Efferent and Afferent Innervation of Primate Trabecular Meshwork and Scleral Spur

Abstract

purpose. To determine the correlation between nerve terminals and cells or extracellular matrix (ECM) components in different portions of the primate trabecular meshwork (TM) and scleral spur (SS).

methods. Serial sagittal and tangential sections through the anterior segments of 10 cynomolgus monkey eyes and 12 human eyes were investigated immunohistochemically with antibodies against the vesicular acetylcholine transporter (VACHT), vasoactive intestinal polypeptide (VIP), tyrosine-hydroxylase (TH), neuropeptide Y (NPY), substance P (SP), calcitonin gene-related peptide (CGRP), and galanin (GAL) and with a reduced nicotinamide adenine dinucleotide phosphate–diaphorase (NADPHd) reaction. The distribution of the terminals was compared with that of α-smooth-muscle actin (SMA) staining in TM and SS. The relationship between terminals and adjacent cells or ECM components was also studied in ultrathin sections through the TM and SS of 11 monkey eyes cut in sagittal, tangential, and frontal planes.

results. NADPHd-positive nerve terminals were present, especially in the outer portion of both human and monkey TM and in the SS. VACHT-immunoreactive (IR) fibers were found in human but not in monkey SS and TM. The fibers were most numerous in the elongated SS and posterior TM where most cells also stained for SMA. SP- and CGRP-IR nerve endings were also more numerous in the outer TM and SS than in the inner TM. Ultrastructurally, staining for SP was seen in nerve endings containing mitochondria and dense core vesicles and was in contact with the cribriform elastic network. In the posterior SS of monkey eyes were large terminals similar to those previously described in human eyes.

conclusions. The results show for the first time that in the primate TM and SS, there are cholinergic and nitrergic nerve terminals that could induce contraction and relaxation of TM and SS cells. Terminals in contact with the elastic-like network of the TM and containing SP-IR resemble afferent mechanoreceptor-like terminals in other parts of the body. These findings raise the possibility that the TM may have some ability to self-regulate aqueous humor outflow.

It is well established that contraction of the ciliary muscle increases outflow facility by spreading the trabecular meshwork (TM) and increasing the filtration area of the cribriform layer. ^{1 2} Contractile elements such as α-smooth muscle actin (SMA) and myosin are also expressed by TM and scleral spur (SS) cells. ^{3 4 5 6 7} In vitro these cells contract in response to different mediators, including acetylcholine. ⁸ The presence of muscarinic receptors has been described in human TM cells. ⁹ Furthermore, strips of bovine meshwork contract in response to muscarinic agonists, an effect that could be completely inhibited by atropine. ¹⁰ In contrast, nitrovasodilators produce a significant relaxation in bovine meshwork strips. ¹¹ In perfused bovine eye anterior segments, substances that contract TM cells decrease outflow facility, whereas substances that induce relaxation increase outflow facility. ¹²  

The mechanism by which contraction or relaxation of TM cells is mediated in vivo is not known. Innervation of the primate outflow pathways has been described by several authors. ^{13 14 15 16 17 18 19 20 21 22} Substance P (SP) is the only neurotransmitter regularly found in nerve terminals of the TM and SS. Only single fibers stain for calcitonin gene-related peptide (CGRP), and almost no fibers stain for tyrosine-hydroxylase (TH), neuropeptide Y (NPY), vasoactive intestinal polypeptide (VIP), or acetylcholinesterase. Experimental studies in monkey eyes show that SP given intracamerally has no effect on outflow facility, and CGRP only slightly enhances aqueous outflow. ²³ If relaxation or contraction of TM cells influences outflow resistance in primate eyes in vivo, as seems likely, substances other than SP or CGRP must be involved. 

In the present study we investigated the distribution of cholinergic nerve fibers in the primate TM and SS using the vesicular acetylcholine transporter (VACHT) antibody. We also studied the distribution of nitrergic fibers, and the relation between nerve fibers and SMA-positive cells. The ultrastructure of the terminals and their relationship to extracellular matrix (ECM) components of the TM and SS were investigated and compared with terminals measuring shear stress in other parts of the body. In addition, mechanoreceptor-like terminals in the monkey eye comparable to those previously described in human SS were sought. ²¹  

Materials and Methods

The eyes of 11 cynomolgus monkeys (Macaca fascicularis; age range, 3–5 years) and 12 human donors (age range, 58–94 years) were investigated. The monkey eyes were obtained from the primate center of the Behringwerke (Marburg, Germany). The animals were killed by an overdose of thiopental in conjunction with other nonocular protocols. The studies conformed with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the Declaration of Helsinki. Human eyes were obtained from donors after autopsy 4 to 10 hours after death. Methods for securing human tissue were humane, included proper consent and approval, and complied with the Declaration of Helsinki. None of the eyes showed pathologic changes in either the posterior or anterior segments. 

Immediately after enucleation, all eyes were cut equatorially behind the ora serrata and dissected into quadrants. Some specimens of all quadrants were immersed either in Zamboni’s fixative for 4 to 12 hours or in paraformaldehyde (4%) for 3 to 4 hours for immunohistochemistry. The other specimens were fixed in Ito’s solution for at least 12 hours for electron microscopic investigation. 

Electron Microscopy

After fixation in Ito’s solution, the specimens were washed in cacodylate buffer, postfixed with 1% osmium tetroxide, dehydrated with graded alcohols, and embedded in Epon (Roth, Karlsruhe, Germany) in the usual way. Serial sagittal, tangential, and frontal semithin and ultrathin sections were cut using a microtome (Ultracut; Reichert, Vienna, Austria). Semithin sections were stained with toluidine blue. Ultrathin sections were treated with uranyl acetate and lead citrate and viewed by electron microscope (EM 109; Carl Zeiss, Oberkochen, Germany). 

Immunohistochemistry

Each quadrant was dissected into 2-mm wide specimens. The specimens were washed for 24 hours in phosphate-buffered saline (PBS) containing 20% sucrose. From 3 to 4 specimens of each quadrant, serial sagittal or tangential cryostat sections (10–20 μm) were cut, the latter in a plane parallel to the inner wall of Schlemm’s canal (SC). The sections were placed on poly-l-lysine–coated slides and initially incubated with Blotto’s dry milk solution at room temperature for 30 minutes to reduce nonspecific background staining. Incubation with the primary antibody (see Table 1 ) was performed in a moist chamber for 12 to 36 hours at room temperature. Afterward, the sections were rinsed in Tris-buffered saline (TBS) three times at 10 minutes each and then incubated for 1 hour with biotinylated secondary antibodies (Dako, Hamburg, Germany). Finally, the antibodies were visualized either with streptavidin-Cy² or -Cy³ (Dianova, Hamburg, Germany) and, after a rinsing in PBS, mounted in Kaiser’s glycerine jelly (Merck KGaA, Darmstadt, Germany). 

Control experiments were performed using either PBS or nonimmune rabbit or mouse serum substituted for the primary antibody. In addition, sections of iris and ciliary muscle were stained as positive control samples, because these tissues are known to be innervated by VACHT-immunoreactive (IR), TH-IR, NPY-IR, SP-IR, and CGRP-IR nerve fibers. The sections were viewed either with an Aristoplan fluorescence microscope (Ernst Leitz, Wetzlar, Germany) or with a confocal laser-scanning microscope (MRC 1000; BioRad Microscience, Hemel Hempstead, UK). 

NADPH-Diaphorase Staining

Nitric oxide synthase (NOS)-reactivity in human and monkey nerve fibers is probably due to the presence of NOS, because the colocalization of reduced nicotinamide adenine dinucleotide phosphate-diaphorase (NADPHd) and NOS is amply documented in the brain and peripheral nervous system. ^{24 25} The antihuman antibodies available against neuronal NOS (NOS I) did not sufficiently cross-react with monkey tissue and thus did not work reliably in monkey eyes. The antibodies also gave less reproducible results in the human donor eyes obtained after death than staining for NADPHd. We therefore used the NADPHd results as a surrogate marker to describe nitrergic innervation. 

For the NADPHd staining, free-floating wholemounts of the anterior segments of three monkey and four human eyes were incubated at 37°C using the following solution: NADPH/tetrasodium salt (Biomol, Hamburg, Germany), 1 mg/ml; nitroblue tetrazolium chloride (Serva, Heidelberg, Germany), 0.1 mg/ml; 0.3% Triton X-100 in 0.1 M phosphate-buffered saline (pH 7.4). The incubation time was 3 to 4 hours. The reaction was stopped by rinsing the specimens in PBS. The specimens then were cut in a tangential plane (16–20 μm). Serial sections were studied with a light microscope. Positive staining of nerve fibers was checked using choroidal tissues of the same eye. 

Immunoelectron Microscopy

For immunoelectron microscopy, specimens were fixed in a solution of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 1 to 2 hours at 4°C. After they were rinsed in cacodylate buffer containing 60% glycerol, specimens were quick frozen in liquid propane (−175°C) and dehydrated in 100% methanol by freeze substitution for 24 hours (16 hours at −90°C, 4 hours at− 70°C, and 4 hours at −50°C), embedded in resin (Lowicryl; Roth, Karlsruhe, Germany; HM 20: 2 hours, 30% resin and 70% methanol; 2 hours, 70% resin and 30% methanol; and overnight, 100% resin), and polymerized by UV light. Ultrathin sections were cut and mounted on coated (Pioloform coating; Plano, Marburg, Germany) nickel grids. 

For blocking of nonspecific binding, grids were floated on drops of PBS containing 0.05 M glycine for 30 minutes (room temperature). The incubation buffer was PBS containing 0.2% gelatin and 0.5% bovine serum albumin. The grids were incubated on drops of rabbit anti-SP diluted 1:100 for 1 hour at 37°C, and, after rinsing in the same buffer (6 × 10 minutes), incubated on drops of the gold-conjugated secondary goat F(ab′)2 anti-rabbit IgG antibody (British Bio Cell, Cardiff, UK) diluted 1:100 for 1 hour at room temperature. Finally, grids were rinsed in incubation buffer (two times, 10 minutes each), PBS (three times, 10 minutes each), and distilled water (two times, 10 minutes each) and stained with uranyl acetate for 5 minutes. 

Results

Immuno- and Enzyme Histochemistry

α-SMA.

In sagittal and tangential sections of the monkey chamber angle, staining for SMA was seen in nearly all cells of the uveal, corneoscleral, and cribriform filtering portion of the TM and in the SS (Table 2) . 

In old human eyes the results were similar to those described previously. ^{3 4 5 6 7} SMA-labeled cells were present in SS and the posterior TM, whereas staining for SMA was present only in single cells in the remaining parts of the TM. 

NADPHd.

In both human and monkey eyes, circumferentially oriented NADPHd-positive varicose nerve fibers were numerous in the SS. In both species NADPHd-positive varicose axons could also be identified in the TM (Figs. 1A 1B ). In the monkey eye they were even more numerous than in human eyes. Most of these fibers were found in the outer corneoscleral TM and in the cribriform meshwork of the TM. Nitrergic nerve fibers were also found adjacent to the inner wall of SC, whereas in the uveal TM and in the anterior nonfiltering portion, no or only single fibers were observed. Almost no NADPHd-positive varicosities were found in the tips of the longitudinal portion of the ciliary muscle. 

VACHT.

In the monkey eye VACHT-IR nerve fibers and terminals were present almost exclusively in the ciliary muscle where each muscle cell was surrounded by positively stained terminals. In only one specimen of four monkey eyes were single fibers found in all layers of the TM (Fig. 1C) . 

In old human eyes, in addition to the ciliary muscle, all parts of the circumference showed abundant VACHT-IR nerves in the SS and TM (Fig. 1D) . Nerve fibers, which in the muscle ran parallel with the longitudinal muscle fibers, bent nearly perpendicular in the SS and TM, so that they were oriented circumferentially (Fig. 1D) . The VACHT-IR nerve fibers in the SS were more numerous than in the TM, where fibers were located mainly in the posterior portion. 

TH, NPY, and VIP.

In general, our results confirm those described of previous investigators. ^{16 17 18 19 20} In both monkey and human eyes in most parts of the SS and in the TM, no TH-IR fibers were present. Only in some parts of the circumference were single axons staining for TH observed in the posterior part of the SS. With the antibody against the vesicular monoamine transporter-2, we obtained the same results. NPY-IR fibers were not seen in the SS or TM. In both species VIP-IR fibers were almost completely absent in the SS and the TM. In only two of four monkeys was a single axon immunoreactive for VIP found. Staining for TH, NPY, and VIP was positive in iris or choroid in both species, showing that the antibodies were working in the material. 

SP, CGRP.

In both monkey and human, numerous varicose axons staining for SP and CGRP could be identified in all quadrants of the SS. Some nerve fibers and terminals were also found in all layers of the TM. They were more numerous in the cribriform and outer corneoscleral layer than in the uveal portion (Fig. 2) . There were more SP- and CGRP-IR nerve endings in the monkey than in the human TM. The distribution of SP-IR was similar to that of CGRP-IR; however, there were slightly more SP-positive varicosities. Galanin (GAL)-IR nerve fibers were not seen in the monkey and human TM or SS, but there were positively stained fibers at the muscle tips. 

Controls.

No staining was seen in the control experiments in which the primary antibody was replaced by either PBS or nonimmune rabbit or mouse serum (not shown). 

Electron Microscopy

Ultrastructural studies were performed only in monkey eyes, where nerve terminals were better preserved due to the postmortem time. In the monkey TM tangential sections through the subendothelial and cribriform layer revealed nerve terminals with dense core vesicles in each specimen, most of them containing numerous mitochondria and empty vesicles. The axons were only partly covered by Schwann cells, whereas the Schwann cell–free portion often was in close contact with an elastic fiber of the cribriform layer. Such fibers were located in the entire cribriform layer up to the subendothelial region of SC (Figs. 3A ). Terminals in direct contact with the inner wall endothelium were not seen. 

Nerve terminals similar to those described in the cribriform layer were also present in the trabecular beams, where the Schwann cell–free part of the axons faced the basement membrane (Fig. 3B) . Terminals were not seen in the intertrabecular spaces. 

In the posterior SS some axons formed large terminals with diameters up to 5 to 10 μm containing granular and agranular vesicles of different sizes, abundant mitochondria, and lysosome-like lamellated structures. They were incompletely ensheathed by glial cells and were found in close contact with elastic-like fibers (Figs. 4A 4B ). These terminals were observed in all quadrants of the circumference. 

Immunogold labeling for SP was observed in dense core vesicles in terminals located in the cribriform region adjacent to elastic and collagen fibers (Figs. 5A 5B ). Of note, such terminals were also seen in septa emanating from the sclera and traversing from the outer to the inner wall of SC (Figs. 6A 6B ). 

Discussion

In the present study we demonstrate for the first time that in the primate TM SP-IR nerve terminals are in intimate contact with the elastic fibers of the TM. In addition, we show that the TM contains abundant nerve fibers and varicosities staining for NADPHd and for VACHT. Because the TM elastic fibers are an important stress-bearing component of the ECM, this arrangement of nerve terminals is consistent with a functional system for regulation of TM tone and thus possibly intraocular pressure (IOP). More specifically, the SP-IR terminals would represent the afferent portion of this pathway (sensing TM stretch), and the NADPHd- and VACHT-positive terminals would represent the efferent portion responsible for inducing relaxation or contraction of TM cells. 

Nitric Oxide Synthase

Numerous NOS-IR ganglion cells have been described in the primate pterygopalatine ganglion and to a lesser extent in the ciliary ganglion. ^{26 27} We therefore assume that these fibers, similar to the fibers in the choroid, belong to the parasympathetic innervation. After removal of the pterygopalatine ganglion in primate eyes, Ruskell described degeneration of nerve fibers in the TM. ¹⁴ We assume that these degenerated nerve fibers represent the nitrergic nerves. According to the studies of Lepple–Wienhues et al. ¹⁰ and Wiederholt et al. ¹¹ in bovine eyes, NO induces relaxation of trabecular cells and increase in aqueous humor outflow facility. In human and monkey eyes, nitrergic nerve terminals were most numerous in the cribriform layer of the TM and in the SS. In this layer the cells are not only connected to each other and to the inner wall of SC, but also to the cribriform elastic network (JMS, unpublished observation). In a recent study we have shown that these cells stain for αB-crystallin, indicating that they may be exposed to mechanical stress. ²⁸ If these cells are relaxed by innervation through nitrergic nerves, aqueous flow could extend the cells and thereby the cribriform region. 

Vesicular Acetylcholine Transporter

Single acetylcholine-esterase–positive nerve fibers have previously been described in the monkey TM, ²⁹ but studies in human eyes and studies using the more specific antibody against the VACHT have not been performed before. VACHT is an antibody directed against a transporter protein and has been shown to be specific for cholinergic neurons in human and monkey central and peripheral nervous system. ³⁰ Distribution of VACHT-IR fibers in elderly human eyes was essentially the same as that seen for α-SMA staining. ⁶ In monkey eyes, all TM and SS cells stained forα -SMA but VACHT-IR fibers were absent. In cynomolgus monkey eyes, the SS is flat, and tendons of the ciliary muscle tips insert directly into the TM. In human eyes, the elongated SS is located between muscle tips and the posterior part of the TM. In addition, in elderly human eyes, the muscle tips are often hyalinized. ³¹ Whether these morphologic differences can explain the differences in VACHT innervation between monkey and human eyes and whether cholinergic innervation in vivo induces contraction of TM cells is not yet known. 

If primate TM contracts in a manner similar to bovine meshwork, cholinergic stimulation of the TM would increase outflow resistance. However, this situation is complicated by the presence of VACHT innervation in the ciliary muscle. Ciliary muscle contraction due to cholinergic stimulation expands the TM, which increases outflow facility. Because ciliary muscle contraction is stronger than TM contraction, the net effect of stimulation of cholinergic nerve fibers would be an increase in outflow facility. The effect of muscarinic agonists on monkey TM facility was studied after disinsertion of the ciliary muscle from the SS and TM. ^{32 33} In these eyes neither intravenous nor intracameral pilocarpine injection had any effect on outflow resistance, suggesting that in primate eyes contraction of TM cells may not modulate outflow resistance. However, caution must be used in interpreting these results, because scar tissue formation, collapse of the disinserted TM, or other postsurgical effects may have produced nonphysiologic responses to cholinergic stimulation in this monkey model. 

Because of the connection between ciliary muscle fibers and the TM, cholinergic stimulation of the ciliary muscle produces maximal force if there is a concurrent stiffening of the TM. Furthermore, such stiffening may be required to avoid disruption of TM by ciliary muscle contraction. Theoretically, this stiffening would occur in several ways. For example, contractile elements in the TM could also be subject to cholinergic stimulation such as seems to be the case in elderly human eyes. A second possibility is that of reactive contraction, which we hypothesize occurs in monkeys. The forceful contraction of ciliary muscle in monkey eyes is expected to induce significant strains in the cells of TM. ³⁴ In tissues with high smooth muscle content (e.g., intestine), such externally imposed mechanical stretching causes a reactive contraction of the stretched muscle fibers. It therefore seems likely that ciliary muscle contraction indirectly induces TM cell contraction in monkey eyes. 

Substance P

A number of the nerve varicosities found in the TM stained for SP. In primates abundant SP-IR ganglion cells have so far been described only in the trigeminal ganglion. ³⁵ In the superior cervical and pterygopalatine ganglion, no nerve cells were SP-IR, whereas in the ciliary ganglion, single SP-IR axons were found, but these axons seemed to traverse the ganglion. ^{36 37 38} We therefore assume that at least part of the SP-IR terminals in the TM represent afferent trigeminal axons. At the electron microscopic level, these terminals contained numerous mitochondria and were in contact with ECM components especially with the elastic-like fibers in the cribriform region. In other parts of the body, axons containing abundant mitochondria and in contact with elastic fibers are characteristic of mechanoreceptors and have been described for the mechanoreceptive nerve endings of the Golgi tendon organ, ³⁹ encapsulated Ruffini corpuscles of the skin, ⁴⁰ and visceral mechanoreceptors such as are present in the respiratory system, ⁴¹ and the dura mater encephali. ⁴² Our findings indicate that there are mechanoreceptor-like terminals, not only in the SS of human ²¹ and monkey eyes, but also in the TM and especially in the cribriform region up to the subendothelial layer of SC and in the scleral septa. Tension of the elastic fibers in the TM changes during contraction of the ciliary muscle and may change because of variation in IOP. It is tempting to speculate that the terminals in the TM and septa measure changes in tension and are, in an as yet unknown way, involved in the regulation of muscle tone and directly or indirectly in regulation of aqueous humor outflow. 

 Supported by grants from the Deutsche Forschungsgemeinschaft (SFB 539).

 Submitted for publication November 3, 1999; revised January 31, 2000; accepted February 24, 2000.

 Commercial relationships policy: N.

 

Table 1.

 

Antibodies Used for Immunohistochemistry

Primary Antibody	Type	Host	Dilution
† >α-SMA*	M (Clone 1A4)	Mouse	1:150
VACHT, †	P	Rabbit	1:2000
TH, ‡	P	Rabbit	1:400
VMAT II, †	P	Rabbit	1:2000
NPY, †	P	Rabbit	1:400
VIP, ‡	P	Rabbit	1:400
SP, ‡	P	Rabbit	1:500
CGRP, ‡	P	Rabbit	1:100
GAL, §	P	Rabbit	1:500

 M, monoclonal antibody; P, polyclonal antibody.

*  Sigma, Deissenhofen, Germany.

†  Phoenix, Mountain View, CA.

‡  Biotrend, Cologne, Germany.

§  Chemicon, Temecula, CA.

Table 2.

 

Staining Pattern of Ciliary Muscle Tips, SS, and TM in Human and Monkey Eyes

	SS	Uveal TM	Corneoscleral TM	Cribriform TM	Nonfiltering Portion
\ {tblsh}MONKEY
NADPHd*	++	+	++	++	+
VACHT	+
TH/VMAT II	+
NPY
VIP
SP	++	+	+	++	+
CGRP	++	+	++	++	+
GAL
\ {tblsh}HUMAN
NADPHd*	++	+	+	+	+
VACHT	++	+	+	+
TH/VMAT II
NPY
VIP
SP	++	+	+	+	+
CGRP	+	+	+	+	+
GAL

 +, single; ++, numerous positively stained axons and varicosities; , positive staining not observed.

*  In addition, diffuse staining of TM and ciliary muscle.

Figure 1.

 

Histologic tangential sections (parallel to the inner wall of SC) through the ciliary muscle (CM) tip, SS, and outer TM of monkey and human eyes. (A, B) Staining for NADPHd. (A) Subendothelial region of a monkey TM and inner wall of SC showing numerous stained nerve fibers and varicosities (arrows). (B) Corneoscleral TM and CM tips of a human eye. Numerous positive nerve fibers with varicosities were oriented circumferentially within the SS and TM. Note that the muscle contained no NADPHd-stained nerve fibers. (C, D) Staining for VACHT. (C) In monkey eyes presence of VACHT-IR nerve fibers and terminals was restricted to the CM. Only in one of four eyes were single varicosities seen (arrow). (D) In human eyes VACHT-IR nerve fibers ran circumferentially within the SS and posterior TM. Varicosities were present not only in the CM but also in the TM and SS. Magnification, (A, B) ×480; (C, D)× 380.

Figure 2.

 

Histologic tangential section through the outer portion of SS and corneoscleral TM. Immunohistochemical staining for SP in a monkey eye. Numerous nerve fibers with varicosities were oriented circumferentially in SS and TM. The distribution of SP-IR nerve fibers was similar in human eyes (not shown). Magnification, ×750.

Figure 3.

 

Electron micrographs of nerve terminals in the TM of monkey eyes. (A) Subendothelial region of SC. Terminals (arrow) were in contact with elastic-like fibers (E). (B) Corneoscleral trabecular lamella. The Schwann cell–free portion of the terminal (arrow) was in close contact with the basement membrane (BM) of the trabecular beam. Magnification, (A) 24,000; (B) 28,000.

Figure 4.

 

Electron micrographs of nerve terminals in the monkey SS. In the posterior portion of SS there were numerous terminals (arrow) containing mitochondria and lysosome-like structures that were in contact with the elastic fibers of the SS (arrowheads). (B) Higher magnification of the area in (A) indicated by an arrow. Magnification, (A) 8,400; (B) 21,600.

Figure 5.

 

(A) Electron micrograph of the cribriform region of a monkey eye. The nerve terminal (arrow) is located adjacent to the elastic network of the cribriform region (arrowheads). (B) Higher magnification of the terminal indicated by arrow in (A) in an adjacent ultrathin section. Dense core vesicles stained for SP. Immunogold stain; magnification, (A) ×9,800; (B) ×66,000.

Figure 6.

 

(A, B) Electron micrographs of a nerve terminal in a scleral septum bridging the inner and outer wall of SC. (A) Overview showing the localization of the nerve fiber in the scleral septum (arrow) of a monkey eye. (B) Higher magnification of the varicosity. Arrows: Immunogold labeling for SP. Magnification, (A) ×3,900; (B)× 32,000.

The authors thank Angelika Hauser, Barbara Teschemacher, Gertrud Link, and Anke Fischer for expert assistance with immunohistochemistry and electron microscopy and Marco Gösswein for excellent preparation of the photographs.