present in the avian choroidal stroma.8 VAChT, calretinin, and NPY. Duck CGCs are surrounded by tyrosine hydroxylase (TH)/ CONCLUSIONS. The data indicate a complex integrative function dopamine-␤-hydrolase (DBH)–immunoreactive nerve fibers of CGCs. The immunohistochemical and ultrastructural char- forming synaptic contacts with the CGCs.9 Evidence has acteristics also indicate that the CGCs may have mechanosen- shown that these fibers derive from the superior cervical gan- sory properties. The complex synaptic information points to a glion. The CGCs in the duck also receive calcitonin gene- specific regulative CGC function in parallel with ciliary muscle related peptide (CGRP)–positive efferent collaterals of trigem- contraction (accommodation). Axons originating from CGCs inal afferents that may indicate precentral reflex arcs.10 mainly supply the choroidal vasculature, thus implicating the Innervation of CGCs by postsynaptic sympathetic fibers in the CGCs as vasodilative neurons, but single CGCs may also inner- avian eye, however, indicates a more complex integrative func- vate other structures such as nonvascular choroidal smooth tion of the ganglion cell plexus, similar to that in the enteric muscle cells. (Invest Ophthalmol Vis Sci. 2004;45:361–367) nervous system.9 DOI:10.1167/iovs.03-0624 It is not yet known whether CGCs in primate eyes have a similar complex function. In the present study we investigated t is well established that the uvea in most species is inner- human CGCs to clarify which neurotransmitters are expressed Ivated by two parasympathetic pathways, namely by nerves by these cells, to characterize their presynaptic input, and to deriving from the ciliary ganglion and by those deriving from define their target tissue. the sphenopalatine ganglion. Within the uvea of human eyes, an additional group of nitric MATERIAL AND METHODS oxide synthase (NOS) and vasoactive intestinal peptide (VIP)- immunoreactive (IR) neurons has recently been discovered.1–3 The studies were performed in 50 human eyes with no previous clinical history of ocular disease, the donors ranging in age between 12 and 95 years. The eyes were obtained during autopsy 4 to 36 hours

1 2 after death. All donors had given permission to use their tissues for From the Departments of Anatomy II, and Anatomy I, Friedrich- research. Tissue observation was in accordance with the Declaration of Alexander University, Erlangen, Germany. Helsinki and the local regulations. Eyes with obvious signs of ocular Supported by German Research Community SFB 539 BII,2. Submitted for publication June 20, 2003; revised September 29, disease (e.g., macular degeneration, loss of retinal pigmented epithe- 2003; accepted October 26, 2003. lium) or ocular treatment other than artiﬁcial lens implantation (e.g., Disclosure: C.A. May, None; W. Neuhuber, None; E. Lu¨tjen- laser-treatment of the retina) were excluded from the study. A further Drecoll, None histologic screening for ocular disease was not performed. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertise- Immunohistochemistry ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Corresponding author: Christian Albrecht May, Anatomisches Thirty-seven eyes of human donors ranging in age between 12 and 95 Institut II, Universita¨tsstr. 19, D-91054 Erlangen, Germany; years were observed 11 to 36 hours after death. All eyes were incised [email protected]. equatorially and ﬁxed either for 4 hours in neutral buffered 4% para-

Downloaded from iovs.arvojournals.org on 09/28/2021 362 May et al. IOVS, February 2004, Vol. 45, No. 2

TABLE 1. Speciﬁcation of the Primary Antibodies Used in the Study

Antibody Host Dilution Source

Panneuronal marker PGP 9.5 Rabbit 1:200 Biotrend, Colone, Germany Tyrosine hydroxylase Rabbit 1:400 Chemicon, Hofheim, Germany Tyrosine hydroxylase Mouse 1:100 Chemicon Dopamine-␤-hydroxylase Rabbit 1:150 Afﬁnity, Nottingham, UK Vesicular monoaminergic transporter 2 Rabbit 1:500 Biomar, Marburg, Germany Vesicular acetylcholine transporter Rabbit 1:1000 Biomar Neuronal nitric oxide synthase Rabbit 1:400 B. Mayer, Graz, Austria Vasoactive intestinal peptide Rabbit 1:400 Biotrend Neuropeptide Y Rabbit 1:200 Biotrend Substance P Rabbit 1:500 Biotrend Calcitonin gene-related peptide Rabbit 1:250 Afﬁnity Galanin Rabbit 1:400 Chemicon Synaptophysin Mouse 1:500 Dako, Glostrup, Denmark Calretinin Rabbit 1:500 Biotrend Calretinin Goat 1:1000 Swant, Bellinzona, CH ␣-Smooth muscle actin Mouse 1:200 Sigma, St. Louis, USA Smooth muscle myosin Rabbit 1:250 Biotrend

formaldehyde (PFA) or for 12 hours in Zamboni fixative, containing 4% thick) were performed and stained with toluidine blue. From localized PFA and 0.01% picric acid. The tissue was then rinsed in phosphate- CGCs, ultrathin sections were performed, stained with leaded citrate buffered saline (PBS, pH 7.4) several times. In 27 eyes, the posterior and uranyl acetate, and viewed with a transmission electron micro- segments were divided into four quadrants and wholemounts of the scope (model EM 902; Carl Zeiss Meditec, Oberkochen, Germany). choroid and sclera were performed. From some of the wholemounts and from the remaining eyes, serial 14- to 16-␮m-thick sagittal and ESULTS tangential sections were cut through the choroid and sclera with a R cryostat (Leica, Bensheim, Germany) and mounted on poly-L-lysine– Choroidal Ganglion Cells coated glass slides. Incubation with the primary antibody was performed overnight at In human choroidal sections and wholemounts incubated with room temperature, with the antibody diluted in PBS containing 1% antibodies against neuronal NOS (nNOS), all 1820 visualized bovine serum albumin and 0.1% Triton X. Control experiments were CGCs showed positive staining (Fig. 1A). No additional CGCs, performed by incubating the sections only with the dilution solution. either identified by their content of autofluorescent lipofuscin The primary antibodies used are listed in Table 1. The sections and or by a nonspecific light fluorescent background in the neuro- wholemounts were then rinsed in PBS and incubated for 1 hour with nal cytoplasm, were detected in the nNOS-stained specimens. an appropriate fluorescent-dye–conjugated secondary antibody, di- The nNOS staining was independent of the size (10–37 ␮min luted in PBS (Dianova, Hamburg, Germany). In the double-staining procedure, the steps were the same, but the incubation time of the primary antibody was reduced to 4 to 6 hours. Most of the antibodies showed sufficient staining in all specimens obtained, regardless of the donor age, time since death, or fixation. For some antibodies, certain special conditions must exist in the tissue to promote sufficient staining. Therefore, the number of tissue specimens was limited for these markers. The antibody against VIP showed sufficient staining of CGCs only in tissue fixed for 24 hours in Zamboni fixative and with a time since death of less than 24 hours. PFA-fixed specimens showed axonal VIP staining but no staining of the CGCs. Staining with the antibody against neuropeptide Y (NPY) was only possible by using cryosections of both PFA- and Zamboni-fixed specimens. Wholemounts did not show sufficient staining of either neurons or nerve fibers. All sections were mounted with Kaiser glycerol jelly and viewed with a fluorescence microscope (Aristoplan; Leica) or with a confocal laser scanning system (MRC 1000; Bio-Rad, Munich, Germany) at- tached to an inverted microscope (Diaphot 300; Nikon, Tokyo, Japan). In estimating the size of single neurons, a diameter was defined as the shortest distance between opposing cell boundaries, when passing through the center of the cell. Electron Microscopy The ultrastructural morphology and the surrounding tissue of the CGCs was investigated in 13 eyes of donors ranging in age between 56 and 91 years. The eyes were obtained 4 to 24 hours after death, incised FIGURE 1. (A, B) Double-immunofluorescent staining of two human equatorially, and fixed in 4% PFA and 1% glutaraldehyde for at least 24 choroidal ganglion cells for nNOS (A) and VIP (B). (C) A human hours. choroidal flatmount stained with antibodies against calretinin shows a Small samples of the choroid and sclera were prepared and embed- positive neuronal network and some immunoreactive neurons. Origi- ded in Epon. Sagittal and tangential semithin serial sections (1 ␮m nal magnification: (A, B) ϫ400; (C) ϫ200.

Downloaded from iovs.arvojournals.org on 09/28/2021 IOVS, February 2004, Vol. 45, No. 2 Human Choroidal Ganglion Cells 363

vealed that 82 of the 146 CGCs evaluated were calretinin IR (Fig. 1C). In some cases, these CGCs were preferentially of larger size (24–36 ␮m in diameter), but there were also other cases showing mainly small calretinin-positive cells (17–21 ␮m in diameter). Only few small CGCs (14–18 ␮m in diameter) were NPY IR (7/152 evaluated CGCs). Single larger CGCs (25–34 ␮min diameter) revealed faint staining with antibodies against galanin (3/189 evaluated CGCs). Most of the neurons in NPY- and galanin-stained sections, however, were unstained. There was no staining of CGCs with the antibodies against TH, DBH, vesicular monoaminergic transporter (VMAT)-2, vesicular acetylcholine transporter (VAChT), substance P (SP), and CGRP (total number of evaluated CGCs: 1340). Ultrastructural investigations showed, that the CGCs appeared as typical neurons with a large nucleus containing light-appearing chromatin and a clear nucleolus and a cytoplasm containing ribosomes, stacks of rough endoplasmic reticulum, and mitochondria (Fig. 2A). In CGCs of aged donor eyes, lipofuscin granules were frequently seen within the cytoplasm of the neurons. numerous vesicle-filled boutons were present around the CGC, forming synaptic contacts with the neuron (Fig. 2B). The surface of the CGCs toward the surrounding connective tissue space was only incompletely covered by glial cell processes. In places, the basement membrane FIGURE 2. Ultrastructural appearance of human CGCs. (A) The nu- of the neurons directly contacted the adjacent collagen fibrils cleus (ZK) contained light-appearing chromatin, and the cytoplasm (Fig. 2C) without forming specialized contacts, such as with contained stacks of rough endoplasmic reticulum (rER) and mitochon- hemidesmosomes. Those CGCs arranged in doublets or clus- dria (M). (B) Numerous vesicle-filled boutons (arrows) were present ters of even more cells were partly apposed to each other around the CNC, forming synaptic contacts (S) with the neuron. (C)In without being completely separated by glial cell processes (Fig. places, the basement membrane of the neurons directly contacted the adjacent collagen fibrils (✱). (D) Those CGCs arranged in doublets or 2D). Neither single cells nor groups of CGCs were supplied by clusters of even more cells were partly apposed to each other without or adjacent to capillaries. All CGCs were embedded in a nerve being separated by glial cell processes (arrows). Original magnifica- fiber plexus. tions: (A) ϫ3,600; (B) ϫ26,000; (C) ϫ17,000; (D) ϫ25,000. Neuronal Boutons Apposed to CGCs diameter) and the location of the neurons. Numerous axons Staining of the choroid with antibodies against synaptophysin within the suprachoroidal nerve plexus and around choroidal revealed the presence of synaptic contacts around each CGC, vessels were also nNOS IR. independent of size (cell diameter) and localization (choroidal Staining with antibodies against VIP revealed that most of versus suprachoroidal; in the choroidal stroma versus in the the CGCs (354/362 CGCs evaluated) were VIP IR. Colocaliza- ciliary nerves). However, not all boutons around the CGCs tion was documented by double staining with antibodies revealed the same neurochemical composition (Table 2). against nNOS and VIP (Fig. 1B). It appeared that the few All CGCs showed appositions of nNOS and VIP IR boutons VIP-negative CGCs, visualized by autofluorescent lipofuscin or (Fig. 3A). Single boutons also stained for NPY (Fig. 3B). Cal- by a nonspecific light fluorescent background, were of smaller retinin IR boutons were seen only at CGCs that were not size (12–18 ␮m in diameter). calretinin IR themselves (Fig. 3C). However, only 42 of the 64 Less frequent IR in CGCs was seen in choroidal tissue calretinin-negative CGCs investigated had close calretinin-pos- stained with antibodies against calretinin. Wholemounts re- itive contacts. The calretinin-negative CGCs that showed no

TABLE 2. Staining of CGCs and Surrounding Boutons with the Neuronal Markers Used in the Study

Antibody CGC (n) % CGC Boutons

Neuronal nitric oxide synthase 1820/1820 100 ϩϩ Vasoactive intestinal peptide 354/362 97,8 ϩ to ϩϩ Calretinin 82/146 56,2 ϩ (29%) Neuropeptide Y 7/152 4,6 ϩ Galanin 3/189 1,6 0 Tyrosine hydroxylase 0/275 0 ϩϩ (50%) Dopamine-beta-hydroxylase 0/78 0 ϩ (50%) Vesicular monoaminergic transporter 2 0/234 0 ϩ to ϩϩ (60%) Vesicular acetylcholine transporter 0/348 0 ϩ to ϩϩ Substance P 0/215 0 0 Calcitonin gene-related peptide 0/190 0 0 synaptophysin 0 0 ϩϩ

The number of positive CGCs in comparison to all CGCs in the sections evaluated and the percentages of positive CGC are given. The number of surrounding boutons was semiquantiﬁed: 0, no boutons; ϩ, single boutons; ϩϩ, numerous boutons. If all CGCs did not have positive boutons apposed, the percentage of CGC with positive boutons is added in brackets.

Downloaded from iovs.arvojournals.org on 09/28/2021 364 May et al. IOVS, February 2004, Vol. 45, No. 2

FIGURE 3. Micrographs of human choroidal ganglion cells stained FIGURE 5. Confocal image of a human choroid section double stained with antibodies against (A) VIP), (B) NPY, and (C, D) calretinin. Note with antibodies against ␣-smooth muscle ␣-actin (green) and NPY the numerous positive appositions of VIP and NPY boutons. (C) Cal- (red), showing the close contact of NPY IR boutons to some of the retinin-positive boutons were seen in proximity with CGCs that did not NVSMCs in the choroid (arrows). stain for calretinin themselves. (D) Two calretinin-negative CGCs (ar- rowheads) that showed no calretinin-positive boutons but lipofuscin granules (arrows) were next to a calretinin-positive CGC (✱). Original magnification, ϫ400 bodies against VAChT. No staining of boutons in close apposi- tion to the CGCs was seen using antibodies against galanin, SP, and CGRP. calretinin IR boutons were always next to calretinin-positive CGCs (Fig. 3D). No colocalization occurred with antibodies Innervation of the Choroidal Vasculature against calretinin and VMAT-2 (Fig. 4A), whereas some but not all boutons showed colocalization with antibodies against cal- A dense perivascular network of nerve fibers was seen around retinin and VAChT (Fig. 4B). the arteries, arterioles, and veins when using antibodies against Approximately half of the CGCs showed an intense inner- VAChT, nNOS, or VIP as the marker for parasympathetic, and vation of TH-positive nerve fibers. One third of the TH-IR TH, DBH, VMAT-2, or NPY as the marker for sympathetic nerve boutons around the CGCs were colocalized with antibodies fibers. The larger arteries were also accompanied by CGRP- against VMAT-2, and others showed colocalization with anti- positive nerve fibers. SP-positive nerve fibers formed a fine network in the outer choroid without showing close affiliation to the vessels but rather contacting cells in the choroidal stroma. In the inner choroid, only single TH/VMAT-2/NPY- positive nerve fibers reached the choriocapillary layer, forming a fine network at the level of the precapillary arterioles.

Innervation of NVSMCs NVSMCs were identified by their general content of ␣-smooth muscle actin. A subgroup of these cells (15%) also stained for smooth muscle myosin, as described previously.11 The general presence of nerve fibers in close proximity to NVSMCs was detected by double staining of choroidal sections and wholemounts with antibodies against ␣-smooth muscle actin and PGP 9.5. However, not every muscle cell was innervated separately. Approximately 20% of all NVSMCs showed close contact to nerve fibers. As the choroid is densely innervated and many NVSMCs also run parallel with nerve fiber bundles, the number of innervated cells could only be estimated. Double staining with antibodies against smooth muscle myosin and PGP 9.5 showed that close contact of the NVSMCs with nerve fibers was not exclusively characteristic of the myosin-containing NVSMCs. FIGURE 4. (A) Double immunostaining for calretinin (red) and vesic- Specific neuronal markers that stained nerve fibers with ular monoaminergic transporter 2 (green) showed no colocalization in boutons in close contact to the NVSMCs included NPY and boutons. (B) Double-immunostaining for calretinin (red) and vesicular VAChT. Choroidal wholemounts double stained for ␣-smooth acetylcholine transporter (green) showed only slight colocalization in ␣ boutons (yellow). The intracellular yellow granules in the double- muscle actin and NPY (Fig. 5) or -smooth muscle actin and stained images refer to autofluorescence of the lipofuscin (arrows), VAChT and investigated by confocal microscopy revealed that also marked by arrows in the single-stained images. Original magnifi- the same estimated number of NVSMCs as for PGP 9.5 received cation: (A) ϫ400; (B) ϫ630. input from NPY- and VAChT-positive nerve fibers. All other

Downloaded from iovs.arvojournals.org on 09/28/2021 IOVS, February 2004, Vol. 45, No. 2 Human Choroidal Ganglion Cells 365

antibodies tested, including nNOS, TH, VMAT-2, VIP, SP, and Another hint of the possible existence of a mechanosensory CGRP, showed no nerve fibers in close relation to the NVSMCs. subpopulation of CGCs is calretinin immunoreactivity in some of them. This calcium-binding protein is typically contained in low-threshold spinal mechanosensors (e.g., muscle spindles) in 38 DISCUSSION both rat and chicken, in vagal mechanosensors in the rat esophagus,39 and in intrinsic primary afferent type II neurons In the present study we showed that numerous CGCs stain not in the human small intestine. However, the predictive value of only for nNOS and VIP,1,2,12 but also show immunoreactivity immunocytochemical neuronal markers for neuronal function for antibodies against calretinin. Single CGCs were NPY IR. should be discussed with great care because of significant Ultrastructurally, the CGCs showed an incomplete glial sheath species differences.41,42 and, in places, close contact to surrounding collagen fibrils. The second major finding of this study was the character- The CGCs were in close contact with numerous boutons stain- ization of the numerous synaptic contacts on the CGCs. Be- ing for different neuronal markers, including synaptophysin, sides calretinin, several neurotransmitters were found in the nNOS, VIP, NPY, TH, VMAT-2, VAChT, calretinin, and NPY. synaptic boutons in close contact with the CGCs. Calretinin-, Because we observed NPY-positive nerve fibers supplying the VIP-, and nNOS-positive boutons probably demonstrate an in- NVSMCs, it is tempting to speculate that the NPY-positive ter-CGCs connection, although they could also derive from the CGCs may regulate the NVSMCs. sphenopalatine ganglion. The distribution of cholinergic bou- In mammals without a fovea centralis, the vasodilative tons differed markedly from that seen for VIP and nNOS. innervation of the choroid derives from the sphenopalatine Because virtually all sphenopalatine ganglion neurons express ganglion,13–19 similar to mammals with foveas, including one of these markers, we concluded that the source of the humans. To some extent, the CGCs in the human choroid cholinergic boutons is most likely not only the sphenopalatine also show similarities to the neurons of the sphenopalatine ganglion. nNOS-positive neurons also exist in the trigeminal ganglion. There, most of the neurons stain for nNOS and ganglion,43 but they do not express VIP and most of the VIP,20–22 and single neurons stain for NPY.22 A colocaliza- trigeminal neurons express SP and CGRP.44 As these markers tion of NPY and VIP is also reported for some neurons in the do not stain CGCs synaptic boutons, the trigeminal ganglion human sphenopalatine ganglion.23 However, in contrast to seems to play no role in regulation of human CGCs. Most of the the mammalian sphenopalatine ganglion, where numerous superior cervical ganglion neurons stain positively for NPY,44 neurons also contain acetylcholinesterase,24 choline acetyl- noradrenalin, TH, and DBH.22,45 Most of the TH/VMAT-2 pos- transferase and acetylcholine,25 no staining for VAChT or itive endings around the CGCs derive from this source, similar choline acetyltransferase was present in human CGCs.4 The to the duck.9 These contacts point to a more complex regula- CGCs therefore differ from sphenopalatine ganglion neu- tion of the CGCs that is generally not common to other intrin- rons in regard to their cholinergic immunohistochemical sic neurons beside the enteric nervous system.46 Presumably staining. Whether this failure of staining in human tissue is these contacts are formed to restrain the vasoconstrictive ef- due to the sensitivity of the transmitters or to a true func- fect of the sympathetic nerve fibers. tional diversity between CGCs and sphenopalatine ganglion Within the ciliary ganglion, various neurons exist that neurons remains to be determined. also show great species differences. In the human, approx- It is unknown, why primates have neurons in the choroid. imately 23% stain positively for TH,43 whereas only single In this context it is interesting to note that the CGCs reveal cells stain for nNOS.47 All ciliary ganglion neurons are glia-free areas, covered only by a basement membrane, that thought to be cholinergic. Therefore, the VAChT-positive– show contact to the surrounding connective tissue. An incom- NOS-negative synaptic contacts on CGCs could also derive plete glial sheath is typical for enteric neurons in the gut and from this ganglion. As the ciliary ganglion mainly supplies gallbladder,26,27 but not for other autonomic ganglia.28 Direct the ciliary muscle, one can speculate that, during ciliary exposure of enteric neurons to periganglionic matrix has been muscle contraction for accommodation, the CGCs receive a discussed in the context of a possible mechanosensory func- copy of this information and therefore can modulate the tion.26 There exist electrophysiological data on direct mech- volume of the choroid to keep the fovea centralis in place. anosensitivity of enteric neurons,29,30 and intrinsic primary This could explain, why choroidal ganglion cells occur only afferent neurons in enteric ganglia have been characterized in species with a well-developed accommodative system and recently by both morphologic and electrophysiological meth- a fovea centralis. In support of this evidence, one can add ods.31 Furthermore, vagal afferent terminals in the gastrointes- that the cholinergic nerve fibers are in close contact with tinal tract are typically associated with enteric ganglia32 where the NVSMCs. In contrast to paraffin-embedded sections de- they are located immediately beneath the basement membrane scribed in the literature,48 our careful examination of frozen facing the periganglionic collagen matrix.33 Their low thresh- serial sections revealed that the NVSMCs in the human old mechanosensory function has been elegantly demonstrated choroid do not receive input from nNOS- or VIP-positive by recent electrophysiological studies.34,35 Thus, it is tempting nerve fibers. Rather, they are innervated by NPY-VAChT– to speculate that, in addition to motor neurons, choroid ganglia positive axons. However, CGCs may not represent the main harbor sensory neurons, as do enteric ganglia, and that their source of these axons. This differs from the avian eye, where intimate relationship to the periganglionic matrix hints of a a clear innervation of NVSMCs deriving from the choroidal mechanosensory subpopulation of CGCs. ganglion cells could be demonstrated.8 Human CGCs also do Many of the processes issued by CGCs are targeted to not have CGRP-positive endings being present in the avian choroidal arteries where they form a dense perivascular eye and leading to the postulation of a local precentral reflex plexus. Although hitherto interpreted in favor of a vasomotor arc.9,10 Regarding these differences, the function of the function,2,5 some of them may represent mechanosensory ter- CGCs in birds may be somewhat different from that in the minals similar to carotid and aortic baroreceptors of the IXth human eye. and Xth cranial nerves, respectively.36,37 In conjunction with In conclusion, we found a rather homogenous staining indications for intrinsic CGC–CGC connections in both pri- pattern of CGCs in the human choroid that showed similarities mates (this study, discussed below) and birds,8 this could be to the pterygopalatine ganglion. In addition, ultrastructural interpreted in terms of an intrinsic neuronal mechanism for findings and immunofluorescence for calretinin point to a regulation of choroidal blood flow. possible mechanoreceptor-like function of the CGCs them-

Downloaded from iovs.arvojournals.org on 09/28/2021 366 May et al. IOVS, February 2004, Vol. 45, No. 2

selves, thus explaining the location of these neurons in the 16. Beckers HJ, Klooster J, Vrensen GF, Lamers WP. Facial parasym- choroid proper. The target of the CGCs in the human eye is pathetic innervation of the rat choroid, lacrimal glands and ciliary mainly the choroidal vasculature. This could recently be con- ganglion: an ultrastructural pterygopalatine tracing and immuno- firmed by intracellular staining of the processes of human histochemical study. Ophthalmic Res. 1993;25:319–330. CGCs with neurobiotin.49 The supply of the CGCs to the 17. Yamamoto R, Bredt DS, Snyder SH, Stone RA. The localization of nitric oxide synthase in the rat eye and related cranial ganglia. choroidal vasculature is regarded as supplementary to a direct Neuroscience. 1993;54:189–200. vasodilative innervation of the sphenopalatine ganglion. Be- 18. Klooster J, Beckers HJ, Ten Tusscher MP, Vrensen GF, van der cause of their complex synaptic input the CGCs seem to form Want JJ, Lamers WP. Sympathetic innervation of the rat choroid: a highly regulated intrinsic system that reacts not only to an autoradiographic tracing and immunohistochemical study. Oph- parasympathetic and sympathetic information, but as an inte- thalmic Res. 1996;28:36–43. grated local vasoregulatory system for foveal vision during 19. Steinle JJ, Krizsan-Agbas D, Smith PG. Regional regulation of cho- accommodation. The importance of the choroid for proper roidal blood flow by autonomic innervation in the rat. Am J vision is also known during recovery from form-deprivation Physiol. 2000;279:R202–R209. myopia in the foveated avian, tree shrew, and primate 20. Motosugi H. Immunohistochemical observation of neurotransmit- eye.50–52 Neuronal involvement of cholinergic, NO-, and VIP- ters in rat and human pterygopalatine ganglia [in Japanese]. Nip- ergic nerve fibers seem to play key roles in this process.53–55 pon Jibiinkoka Gakkai Kaiho. 1993;96:936–945. Their source is not defined and may also include the CGCs. 21. Hanazawa T, Tanaka K, Chiba T, Konno A. Distribution and origin of nitric oxide synthase-containing nerve fibers in human nasal mucosa. Acta Otolaryngol. 1997;117:735–737. References 22. Grimes PA, Koeberlein B, Tigges M, Stone RA. Neuropeptide Y-like immunoreactivity localizes to preganglionic axon terminals in the 1. Bergua A, Ju¨nemann A, Naumann GOH. NADPH-D reaktive cho- rhesus monkey ciliary ganglion. Invest Ophthalmol Vis Sci. 1998; roidale Ganglienzellen beim Menschen. Klin Monatsbl Augen- 39:227–232. heilk. 1993;203:77–82. 23. Lacroix JS, Anggard A, Hokfelt T, O’Hare MM, Fahrenkrug J, Lun- 2. Flu¨gel C, Tamm ER, Mayer B, Lu¨tjen-Drecoll E. Species differences dberg JM. Neuropeptide Y: presence in sympathetic and parasym- in choroidal vasodilative innervation: evidence for specific intrin- pathetic innervation of the nasal mucosa. Cell Tissue Res. 1990; sic nitrergic and VIP-positive neurons in the human eye. Invest 259:119–128. Ophthalmol Vis Sci. 1994;35:592–599. 24. Uddman R, Malm L, Sundler F. Peptide containing nerves in the 3. Tamm ER, Flu¨gel-Koch C, Mayer B, Lu¨tjen-Drecoll E. Nerve cells nasal mucosa. Rhinology. 1981;19:75–79. in the human ciliary muscle: ultrastructural and immunocyto- 25. Suzuki N, Hardebo JE. The cerebrovascular parasympathetic inner- chemical characterization. Invest Ophthalmol Vis Sci. 1995;36: vation. Cerebrovasc Brain Metab Rev. 1993;5:33–46. 414– 426. 26. Gabella G. On the ultrastructure of the enteric nerve ganglia. 4. May CA, Fuchs AV, Scheib M, Lu¨tjen-Drecoll E. Characterization of Scand J Gastroenterol Suppl. 1982;71:15–25. nitrergic neurons in the porcine and human ciliary nerves. Invest 27. Cornbrooks EB, Pouliot WA, Mawe GM. Structure of neurons and Ophthalmol Vis Sci. 2002;43:581–586. ganglia of the guinea-pig gallbladder: an immunohistochemical 5. Flu¨gel-Koch C, Kaufman P, Lu¨tjen-Drecoll E. Association of a cho- study. J Comp Neurol. 1992;317:45–56. roidal ganglion cell plexus with the fovea centralis. Invest Oph- 28. Baluk P, Fujiwara T, Matsuda S. The fine structure of the ganglia of thalmol Vis Sci. 1994;35:4268–4272. the guinea-pig trachea. Cell Tissue Res. 1985;239:51–60. 6. Bergua A, Mayer B, Neuhuber WL. Nitrergic and VIPergic neurons 29. Wood JD. Physiology of the enteric nervous system. In: LR John- in the choroid and ciliary ganglion of the duck Anis carina. Anat son, ed. Physiology of the Gastrointestinal Tract. Vol 1. 2nd ed. Embryol. 1996;193:239–248. New York: Raven Press; 1987;67–109. 7. Cuthbertson S, Jackson B, Toledo C, et al. Innervation of orbital 30. Kunze WA, Clerc N, Furness JB, Gola M. The soma and neurites of and choroidal blood vessels by the pterygopalatine ganglion in primary afferent neurons in the guinea-pig intestine respond dif- pigeons. J Comp Neurol. 1997;386:422–442. ferentially to deformation. J Physiol. 2000;526:375–385. 8. Schro¨dl F, Brehmer A, Neuhuber WL. Intrinsic choroidal neurons 31. Furness JB, Kunze WAA, Bertrand PP, Clerc N, Bornstein JC. in the duck eye express galanin. J Comp Neurol. 2000;425:24–33. Intrinsic primary afferent neurons of the intestine. Prog Neuro- 9. Schro¨dl F, Tines R, Brehmer A, Neuhuber WL. Intrinsic choroidal biol. 1998;54:1–18. neurons in the duck eye receive sympathetic input: anatomical 32. Phillips RJ, Powley TL. Tension and stretch receptors in gastroin- evidence for adrenergic modulation of nitrergic functions in the testinal smooth muscle: re-evaluating vagal mechanoreceptor elec- choroid. Cell Tissue Res. 2001;304:175–184. trophysiology. Brain Res Brain Res Rev. 2000;34:1–26. 10. Schro¨dl F, Schweigert M, Brehmer A, Neuhuber WL. Intrinsic 33. Neuhuber WL. Sensory vagal innervation of the rat esophagus and neurons in the duck choroid are contacted by CGRP-immunoreac- cardia: a light and electron microscopic anterograde tracing study. tive nerve fibres: evidence for a local pre-central reflex arc in the J Auton Nerv Syst. 1987;20:243–255. eye. Exp Eye Res. 2001;72:137–146. 34. Zagorodnyuk VP, Brookes SJ. Transduction sites of vagal mechano- 11. Flu¨gel-Koch C, May CA, Lu¨tjen-Drecoll E. Presence of a contractile receptors in the guinea pig esophagus. J Neurosci. 2000;20:6249– cell network in the human choroid. Ophthalmologica. 1996;210: 6255. 296–302. 35. Zagorodnyuk VP, Chen BN, Brookes SJ. Intraganglionic laminar 12. Miller AS, Coster DJ, Costa M, Furness JB. Substance P and vaso- endings are mechano-transduction sites of vagal tension receptors active intestinal peptide immunoreactive nerves in the human and in the guinea-pig stomach. J Physiol. 2001;534:255–268. rat eye (Abstract). Neurosci Lett. 1982;8:S67. 36. Hohler B, Mayer B, Kummer W. Nitric oxide synthase in the rat 13. Uddman R, Alumets J, Ehinger B, Hakanson R, Loren I, Sundler carotid body and carotid sinus. Cell Tissue Res. 1994;276:559– F. Vasoactive intestinal peptide nerves in ocular and orbital 564. structures of the cat. Invest Ophthalmol Vis Sci. 1980;19:878– 37. Thames MD, Smith ML, Dibner-Dunlap ME, Minisi AJ. Reflexes 885. governing autonomic outflow to the heart. In: Nervous Control of 14. Butler JM, Ruskell GL, Cole DF, et al. Effects of VIIth (facial) nerve the Heart. Shepherd JT, Vatner SF, eds. Amsterdam: Harwood degeneration on vasoactive intestinal polypeptide and substance P Academic Publishers; 1996;295–327. levels in ocular and orbital tissues of the rabbit. Exp Eye Res. 38. Duc C, Barakat-Walter I, Droz B. Innervation of putative rapidly 1984;39:523–532. adapting mechanoreceptors by calbindin- and calretinin-immuno- 15. Ten Tusscher MP, Klooster J, Baljet B, Van der Werf F, Vrensen GF. reactive primary sensory neurons in the rat. Eur J Neurosci. 1994; Pre- and post-ganglionic nerve fibres of the pterygopalatine gan- 6:264–271. glion and their allocation to the eyeball of rats. Brain Res. 1990; 39. Du¨tsch M, Eichhorn U, Wo¨rl J, Wank M, Berthoud HR, Neuhuber 517:315–323. WL. Vagal and spinal afferent innervation of the rat esophagus: a

Downloaded from iovs.arvojournals.org on 09/28/2021 IOVS, February 2004, Vol. 45, No. 2 Human Choroidal Ganglion Cells 367

combined retrograde tracing and immunocytochemical study with 47. Demer JL, Poukens V, Miller JM, Micevych P. Innervation of ex- special emphasis on calcium-binding proteins. J Comp Neurol. traocular pulley smooth muscle in monkeys and humans. Invest 1998;398:289–307. Ophthalmol Vis Sci. 1997;38:1774–1785. 40. Brehmer A, Schro¨dl F, Neuhuber W. Quantitative morphological 48. Poukens V, Glasgow BJ, Demer JL. Nonvascular contractile cells in and immunohistochemical assessment of putative primary afferent sclera and choroid of humans and monkeys. Invest Ophthalmol (type II) versus type V myenteric neurons of human small intestine Vis Sci. 1998;39:1765–1774. (Abstract). Neurogastroenterol Motil. 2003;15:217A. 49. Schro¨dl F, De Laet A, Tassignon MJ, et al. Intrinsic choroidal 41. Timmermans JP, Hens J, Adriaensen D. Outer submucous plexus: neurons in the human eye: projections, targets and basic electro- an intrinsic nerve network involved in both secretory and motility physiological data. Invest Ophthalmol Vis Sci. 2003;44:3705– processes in the intestine of large mammals and humans. Anat 3712. Rec. 2001;262:71–78. 50. Wildsoet C, Wallman J. Choroidal and scleral mechanisms of com- 42. Brehmer A, Schro¨dl F, Neuhuber W. Morphological phenotyping pensation for spectacle lenses in chicks. Vision Res. 1995;35: of enteric neurons using neuroﬁlament immunohistochemistry 1175–1194. renders chemical phenotyping more precise in porcine ileum. 51. Wallman J, Wildsoet C, Xu A, et al. Moving the retina: choroidal Histochem Cell Biol. 2002;117:257–263. modulation of refractive state. Vision Res. 1995;35:37–50. 43. Kirch W, Neuhuber W, Tamm ER. Immunohistochemical localiza- 52. Hung LF, Wallman J, Smith EL III. Vision-dependent changes in the tion of neuropeptides in the human ciliary ganglion. Brain Res. choroidal thickness of macaque monkeys. Invest Ophthalmol Vis 1995;681:229–234. Sci. 2000;41:1259–1269. 44. Matsumoto Y, Tanabe T, Ueda S, Kawata M. Immunohistochemical 53. Fujikado T, Kawasaki Y, Fujii J, et al. The effect of nitric oxide and enzyme histochemical studies of peptidergic, aminergic and synthase inhibitor on form-deprivation myopia. Curr Eye Res. cholinergic innervation of the lacrimal gland of the monkey (Ma- 1997;16:992–996. caca fuscata). J Autonom Nerv Syst. 1992;37:207–214. 54. Seltner RL, Stell WK. The effect of vasoactive intestinal peptide on 45. Tajti J, Moller S, Uddman R, Bodi I, Edvinsson L. The human development of form deprivation myopia in the chick: a pharma- superior cervical ganglion: neuropeptides and peptide receptors. cological and immunocytochemical study. Vision Res. 1995;35: Neurosci Lett. 1999;263:121–124. 1265–1270. 46. Baluk P, Gabella G. Innervation of the guinea pig trachea: a quan- 55. Pendrak K, Lin T, Stone RA. Ciliary ganglion choline acetyltrans- titative morphological study of intrinsic neurons and extrinsic ferase activity in avian macrophthalmos. Exp Eye Res. 1995;60: nerves. J Comp Neurol. 1989;285:117–132. 237–243.

ERRATUM

Erratum in: “Characterization of Vitreous B-Cell Inﬁltrates in Patients with Primary Ocular Lymphoma, Using CDR3 Size Polymorphism Analysis of Antibody Transcripts” by Gorochov et al. (Invest Ophthalmol Vis Sci. 2003;44:5235–5241). It is erroneously mentioned at the end of our general introduction on page 5235 that “immunoglobulins are composed of heavy (H) and light (L) chains, both encoded by rear- ranged genes assembled from sets of variable (V), diversity (D), joining (J), and constant (C) germline gene segments.” It should be noted that D segments are found only in the heavy chain locus.

Downloaded from iovs.arvojournals.org on 09/28/2021