Localization of Preganglionic Neurons That Innervate Choroidal Neurons of Pterygopalatine Ganglion

Sherry Cuthbertson,1 Mark S. LeDoux,1,2 Seth Jones,1 Julia Jones,1 Qihong Zhou,2 Suzhen Gong,2 Patrick Ryan,3 and Anton Reiner1

PURPOSE. The pterygopalatine ganglion (PPG) receives pregan- mammals is innervated by parasympathetic, sympathetic, glionic input from the superior salivatory nucleus (SSN) of the and sensory nerve fibers that adaptively regulate ChBF accord- facial motor complex and is the main source of parasympa- ing to retinal needs.4–11 Such adaptive control may be impor- thetic input to the choroid in mammals. The present study was tant for maintaining the health of retinal photoreceptors and undertaken to determine in rats the location and neurotrans- maintaining normal visual functioning.1,12–15 Numerous stud- mitters of SSN neurons innervating those PPG neurons that ies have shown that the pterygopalatine ganglion (PPG) is the target the choroid and to determine the location and neuro- major source of parasympathetic input to the choroid in mam- transmitters of the PPG choroidal neurons themselves. mals.6–9,16–18 The PPG efferent fibers to the choroid contain METHODS. Retrograde labeling from rat choroid using a fluores- the vasodilators VIP and nitric oxide (NO).9,17,19,20 These fi- cent tracer, in combination with immunofluorescence labeling bers also appear to be, at least in part, cholinergic.21–24 for nitric oxide synthase (NOS), vasoactive intestinal polypep- The PPG receives its preganglionic input from the superior tide (VIP), and choline acetyltransferase (ChAT), was used to salivatory nucleus (SSN) of the hindbrain through the greater characterize the location and neurotransmitters of choroidal petrosal branch of the facial nerve.25–29 The SSN itself is lo- PPG neurons. To identify SSN neurons that innervate the cho- cated dorsolateral to the facial motor nucleus. The SSN neu- roidal PPG neurons, the Bartha strain of the retrograde trans- rons, which are somewhat intermingled among and sur- neuronal tracer pseudorabies virus (PRV-Ba) was injected into rounded by noradrenergic neurons of the A5 cell group, are rat choroid, and immunolabeling for NOS or ChAT was used to cholinergic, and, in rabbits and humans, some have been re- characterize their neurochemistry. ported to contain nitric oxide synthase (NOS) as well.30–32 The RESULTS. Fluorescent retrograde labeling showed that PPG neu- SSN also provides preganglionic input through the chorda rons projecting to the choroid contained NOS, VIP, and ChAT tympani nerve to the submandibular ganglion,25,26,28,33 which and were widely distributed in PPG and its preganglionic root, ϩ sends postganglionic fibers to the submandibular and sublin- the greater petrosal nerve. SSN neurons were ChAT , and a gual glands and thereby regulates blood flow and salivary subset of them was found to contain NOS. PRV-Ba transneuro- secretion within these glands. The PPG, in addition to its nal retrograde labeling revealed that choroidal preganglionic innervation of choroidal blood vessels, innervates orbital blood neurons were localized to the rostral medioventral part of the ϩ vessels, the meibomian glands, the lacrimal gland, the hard- ipsilateral SSN. The choroidal SSN neurons were ChAT and ϩ erian gland, blood vessels of the nasal mucosa and palate, and appeared largely to correspond to the NOS neurons of the 34–39 SSN. cerebral blood vessels. Thus, functionally diverse types of preganglionic neurons CONCLUSIONS. These results show that preganglionic neurons in may be present within the SSN. Although the location within rats that are presumed to regulate choroidal blood flow the SSN of the preganglionic neurons controlling the mei- through the PPG reside within the rostral medioventral SSN, bomian glands40 and the lacrimal gland29 has been de- and that NOS is a marker for these SSN neurons. (Invest scribed, the location within the SSN of the preganglionic Ophthalmol Vis Sci. 2003;44:3713–3724) DOI:10.1167/ neurons controlling choroid is unknown. In the present iovs.02-1207 study, we sought to determine the location within the SSN he choroid is the blood supply to the outer retina (includ- in rats of those neurons that innervate the PPG neurons Ting photoreceptors), and disturbances in choroidal blood innervating the choroid (prechoroidal neurons). This study flow (ChBF) can lead to impaired retinal function and dam- also sought to determine whether these SSN neurons con- age to photoreceptors.1–3 The choroid in both birds and tain NOS and can be distinguished by morphology or location from the nearby A5 adrenergic neurons. To this end, a transneuronal retrograde tracer, the Bartha strain of pseudo- From the Departments of 1Anatomy and Neurobiology, 2Neurol- rabies virus (PRV-Ba), was injected into the choroid of adult ogy, and 3Molecular Sciences, University of Tennessee, Memphis, Ten- rats to identify prechoroidal neurons of SSN, and immuno- nessee. labeling was used to further characterize these neurons. Our Supported by the Benign Essential Blepharospasm Research Foun- results show that the prechoroidal neurons of rat SSN reside dation Inc. (MSL) and National Eye Institute Grants EY12232 (MSL) and within a characteristic location within the SSN and largely EY05298 (AR). ϩ Submitted for publication November 25, 2002; revised April 29, coincide with NOS preganglionic neurons within the SSN. 2003; accepted May 19, 2003. In addition, retrograde labeling from the choroid using flu- Disclosure: S. Cuthbertson, None; M.S. LeDoux, None; S. orescent tracer (FG; Fluorogold; Flurochrome, Englewood, Jones, None; J. Jones, None; Q. Zhou, None; S. Gong, None; P. CO), in combination with immunofluorescence, was used to Ryan, None; A. Reiner, None show that PPG neurons projecting to choroid contain NOS, The publication costs of this article were defrayed in part by page VIP, and ChAT and are widely distributed in the PPG and its charge payment. This article must therefore be marked “advertise- preganglionic root, the greater petrosal nerve. The results of ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Corresponding author: Anton Reiner, Department of Anatomy and the present studies will aid in elucidating the central sources Neurobiology, University of Tennessee Health Science Center, 855 of input to these SSN neurons and thereby help clarify the Monroe Avenue, Memphis, TN 38163; [email protected]. signals driving parasympathetic control of ChBF.

Downloaded from iovs.arvojournals.org on 10/02/2021 3714 Cuthbertson et al. IOVS, September 2003, Vol. 44, No. 9

MATERIALS AND METHODS The experimental design of the present studies using PRV-Ba, therefore, incorporated these considerations. Twenty-three rats received Subjects and Approach PRV-Ba injections into the choroid. These rats were anesthetized with an intraperitoneal injection of a ketamine-xylazine mix of 87 and 13 The results reported are based on studies in 40 adult Sprague-Dawley mg/kg, and the right superior–temporal choroid was injected with 1.0 rats (220–550 g; Harlan Inc., Indianapolis, IN). All experiments per- to 2.0 ␮L of PRV-Ba (3 ϫ 108 plaque forming units/mL). The same formed were in compliance with the ARVO Statement for the Use of injection approach was used as for the above described intrachoroidal Animals in Ophthalmic and Vision Research and with National Insti- injections of FG. In 7 of the 23 cases, a strain of PRV-Ba bearing a LacZ tutes of Health and institutional guidelines. To identify SSN neurons construct (which codes for the enzyme ␤-galactosidase) was used. The involved in the control of ChBF, we injected a retrograde transneuro- animals were allowed to survive between 52 and 88 hours after virus nal tracer, the Bartha strain of pseudorabies virus (PRV-Ba), into the injection. All PRV-Ba injections were performed with a syringe (Ham- choroid of rats. Brains from these injected rats, or from normal animals, ilton) with a 30-gauge needle. The rats receiving an intrachoroidal were fixed by transcardial perfusion and processed by immunohisto- injection of PRV-Ba had already received bilateral resections of the chemistry to visualize PRV-Ba, NOS, ChAT, tyrosine hydroxylase (TH), superior cervical ganglia (SCG) to prevent retrograde transneuronal or pairs of the aforementioned (as well as by histochemistry to visual- labeling along sympathetic circuitry.40 Bilateral resections were per- ize reduced nicotinamide adenine dinucleotide phosphate diaphorase formed rather than unilateral because of evidence that the superior [NADPHd]), for the purpose of characterizing the location and neuro- cervical ganglion has a contralateral orbital projection.37,39 The SCG chemistry of neurons in the rat SSN that innervate PPG neurons that lies immediately dorsal to the bifurcation of the common carotid target choroid. To characterize the PPG neurons themselves involved artery. A single ventral midline neck incision allowed access to both in the control of ChBF, a retrograde tracer, FG (FluoroGold, Fluoro- the right and left SCG. Careful blunt and sharp dissection was used to chrome) was injected into the choroid of additional rats. These animals localize the common carotid artery and then the cervical portion of the were fixed by transcardial perfusion. The eyes, the greater petrosal sympathetic trunk. The cervical portion of the sympathetic trunk and nerves, and the PPG from the injected side were sectioned and FG in SCG were freed from the carotid artery and excised in toto. Because the PPG was visualized by fluorescence microscopy. PPG sections with ϩ PRV-Ba does not typically demonstrate transganglionic transport FG neurons were processed by immunofluorescence to visualize through sensory ganglia,33,40,42 there was no reason to transect the NOS, ChAT, or vasoactive intestinal polypeptide (VIP), to characterize ophthalmic nerve to prevent central transport of PRV through trigem- the neurochemistry of PPG neurons innervating the choroid. inal circuitry. The absence of PRV-Ba–labeled neuronal perikarya or FG Injection terminals within the trigeminal nuclear complex in our studies confirmed that there was no transganglionic transport of virus through the Seven rats received bilateral FG injections into the choroid. The rats trigeminal nerve in our rats. Thus, our bilateral SCG-ectomies in con- were anesthetized with an intraperitoneal injection of a ketamine- junction with the absence of central PRV-Ba transport through the xylazine mix of 87 and 13 mg/kg, and both the right and left choroid trigeminal nerve served to prevent virus labeling in the central nervous were injected with 1 ␮L of 2% to 4% FG. The FG was injected into the system (CNS) in sympathetic preganglionic neurons, in the hindbrain superior–temporal sector of the choroid. The animals were allowed to trigeminal nerve targets, and (in the case of longer survival times) in survive approximately 90 hours after the injections. All FG injections the higher order neurons projecting to them.40 were performed with a syringe (Hamilton, Reno, NV) connected to a 30-gauge needle. The needle tip punctured the conjunctiva and then Histochemical Studies and Colchicine Treatment the sclera posterior to the ciliary complex before entering the choroidal space. Tracer was slowly injected into the choroid over 3 to 5 Ten rats were used exclusively in immunohistochemical and/or histo- minutes. The conjunctival puncture site was carefully monitored chemical studies of the SSN region. None of these rats received ocular through a surgical microscope for efflux of tracer. In all cases, there or orbital virus injections or had the superior cervical ganglia resected. was minimal or no evident efflux, and any efflux was blotted with a Two of them, however, were treated with the axonal transport blocker sterile cotton swab. The conjunctival sac was rinsed with sterile nor- colchicine to enhance visualization of neurotransmitter-related sub- mal saline at the conclusion of the injection procedure to prevent stances, notably NOS, in the SSN neuronal perikarya, as described 43 further spread of tracer to extraocular tissues. Inadvertent spread of FG previously. These rats were anesthetized with ketamine (0.66 mL/kg) from our intrachoroidal injections to the lacrimal glands can be ruled and xylazine (0.33 mL/kg) and secured in a stereotaxic device. Body ␮ ␮ out in all cases, because the main lacrimal gland in rats is extraorbital, temperature was maintained at 38°C, and colchicine (45 g/1 L; and the secondary lacrimal gland—the infraorbital gland—is located Sigma-Aldrich, St. Louis, MO) was injected into the fourth ventricle. 44 below the eyeball, well removed from the site at which we injected Coordinates for the injection site were from Paxinos and Watson. ␮ into the choroid.41 Similarly, the meibomian glands of the eyelids were Four microliters of colchicine was injected with a 10- L syringe (Ham- ␮ remote from our intrachoroidal penetration site,40 and thus accidental ilton) at a rate of 1 L every 10 minutes. Animals were allowed to spread of tracer to them could be ruled out as well. survive for 30 to 36 hours after the colchicine injection. All 10 animals were then processed for histologic analysis. PRV-Ba Injections Histologic Tissue Preparation Viruses such as PRV-Ba can be uniquely valuable as pathway tracing agents for delineating central circuits, because (unlike a conventional Normal rats (n ϭ 8), colchicine-treated rats (n ϭ 2), rats that had retrograde tracer such as FG) they are transported retrogradely trans- received PRV-Ba injections (n ϭ 23), and rats that had received FG neuronally (i.e., across synapses) and provide robust labeling in recip- injections (n ϭ 7) were anesthetized with an intraperitoneal injection ient neurons due to virus replication.40 Nonetheless, use of viruses as of 0.1 mL/100g of a ketamine-xylazine mixture (87 and 13 mg/kg). In neuronal tracers is not without potential pitfalls, which include virus- all rats, 0.4 mL of heparinized saline was injected into the heart, and induced neuronal degeneration with longer postinoculation survival then they were transcardially perfused with 0.9% saline followed by times, failure of certain sets of neurons to transport the virus, and either 4% paraformaldehyde in 0.1 M sodium phosphate buffer (PB; pH variability between cases in the extent of transneuronal labeling. Use of 7.2–7.4) or 4% paraformaldehyde in 0.1 M lysine-0.01 M sodium perio- the smallest effective doses of virus and an attenuated strain such as the date in 0.1 M PB (pH 7.2–7.4). Brains were postfixed for 1 to 4 hours PRV-Ba can mitigate neuronal injury caused by the virus.40 Use of at room temperature in the same fixative used for perfusion and then several animals at each of several postinoculation survival times and cryoprotected at 4°C for at least 24 hours in a 20% sucrose, 10% comparison with the results obtained with conventional tracers can glycerol, and 0.02% sodium azide in 0.1 M PB solution. Once the brains help overcome case-to-case variability and possible labeling failures.40 were cryoprotected, they were frozen with dry ice and sectioned on a

Downloaded from iovs.arvojournals.org on 10/02/2021 IOVS, September 2003, Vol. 44, No. 9 Preganglionic Neurons Controlling Choroidal Blood Flow 3715

sliding microtome at 40 ␮m. Sections were collected as six parallel brief, tissue from normal rats was incubated in a primary antibody series, and one series was mounted immediately during sectioning on cocktail of rabbit anti-chicken ChAT (1:1000) and mouse anti-TH (1: gelatin-coated slides, allowed to dry, and stained with cresyl violet. The 1000) diluted with PB/Tx/Azϩ5% normal horse serum for 48 to 72 cresyl violet series for each case made it possible to identify unambig- hours at 4°C. Tissue was rinsed in 0.1 M PB and then incubated at room uously the rostral-to-caudal order of all sections from that case. The temperature for 1 hour in a donkey anti-mouse secondary antibody remaining free-floating sections were stored at 4°C in a 0.02% sodium (1:50, Jackson ImmunoResearch Laboratories). Sections were then azide and 0.02% imidazole in 0.1 M PB solution until they were labeled rinsed in 0.1 M PB and incubated in mouse PAP for 1 hour at room for PRV-Ba, ␤-galactosidase (␤-gal), ChAT, NOS, and/or TH by immu- temperature (1:200; Sternberger Monoclonals). The tissue was next nohistochemistry, or for NADPHd by histochemistry. rinsed in 0.1 M PB (pH 7.2–7.4), and the TH labeling visualized using DAB in 0.1 M PB with 0.04% nickel ammonium sulfate (pH 7.2–7.4), ϩ Peroxidase-Antiperoxidase Single-Labeling resulting in brown-black TH cells, as described in our prior stud- 50,52 Immunohistochemistry ies. The tissue was then rinsed extensively, incubated in donkey anti-rabbit (1:50, Jackson ImmunoResearch Laboratories) for 1 hour at Immunohistochemical single labeling was performed as described pre- room temperature, rinsed three times for five minutes each in 0.1 M viously.11,43 The primary antibodies used were goat anti-PRV-Ba diluted PB, and incubated in rabbit PAP (1:200; Sternberger Monoclonals) for 1:15,000 to 1:50,000,40,45 rabbit anti-␤-gal diluted 1:50,000 (Rockland 1 hour at room temperature. The sections were subsequently rinsed in Immunochemicals, Gilbertsville, PA), rabbit anti-chicken ChAT diluted 0.1 M PB (pH 7.2–7.4), and the ChAT labeling visualized using DAB in 1:1,000 (generously provided by Miles Epstein and Carl D. Johnson, a 0.2 M sodium cacodylate buffer (pH 7.2–7.4), resulting in brown DAB University of Wisconsin), goat anti-ChAT diluted 1:250 (Chemicon labeling of ChATϩ neurons, as described previously.49,50,52 The sec- International, Inc., Temecula, CA), rabbit anti-NOS diluted 1:1,000 to tions were subsequently rinsed in 0.1 M PB, mounted on gelatin-coated 1:2,000 (Alexis Biochemicals, San Diego, CA), rabbit anti-NOS diluted slides, air-dried, dehydrated, and coverslipped in nonaqueous medium 1:400 to 1:1,000 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), and (Permount; Fisher Scientific). The sections were examined with a mouse anti-TH diluted 1:1,000 (DiaSorin Inc., Stillwater, MN). The microscope (BHS; Olympus) with standard transmitted light or differ- diluent for all antisera or antibodies was 0.1 M phosphate buffer, 0.3% ential interference contrast optics. Camera lucida drawings were made ϩ Triton X-100, and 0.01% sodium azide solution (PB/Tx/Az) 5% normal through the levels of SSN/A5 of labeled neurons in sections double- ␤ horse serum. The anti-PRV-Ba or anti- -gal was used to detect trans- labeled for ChAT and TH by the two-color DAB immunolabeling neuronal retrograde labeling with the virus. The anti-ChAT was used to method. identify PPG preganglionic neurons in the SSN, the anti-NOS was used to determine whether SSN neurons contain this enzyme (which pro- duces nitric oxide), and the anti-TH was used to identify A5 neurons. Immunofluorescence Double Labeling The specificity of these primary antisera has been demonstrated pre- Immunofluorescence double labeling was performed to determine viously.45–48 whether PRV-Ba–labeled neurons in the SSN were cholinergic (and For the immunolabeling studies of brain, free-floating sections were thus preganglionic) and whether the PRV-Ba–labeled neurons of SSN rinsed and pretreated in 1% NaOH with 0.5% H2O2 in 0.1 M PB for 15 minutes followed by 1% nonfat dry milk for 15 to 30 minutes. The also contain NOS. Animals with intrachoroidal injection of PRV-Ba NaOH enhances antigenicity (especially for ChAT), the H O inacti- were used for this analysis. Immunofluorescence double labeling was 2 2 also used to determine whether the NOS-immunolabeled neurons of vates endogenous peroxidases, and the 1% nonfat dry milk reduces ϩ nonspecific background immunostaining. The sections were incubated SSN contain ChAT, as a means of further assessing whether the NOS in primary antisera for 48 to 72 hours at 4°C in plastic 5 mL vials.43 neurons might be preganglionic. Colchicine-treated sections and sec- Sections were then rinsed in 0.1 M PB and incubated for 1 hour at room tions with intrachoroidal injection of PRV-Ba were used for this anal- temperature in a bridging secondary antiserum directed against IgG of ysis. The immunofluorescence double-labeling method was performed 11,43 the host in which the primary antibody was raised (1:50, secondary as described previously. In brief, tissue was pretreated by incuba- antibodies; Jackson ImmunoResearch Laboratories, Inc., West Grove, tion in 1% NaOH with 0.5% H2O2 in 0.1 M PB for 15 minutes followed PA). The sections were then rinsed in 0.1 M PB and incubated for 1 by 1% nonfat dry milk for 15 to 30 minutes and then incubated in a hour at room temperature in rabbit, goat, or mouse peroxidase-anti- primary antibody cocktail for 48 to 72 hours at 4°C. To examine the peroxidase (PAP, all at 1:200, rabbit and mouse PAP from Sternberger localization of ChAT or NOS in the PRV-Ba–labeled SSN neurons, the Monoclonals, Lutherville, MD, and goat PAP from Jackson Immuno- primary antibody cocktail contained goat anti-PRV-Ba (diluted Research Laboratories). The sections were then rinsed in 0.1 M PB (pH 1:15,000–1:50,000) and either rabbit anti-nNOS (1:1,000–1:2,000) or 7.2–7.4), and the labeling visualized using diaminobenzidine tetrahy- rabbit anti-chicken ChAT (1:1,000). In those cases in which PRV-Ba drochloride (DAB) in a 0.2 M sodium cacodylate buffer (pH 7.2–7.4), as bearing a LacZ construct had been injected intrachoroidally, some in our previous work.11,43,49 The sections were subsequently rinsed, additional sections were processed for immunofluorescence double mounted on gelatin-coated slides, air dried, dehydrated and cover- labeling with the rabbit anti-␤-gal and the goat anti-ChAT antibodies. To slipped with nonaqueous mounting medium (Permount; Fisher Scien- examine the colocalization of ChAT and NOS in the SSN neurons, we tific, Pittsburgh, PA). The sections were examined with a microscope used a primary antibody cocktail containing rabbit anti-nNOS (BHS; Olympus, Lake Success, NY) with standard transmitted light or (1:1,000–1:2,000) and goat anti-ChAT (1:250). The diluent in all cases differential interference contrast optics. Camera lucida drawings were was PB/Tx/Azϩ5% normal horse serum. made of sections through the levels of SSN and A5 in representative After incubation in primary antibodies, the sections used for detec- cases to characterize the location and extent of the perikaryal labeling tion of ChAT or NOS in virus-labeled neurons were rinsed in 0.1 M PB for PRV-Ba, ChAT, TH, and/or NOS, as well as for NADPHd (NADPHd and incubated for 1 to 2 hours at room temperature in a secondary labeling methods described later). antisera cocktail containing a green fluorophore (Alexa 488)–conjugated donkey anti-goat IgG antiserum (1:100–1:500; Molecular Probes, Peroxidase-Antiperoxidase: Two-color DAB Inc., Eugene, OR), and tetramethylrhodamine isothiocyanate (TRITC)– Double-Label Immunohistochemistry conjugated donkey anti-rabbit (1:100; Jackson ImmunoResearch Labo- ratories). After primary antibody incubation, sections used to assess the Immunohistochemical two-color DAB double labeling was performed colocalization of ChAT and NOS were incubated in an antiserum to map the relative locations of the THϩ noradrenergic neurons of the cocktail containing TRITC-conjugated donkey anti-rabbit IgG (1:100; A5 cell group and the cholinergic neurons of SSN in the same sections. Jackson ImmunoResearch Laboratories) and green fluorophore–conju- This labeling method was performed as described previously.50–52 In gated donkey anti-goat IgG (1:100–1:500, Alexa 488; Molecular Probes,

Downloaded from iovs.arvojournals.org on 10/02/2021 3716 Cuthbertson et al. IOVS, September 2003, Vol. 44, No. 9

Inc.). The sections were then rinsed, mounted on gelatin-coated slides, NADPHd. For NADPHd histochemistry, free-floating brain sections and air dried, and coverslipped with several drops of glycerol carbonate slide-mounted eye sections were allowed to warm to room tempera- buffer (pH 10.5) or several drops of anti-fade coverslip-mounting me- ture, rinsed in 0.1 M PB three times for 5 minutes each, and placed in dium (ProLong; Molecular Probes, Inc.). the incubation medium containing 0.1 M Tris-HCl (pH 8.0), 1 mM In some PRV-Ba cases, the brains were sectioned transversely at 20 ␤-NADPH, 0.2 mM nitroblue tetrazolium (NBT) and 0.2% Triton X-100. ␮m using a cryostat (Leica, Deerfield, IL), and sections were collected Both free-floating and slide-mounted sections were incubated for 4 to on slides. For immunofluorescence double labeling of this tissue, the 20 minutes at 37°C. Progress of the reaction was assessed by micro- slide-mounted sections were circled with a hydrophobic slide-marking scopic examination of the tissue. Free-floating sections were mounted pen (PAP pen; Electron Microscopy Sciences, Fort Washington, PA) on gelatin-coated slides when labeling was complete. All sections were and dried on a slide warmer. Sections were then rinsed in 0.02 M dried, dehydrated, cleared, coverslipped, and examined with a trans- phosphate-buffered saline (PBS) with 0.1% sodium azide. Endogenous mitted-light microscope (BHS; Olympus). To determine the specificity peroxidases were inactivated by a 20-minute incubation in 10% meth- of NADPHd staining, sections of brain tissue through the SSN were anol and 3% hydrogen peroxide in PBS. Sections were next rinsed in stained with a solution that contained all the ingredients for the PBS and pretreated with 2% nonfat milk and 0.3% Triton X-100 for 1 NADPHd histochemical procedure, except the NADPH itself.58 Adja- hour at room temperature and incubated overnight in droplets of a cent sections from the same animal were stained in parallel for primary antibody cocktail containing goat anti-PRV-Ba (diluted NADPHd with a solution containing all ingredients. This control pro- 1:100,000 with PBSϩ0.1% Tritonϩ3% normal donkey serum) and ei- cedure yielded no SSN labeling. ther rabbit anti-nNOS (1:500; Santa Cruz Biotechnology, Inc.), rabbit anti-nNOS (1:1000; Alexis Biochemicals), or rabbit anti-chicken ChAT (1:1000). Secondary antibodies—namely, donkey anti-goat IgG conjugated to Cy2 and donkey anti-rabbit IgG conjugated to Cy5—were used RESULTS at 1:250 dilutions. Tissue was incubated in secondary antisera for 4 hours, rinsed, dehydrated, cleared, and coverslipped with 1,3-diethyl- Labeling in PPG after Intrachoroidal FG Injection 8-phenylxanthine (DPX; Sigma-Aldrich). Labeling was observed in the ipsilateral PPG for five of the eyes The sections double-labeled by immunofluorescence were viewed that received an intrachoroidal FG injection. The FG-labeled with an epi-illumination fluorescence microscope (Olympus), as 49 neurons in each case were found within the proximal part of described previously, or with a confocal laser scanning micro- the ipsilateral PPG and scattered along the greater petrosal scope (CLSM; MRC-1000; Bio-Rad, Richmond, CA), as described previ- 53,54 ϫ nerve itself, in which the preganglionic fibers to PPG travel. ously. For CLSM examination, a 20 objective (Euplan; Olympus) The FG labeling at the intrachoroidal injection sites showed was used, and sections were scanned with a krypton-argon laser, with that these intrachoroidal injections were largely restricted to specific excitation wavelength settings for TRITC (568 nm) and dichlo- choroid and typically involved no more than approximately rotriazinylamino fluorescein (DTAF; 488 nm for green fluorescence 25% to 50% of the choroidal circumference (Fig. 1). Whereas visualization), or for far red (647 nm for Cy5) and DTAF (for Cy2). slight leakage of FG into the musculature surrounding the Images of the labeling for individual fluorophores were captured se- temporal side of the eye was observed in all five of the eyes quentially for examination and analysis. yielding PPG neuron labeling, similar extrachoroidal labeling was observed in six eyes in which the attempted intrachoroidal Immunofluorescence Combined with FG FG injections failed to yield PPG labeling and showed no Retrograde Labeling evident labeling of the choroid at the injection site. Similarly, in two of the cases with neuronal labeling in the PPG, FG spread In animals that had intrachoroidal injections of FG, the injected eyes to a small laterally directed portion of the harderian gland and attached nerves were removed from the orbit and sectioned on a (constituting no more than 1%–2% of its volume) was ob- cryostat (Hacker-Bright Instruments, Fairfield, NJ) at 20 ␮minthe served. The pattern of PPG labeling was, however, the same as horizontal plane. Sections were then collected on slides (Superfrost in the successful intrachoroidal cases without spread to the Plus; Fisher Scientific). The eyes were similarly sectioned to assess the harderian gland. Moreover, one of the failed choroidal injec- FG injection sites. The sections were then viewed with an epi-illumi- tions showed slight spread to a small temporal portion of the nation fluorescence microscope (Olympus), as described previously.49 harderian gland, but yielded no PPG labeling. Thus, it seems In some cases in which FG labeling had been identified in PPG neu- unlikely that FG spread outside the choroid contributed signif- ϩ rons, sections through the PPG with FG neurons were processed for icantly to the PPG labeling. Similarly, whereas some FG spread immunofluorescence. The sections were incubated overnight in a into temporal retina was evident in four of the cases with PPG humid chamber at 4°C in either rabbit anti-nNOS (1:1000; Alexis labeling, no anterograde labeling was observed in the optic Biochemicals), rabbit anti-VIP (1:1000; DiaSorin), or goat anti-ChAT nerve, tract, or central retinal target areas with retinal spread, (1:250; Chemicon International, Inc.). After incubation in the primary suggesting that FG had not leaked significantly into the vitreous antibodies, the sections were rinsed in 0.1 M PB and incubated for 1 to from the choroidal injection site. Moreover, similar leakage 2 hours at room temperature in TRITC-conjugated donkey anti-rabbit into the retina was evident in the failed intrachoroidal injected IgG in the case of the tissue incubated in either anti-NOS or anti-VIP eyes without PPG retrograde labeling, thus further indicating (1:100; Jackson ImmunoResearch Laboratories) or in green fluoro- that such leakage was not the basis of the retrograde PPG phore (Alexa 488)–conjugated donkey anti-goat IgG in the case of labeling in the four eyes with effective intrachoroidal injec- tissue incubated in anti-ChAT (1:500; Molecular Probes, Inc.). tions. Based on the maximum number of retrogradely labeled neurons observed in the PPG and greater petrosal nerve after NADPHd Histochemistry intrachoroidal FG injection, our results suggest that at least 200 to 400 PPG neurons innervate the choroid in the rat. Our A standard NADPHd histochemical procedure was used to localize immunolabeling studies on the FG-labeled sections through NADPHd activity in eye tissue sectioned in the frontal plane and in PPG confirmed that PPG neurons in general and those inner- free-floating sections through the pons of rats.55–58 The eyes were vating choroid in particular contain NOS (Figs. 2A, 2D), VIP removed from the orbit and sectioned on a cryostat (Hacker-Bright) at (Figs. 2B, 2E), and ChAT (Figs. 2C, 2F) in rats. Moreover, we 20 ␮m and the sections collected on slides (Superfrost Plus; Fisher confirmed that the choroid is rich in nerve fibers containing Scientific) and stored at Ϫ20°C until histochemically processed for the NOS marker NADPHd (Fig. 1C), as reported previously by

Downloaded from iovs.arvojournals.org on 10/02/2021 IOVS, September 2003, Vol. 44, No. 9 Preganglionic Neurons Controlling Choroidal Blood Flow 3717

brown DAB for ChAT) had been used to distinguish the two neuron types (Fig. 3).

Labeling in the SSN after Intrachoroidal PRV-Ba Injection Retrograde transneuronal PRV-Ba labeling was observed in the SSN after intrachoroidal injection of the virus, after various postinjection survival periods (52–88 hours). With longer survival times, more neurons in and around the SSN were labeled, whereas with short survival times the labeling was more con- FIGURE 1. (A) Temporal sector of the retina, choroid, and sclera from a rat eye that had an FG injection into the temporal choroid. The FG fined to a limited area within the SSN. For example, in six successful cases with less than 70 hours’ survival, an average of labeling is evident as a strong white glow in the choroid and sclera. ϩ Note the absence of spread into the retina. (B) The nasal sector of the 16.8 PRV-Ba neurons was observed in the SSN at the 10.5 mm same eye shows that FG labeling did not spread to the nasal choroid or level (Fig. 3), and these were confined to a small cluster at the retina. (C) NADPHd histochemical staining of a section through the rat rostral ventromedial aspect of the SSN. By contrast, in five outer retina and choroid. A dense meshwork of fibers (arrows) within successful cases with more than 70 hours’ survival, an average the choroid just external to the choriocapillaris (ChCap) is labeled. ϩ ϩ of 33.8 PRV-Ba neurons was observed in the SSN at the NADPHd is a marker for NOS fibers and (C) shows that the rat 10.5-mm level (Fig. 3), and these were more widely distributed choroid is rich in nitrergic innervation. Prior studies by others show in the SSN. In addition, with survival times beyond 70 hours, that these fibers primarily arise from the PPG, and co-contain VIP.17,18 (A, B) Same magnification. perikaryal labeling for PRV-Ba became evident within the A5 region. Given its late appearance, the PRV-Ba–labeled neurons within the A5 after intrachoroidal injection of the virus may be others, which are known to arise from the PPG and to co- higher-order labeling from the SSN cluster. contain VIP.17,18 The cluster of PRV-Ba–labeled neurons observed in the rostral ventromedial SSN with short survival times (Ͻ70 hours) Neurochemistry of SSN Neurons: ChAT versus TH appears attributable to transneuronal retrograde transport from The distributions of ChATϩ neurons and THϩ neurons in and the choroid through the PPG neurons innervating the choroid. around the SSN were somewhat overlapping but nonetheless For example, although there was evidence of some variable distinct (Fig. 3). The ChATϩ neurons in this region included spread of the PRV-Ba from the injection site to nearby perior- the somatomotor neurons of the facial nucleus as well as bital facial muscles (such as the orbicularis oculi) or into the neurons of the SSN. The somatomotor facial nucleus neurons extraocular muscles, as evidenced by labeling in the somato- are found within a circumscribed oval-shaped field that begins motor neurons of the facial nucleus or the nuclei innervating just caudal to the superior olive and the exit of the facial nerve the extraocular muscles (oculomotor, trochlear, and abdu- itself, and it extends to just rostral to the level of the inferior cens), the size and location of the PRV-Ba–labeled neuronal olive. The somatomotor neurons are relatively large (20–30 cluster in the SSN was invariant at short survival times, regard- ␮m) and labeled intensely for ChAT. The neurons of the SSN, less of the occurrence of spread to any of this somatic muscu- by contrast, were smaller (10–15 ␮m) and more lightly labeled lature. Moreover, there is no known route by which SSN for ChAT, and they occupied a field at the dorsolateral margin labeling could arise transneuronally after retrograde labeling of of the somatomotor facial nucleus, which extended from the lateral aspect of the somatomotor facial nucleus toward the periventricular gray (Figs. 3, 4). A small rostral part of this field was observed to extend over the dorsolateral margin of the caudalmost part of the superior olive, flanking the medial aspect of the facial nerve as it exits the brain stem. The THϩ neurons within the vicinity of the facial somatomotor nucleus and the SSN constitute the A5 noradrenergic cell group.29,59,60 Both the SSN and A5 are parts of central circuitry controlling the peripheral vasculature. For this reason and because of their proximity, it was important to distinguish the two neuronal populations. The THϩ perikarya were smaller (15–20 ␮m) and more spindle-shaped than the cholinergic neurons of the facial somatomotor nucleus, but slightly larger than the cholinergic neurons of the SSN. In addition, the THϩ FIGURE 2. A series of images of fluorescent labeling showing that neurons of A5 were most abundant at the level of the superior neurons of the PPG that project to the choroid contain NOS, VIP, and olive and facial nerve root (Fig. 3). Although the most rostral ChAT. (A) Neurons of the PPG that had been retrogradely labeled by part of the cholinergic SSN neuron field overlapped the caudal ϩ intrachoroidal FG injection into the temporal sector of the choroid, part of the TH neuron field, at this point of overlap the bulk whereas (D) shows NOSϩ immunolabeling in this same field. NOS is ϩ of the TH neurons were located more medially and ventrally present in most of the FG-labeled PPG neurons. (A, D, arrows) Some than the bulk of the SSN neurons. At more caudal levels, the of the FG-labeled neurons that are also NOSϩ.(B) Neurons of the PPG ϩ that were retrogradely labeled by the same intrachoroidal injection of few TH neurons observed tended to be ventral to the SSN ϩ neurons. Thus, although the THϩ neurons of A5 and the cho- FG. (E) VIP immunolabeling in this same field. VIP is also present in ϩ most of the FG-labeled PPG neurons. (B, E, arrows) Some of the linergic ChAT neurons of the SSN overlap slightly in their ϩ FG-labeled neurons that were also VIP .(C) Neurons of the PPG that distributions, the two cell types occupy distinct territories and had been retrogradely labeled by the same intrachoroidal injection of differ in their size and shape. This was evident in both adjacent FG as in (A) and (B). (F) ChATϩ immunolabeling in the same field. sections that had been single-labeled for TH and ChAT, respec- ChAT was present in most of the FG-labeled PPG neurons. (C, F, tively, and in individual sections in which two-color DAB dou- arrows) Some of the FG-labeled neurons that are also ChATϩ. All ble labeling (nickel intensified brown-black DAB for TH and images are at the same magnification.

Downloaded from iovs.arvojournals.org on 10/02/2021 3718 Cuthbertson et al. IOVS, September 2003, Vol. 44, No. 9

FIGURE 3. Camera lucida drawings of the right side of the brain in a rostral (top row)-to-caudal (bottom row) series of transverse sections showing the distribution within the rat SSN region of cholinergic neurons immunolabeled for ChAT (illustrated by dots in the drawings in the left column), noradrenergic neurons immunolabeled for TH (illustrated by triangles in the middle column), and neurons transneuronally retrogradely labeled from the choroid with PRV-Ba (᭜, right column). Numbers at right: rostrocaudal level of each section in the respective row in stereotaxic coordinates,44 with the rostrocaudal position speciﬁed in terms of the distance behind (P for posterior) the skull suture Bregma. The ChAT and TH immunolabeled sections are from one normal rat, and the ChAT and TH labeling were mapped from the same individual sections that had been double-labeled using the two-color DAB method (brown DAB for ChAT and brown-black DAB for TH). The PRV- Ba–labeled sections are from a different rat, which received bilateral resections of the superior cervical ganglia and survived 63 hours after right intrachoroidal PRV-Ba injection. The ﬁrst two columns show that the THϩ neurons of the A5 cell group and the cholinergic neurons of SSN had only slight overlap, with the SSN neurons largely lying caudomedial to the A5. The PRV-Baϩ neurons transneuronally retrogradely labeled from the choroid were restricted in their location to the rostral and ventromedial part of the SSN (at P10.5). g7, genu of the facial nerve; M5, trigeminal motor nucleus; mlf, medial lon- gitudinal fasciculus; n6, abducens nucleus; n7, facial motor nucleus; N7, facial nerve root; nTTd, nucleus of the descending trigeminal tract; Pr5, principal trigeminal nucleus; py, py- ramidal tract; SO, superior olivary nucleus; SSN, superior salivatory nucleus; TTd, descending trigeminal tract; A5, adrenergic cell group 5.

these somatomotor neuron pools.27 Note that these cases were Neurochemistry of PRV-Ba–labeled SSN also not confounded by leakage of virus into the vitreous, Prechoroidal Neurons: ChAT because this would have resulted in labeling of neurons in the nucleus of Edinger-Westphal61 and several hypothalamic and Double labeling for ChAT and PRV-Ba was performed to deter- ϩ pretectal retinorecipient groups, and no such labeling was mine whether the PRV-Ba neurons within the SSN cluster observed. Thus, the prechoroidal neurons of the SSN (i.e., observed in the single-labeling studies of virus localization those innervating PPG neurons targeting ipsilateral temporal (P10.5 level in Fig. 3) in fact were cholinergic (Table 1). choroid) appear to be restricted to a ventromedial and some- Examination of short survival cases (Ͻ70 hours postinjection what rostral part of the SSN field, as the field is defined by survival) revealed that nearly all of the PRV-Baϩ neurons in the ChAT immunolabeling (Figs. 3, 4, 5). small cluster of PRV-Baϩ neurons in the SSN contained ChAT

Downloaded from iovs.arvojournals.org on 10/02/2021 IOVS, September 2003, Vol. 44, No. 9 Preganglionic Neurons Controlling Choroidal Blood Flow 3719

FIGURE 4. A series of images at two different magnifications of transverse sections of the SSN at the rostral level of the facial somatomotor nucleus (n7), showing the location of the neurons within the SSN that appear to be the preganglionic neurons for the choroidal neurons of the PPG. (A, C) Sections from a normal rat at the P10.5 level (Fig. 3) immunolabeled for ChAT. This pair of images shows that SSN and n7 neurons were ChATϩ, and SSN was located at the dorsolateral edge of n7. (B, D) PRV-Baϩ neurons in the SSN. These images are from the same animal, which survived 63 hours after intrachoroidal PRV-Ba, as shown in the drawings in Figure 3. The PRV-Baϩ prechoroidal neurons were restricted to a region at the dorsolateral edge of the somatomotor facial nucleus, in the more ventromedial part of the SSN field at this level. Note that this region contained numerous fiber bundles (pale, round areas) that coursed longitudinally through the brain stem, with the dendrites of SSN neurons surrounding these fiber bundles. All four images are of the right side of the brain, with medial to the left and dorsal toward the top.(A, B) Same magnification; (C, D) same magnification.

(Fig. 5). For example, counts of 91 PRV-Baϩ neurons in the SSN cluster in four cases with short survival (Ͻ70 hours) revealed that 94.2% of these PRV-Baϩ contained ChAT. The observation FIGURE 5. CLSM images of a single field of view of the SSN from tissue ϩ double labeled with immunofluorescence for PRV-Ba (A) and ChAT that the PRV-Ba neurons in the small cluster within the SSN (B), from an animal that underwent bilateral resection of the superior (at least at short survival times) are typically cholinergic sup- cervical ganglia and survived 65 hours after unilateral virus injection ports the view that they are parasympathetic preganglionic into the choroid. Arrows: neurons within the SSN that were double- neurons that project to those PPG neurons that innervate the labeled for PRV-Ba and ChAT. Most of the PRV-Baϩ neurons shown choroid. By contrast, within the A5 field, PRV-Baϩ neurons were ChATϩ, but not all ChATϩ neurons were PRV-Baϩ. These results ϩ rarely were ChATϩ, supporting the view that virus-labeled indicate that the PRV-Ba neurons within the SSN labeled transneuro- neurons within this area are not SSN neurons. nally from the choroid were cholinergic preganglionic neurons, and The PRV-Baϩ neurons in the SSN labeled by intrachoroidal they represented only a subset of SSN neurons. The field shown is injection did not include all SSN neurons, because in the four of the right side of the brain, with medial to the left and dorsal toward the top. Both images are at the same magnification. The cases with the shortest survival times examined by double ChAT labeling was visualized with a secondary antibody conjugated labeling for ChAT and PRV-Ba (which provide the most reliable to CY5 and pseudocolored red for ease of visualization with an delineation of the prechoroidal neurons of the SSN), only image-analysis program (Photoshop; Adobe Systems, Mountain ϩ 27.5% of the ChAT neurons within the SSN at the level of the View, CA). cluster were virus-labeled. With longer survival times, which yields a greater number of SSN neurons labeled for virus, the

TABLE 1. Co-localization in SSN Neurons

Localization of ChAT or NOS in PRV-Ba؉ SSN Neurons

Less Than 70-h Survival More Than 70-h Survival

94.2% of PRV-Baϩ neurons contain ChAT (n ϭ 4) 76.7% of PRV-Baϩ neurons contain ChAT (n ϭ 3) 27.5% of ChATϩ neurons are PRV-Baϩ (n ϭ 2) 30.5% of ChATϩ neurons are PRV-Baϩ (n ϭ 2) 28.9% of PRV-Baϩ neurons contain NOS (n ϭ 3) 40.0% of PRV-Baϩ neurons contain NOS (n ϭ 1) 33.4% of NOSϩ neurons are PRV-Baϩ (n ϭ 3) 40.0% of NOSϩ neurons are PRV-Baϩ (n ϭ 1)

Co-localization of ChAT and NOS in SSN Neurons

36.4% of ChATϩ neurons contain NOS (n ϭ 2) 79.1% of NOSϩ neurons contain ChAT (n ϭ 2)

Downloaded from iovs.arvojournals.org on 10/02/2021 3720 Cuthbertson et al. IOVS, September 2003, Vol. 44, No. 9

stricted subset of the SSN neurons defined by ChAT immunolabeling. Distinct NADPHdϩ and NOSϩ perikarya were not observed in the facial somatomotor nucleus in either normal or in colchicine-treated rats. Counts of 52 NOS-labeled neurons in two cases double-labeled by immunofluorescence for ChAT and NOS (Table 1) revealed that 79.1% of the NOSϩ perikarya within the SSN were ChATϩ (Fig. 7), thereby suggesting most NOSϩ neurons to be parasympathetic preganglionic SSN neurons. Because of the limited extent of the NOSϩ field, however, only approximately 36.4% of the ChATϩ SSN neurons at the level of the cluster were NOSϩ.

Neurochemistry of PRV-Ba–Labeled SSN Prechoroidal Neurons: NOS Immunofluorescence double labeling for NOS and PRV-Ba in sections from three animals with bilateral resection of the superior cervical ganglia that survived less than 70 hours after virus injection was used to determine whether the NOSϩ neurons in the SSN were parasympathetic preganglionic neurons that project to prechoroidal PPG neurons (Table 1). The double labeling for NOS and PRV-Ba (Fig. 8) in these animals ϩ ϩ ϩ FIGURE 6. Images and drawings showing that nitrergic (i.e, NAD- showed that 28.9% of the PRV-Ba were NOS . For all NOS PHdϩ/NOSϩ) neurons are located within the ventromedial part of the neurons observed in the SSN cluster, 33.4% were observed to ϩ ϩ superior salivatory nucleus (SSN). (A, C) Transverse section from a be PRV-Ba . These results confirm that many NOS neurons of normal rat at the level of the facial somatomotor nucleus (n7) and SSN, the SSN are prechoroidal parasympathetic neurons. It is possi- labeled histochemically for NADPH-diaphorase to reveal nitrergic neu- ble that those NOSϩ neurons not labeled for PRV-Ba either are rons. A small number of neurons within the confines of SSN label for not prechoroidal SSN neurons or innervate PPG neurons pro- NADPHd, some of which are indicated by arrows in (A). This popu- jecting to parts of the choroid not injected with virus in this lation was located within ventromedial SSN at the rostral dorsolateral edge of n7. (B, D) Transverse section from a colchicine-treated rat at study (for example, the inferior–nasal choroid). the level of n7 and SSN, immunolabeled for NOS. Arrows: some of the neurons within the confines of SSN labeled for NOS. This population of NOSϩ neurons was located within the rostral, ventromedial part of SSN DISCUSSION at the dorsolateral edge of n7, and they appeared to be the same population that labeled for NADPHd. The fields shown in (A) and (B) The SSN contains preganglionic neurons projecting to two are of the right side of the brain, with medial to the left and dorsal different cranial parasympathetic ganglia: the PPG through the toward the top. Abbreviations are as in Figure 3. (A, B) same magnifi- greater petrosal nerve and the submandibular ganglion through 25–29,33 cation; (C, D) drawings of the right side of the sections from which the chorda tympani. The PPG and submandibular gan- images (A) and (B), respectively, were taken. glion innervate different targets, with the PPG sending projections to orbital, nasal, and brain targets, and the submandibular ganglion innervating the submandibular and sublingual salivary percentage of cholinergic neurons labeled for virus was slightly glands. Consistent with the difference in peripheral targets, a increased (Table 1). It was uncertain whether this reflected number of neuroanatomical studies have shown that the lateral spread of the virus, delayed transneuronal transport preganglionic neurons innervating the PPG are segregated from more distant parts of the intrachoroidal injection site, or from those projecting to the submandibular ganglion, with the slower accumulation of detectible virus in more weakly labeled ϩ former located ventrally within the SSN and the latter dor- neurons. Note that not all PRV-Ba neurons in the SSN labeled sally.25–29,33 The PPG itself, however, innervates diverse cra- for ChAT. This may have occurred due to collateral spread of nial targets, including the lacrimal gland, meibomian glands, virus to nonpreganglionic neurons within the confines of the orbital conjunctiva, choroidal blood vessels, cerebral vascula- SSN or due to low levels of ChAT in some preganglionic 34–39 ϩ ture, and nasal and palatal mucosa. Prior studies in which neurons. The tendency for the percentage of PRV-Ba neurons PRV-Ba was injected into different peripheral PPG targets have containing ChAT to be higher with shorter survival times is suggested that the SSN populations responsible for innervation consistent with the former possibility. of the different PPG targets may differ in their location within Neurochemistry of SSN Neurons: the SSN. For example, SSN neurons controlling lacrimal gland may be centered slightly caudal to those responsible for mei- NOS and NADPHd bomian gland control.29,40 Nonetheless, the degree to which Single-label NADPHd histochemistry and single-label NOS im- the various functionally distinct types of SSN preganglionic munohistochemistry were used to assess the possibility that neurons might be spatially segregated has not been fully deter- SSN neurons that project to the choroid contain NOS and mined. Similarly, although there is some evidence for segrega- therefore perhaps use NO as a transmitter. The SSN in normal tion of function among PPG neurons innervating different and colchicine-treated rats contained a small but distinct pop- targets,37,40,62 their spatial segregation is not yet fully resolved. ulation of neurons in both the tissue labeled by NADPHd The present results suggest that preganglionic neurons con- histochemistry and that labeled by NOS immunolabeling (Fig. trolling PPG choroidal neurons in rat reside within a specific 6). The location of this cluster matched the PRV-Baϩ cluster region of SSN, and that NOS within this region defines the observed within the SSN after intrachoroidal injection of the location of and is a marker for prechoroidal SSN neurons. The virus (Fig. 3; P10.5 level), and the labeling of perikarya for NOS neurons preganglionic to PPG neurons that innervate the cho- was more prominent with colchicine-treatment. The NAD- roid appear to be localized to a specific subregion of the SSN, PHdϩ/NOSϩ neurons, like the PRV-Baϩ neurons from intracho- because they are slightly more rostral or medial in location than roidally injected animals, included only a small, spatially re- those observed after injection of PRV-Ba into other peripheral

Downloaded from iovs.arvojournals.org on 10/02/2021 IOVS, September 2003, Vol. 44, No. 9 Preganglionic Neurons Controlling Choroidal Blood Flow 3721

PPG neurons innervating these different targets does not establish that they project to the same targets, it nonetheless is not consistent with the view that different PPG populations innervate different targets. Thus, it is uncertain whether the PPG neurons that innervate the choroid exclusively subserve the choroid or whether they also innervate such vascular targets as the orbital and cerebral blood vessels, for example. Although our studies thus cannot establish whether populations of PPG neurons that innervate the choroid in rats and the SSN neurons that innervate those PPG neurons control only ChBF, they establish a number of points regarding these neuronal populations. First, these studies have directly shown that the PPG neurons that innervate the choroid contain NOS, VIP, and ChAT. This is consistent with prior evidence that VIPϩ and NOSϩ fibers arising from the PPG innervate the choroid.9,19 Published data suggest that the same PPG neurons that innervate the choroid or additional PPG neurons may innervate orbital vessels that feed into the choroid and thereby exert a further influence on ChBF.9 The identification of prechoroidal neurons in the SSN is consistent with prior studies showing that facial nerve or SSN activation yields choroidal vasodilation, which appears to be mediated by NOS and VIP.63,64 Our

FIGURE 7. CLSM images of a single field of view of the SSN from tissue double labeled by immunofluorescence for NOS (A) and ChAT (B), from a normal rat. Most of the NOSϩ neurons also contained ChAT, suggesting they were cholinergic preganglionic neurons. Arrows: ex- amples of these double-labeled neurons. Note that there were many more ChATϩ neurons than there were NOS neurons, suggesting that NOSϩ neurons are a subset of cholinergic preganglionic SSN neurons. The field shown is of the right side of the brain, with medial to the left and dorsal toward the top. Both images are at the same magnification.

PPG targets, such as the meibomian glands40 and the lacrimal gland.29 The possibility remains, nonetheless, that the same population of preganglionic SSN neurons may control more than one of the target structures of the PPG. For example, it may be the case that, given the similar metabolic needs of the brain and eye,6,7 the same SSN neurons mediate vasodilation of brain and choroidal vessels. Studies using two or more different transneuronal tracers are needed to determine whether the FIGURE 8. CLSM images of a single field of view of the SSN from tissue ventral SSN is divided into spatially distinct subpopulations of double labeled by immunofluorescence for PRV-Ba (A) and NOS (B), preganglionic neurons for each PPG target.40 from an animal that received bilateral resection of the superior cervical The distribution of the PPG neurons innervating different ganglia and survived for 65 hours after unilateral intrachoroidal virus injection. Many of the PRV-Baϩ neurons in the SSN were labeled for cranial targets, however, appears to suggest considerable over- ϩ lap of functionally distinct neuron types at the level of the NOS (arrows) and many of the NOS neurons were labeled for PRV-Ba ganglion, at least. For example, although the PPG neurons (arrows). These results indicate that NOS within the SSN is found in preganglionic neurons that innervate PPG neurons projecting to the projecting to the lacrimal gland appear to arise from a different 37 choroid. The NOS labeling was visualized with a secondary antibody part of the PPG than those to the iris, the distribution of PPG conjugated to CY5 and pseudocolored red with an image-analysis neurons that innervate the choroid overlaps those of PPG program (Photoshop; Adobe Systems, Mountain View, CA). The field 37 40 neurons innervating iris, meibomian glands, and cerebral shown is of the right side of the brain, with medial to the left and dorsal vasculature.62 Although the overlapping distributions of the toward the top. Both images are at the same magnification.

Downloaded from iovs.arvojournals.org on 10/02/2021 3722 Cuthbertson et al. IOVS, September 2003, Vol. 44, No. 9

findings on the localization of ChAT in PPG neurons that flow: Mu¨ller cells express GFAP following lesions of the nucleus of innervate the choroid also suggest a possible role of cholin- Edinger-Westphal in pigeons. Curr Eye Res. 1990;9:583–598. ergic PPG mechanisms in control of ChBF in rats, although the 2. Fitzgerald MEC, Tolley E, Frase S, et al. Functional and morpho- physiological evidence for such a mechanism is equivocal.65,66 logical assessment of age-related changes in the choroid and outer It is important to note that, in addition to the parasympathetic retina in pigeons. Vis Neurosci. 2001;18:299–317. influence on the choroid mediated by the PPG, the choroid 3. Hodos W, Miller RF, Ghim MM, Fitzgerald MEC, Toledo, C, Reiner, also is regulated by sensory fibers from the ophthalmic nerve A. Visual acuity losses in pigeons with lesions of the nucleus of and sympathetic fibers from the superior cervical ganglion.4–9 Edinger-Westphal that disrupt the adaptive regulation of choroidal blood flow. Vis Neurosci. 1998;15:273–287. In any event, the location of the prechoroidal SSN neurons 4. Kirby ML, Diab IM, Mattio TG. Development of adrenergic inner- revealed by the present study can be of use in determining the vation of the iris and fluorescent ganglion cells in the choroid of central circuitry that governs control of ChBF. Prior studies the chick eye. Anat Rec. 1978;191:311–320. have shown that the paraventricular nucleus (PVN) of the 5. Guglielmone R, Cantino D. Autonomic innervation of the ocular hypothalamus and the nucleus of the solitary tract (NTS) are choroid membrane in the chicken. Cell Tissue Res. 1982;222:417– 27,65,66 major sources of input to the SSN. The PVN region of 431. the diencephalon is known to be responsive to systemic blood 6. Bill A. The circulation in the eye. In: Renkin EM, Michel CC, eds. pressure (BP) and to exert a vasodilatory influence on cerebral The Microcirculation Part 2: Handbook of Physiology, Section 2. blood flow.27,67,68 The part of the NTS that projects to the SSN Washington DC: The American Physiological Society; 1984;1001– is known to receive aortic baroreceptor input through the 1035. vagus nerve and to respond to BP fluctuation.69,70 Similar to 7. Bill A. Some aspects of the ocular circulation. Invest Ophthalmol the PVN, the NTS exerts a vasodilatory influence on cerebral Vis Sci. 1985;26:411–424. blood flow.71 In addition, the NTS projects directly and indi- 8. Bill A. Effects of some neuropeptides on the uvea. Exp Eye Res. rectly through the parabrachial region to the PVN.72,73 It is 1991;53:3–11. unknown, however, whether the NTS or the PVN region has an 9. Stone RA, Kuwayama Y, Laties AM. Regulatory peptides in the eye. impact on ChBF, and it is not established that the NTS and the Experientia. 1987;43791–800. PVN projections to the SSN include among their targets the 10. Corvetti G, Pignocchino, P, Sisto Daneo, L. Distribution and devel- prechoroidal neurons of the SSN. The fact that we observed opment of substance-P immunoreactive axons in the chick cornea and uvea. Basic Appl Histochem. 1988;32:187–192. higher-order labeling in both the PVN and the NTS after intrachoroidal injection of PRV-Ba (Reiner A, LeDoux MS, Cuthbert- 11. Cuthbertson S, FitzgeraldMEC, Shih YF, White J, Reiner A. Distri- bution of ciliary ganglion nerve fibers in the avian choroid. Vision son S, unpublished observations, 2000) is consistent with the Res. 1996;36:775–786. possibility that these sites innervate prechoroidal neurons of 12. Potts AM. An hypothesis on macular disease. Trans Am Acad the SSN. Because the present study shows that the prechoroi- Ophthalmol Otolaryngol. 1966;70:1058–1062. dal neurons of the SSN can be identified by PRV-Ba transneu- 13. Reiner A, Karten HJ, Gamlin PDR, Erichsen JT. Parasympathetic ronal retrograde labeling or by NOS immunolabeling, it should ocular control: functional subdivisions and circuitry of the avian be possible to combine this means of detecting prechoroidal nucleus of Edinger-Westphal. Trends Neurosci. 1983;6:140–145. SSN neurons with anterograde labeling from the PVN or the 14. Shih YF, Fitzgerald MEC, Reiner A. Effects of choroidal and ciliary NTS to confirm that the PVN and/or the NTS innervate precho- nerve transection on choroidal blood flow, retinal health and roidal SSN neurons. This then would provide insight into the ocular enlargement. Vis Neurosci. 1993;10:969–979. higher-order brain regions involved in regulation of ChBF 15. Shih YF, Fitzgerald MEC, Reiner A. The effects of choroidal or through the PPG. ciliary nerve transection on myopic eye growth induced by gog- Given the apparent role of the PVN and the NTS in medi- gles. Invest Ophthalmol Vis Sci. 1994;35:3691–3701. ating vascular responses to fluctuations in systemic BP, if the 16. Ruskell GL. Facial parasympathetic innervation of the choroidal PVN and the NTS in fact project to the prechoroidal neurons of blood vessels in monkeys. Exp Eye Res. 1971;12;166–172. the SSN, it would suggest the possibility that these inputs 17. Uddman R, Alumets J, Ehinger B, Hakanson R, Loren I, Sundler F. regulate ChBF, in part, as a function of systemic BP. Alterna- Vasoactive intestinal peptide nerves in ocular and orbital struc- tively or in addition, the PVN input to the prechoroidal SSN tures of the cat. Invest Ophthalmol Vis Sci. 1980:19:878–885. may be involved in light-mediated control of ChBF. Light- 18. Stone RA. Vasoactive intestinal polypeptide and the ocular innervation. Invest Ophthalmol Vis Sci. 1986;27:951–957. mediated and flicker-mediated ChBF regulation, which may be 19. Yamamoto R, Bredt DS, Snyder SH, Stone RA. The localization of adaptive responses to the thermal or metabolic demands nitric oxide synthase in the rat eye and related cranial ganglia. placed by such stimuli on the retina,63,74,75 have been demon- 76 77 63,78,79 Neuroscience. 1993;54:189–200. strated in pigeons, chickens, monkeys, and humans, 20. Alm A, Uvelius B, Ekstrom J, Holmqvist B, Larsson B, Andersson. and the suprachiasmatic nucleus (which receives retinal input) KE. Nitric oxide synthase-containing neurons in rat parasympa- 80 is known to project to the PVN. It may be, therefore, that thetic sympathetic and sensory ganglia: a comparative study. His- prechoroidal neurons of the SSN receive central inputs by tochem J. 1995;27:819–831. which facial nucleus parasympathetic outflow to the choroid is 21. Johansson O, Lundberg JM. Ultrastructure localization of VIP-like involved in light-mediated and/or systemic BP-mediated con- immunoreactivity in large dense-core vesicles of “cholinergic” type trol of ChBF. nerve terminals in cat exocrine glands. Neuroscience. 1981;6:847– 862. 22. Lundberg JM, Anggard A, Emson P, Fahrenkrug J, Hokfelt T. Vaso- Acknowledgments active intestinal polypeptide and cholinergic mechanisms in cat nasal mucosa: Studies on choline acetyltransferase and release of The authors thank Rebeca-Ann Weinstock, Raven Babcock, Amanda vasoactive intestinal polypeptide. Proc Natl Acad Sci USA. 1981; Valencia, Malinda Fitzgerald, Christopher Meade, Nobel Del Mar, Yun 78:5255–5259. Jiao, Felicia Covington, Karen Fife, Shani Bell, and Christy Loggins for 23. Lundberg JM, Anggard A, Fahrenkrug J, Lundgren G, Holmstedt B. excellent assistance. Corelease of VIP and acetylcholine in relation to blood flow and salivary secretion in cat submandibular salivary gland. Acta Physiol References Scand. 1982;115:525–528. 24. Suzuki N, Hardebo JE, Kahstrom J, Owman C. Neuropeptide Y 1. Fitzgerald MEC, Vana B, Reiner A. Evidence for retinal pathology co-exists with vasoactive intestinal polypeptide and acetylcholine following interruption of neural regulation of choroidal blood in parasympathetic cerebrovascular nerves originating in the sphe-

Downloaded from iovs.arvojournals.org on 10/02/2021 IOVS, September 2003, Vol. 44, No. 9 Preganglionic Neurons Controlling Choroidal Blood Flow 3723

nopalatine, otic, and internal carotid ganglia of the rat. Neuro- 47. Johnson CD, Epstein ML. Monoclonal antibodies and polyvalent science. 1990;36:507–519. antiserum to chicken choline acetyltransferase. J Neurochem. 25. Contreras RJ, Gomez MM, Norgren R. Central origins of cranial 1986;46:968–976. nerve parasympathetic neurons in the rat. J Comp Neurol. 1980; 48. Reiner A. Catecholiminergic innervation of the basal ganglia in 190:373–394. mammals. In: Smeets WJAJ, Reiner A, eds. Phylogeny and Devel- 26. Nicholson JE, Severin CM. The superior and inferior salivatory opment of Catecholamine Systems in the CNS of Vertebrates. nuclei in the rat. Neurosci Lett. 1981;21:149–154. Cambridge, UK: Cambridge University Press; 1994:247–272. 27. Spencer SE, Sawyer WB, Wada H, Platt KB, Loewy AD. CNS pro- 49. Anderson KD, Reiner A. The extensive co-occurrence of substance jections to the pterygopalatine parasympathetic preganglionic P and dynorphin in striatal projection neurons: an evolutionarily neurons in the rat: a retrograde transneuronal viral cell body conserved feature of basal ganglia organization. J Comp Neurol. labeling study. Brain Res. 1990;534:149–169. 1990;25:339–369. 28. Ng YK, Wong WC, Ling EA. A light and electron microscopical 50. Medina L, Veenman CL, Reiner A. Evidence for a possible avian localisation of the superior salivatory nucleus of the rat. J Brain dorsal thalamic region comparable to the mammalian ventral an- Res. 1994;35:39–48. terior, ventral lateral, and oral ventroposterolateral nuclei. J Comp 29. To´th IE, Boldogkoi Z, Medveczky I, Palkovits M. Lacrimal pregan- Neurol. 1997;348:86–108. glionic neurons form a subdivision of the superior salivatory nu- 51. Zin-Ka-Ieu S, Roger M, Arnault P. Direct contacts between fibers cleus of rat: transneuronal labelling by pseudorabies virus. J Auton from the ventrolateral thalamic nucleus and frontal cortical neu- Nerv Syst. 1999;77:45–54. rons projecting to the striatum: a light-microscopy study in the rat. 30. Zhu BS, Gai WP, Yu YH, Gibbins IL, Blessing WW. Preganglionic Anat Embryol. 1998;197:77–87. parasympathetic salivatory neurons in the brainstem contain mark- 52. Reiner A, Veenman CL, Medina L, Jiao Y, Del Mar N, Honig MG. ers for nitric oxide synthesis in the rabbit. Neurosci Lett. 1996; Pathway tracing using biotinylated dextran amines. J Neurosci 204:128–132. Methods. 2000;103:23–37. 31. Zhu BS, Gibbons IL, Blessing WW. Preganglionic parasympathetic 53. Medina L, Figueredo-Cardenas G, Reiner A. Differential abundance neurons projecting to the sphenopalatine ganglion contain nitric of superoxide dismutase in interneurons versus projection neu- oxide synthase in the rabbit. Brain Res. 1997;769:168–172. rons in patch versus matrix neurons in monkey striatum. Brain Res. 1996;708:59–70. 32. Gai WP, Blessing WW. Human brainstem preganglionic parasympathetic neurons localized by markers for nitric oxide synthesis. 54. Fusco FR, Chen Q, Lamoreaux WJ, et al. Cellular localization of Brain. 1996;119:1145–1152. huntingtin in striatal and cortical neurons in rats: lack of correla- tion with neuronal vulnerability in Huntington’s disease. J Neuro- 33. Jansen ASP, Ter Horst GJ, Mettenleiter TC, Loewy AD. CNS cell sci. 1999;19:1189–1202. groups projecting to the submandibular parasympathetic preganglionic neurons in the rat: a retrograde transneuronal viral cell 55. Vincent SR, Johansson O, Hokfelt T, et al. NADPH-diaphorase: a body labeling study. Brain Res. 1992;572:253–260. selective histochemical marker for striatal neurons containing both somatostatin- and avian pancreatic polypeptide (APP)-like 34. Ruskell GL. The orbital distribution of the sphenopalatine ganglion immunoreactivities. J Comp Neurol. 1983;217:252–263. in the rabbit. In: Rohen JW, ed. The Structure of the Eye. Stuttgart, 56. Sancesario G, Reiner A, Figueredo-Cardenas G, Morello M, Bernardi Germany: Schattauer-Verlag; 1965:335–368. G. Differential distribution of NADPH-diaphorase and neural nitric 35. Ruskell GL. The distribution of autonomic post-ganglionic nerve oxide synthase in the rat choroid plexus: a histochemical and fibers to the lacrimal gland in monkeys. J Anat. 1971;109:229– immunocytochemical study. Neuroscience. 1996;72:365–375. 242. 57. Cuthbertson S, Jackson BJ, Toledo C, et al. Innervation of orbital 36. Uddman R, Malm L, Sundler F. The origin of vasoactive intestinal and choroidal blood vessels by the pterygopalatine ganglion in polypeptide (VIP) nerves in the feline nasal mucosa. Acta Otolar- pigeons. J Comp Neurol. 1997;386:422–442. yngol. 1980;89:152–156. 58. Cuthbertson S, Zagvazdin YS, Kimble TDH, et al. Preganglionic 37. Ten Tusscher MPM, Klooster J, Baljet B, Van der Werf F, Vrensen endings form nucleus of Edinger-Westphal in pigeon ciliary gan- GFJM. Pre- and post-ganglionic nerve fibers of the pterygopalatine glion contain nitric oxide synthase. Vis Neurosci. 1999;16:819– ganglion and their allocation to the eyeball of rats. Brain Res. 834. 1990;170:315–323. 59. Smeets WJAJ, Reiner A. Phylogeny and Development of Catechol- 38. Nakai M, Tamaki K, Ogata J, Matsui Y, Maeda M. Parasympathetic amine Systems in the CNS of Vertebrates. New York: Cambridge cerebrovasodilator center of the facial nerve. Circ Res. 1993;72: University Press; 1994. 470–475. 60. Nemoto T, Konno A, Chiba T. Synaptic contact of neuropeptide 39. Van der Werf F, Baljet B, Prins M, Otto JA. Innervation of the and amine-containing axons on parasympathetic preganglionic lacrimal gland in the cynomolgous monkey: a retrograde tracing neurons in the superior salivatory nucleus of the rat. Brain Res. study. J Anat. 1996;188:591–601. 1995;685:33–45. 40. LeDoux MS, Zhou Q, Murphy RB, Greene ML, Ryan P. Parasympa- 61. Pickard GE, Smeraski CA, Tomlinson CC, et al. Intravitreal injec- thetic innervation of the meibomian glands in rats. Invest Ophthal- tion of the attenuated pseudorabies virus PRV Bartha results in mol Vis Sci. 2001;42:2434–2441. infection of the hamster suprachiasmatic nucleus only by retro- 41. Hebel R, Stromberg MW. The Anatomy of the Laboratory Rat. grade transsynaptic transport via autonomic circuits. J Neurosci. Baltimore: William & Wilkins; 1976. 2001;22:2701–2710. 42. Rotto-Percelay DM, Wheeler JG, Osorio FA, Platt KB, Loewy AD. 62. Suzuki N, Hardebo JE Owman C. Origins and pathways of cere- Transneuronal labeling of spinal interneurons and sympathetic brovascular vasoactive intestinal polypeptide-positive nerves in preganglionic neurons after pseudorabies virus injections in the rat the rat: a retrograde transneuronal viral cell body labeling study. medial gastrocnemius muscle. Brain Res. 1992;574:291–306. Brain Res. 1988;534:149–169. 43. Reiner A, Erichsen JT, Cabot JB, Evinger C, Fitzgerald MEC, Karten 63. Bill A, Sperber GO. Control of retinal and choroidal blood flow. HJ. Neurotransmitter organization of the nucleus of Edinger-West- Eye. 1990;4:319–325. phal and its projection to the avian ciliary ganglion. Vis Neurosci. 64. Steinle JJ, Krizsan-Agbas D, Smith PG. Regional regulation of cho- 1991;6:451–472. roidal blood flow by autonomic innervation in the rat. Am J 44. Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates. Physiol. 2000;279:R202–R209. 4th ed. New York: Academic Press; 1994. 65. Agassandian K, Fazan VPS, Cassell LH, Lin LH, Talman WT. Antero- 45. LeDoux MS, Ryan P. Motor and trigeminal blink circuitry in the rat grade and retrograde tracers define projections between nucleus (Abstract). Soc Neurosci. 1999;25:564.12. tractus solitarii and superior salivatory nucleus in rats. Soc Neuro- 46. Mayer B, Mathias J, Bohme E. Purification of Ca2ϩ/calmodulin sci Abstr. 1999:25:774.14. dependent nitric oxide synthase from porcine cerebellum. Fed 66. Hosoya Y, Matsushita M, Sugira Y. Hypothalamic descending affer- Eur Biochem Soc. 1990;277:215–219. ents to cells of origin of the greater petrosal nerve in the rat, as

Downloaded from iovs.arvojournals.org on 10/02/2021 3724 Cuthbertson et al. IOVS, September 2003, Vol. 44, No. 9

revealed by a combination of retrograde HRP and anterograde ogy, Diagnosis and Management. New York: Raven Press; 1990: autoradiographic techniques. Brain Res. 1984;290:141–145. 711–747. 67. Krukoff TL, Mactavish D, Jhamandas JH. Activation by hypotension 74. Steinberg RH. Monitoring communications between photorecep- of neurons in the hypothalamic paraventricular nucleus that tors and pigment epithelial cells: effects of “mild” systemic hyp- project to the brain stem. J Comp Neurol. 1997;385:285–296. oxia. Invest Ophthalmol Vis Sci. 1987;28:1888–1904. 68. Golanov EV, Christensen JR, Reis DJ. Neurons of a subthalamic 75. Linsenmeier RA. Effects of light and darkness on oxygen distribu- cerebrovasodilator area (SCA) mediate elevations in cerebral blood tion and consumption in the cat retina. J Gen Physiol. 1986;88: flow evoked by excitation of neurons of the rostral ventrolateral 521–542. medulla. Soc Neurosci Abstr. 1999;25:11.9. 76. Fitzgerald MEC, Gamlin PDR, Zagvazidin Y, Reiner A. Central 69. Rogers RF, Patton JFR, Schwaber JS. NTS Neuronal responses to neural circuits for the light-mediated reflexive control of choroidal blood flow in the pigeon eye: a laser Doppler study. Vis Neurosci. arterial pressure and pressure changes in rat. Am J Physiol. 1993; 1996;13:655–669. 265:R1355–R1368. 77. Shih YF, Lin SY, Huang JK, Jian SW, Lin LLK, Hung PT. The 70. Criello J. Brainstem projections of aortic baroreceptors afferent choroidal blood flow response after flicker stimulation in chicks. J fibers in the rat. Neurosci Lett. 1983;36:37–42. Ocul Pharmacol Ther. 1997;13:213–218. 71. Nakai M, Ogino K. Relevance of cardiopulmonary vascular reflex 78. Parver LM, Auker CR, Carpenter DO, Doyle T. Choroidal blood to the regulation of brain vessels. Jpn J Physiol. 1984;34:193–197. flow. II. Reflexive control in the monkey. Arch Ophthalmol. 1982; 72. Calarescu FR, Criello J, Caverson MM, Cechetto DF, Krukoff TL. 100:1327–1330. Functional neuroanatomy of ventral pathways controlling the cir- 79. Parver LM, Auker CR, Carpenter DO. Choroidal blood flow. III. culation. In: Kochen TA, Guthrie CP, eds. Hypertension and the Reflexive control in human eyes. Arch Ophthalmol. 1983;101: Brain. Mt. Kisco, NY: Futura Publications; 1984:3–21. 1604–1606. 73. Goldstein DS, Kopin IJ. The autonomic nervous system and cat- 80. Hermes MLHJ, Buijs RM, Renaud LP. Electrophysiology of supra- echolamines in normal blood pressure control and in hyperten- chiasmatic nucleus projections to hypothalamic paraventricular sion. In: Laragh JH, Brener DM, eds. Hypertension: Pathophysiol- nucleus neurons. Prog Brain Res. 1996;111:241–252.

Downloaded from iovs.arvojournals.org on 10/02/2021