Intrinsic Choroidal Neurons in the Human Eye: Projections, Targets, and Basic Electrophysiological Data

Abstract

purpose. The chemical coding of intrinsic choroidal neurons (ICNs) has features in common with extrinsic fibers (e.g., from the pterygopalatine ganglion) making it impossible to assess whether a neuronal nitric oxide synthase (nNOS)/vasoactive intestinal polypeptide (VIP)–immunoreactive nerve fiber is of intrinsic or extrinsic origin. Neurobiotin injections into single neurons allow the visualization of projections of these cells and the determination of the origin of target innervation. Thus, this technique was used in the present study to help characterize the organization of the ICN in the human eye.

methods. ICNs were visualized with the fluorescent vital dye 4-Di-2-ASP. Electrophysiological properties were determined by means of intracellular recordings. The impaled neurons were iontophoretically filled with neurobiotin. After fixation, immunohistochemistry for neuronal nitric oxide synthase (nNOS), α-smooth muscle actin, and calcitonin gene-related peptide (CGRP) was conducted.

results. ICN processes were traced over distances of up to 2.612 μm. They were found in the immediate vicinity of other nNOS-positive or -negative ICNs and were also found apposed to smooth muscle fibers (vascular and stromal nonvascular). CGRP-positive fibers forming boutons were observed closely associated with ICNs. Electrophysiological recording showed phasic firing without slow afterhyperpolarization, no spontaneous activity, an input resistance of 136 ±73 MΩ, and a membrane time constant of 7 ± 1 ms.

conclusions. Apart from the first functional characterization of ICNs, this study provided more precise evidence of reciprocal ICN-to-ICN contacts and innervation of both choroidal nonvascular and vascular smooth muscle. The presented technique offers promising perspectives to further investigate the function of ICNs in ocular homeostasis.

The autonomic innervation of the vertebrate eye is mediated through sympathetic and parasympathetic pathways originating in superior cervical, ciliary, and pterygopalatine ganglia, respectively. As a third component, trigeminal primary afferent neurons release active substances from their peripheral terminals in the choroid. ^{1 2} This classic scheme does not take into account the presence of autonomic nerve cells residing within the choroid of the eye, the intrinsic choroidal neurons (ICNs). Though first described some 150 years ago, ³ they remained poorly recognized despite repeated reports. ^{4 5 6} These neurons received renewed interest when it was found that they contained vasoactive intestinal polypeptide (VIP) ⁷ and neuronal nitric oxide synthase (nNOS) ^{8 9} and when they were commonly observed in reduced nicotinamide adenine dinucleotide phosphate diaphorase (NADPHd)–stained choroidal wholemounts in both higher primates ⁸ and birds. ¹⁰  

Putative targets that might be innervated by ICNs include arteries and nonvascular smooth muscle fibers of the choroid. ^{8 9 11 12} Thus, ICN may be involved in the control of ocular blood flow, choroidal thickness, and intraocular pressure. In primate eyes, no clear-cut marker substance of ICN is known. Nerve terminals on ICN costaining for VIP and nNOS may be of either intrinsic or extrinsic origin, such as from pterygopalatine ganglion cells, because in the latter neurons exhibit a neurochemical coding similar to that of ICNs. ^{13 14}  

The purposes of the present study were to develop a method for intracellular recording and dye filling of single ICNs in the human choroid and to perform a basic morphologic and electrophysiological characterization of the neurons identified by this method. 

Materials and Methods

Tissue Preparation

In this study, 15 choroids of human donors (65–89 years of age; both sexes) were investigated, according to the Declaration of Helsinki for the use of human tissue in research. Donor bulbi were obtained from the cornea bank of the University of Antwerp (UIA), Belgium, within 4 to 12 hours after death. Eyes were cut open circumferentially around the ora serrata, and retina and vitreous body were removed. Residual scleral cups with the choroid attached were gently rinsed in a standard ice-cooled oxygenated (95% O_2, 5% CO₂) Krebs-Ringer solution of the following composition (in mM): 118.0 NaCl, 4.75 KCl, 2.54 CaCl₂, 1.2 MgSO₄, 1.0 NaH₂PO₄, 25.0 NaHCO₃, and 11.1 glucose. Retinal pigment epithelium was detached carefully using cotton-wool tips. Choroids were dissected and transferred into ice-cooled oxygenated Krebs-Ringer solution. Before the electrophysiological experiments began, the choroids were pinned flat in plastic dishes coated with a silicone elastomer (Sylgard; Dow-Corning Europe, La Hulpe, Belgium). 

Fixation of the Tissue with Low-Gelling-Temperature Agarose

Intracellular impalements of human ICNs were hampered by loose fibers of the lamina suprachoroidea floating on top of the preparation. Fixation of these loose fibers was achieved by covering the preparation with a thin layer of agarose, using the following protocol. First, a stock solution of 1.6% agarose (SeaPlaque; Duchefa, Haarlem, The Netherlands) was made and kept at a temperature of more than 40°C. This stock solution was mixed with Krebs-Ringer solution (1:1) and heated up to 60°C for several minutes. The Krebs-agarose mixture was then cooled down to 40°C to 50°C, and a thin layer was put on top of the preparation. To allow the agarose to gel, the preparations were incubated for another 30 minutes at 4°C. Once the agarose had completely gelled, the recording dish was transferred into ice-cooled oxygenated Krebs-Ringer solution. 

Visualization of ICNs

Therefore, a 4 μM solution of the fluorescent vital dye 4-(4-diethylaminostyryl)-N-methylpyridinium iodide (4-Di-2-ASP; Molecular Probes, Eugene, OR) in oxygenated Krebs-Ringer solution was used (20 minutes at 37°C). The validity of this vital staining method has been demonstrated in different tissues, ^{15 16} including human choroid. ¹⁷ During the staining procedure and afterward, the specimens were shielded from light and stored in ice-cooled oxygenated Krebs-Ringer solution. 

Intracellular Electrophysiological Recordings

Silicone-coated dishes were continuously superfused (10 mL/min) with oxygenated Krebs-Ringer solution (37°C). Nicardipine (1 μM; Sigma-Aldrich, St Louis, MO) was added to the solution to prevent smooth muscle contractions during recordings. Recording chambers were placed on the stage of an inverted microscope (Diaphot; Nikon, Tokyo, Japan) equipped with epifluorescent illumination (HBO 50-W mercury lamp; Osram, Munich, Germany) and a B-2A filter combination. Brief epifluorescent illumination was used to localize individual neuronal cell bodies, and impalements were performed under transmitted light conditions. The recording chamber was grounded through the HS-2L headstage (gain 0.1) of a current–voltage amplifier (Axonclamp 2A), which was connected to an interface (Labmaster TL-1 DMA Interface; Axon Instruments, Foster City, CA). Intracellular recordings were made with borosilicate glass microelectrodes (1 mm outer diameter; Clarc Elecromedical Instruments, Reading, UK) pulled on a horizontal micropipette puller (P-97 Brown-Flaming puller; Sutter Instrument Co., Novato, CA). The electrodes were back filled with 1 M KCl containing 2% neurobiotin (Vector Laboratories, Burlingame, CA; resistance 60–100 MΩ). The amplifier was equipped with a bridge circuit that was balanced for each electrode before impalement, and compensation of the capacitance was accomplished during injection of rectangular electrical current pulses (−0.2 nA, 7 ms). 

Passive and Active Membrane Characteristics

Measurements were made after allowing the impalement to stabilize for a few minutes (until the resting potential of the cells stabilized) without applying intracellular current. To investigate passive and active membrane properties, current step commands were created on computer (pClamp ver. 6.0.2; Axon Instruments). The data were low-pass filtered online (3 kHz), digitized (sample frequency 5 kHz), and stored. The input resistance and membrane time constant of the impaled neurons were estimated by passing small hyperpolarizing current pulses of variable amplitude (−0.05 to −0.3 nA) through the intracellular recording electrode. The resulting membrane potential change was measured so that current–voltage curves could be constructed. Each step was repeated three times, and the average value was used for further calculations. Durations of action potentials were measured as half widths (i.e., the time interval between the point on the upstroke at which the amplitude of the action potential is halfway between the membrane resting potential and the maximum potential, and the equivalent point on the downstroke). To avoid interference of voltage changes due to the depolarizing current applied, short pulses of 2 ms were used to evoke a single action potential, whenever individual action potential characteristics were measured. Data analysis was performed on computer (pClamp, ver. 6.0.2; Axon Instruments, Excel 97; Microsoft, Redmond, WA; and SigmaPlot, ver. 5.0; SSPS Sciences, Chicago, IL). All data are presented as the mean ± SD. 

Immunohistochemistry

During electrophysiological recordings, impaled neurons were iontophoretically filled with neurobiotin by passing depolarizing current pulses (0.5–1 nA, 100–500 ms duration) through the recording electrode. After being electrophysiologically recorded, the impaled cell was photographed and its position in the choroid, mainly in the temporal quadrants, was mapped for later localization and morphologic identification. The preparations were eventually fixed in a modified Zamboni solution (4% paraformaldehyde, 0.2% picric acid, and 0.1 M sodium phosphate buffer) for 2 hours at room temperature. After fixation, the preparations were further processed to improve conditions for immunocytochemistry. To this end, they were first rinsed in phosphate-buffered saline (PBS; pH 7.4, 0.01 M) for 10 minutes, followed by a rinse in 50% ethanol (5 × 8 minutes) and rehydrated in PBS (5 × 8 minutes). The preparations were then incubated in 0.05% thimerosal (Sigma-Aldrich) for 30 minutes and rinsed again in PBS (3 × 10 minutes). To enhance antibody penetration and to prevent nonspecific binding of antibodies, the preparations were incubated in PBS (30 minutes at room temperature) containing 10% normal goat serum (Dako, Glostrup, Denmark) or normal donkey serum (Jackson ImmunoResearch Laboratories, West Grove, PA), depending on the secondary antibody used, and 0.1% bovine serum albumin, 0.05% thimerosal, 0.01% NaN₃, and 1% Triton X- 100 (all Sigma-Aldrich). The preparations were then incubated (36 hours; room temperature) for triple labeling in primary antisera and antibodies against nNOS (1:1000, raised in rabbit; EuroDiagnostica, Malmö, Sweden) and α-smooth muscle actin (clone1A4, 1:2000, mouse; Sigma-Aldrich, or nNOS and human calcitonin gene-related peptide [CGRP], 1:1000, guinea-pig; Euro-Diagnostica), respectively. After six washes (5 minutes each) in PBS, the preparations were incubated in secondary antisera for 6 hours at room temperature in corresponding fluorescein isothiocyanate (FITC; for nNOS; 1:100) or Cy5 (for SMA and CGRP, 1:200; both from Jackson ImmunoResearch Laboratories) tagged antibodies. To reveal the presence of neurobiotin in the impaled neurons, the tissue was incubated with streptavidin coupled to Cy3 (1:2000; Jackson ImmunoResearch Laboratories). After a final wash to remove the unbound secondary antibodies, the preparations were mounted in antifade medium (Vectashield; Vector Laboratories). 

Microphotography

Projections of the impaled neurons were photographed with a fluorescence microscope (40× air objective; BX 50; Olympus, Birkeroed, Denmark) equipped with a motorized stage and charge-coupled device (CCD) color video camera. Images were taken, and mapping of neuronal processes was performed on computer (Neurolucida, ver. 3.0; MicroBrightField, Magdeburg, Germany). Combined immunolabeling was examined in a confocal microscope (LSM 510; Carl Zeiss Meditec, Jena, Germany) equipped with an Argon (488 nm) and He/Ne laser (543 and 633 nm). Two water-immersion objectives (25×, numeric aperture 0.8 and 40× numeric aperture 1.2; Carl Zeiss Meditec) were used. 

Results

Immunohistochemistry

From the group of 15 choroids in this study, 9 neurons were successfully impaled. Five of them appeared to be immunoreactive for nNOS. 

All impaled neurons showed a clear labeling for neurobiotin in both cell soma and processes. Largest cell body diameter of these group of neurons, as measured in projected confocal Z-stacks, ranged from 19 to 35 μm. Confocal images showed eight cells with a clearly discernible, eccentrically positioned nucleus. Cell bodies were ovoid or droplike in shape and mainly had a smooth contour with excavations of the cell body presumably caused by surrounding satellite cells. These neurons gave rise to up to seven processes: up to four short processes less than 60 μm in length and one to three long processes up to 2.612 μm (Fig. 1G) . Processes arose either radially or polarly from the impaled cell bodies. Only in one case was a process visualized in the confocal microscope thought to be an axon based on the presence of a hillocklike appearance (Figs. 1F 2B) . The projection of this axon-like process was observed over a distance of 887.7 μm, giving rise to a short (64.7 μm; 215 μm off the cell body) and a long (357.7 μm; 341.6 μm off the cell body) collateral. A detailed analysis of the length of these processes, however, did not allow a further subclassification of these neurons (see Table 1 ). 

Neurobiotin-filled varicose fibers forming boutons could be followed over long distances, sometimes crossing their own pathways in wide turns (Figs. 1B 1G) and running parallel to choroidal blood vessels (Fig. 2A) . Detailed analysis of confocal Z-series, using antibodies against α-smooth muscle actin, revealed neurobiotin-positive fibers in close apposition to the vessel wall, sometimes coursing along the media (Figs. 3A 3B 3C 3D) . 

Neurobiotin-filled fibers were found to run parallel to nonvascular smooth muscle fibers of the choroid (Fig. 3E 3F 3G) , forming bouton-like approaches. Neurobiotin-filled processes of impaled cells showed boutons and closely embracing nerve fibers on other nitrergic (Fig. 4A) and nonnitrergic cells in the same ganglion, or with neurons in ganglia at a certain distance (up to 800 μm; Fig. 4B ). In contrast, close appositions of intrinsic nitrergic neurons onto neurobiotin-filled ICNs up to 400 μm removed from the impaled cells were also detected (Figs. 4C 4D) . A neurobiotin-filled ICN immunonegative for nNOS is depicted in Figures 4E and 4E ′. Neurobiotin-positive fibers forming varicosities projected along with nNOS-immunoreactive fibers in the same nerve bundles (Fig. 4F) . 

CGRP-immunoreactive fibers coursed through the choroid, forming varicose fibers with small boutons. These fibers were found to run parallel to nNOS-immunoreactive fibers (Fig. 4G) , presumably originating from ICNs, but an extrinsic origin cannot be excluded. Small CGRP-immunoreactive boutons were closely attached to cell bodies of both impaled and nonimpaled nNOS-negative and nNOS-positive ICNs (Fig. 4H) . Close appositions of CGRP-positive boutons on processes of nNOS-positive ICNs were also observed (Fig. 4I) . 

Electrophysiology

Approximately 30% of the impaled neurons of the human choroid gave stable electrophysiological recordings (resting membrane potential: −52 ± 10 mV) and fired action potentials after direct somal depolarization, with a clear reversal (“overshoot”) of the resting membrane potential (Fig. 5A) . The neuron shown in Figures 5E and 5E ′ had a stable membrane resting potential (−50 mV) and an input resistance of 61 MΩ but was not able to fire action potentials after direct somal depolarization, not even during long-lasting (200–500 ms) depolarizing current pulses. None of the impaled neurons showed spontaneous activity. 

The neurons that fired action potentials all displayed rather brief spikes (action potential half-width: 1.6 ± 0.6 ms) with a monophasic repolarization, which is reflected in the first-time derivative of the voltage trace without inflections (Figs. 5B 5b ′). During the application of long-lasting (200–500 ms) suprathreshold depolarizing current pulses, the neurons fired action potentials only at the onset of the depolarization, irrespective of the applied current strength (Fig. 5C) , which is indicative of phasic firing behavior. The action potentials were not followed by a slow afterhyperpolarization (Fig. 5D) . 

The mean input resistance of the neurons was 136 ± 73 MΩ, and the mean membrane time constant of these neurons amounted to 7 ± 1 ms. 

Discussion

This study presents for the first time electrophysiological recordings from intrinsic neurons of the human choroid. They show phasic firing behavior but no spontaneous activity. Iontophoretic filling of impaled ICNs using neurobiotin revealed single nerve cells and their projections over distances as long as 2.612 μm. Two additional procedures were crucial for successful impalements with microelectrodes: (1) the use of the low-gelling-temperature agarose covering the tissue immobilized the neurons and prevented the loose connective tissue fibers from adhering to the microelectrode, and (2) the application of the supravital 4-Di-2-ASP staining enabled the microscopic visualization of the ICNs. This dye has no influence on basic electrophysiological neuronal features. ^{15 16 18 19} Although agarose has been used earlier to prevent single neurons or tissue slices from moving in the recording chamber, ^{20 21 22 23} we used it, to our knowledge, for the first time on vital wholemount preparations followed by immunohistochemistry. Comparison of the staining technique using the vital fluorescent dye 4-Di-2-ASP in agarose-treated versus nontreated tissue yielded no differences. In addition, the penetration of antibodies used in our immunohistochemical protocols was not hampered by the agarose technique, and, as reported earlier, ^{20 23} no interference of agarose with electrophysiological recordings was observed. Therefore, this technique appears to be a valid tool for future pharmacofunctional studies of ICNs. Nevertheless, the intracellular recording technique used in this study can cause a leak current at the pipette–membrane interface and can cause cell damage, either of which could result in depolarization. Future studies involving various configurations of a patch–clamp approach may be helpful to the characterization of the electrophysiology of these neurons. 

Donor eyes available in this study were those of elderly persons; hence, age-related alterations of the tissue cannot be ruled out. ^{24 25 26} However, one advantage pertains to the reduced choroidal pigmentation observed in elderly compared with younger choroids, rendering identification of the supravitally stained ICNs easier in the inverted microscope setup. 

Application of neurobiotin as an intracellular morphologic marker revealed a rather “simple” morphology of human ICNs, showing neurons with a bi- or multipolar appearance. This observation is in accordance with earlier studies using neurochemical markers, such asVIP, nNOS or NADPHd. ^{7 8 9} Recently, antibodies against different neurofilament subunits have been used in a study of ICNs. ²⁷ Iontophoretic filling, however, has the advantage that the intrinsic origin of the neurobiotin-positive fibers can unequivocally be determined, even when the neurochemical features of intrinsic and extrinsic fibers overlap, as is the case in the human choroid. Apart from the ICNs, the parasympathetic postganglionic neurons of the pterygopalatine ganglion supplying the eye ^{2 28} also contain VIP and nNOS/NADPHd. ^{13 14} Other extrinsic fibers in the human choroid contain only one or even none of these two markers: no VIP-positive neurons were detected in the ciliary ganglion, ¹⁴ whereas in the superior cervical ganglion, few neurons show immunoreactivity for VIP, but not for nNOS. ²⁹ In the trigeminal ganglion, a minority of the cells is positive for nNOS, but colocalize CGRP. ³⁰  

Contacts of ICNs to vascular smooth muscle cells have been suggested in earlier studies, ^{8 9} but direct evidence has not been provided yet. By contacting vascular smooth muscle cells, as our results show, ICN may influence the muscle tone of choroidal blood vessels and thus be considered an important component in the autonomic regulation of choroidal blood flow; thus, they may play an important role in the autonomic regulation of intraocular pressure. ^{31 32 33} The function of the other nonneuronal target of ICNs (i.e., stromal nonvascular smooth muscle cells) is less obvious. Nonvascular smooth muscle cells were described in the choroid of humans, ^{3 11 34} other mammalians, ^{11 35 36} and various avian species. ¹² In humans, they are mainly concentrated in the temporal and submacular regions of the choroid. ^{11 34} Of note, these are the regions where ICNs also accumulate. ^{8 27 37} Influencing the contractile state of these nonvascular smooth muscle fibers may result in a change in the thickness of the choroid, leading to a shift of the adjacent retina and thus altering the optic properties of the eye. However, whether this is also the case in human eyes still must be elucidated. This mechanism has been extensively investigated in the avian eye ^{38 39} but the underlying mechanisms are poorly understood. ICNs may be suitable candidates for the neuronal control of these refractive changes. ^{12 40 41} Possible targets of ICNs other than those expressing smooth muscle actin are still to be determined. 

The observed close appositions of CGRP-immunoreactive fibers, most likely of trigeminal primary afferent origin, with ICNs may represent a precentral reflex arrangement similar to the spinal visceral afferent contacts on neurons in prevertebral ganglia. ⁴² Contacts of CGRP-positive nerve fibers were not described in the human pterygopalatine ganglion, nor were CGRP1 receptors expressed. ⁴³ In the avian eye, CGRP-positive contacts on ICNs have been recently reported, presumably forming the anatomic basis of a precentral reflex arc. ⁴⁰ Therefore, it is reasonable to assume that CGRP-positive fibers in close apposition to human ICNs are also integrated in a precentral reflex arc. Our experiments with neurobiotin injections revealed that nerve fibers of intrinsic origin closely appose other ICNs. Although the appositions of both intrinsic and extrinsic nerve fibers on ICNs are suggestive of synaptic contacts, this has to be verified at the ultrastructural level. Further electrophysiological recordings are necessary to determine whether different functional types of ICNs are present in the human eye. Nevertheless, the technique developed herein enables functional studies of human ICNs in vitro, which will yield a better insight into the mechanisms regulating ocular homeostasis. 

 Submitted for publication March 6, 2003; revised April 30, 2003; accepted May 13, 2003.

 Disclosure: F. Schrödl, None; A. De Laet, None; M.J. Tassignon, None; P.-P. Van Bogaert, None; A. Brehmer, None; W.L. Neuhuber, None; J.-P. Timmermans, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A–I) Camera lucida drawings illustrating the projection pattern and morphologic features of neurobiotin-filled human ICNs.

Table 1.

 

Morphologic Properties of ICNs as Revealed by Neurobiotin Injection

Impaled Neurons (Fig. 1)	Soma Shape	Soma Diameter (μm)*	Cell Body Perimeter (μm)	Cell Body Area (μm2)	Arrangement of Processes	Length of Processes Arising from Soma (μm), †			Longest Process (μm), ‡	nNOS ImmunoReactivity
A	Droplike, smooth-	25.5 ⊥ 15.9	109.7	637.3	Radial	A: 289.1	B: 256.8	389.5	+
	contoured					C: 389.5	D: 44.1
						E: 18.6	F: 52.1
						G: 24.7
B	Ovoid	26.9 ⊥ 20.6	117.9	706.3	Polar	A: 5016.3	B: 14.6	2592.2	−
						C: 738.7	D: 194.9
C	Ovoid, smooth-contoured	27.7 ⊥ 25.0	29.0	53.2	Radial	A: 14.6	B: 274.8	274.8	+
						C: 80.7	D: 78.9
D	Drop-like, smooth-	26.5 ⊥ 21.7	96.6	481.4	Radial	A: 110.2	B: 798.0	743.5	−
	contoured					C: 138.6	D: 55.8
E	Ovoid, smooth-contoured	35.2 ⊥ 25.1	109.4	880.8	Polar	A: 206.1	B: 386.8	659.8	+
						C: 956.2
F	Ovoid	18.9 ⊥ 11.1	76.2	240.6	Polar	A: 1111.2	B: 1362.7	887.6	−
G	Ovoid, smooth-contoured	34.8 ⊥ 21.2	107.0	794.1	Polar	A: 53.4	B: 109.6	2612.5	+
						C: 2222.9	D: 2722.7
H	Ovoid, smooth-contoured	23.5 ⊥ 20.4	83.2	468.2	Radial	A: 3157.4	B: 424.0	2566.9	+
I	Ovoid, smooth-contoured	30.5 ⊥ 21.3	85.0	487.2	Radial	A: 721.5	B: 324.5	552.3	−
						C: 303.4	D: 35.7
						E: 30.8	F: 20.0
						G: 131.1

Figure 2.

 

(A) Photomontage generated from two confocal images in extended-focus mode: origin (filled arrowhead) and coursing of a process (arrows) of a neurobiotin-filled ICNs (open arrowhead) along a choroidal blood vessel (stars). ICNs corresponds to Figure 1C . (B) Axon-like process (arrows) of a neurobiotin-filled ICN, immunonegative for nNOS (ICN corresponds to Fig. 1F ). Scale bars, 10 μm.

Figure 3.

 

(A–D) Four consecutive single optical sections from a stack of 31 confocal images (Z-step, 0.6 μm): neurobiotin-filled process (red) of an ICN tangentially bypassing the smooth muscle fibers of a choroidal blood vessel (blue), forming en passant bouton-like contacts with smooth muscle fibers (arrows). A second fiber (arrowhead) terminates on the vessel. (E–G) Three consecutive single optical sections from a stack of 16 confocal images (Z-step 0.5 μm): neurobiotin-filled process (red) of an ICN closely approaching a nonvascular smooth muscle cell immunoreactive for α-smooth muscle actin (blue) forming a boutonlike contact (arrow). Scale bars, 10 μm.

Figure 4.

 

(A) Neurobiotin-filled ICN (red; corresponds to Fig. 1H ) giving rise to a process closely embracing (arrow) an nNOS-positive ICN (green). Stars: cell bodies of two nNOS-negative ICN in the same ganglion (single optical section; scale bar 20 μm). (B) Neurobiotin-filled process of an ICN (red) closely apposed to a nitrergic ICN (green), forming a boutonlike contact (arrow; single optical section). (C) nNOS-immunoreactive ICN (green; arrow) approaching a nitrergic neurobiotin-filled ICN (resulting in a yellow color, boxed area). (D) Higher magnification of boxed area in (C) shows process of an nNOS-positive ICN (green) forming close contacts on an nNOS-positive neurobiotin-filled ICN (yellow). The nNOS-positive fiber contacts the soma (arrowhead) and intertwines with the neurobiotin-filled axon (arrow). ICN corresponds to Figure 1E (single optical section). (E, E′) Neurobiotin-filled ICN (E corresponds to the camera lucida drawing in Fig. 1I ), immunonegative for nNOS (E′). Star: position of the cell body. (F) Fibers of neurobiotin-filled ICN (red) intertwining with fibers immunoreactive for nNOS (green) in the same nerve fiber bundle (single optical section). (G) CGRP-positive nerve fibers (blue; arrows) coursing together with nNOS-immunoreactive nerve fibers (green) in the same nerve fiber bundle and forming close contacts with nitrergic (arrowheads) and nonnitrergic ICN (open arrowheads). (H, I) Nerve fibers immunoreactive for CGRP (blue) forming close contacts on ICN immunoreactive for nNOS (green) on soma (H) and dendrites (I; single optical sections). Scale bars: (A, B, D, E, E′, G, I) 20 μm; (C) 50 μm; (F, H) 10 μm.

Figure 5.

 

Neurobiotin-filled ICN (A), immunoreactive for nNOS (A′; corresponds to Fig. 1G ). (B–D) Electrophysiological recordings of this neuron (in all panels, zero potential is indicated with a small dash). (B, B′) Action potentials with brief spikes (AP half-width 1.6 ± 0.6 ms) with monophasic repolarization, reflected in the first-time derivative of the voltage trace without inflections. (C) Phasic firing behavior of the impaled neuron: firing of action potentials occurred only at onset of depolarization. (D) Action potentials were not followed by a slow afterhyperpolarization. (C, D) Top trace: transmembrane voltage; bottom trace: represents injected current. Neurobiotin-filled cell (E) immunoreactive for nNOS (E′). This cell had a stable resting membrane potential (−50 mV) but was unable to fire action potentials after direct somal depolarization. Neurobiotin injection showed that the impaled cell was indeed an ICN (the cell soma depicted corresponds to Fig. 1A ). Scale bars: (A, E) 20 μm.

The authors thank Alfons B. A. Kroese for technical advice and Nezahat Bostan, Dominique De Rijck, Lieve Svensson, Francis Terloo, and Jan Van Daele for excellent technical assistance.