- Email: [email protected]
CHOLINERGIC NERVES IN HUMAN CORPUS CAVERNOSUM AND SPONGIOSUM CONTAIN NITRIC OXIDE SYNTHASE AND HEME OXYGENASE
0022-5347/00/1643-0868/0 THE JOURNAL OF UROLOGY® Copyright © 2000 by AMERICAN UROLOGICAL ASSOCIATION, INC.®
Vol. 164, 868 – 875, September 2000 Printed in U.S.A.
CHOLINERGIC NERVES IN HUMAN CORPUS CAVERNOSUM AND SPONGIOSUM CONTAIN NITRIC OXIDE SYNTHASE AND HEME OXYGENASE PETTER HEDLUND, LARS NY, PER ALM
AND
KARL-ERIK ANDERSSON*
From the Departments of Clinical Pharmacology and Pathology, Lund University Hospital, Lund, Sweden
ABSTRACT
Purpose: To characterize the distribution of cholinergic nerves in the human corpus cavernosum (CC) and spongiosum (CS) using antibodies to the vesicular acetylcholine transporter (VAChT), and to compare this distribution to those of other transmitters/mediators or transmitter/mediator generating enzymes (heme oxygenases: HO-1 and HO-2; neuronal and endothelial NO synthases: nNOS and eNOS; vasoactive intestinal polypeptide: VIP; and tyrosine hydroxylase: TH), and to investigate NO- and carbon monoxide (CO)-mediated effects. Materials and Methods: Immunocytochemistry, confocal laser scanning microscopy, radioimmunoassay, and functional in vitro studies. Results: Along strands of smooth muscle in the CC and CS, rich numbers of VAChT-, nNOS-, VIP-, TH-, and very few HO-1-immunoreactive (-IR) nerve fibers were observed. Immunoreactivities for VAChT and nNOS, VAChT and VIP, and nNOS and VIP, were generally found in the same varicose nerve terminals. TH-IR nerve fibers or terminals did not contain immunoreactivities for VAChT, NOS or VIP. In the endothelium lining penile arteries, immunoreactivities for eNOS, HO-1, and HO-2 were detected. Single endothelial cells, lining the sinusoidal walls of the CC and CS, were found also to contain eNOS and HO-immunoreactivities. Noradrenaline (NA)-contracted preparations of CC and CS were relaxed by NO, CO, carbachol and by electrical stimulation of nerves. Inhibition of NO synthesis abolished electrically- and carbachol-induced relaxation. In NA-activated strips, relaxation induced by exogenously applied NO, but not those by CO, were accompanied by increases in intracellular levels of cyclic GMP. Conclusions: VAChT, NOS and VIP are found in the same nerve terminals within the human CC and CS, suggesting that these terminals comprise a distinct population of parasympathetic, cholinergic nerves. Endothelially derived NO and the HO/CO system may have a complementary role in penile erection. KEY WORDS: penile erection, innervation, co-localization, confocal microscopy
Pelvic parasympathetic nerves provide a pathway for the vasodilator responses in the penile vasculature, and for the relaxation of corpus cavernosum (CC) smooth muscle necessary for erection. Acetylcholine (ACh), released from cholinergic nerves, may via stimulation of muscarinic receptors on adrenergic nerves decrease the release of noradrenaline (NA), and also release vasoactive agents from the vascular endothelium. In addition, stimulation of cholinergic nerves can lead to release of other relaxation-producing transmitters.1 Numerous nerves in the erectile tissue contain ACh esterase (AChE). However, this enzyme is not confined to cholinergic nerves, and the presence of AChE is not considered sufficient for identification of cholinergic nerves.2 ACh is transferred into vesicles in the nerves by the vesicular acetylcholine transporter (VAChT), and immunohistochemical demonstration of VAChT is considered to be a specific approach for visualization of cholinergic nerves in both the central and peripheral nervous system.3, 4 Nitric oxide synthase (NOS) has been localized to nerves in penile erectile tissues from several species, including humans, and effects mediated by NO have been extensively
studied in isolated CC tissue and in several animal models in vivo.1 In the rat, the presence of NOS in the sinusoidal endothelium (eNOS) is disputed,5 but endothelial NO-dependent relaxations have been demonstrated functionally in human CC preparations.1 However, immunohistochemical demonstration of eNOS in human sinusoidal endothelium has so far not been convincing. Heme oxygenase (HO), a carbon monoxide (CO) producing enzyme, has been demonstrated in neurons and in endothelial and smooth muscle cells in various peripheral organs.6 –9 CO has been shown to induce relaxation of vascular and gastrointestinal smooth muscle preparations.7–10 If CO is involved in smooth muscle regulatory functions in the human penis, this has, to the best of our knowledge, not been clarified. In the present study, we wanted to study and compare in isolated human CC and CS tissue: 1) the distribution of VAChT-immunoreactive (-IR) nerves, and relate it to the total distribution of nerves as well as to populations of nerves containing nNOS, HO, vasoactive intestinal polypeptide (VIP), and tyrosine hydroxylase (TH); 2) the distribution of eNOS immunoreactivity; and 3) the functional effects of NO and CO in isolated preparations of the two tissues.
Accepted for publication April 4, 2000. * Requests for reprints: Department of Clinical Pharmacology, Lund University Hospital, S-221 85 Lund, Sweden. Supported by the Swedish Medical Research Council (Grants no. MATERIALS AND METHODS 6837 and 11205), the Royal Physiographic Society, the Foundations Preparations. Specimens of penile erectile tissue were obof Crafoord, Magnus Bergvall, and Anna-Lisa and Sven-Erik Luntained from 12 men, aged 21 to 71 years (mean 55 years). Ten dgren, and the Medical Faculty, University of Lund, Sweden. 868
869
CHOLINERGIC NERVES CONTAIN NO AND HO
(with normal erectile capacity) underwent cystourethrectomy (no preoperative external radiation) because of bladder malignancy. Cavernous tissue was taken by biopsy peroperatively after consent from the patients had been obtained. One patient underwent a gender reassignment operation. Spongious tissue was also obtained from one donor patient with brain damage after an intracranial hemorrhage. The procedure for obtaining CC and CS tissue was approved by the Ethics Committee, Lund University. Immunocytochemistry. Tissue specimens were fixed and further processed for immunocytochemistry as previously described.11 Cryostat sections were incubated overnight in the presence of a primary antiserum. The primary antisera used (table 1) were rabbit antisera to protein gene product 9.5 (PGP), inducible heme oxygenase (HO-1/Hsp32), constitutive heme oxygenase (HO-2), endothelial NOS (eNOS), and a C-terminal dodecapeptide to cloned human vesicular acetylcholine transporter (VAChT), a sheep antiserum to neuronal NOS (nNOS), a mouse monoclonal tyrosine hydroxylase (TH) antiserum, and a guinea pig antiserum to vasoactive intestinal peptide (VIP). After rinsing, the sections were incubated for 90 minutes in the presence of an appropriate secondary antiserum (see table 2), whereupon, the sections were rinsed and mounted. For the simultaneous demonstration of two antigens,12 some sections were incubated overnight in rabbit antisera to HO-1, or HO-2, rinsed, and then incubated overnight with sheep antiserum to nNOS, or mouse monoclonal THantiserum, or guinea pig antiserum to VIP. Some sections were incubated overnight in guinea-pig VIP or sheep nNOS antisera, rinsed, and then incubated overnight in rabbit VAChT- antiserum. Some sections were incubated overnight in in sheep nNOS antiserum, rinsed, and then incubated overnight in guinea-pig VIP antiserum, or mouse monoclonal TH-antiserum. After rinsing, the sections were incubated in appropriate combinations of secondary antisera, whereupon they were rinsed and mounted. The primary and secondary antisera used, and their combinations, are described in tables 1-4. The characteristics of the rabbit VAChT and sheep nNOS antisera have previously been described.13, 14 In control experiments, no immunoreactivity could be detected in sections incubated in the absence of primary antisera, or incubated with HO-1, HO-2, VAChT or VIP antisera absorbed with excess of the respective antigens (100 g.ml.⫺1). No absorption tests were performed with the TH, eNOS, or nNOS antisera, as antigenic substances were not available. The immunoreactive nerve structures were independently evaluated by the authors with respect to type (nerve trunks and nerve terminals) and relative frequency (arbitrarily graded as very numerous, numerous, moderate number, few and absent.15 In the evaluation of the latter parameter, immunoreactivity to PGP was used as an internal standard as a marker for all populations of nerve fibers.15, 16 In the present study, approximately 250 glasses, each with 4 to 6
sections, were used for the immunohistochemical characterization of the investigated structures. The structures related are referred to as PGP-, HO-1-, HO-2-, nNOS-, eNOS-, TH-, VAChT- and VIP-IR, as cross-reactions with other antigens sharing similar amino acid sequences cannot be completely excluded. The sections were examined in an Olympus BX 40 fluorescence microscope with epi-illumination and filter settings for Fluorescein 5⬘-IsoThioCyanate (FITC)- and Texas Red (TR)- immunofluorescence.17 Confocal laser scanning microscopy. To evaluate if two immunoreactivities were located in the same nerve structures, sections were analyzed in a Bio-Rad MRC-1024 confocal laser scanning equipment (Bio-Rad Laboratories, UK) attached to a Nikon Eclipse E800 upright microscope (Nikon, Japan). The sections were scanned with a 60 x (1.40) oil immersion plan apochromat lens. Sequential scanning using excitation wavelengths of 488 and 568 nm from a KryptonArgon laser. FITC immunofluorescence was detected with a 522 ⫾ 30 nm band pass filter and TR immunofluorescence was detected with a 605 ⫾ 30 nm band pass filter. Functional studies. Strip preparations (1 ⫻ 2 x 5 mm.) were prepared and then suspended in aerated, thermostatically controlled 5 ml. tissue baths (37C, pH 7.4) containing Krebs solution, which was routinely replaced with fresh Krebs solution every 30 minutes. Isometric tension was recorded by means of a Grass Instruments FTO3C force-transducer connected to a Grass 7D polygraph (Grass Instruments Co., USA). An equilibration period of 45 minutes was allowed until mean final stable tensions of 3.8 ⫾ 0.4 mN (n ⫽ 6, n ⫽ 10) for CC, and 3.0 ⫾ 0.3 mN (n ⫽ 12, n ⫽ 30) for CS were obtained. The viability of the preparations was verified by addition of a high K⫹ solution (124 mM) to the organ baths, and contractile responses amounting to 12.4 ⫾ 3.8 mN for CC, and 8.3 ⫾ 1.0 mN for CS, were recorded. The l-noradrenaline (NA) concentration used corresponded to the approximate EC50 value for NA in this tissue (CC, 10⫺6 M; CS 3 ⫻ 10⫺6 M). Electrical field stimulation (EFS) was performed and inhibitory frequency-response relationships were investigated as previously described.11 The effects of NO, CO, carbachol, and EFS in NA-contracted preparations were studied, and the degree of relaxation was expressed as a percentage of the NA-induced contraction. Some preparations were pretreated for 20 minutes with the NO-synthase inhibitor NG-nitro-L-arginine (L-NNA; 10⫺4 M). Radioimmunoassays. Guanosine and adenosine 3⬘: 5⬘cyclic monophosphate (cGMP and cAMP) were analyzed in NA-contracted preparations of erectile tissue (controls) and in NA-contracted preparations exposed to NO and CO. When the effects of the agents reached a maximal and stable level, approximately after 1 minute, the strips were immediately frozen in liquid nitrogen and then stored in 0.5 ml. of 10% trichloracetic acid (TCA) at ⫺70C. The tissue was then homogenized, further processed and the amounts of cGMP and cAMP were quantified as previously described.18 Drugs and solutions. A Krebs solution of the following
TABLE 1. Primary antisera Anti-Serum
Host
Working Dilution
Code
Source
HO-1 (Hsp32) (ra) 1:500 SPA895 StressGen, Victoria, Canada HO-1 (ra) 1:500 OSA100 StressGen, Victoria, Canada HO-2 (ra) 1:500 OSA200 StressGen, Victoria, Canada eNOS (ra) 1:500 Sancta Cruz Biotech, Heidelberg, Germany nNOS (sh) 1:6000 P. Emson, Babraham Institute, Cambridge, England PGP (ra) 1:2000 UltraClone, Wellow, Isle of Wight, England TH (mo) 1:200 Incstar, Stillwater, MO, USA VAChT (ra) 1:1600 E. Weihe, Marburg, Germany VIP (gp) 1:640 Euro-Diagnostica, Malmo¨, Sweden (ra) ⫽ rabbit, (go) ⫽ goat, (sh) ⫽ sheep, (gp) ⫽ guinea pig, (mo) ⫽ mouse. HO(Hsp32) ⫽ Heme Oxygenase, eNOS ⫽ endothelial NO Synthase, nNOS ⫽ neuronal NO Synthase, PGP ⫽ Protein Gene Product 9.5, TH ⫽ Tyrosine Hydroxylase, VAChT ⫽ Vesicular Acetylcholine Transporter, VIP ⫽ Vasoactive Intestinal Polypeptide
870
CHOLINERGIC NERVES CONTAIN NO AND HO TABLE 2. Secondary antisera
Anti-Serum
Host
1.FITC swine anti-rabbit 2.FITC goat anti-guinea pig 3.FITC donkey anti-sheep 4.TR donkey anti-mouse 5.TR donkey anti-rabbit 6.TR donkey anti-sheep FITC ⫽ Fluorescein 5⬘-IsoThioCyanate, TR ⫽ Texas Red
Working Dilution
Code
Source
1:80 1:80 1:80 1:125 1:125 1:125
F 0205 F 6261 T 5393 715-076-151 711-075-152 713-075-147
Dakopatts, Stockholm, Sweden Sigma, St Louis, MO, USA Sigma, St Louis, MO, USA Jackson ImmunoResearch, West Grove, PA, USA Jackson ImmunoResearch, West Grove, PA, USA Jackson ImmunoResearch, West Grove, PA, USA
composition was used (mM): NaCl 119, KCl 4.6, CaCl2 1.5, MgCl2 1.2, NaHCO3 15, NaH2PO4 1.2, glucose 11. An isotonic K⫹ solution (124 mM) was used, in which the NaCl in the normal Krebs solution was replaced by equimolar amounts of KCl. The following drugs were used: NA (Aldrich-Chemie GmbH & Co., Germany), carbachol, tetrodotoxin (TTX) and L-NNA (Sigma Chemical Co., USA). NO and CO were freshly prepared at each experiment. Air-tight glass beakers containing 20 ml. of distilled water were deoxygenated for 1 hour with helium gas. The beakers were then aerated with medical NO or CO gas (purity ⬎99.5%) for 15 minutes until saturated solutions were obtained (NO 3 ⫻ 10⫺3 M, CO 10⫺3 M). Calculations. Student’s paired and unpaired two-tailed t tests were used for statistical comparison of two means. For multiple comparisons, analysis of variance with Bonferroni correction was used. A probability of p ⬍0.05 was accepted as significant. Results are given as mean values ⫾ standard error of the mean (SEM). N denotes the number of individuals, and n denotes the number of strip preparations. All statistical comparisons were made using N. RESULTS
Immunocytochemical findings. In the trabecular smooth muscle tissue, and around arteries in the CC and CS, no nerve trunks displaying VAChT-IR nerve fibers could be observed. However, nerve trunks with immunoreactivities for PGP, nNOS, HO-1, or HO-2 were found. nNOS-IR nerve trunks also contained immunoreactivity for VIP, and single nerve fibers within the trunks contained both nNOS and VIP. TH-IR nerve fibers were also seen in nerve trunks which contained nNOS-IR fibers. Within these nerve trunks, single nNOS- and TH-IR nerve fibers were running close together. As seen in double filter settings and in confocal sections, immunoreactivities for nNOS and TH were found in separate nerve fibers. Nerve trunks with HO-1- and HO-2-IR fibers also contained nNOS- VIP-, or TH-immunoreactivity. Within these nerve trunks, the HO-1- or HO-2-IR fibers were found to exhibit profiles which were similar but not identical to corresponding nNOS-, VIP-, or TH-IR fibers. TABLE 3. Mono-immunolabelling Primary Antiserum (see Table 1)
Secondary Antiserum (see table 2)
HO-1 (Hsp32) (ra) 5 HO-1 (ra) 5 HO-2 (ra) 5 eNOS (ra) 5 nNOS (sh) 3 PGP (ra) 5 TH (mo) 4 VAChT (ra) 5 VIP (gp) 2 After incubation with the two primary antisera, the sections were rinsed, and then incubated for 90 min with first secondary antiserum, rinsed, and then incubated for 90 min with other secondary antiserum. After rinsing, the sections were rinsed and mounted. (ra) ⫽ rabbit, (go) ⫽ goat, (sh) ⫽ sheep, (gp) ⫽ guinea pig, (mo) ⫽ mouse. HO(Hsp32) ⫽ Heme Oxygenase, eNOS ⫽ endothelial NO Synthase, nNOS ⫽ neuronal NO Synthase, PGP ⫽ Protein Gene Product 9.5, TH ⫽ Tyrosine Hydroxylase, VAChT ⫽ Vesicular Acetylcholine Transporter, VIP ⫽ Vasoactive Intestinal Polypeptide.
TABLE 4. Double immunolabelling Combinations of Primary Antisera (see Table 1)
Combinations of Secondary Antisera (see Table 2)
HO-1/Hsp32 (ra) ⫹ nNOS (sh) 3⫹5 HO-1/Hsp32 (ra) ⫹ TH (mo) 1⫹4 HO-1/Hsp32 (ra) ⫹ VIP (gp) 2⫹5 HO-2 (ra) ⫹ nNOS (sh) 3⫹5 HO-2 (ra) ⫹ TH (mo) 1⫹4 HO-2 (ra) ⫹ VIP (gp) 2⫹5 nNOS (sh) ⫹ TH (mo) 3⫹4 nNOS (sh) ⫹ VAChT (ra) 3⫹5 VIP (gp) ⫹ nNOS (sh) 2⫹6 VIP (gp) ⫹ VAChT (ra) 2⫹5 (ra) ⫽ rabbit, (go) ⫽ goat, (sh) ⫽ sheep, (gp) ⫽ guinea pig, (mo) ⫽ mouse. HO(Hsp32) ⫽ Heme Oxygenase, nNOS ⫽ neuronal NO Synthase, PGP ⫽ Protein Gene Product 9.5, TH ⫽ Tyrosine Hydroxylase, VAChT ⫽ Vesicular Acetylcholine Transporter, VIP ⫽ Vasoactive Intestinal Polypeptide
Gracile nerve fibers and varicose nerve terminals containing VAChT-, nNOS- and VIP-immunoreactivities were found in numerous numbers interspersed among trabecular smooth muscle cells. Single fibers containing VAChT-, nNOS-, or VIP-immunoreactivities were observed. In comparison, smooth muscle related PGP-IR varicose nerve fibers were very numerous. The distribution pattern and amount of the different immunoreactivities were similar in specimens from CC and CS (fig. 1, A-D). Very few HO-1-IR, but no HO-2-IR, nerve fibers were seen in the trabecular smooth muscle. In their adventitial layers, arteries were encircled by plexa of VAChT-, nNOS-, VIP-, and TH-IR terminals. Double immunolabeling and confocal microscopy disclosed that the majority of nNOS-IR varicose nerve terminals also contained VAChT- and VIP-immunoreactivity, and that VAChT-IR terminals also contained VIP-immunoreactivity (figs. 2 and 3). nNOS-and TH-IR smooth muscle-related nerve terminals were similarly distributed, but the profiles of the respective nerves terminals were not identical (fig. 3, A). In the CC, the cavernous artery was lined by eNOS-IR endothelial cells, as were small arteries and veins in the CC and CS. The sinusoidal spaces in both tissues were lined by a thin and sometimes disrupted layer of eNOS-IR endothelial cells. The eNOS-IR endothelial cells contained a central nucleus that protruded toward the vascular lumen, and was surrounded by cytoplasm, which extended sideways (fig. 4, A). Immunoreactivities for HO-1 (both HO-1 antisera used) and HO-2 were expressed in endothelial cells of different vessels, including the cavernous artery, and the endothelium lining the walls of the sinusoidal spaces of the CC and CS (fig. 4, B) also contained immunoreactivity for HO-1 (both antisera used) and HO-2. Functional effects. The NA-induced contraction (10⫺6 M for CC and 3 ⫻ 10⫺6 M for CS) amounted to 8.9 ⫾ 1.2 mN (n ⫽ 5, n ⫽ 10) and 6.6 ⫾ 0.8 (n ⫽ 12, n ⫽ 29) in CC and CS, respectively. Addition of NO to precontracted strips, produced concentration-dependent relaxations in both CC and CS (fig. 5). The -logIC50 values for NO in CC and CS were 6.1 ⫾ 0.1 (n ⫽ 5, n ⫽ 5) and 5.6 ⫾ 0.2 (n ⫽ 6, n ⫽ 6; ns). At the same concentrations as used for NO, CO also relaxed the preparations, but not as potently and not with the same efficacy as NO (fig. 5). The corresponding -logIC50 values for CO in CC and CS were 4.3 ⫾ 0.1 (n ⫽ 5, n ⫽ 5) and 4.6 ⫾ 0.1
871
CHOLINERGIC NERVES CONTAIN NO AND HO
FIG. 1. Protein Gene Product (PGP)-IR terminals along smooth muscle bundles in A, corpus cavernosum and B, corpus spongiosum, Texas Red (TR) immunofluorescence. Vesicular Acetylcholine Transporter (VAChT)-IR terminals along smooth muscle bundles in C, corpus cavernosum and D, corpus spongiosum, TR immunofluorescence. Bars ⫽ 100 m.
(n ⫽ 6, n ⫽ 6; ns). The -logIC50 values for NO and CO in the respective erectile tissue were significantly different (p ⬍0.001). In control experiments, addition of vehicle did not produce any effects. Carbachol 10⫺5 M produced relaxant responses in NAcontracted preparations amounting to 58 ⫾ 8% (n ⫽ 5, n ⫽ 4) for CC and 56 ⫾ 7% (n ⫽ 8, n ⫽ 5) for CS. After pretreatment with L-NNA, the carbachol-induced responses were reduced (p ⬍0.01) to 22 ⫾ 6% and 25 ⫾ 5%, respectively. In NA-contracted preparations, EFS produced frequencydependent relaxations, which were abolished by TT ⫻ 10⫺6 M Mean maximal relaxations were obtained at 18 Hz in both CC and CS and amounted to 47 ⫾ 5% (n ⫽ 5, n ⫽ 5) and 53 n˜ 8% (n ⫽ 7, n ⫽ 6; ns). The EFS-induced relaxant responses were at any of the investigated frequencies abolished by pretreatment with L-NNA in both CC and CS preparations (fig. 6). Cyclic nucleotides. The relaxant responses to exogenously applied NO in NA-contracted preparations of CC and CS, were accompanied by increases in the tissue concentration of cGMP. The mean tissue concentrations of cGMP in controls were 0.8 ⫾ 0.1 pmol./mg. protein⫺1 (n ⫽ 5, n ⫽ 5) for CC, and 1.1 ⫾ 0.2 pmol./mg. protein⫺1 (n ⫽ 6, n ⫽ 6; ns) for CS. After administration of NO (10⫺4 M), corresponding values were 4.7 ⫾ 1.0 pmol./mg. protein⫺1 (n ⫽ 5, n ⫽ 5; p ⬍0.01) and 5.8 ⫾ 1.1 pmol./mg. protein⫺1 (n ⫽ 6, n ⫽ 6; p ⬍0.001), respectively. CO, in the same concentration as NO, had no significant effect on the levels of cGMP (0.7 ⫾ 0.1 pmol./mg. protein⫺1 in CC; n ⫽ 5, n ⫽ 5, and 1.3 ⫾ 0.2 pmol./ mg. protein⫺1 in CS; n ⫽ 6, n ⫽ 6). In control specimens, the
mean concentrations of cAMP in CC and CS (25.5 ⫾ 2.1 and 23.7 ⫾ 4.0 pmol./mg. protein⫺1), were not significantly altered after administration of NO or CO. DISCUSSION
The erectile tissues of the corpus cavernosum (CC) and corpus spongiosum (CS) are morphologically similar with endothelium-lined trabecular smooth muscle, supported by fibroconnective tissue, nerves and vessels.1 As seen in the present study, both the CC and CS receive a dense innervation by varicose nerve terminals, gracile nerve fibers and numerous nerve trunks. There was no obvious difference in the amounts or distribution patterns of neuronal structures between the two tissues. Dail et al5 suggested that most neurons derived from the pelvic plexus and destined for the penis comprise a uniform population, demonstrating cholinergic characteristics, and capable of synthesizing NO. Their conclusion was partly based on the distribution of AChE staining. However, as mentioned previously, AChE staining is not specific for cholinergic nerves. We have used antibodies to VAChT, since VAChT is specific for cholinergic nerves. We can conclusively confirm that both human CC and CS receive a rich cholinergic innervation. As previously demonstrated by Weihe et al,3 VAChT immunoreactivity was most prominent in nerve terminals, consistent with the clustering of VAChTcontaining synaptic vesicles at the nerve terminal. In double immunolabeled sections of CC and CS, we found that VAChT- and NOS-, VAChT- and VIP-, and NOS- and VIP-IR
872
CHOLINERGIC NERVES CONTAIN NO AND HO
FIG. 2. Corpus cavernosum with varicose terminals accompanying smooth muscle bundles. A, Vesicular Acetylcholine Transporter (VAChT)-IR terminals, Texas Red (TR) immunofluorescence. B, same section as in A. Vasoactive Intestinal Polypeptide (VIP)-IR terminals with similar profiles as VAChT-IR terminals in A. Fluorescein 5⬘-IsoThioCyanate (FITC) immunofluorescence. C, Nitric Oxide Synthase (NOS)-IR terminals, TR immunofluorescence. D, same section as in C. VIP-IR terminals with similar profiles as NOS-IR terminals in C. FITC immunofluorescence. Bars ⫽ 100 m.
terminals exhibited coinciding profiles, and co-localization was confirmed by confocal microscopy. Thus, nerves containing not only VAChT, but also VIP and NOS, seem to comprise a distinct nerve population in human CC and CS. This is in agreement with what has previously been found in rat CC,19 underlining the similarity in CC innervation between the two species. The reason why NOS and VIP are located to the same terminals is unclear. Since erection and reproduction are basal functions, more than one transmitter system may be required to secure proper function. Single nerve fibers, containing VAChT-, NOS-, or VIP-immunoreactivities, may represent populations of nerves with possible afferent function. Jen et al20 reported that NOS and TH were co-localized in nerves supplying the postnatal human penis. Tamura et al21 reported that in the adult human penis, NOS could be found in nerves containing TH, suggesting that NO may be generated by adrenergic nerves. We found by double immunolabeling that VAChT- and NOS-IR terminals, VAChT- and VIP-IR terminals, and NOS- and VIP-IR terminals generally showed coinciding profiles, whereas profiles of VAChT- and TH-, and NOS- and TH-IR terminals were running closely together, although not coinciding. As verified with confocal laser scanning microscopy, VAChT- and TH-, nNOS- and TH-, and VIPand TH- immunoreactivities were found in separate nerve fibers and terminals. The described morphological arrangement would make possible the cholinergic modulation of NArelease from adjacent adrenergic nerves, as well as simulta-
neous direct smooth muscle relaxant action mediated by the cGMP and cAMP pathways. In the CC, a veno-occlusive mechanism produces a “closed” compartment which generates the intracavernosal pressure needed for penile erection. With a free drainage to penile veins, the CS is a “flow-through” system which is responsible for the erection of the glans.1 Both events depend upon trabecular smooth muscle relaxation, which in the CC is mainly mediated by the NO/cGMP pathway. In the dog CS, NOmediated effects were found to be similar to those seen in the CC, and a functional role for NO in the control of smooth muscle function in the CS was proposed.11 This is in agreement with the present findings, which demonstrate similar NO-mediated relaxant effects in the human CS and CC, and further support a role for NO in the functional control of CS smooth muscle. The endothelium may provide a complementary source of NO during penile erection. ACh-induced relaxation of erectile tissue has been shown to involve the release of NO from endothelial cells.1 No immunoreactivity for eNOS could be demonstrated in endothelial cells of the sinusoidal endothelium of the rat CC.5, 19, 22 Furthermore, carbachol induced only a small relaxant effect in NA-contracted rat CC preparations,19 which suggests that NO released from the endothelium lining the cavernous sinuses is not of major importance for erection in the rat. In contrast, the present study showed that the sinusoidal spaces in both CC and CS from humans were lined by a thin, and sometimes disrupted layer
CHOLINERGIC NERVES CONTAIN NO AND HO
873
FIG. 3. A, confocal microscopy. Corpus cavernosum with terminals showing Tyrosine Hydroxylase (TH)-IR (red) and Nitric Oxide Synthase (NOS)-IR (green) varicosities, which do not coincide. Bar ⫽ 10 m. B, corpus cavernosum with Vesicular Acetylcholine Transporter (VAChT)-IR terminals along smooth muscle bundles. Texas Red immunofluorescence. C, same section as in B. NOS-IR terminals with similar profiles as VAChT-IR terminals in A. Fluorescein 5⬘-IsoThioCyanate (FITC) immunofluorescence. Bars ⫽ 100 m.
FIG. 4. Endothelial cells along sinusoid in corpus cavernosum expressing A, endothelial Nitric Oxide Synthase (NOS)-immunoreactivity and B, Hemoxygenase-1 (HO-1)-immunoreactivity (B). Bars ⫽ 30 m (A) and 50 m (B).
of eNOS-IR endothelial cells. This is also in agreement with the functional findings, showing that carbachol had a good relaxant effect in the preparations, and supports the view that NO produced by eNOS in the sinusoids contributes to erection. CO has been shown to produce relaxation of smooth muscle by stimulating guanylate cyclase,23 and both CO and NO have been implicated as messenger molecules in the central
and peripheral nervous system.24 The present study revealed that relaxations produced by CO were similar in appearance in NA-contracted preparations of human CC and CS. In comparison with the effects obtained with NO in the same tissues, the maximal CO-induced relaxation was about 40% lower than that produced by NO. The tissue-levels of cyclic nucleotides were unaffected by CO, suggesting that other mechanisms than stimulation of guanylate cyclase (or ade-
874
CHOLINERGIC NERVES CONTAIN NO AND HO
of exogenously applied CO, suggest that CO may be involved as a messenger molecule in the regulation of smooth muscle tone in the penis. Conclusion. This study demonstrates that the human CC and CS receive a dense cholinergic innervation. VAChT, NOS and VIP are found in the same nerve terminals within the human CC and CS, suggesting that these nerves comprise a distinct population of parasympathetic, cholinergic nerves. In both CC and CS, HO-can be demonstrated in nerves and in endothelium, and exogenous CO produces relaxation. It seems likely that both the NOS/NO and CO/HO-systems have functional roles in the control of smooth muscle tone of CC and CS. We thank Dr. E. Weihe and Dr. P. Emson for a generous supply of antibodies, and Dr. Anders Håkansson for expert help with the confocal microscopy. REFERENCES
FIG. 5. Addition of NO (squares) or CO (circles) to noradrenaline (l-NA)-contracted strips produced concentration-dependent relaxation in both corpus cavernosum (CC: filled symbols) and corpus spongiosum (CS: open symbols).
FIG. 6. In noradrenaline (l-NA)-contracted preparations, electrical field stimulation produced frequency-dependent relaxation in both corpus cavernosum (CC: open squares) and corpus spongiosum (CS: open circles) preparations, which was effectively suppressed by pretreatment with NG-nitro-L-arginine (10⫺4 M) in both CC (filled squares) and CS (filled circles).
nylate cyclase) are responsible for the relaxant activity of CO. Involvement of K⫹ channels in the action of CO on smooth muscle preparations from the gastrointestinal tract has been proposed.25, 26 If this is the case also in human penile erectile tissues cannot be decided on the basis of the present results. In CC tissue from the rabbit, Kim et al27 found zinc deuteroporphyrin to be without effect on electrically-induced inhibitory responses, and suggested that CO was not involved in neuronally mediated relaxations of cavernous erectile tissue. However, non-specific effects, other than inhibition of HO, have been described for the metalloporphyrins,28 which questions the use of these compounds for the study of the activity of the CO/HO-system. The presently found immunoreactivity for HO isozymes in nerve structures and endothelium of human CC and CS, and the functional effects
1. Andersson, K-E. and Wagner, G.: Physiology of penile erection. Physiol Rev, 75: 191, 1995 2. Lincoln, J. and Burnstock, G.: Autonomic innervation of the urinary bladder and urethra. In: The Autonomic Nervous System. Vol. 6, Chapter 8, Nervous Control of the Urogenital System, edited by C. A. Maggi. Harwood Academic Publishers: London, p. 33, 1993 3. Weihe, E., Tao-Cheng, J-H., Scha¨fer, M. K-H., Erickson, J. D. and Eiden, L. E.: Visualization of the vesicular acetylcholine transporter in cholinergic nerve terminals and its targeting to a specific population of small synaptic vesicles. Proc Natl Acad Sci USA, 93: 3547, 1996 4. Arvidsson, U., Reidl, M., Elde, R. and Meister, B.: Vesicular acetylcholine transporter (VAChT) protein: A novel and unique marker for cholinergic neurons in the central and peripheral nervous systems. J Comp Neurol, 378: 454, 1997 5. Dail, W. G., Barba, V., Leyba, L. and Galindo, R.: Neuronal and endothelial nitric oxide synthase activity in rat penile erectile tissue. Cell Tissue Res, 282: 109, 1995 6. Cook, M. N., Nakatsu, K., Marks, G. S., McLaughlin, B. E., Vreman, H. J., Stevensen, D. K. and Brien, J. F.: Heme oxygenase activity in the adult rat aorta and liver as measured by carbon monoxide formation. Can J Physiol Pharmacol, 73: 515, 1995 7. Morita, T., Perella, M. A., Lee, M-E. and Kourembanas, S.: Smooth muscle cell derived carbon monoxide is a regulator of vascular cGMP. Proc Natl Acad Sci USA, 92: 1475, 1995 8. Ny, L., Alm, P., Ekstro¨m, P., Larsson, B., Grundemar, L. and Andersson, K-E.: Localization and activity of haem oxygenase and functional effects of carbon monoxide in the feline lower oesophageal sphincter. Br J Pharmacol, 118: 392, 1996 9. Zakhary, R., Gaine, S. P., Dinerman, J. L., Ruat, M., Flavahan, N. A. and Snyder, S. H.: Heme oxygenase-2: Endothelial and neuronal localization and role in endothelium-dependent relaxation. Proc Natl Acad Sci USA, 93: 795, 1996 10. Rattan, S. and Chadker, S.: Inhibitory effect of CO on internal anal sphincter: heme oxygenase inhibitor inhibits NANC relaxation. Am J Physiol, 265: G799, 1993 11. Hedlund, P., Larsson, B., Alm, P. and Andersson K-E.: Distribution and function of nitric oxide-containing nerves in canine corpus cavernosum and spongiosum. Acta Physiol Scand, 155: 445, 1995 12. Wessendorf, M. W. and Elde, R. P.: Characterization of an immunofluorescence technique for the demonstration of coexisting neurotransmitters within nerve fibres and terminals. J Histochem Cytochem, 33: 984, 1985 13. Scha¨fer, K-H., Weihe, E., Erickson, J. D. and Eiden, L. E.: Human and monkey cholinergic neurons visualized in paraffin-embedded tissues by immunoreactivity for VAChT, the vesicular acetylcholine transporter. J Mol Neurosci, 6: 225, 1996 14. Herbison, A. E., Simonian, S. X., Norris, P. J. and Emson, P. C.: Relationship of neuronal nitric oxide synthase immunoreactivity to GnRH neurons in the ovariectomized and intact female rat. J Neuroendocrinol, 8: 73, 1996 15. Lundberg, L. M., Alm, P., Wharton, J. and Polak, J. M.: Protein gene product 9.5 (PGP 9.5) A new neuronal marker visualizing
CHOLINERGIC NERVES CONTAIN NO AND HO
16.
17.
18.
19. 20.
21.
the whole uterine innervation and pregnancy-induced and developmental changes in the guinea-pig. Histochemistry, 90: 9, 1988 Gulbenkian, S., Wharton, J. and Polak, J. M.: The visualization of cardiovascular innervation in the guinea pig using an antiserum to protein gene product 9.5 (PGP 9.5). J Auton Nerv Syst, 18: 235, 1987 Ny, L., Alm, P., Ekstro¨m, P., Hannibal, J., Larsson, B. and Andersson, K-E.: Nitric oxide synthase-containing, peptidecontaining, and acetylcholinesterase-positive nerves in the cat lower oesophagus. Histochem J, 26: 721, 1994 Hedlund, P., Alm, P., Ekstro¨m, P., Fahrenkrug, J., Hannibal, J., Hedlund, H., Larsson, B. and Andersson, K-E.: Pituitary adenylate cyclase-activating polypeptide, helospectin, and vasoactive intestinal polypeptide in the human corpus cavernosum. Br J Pharmacol, 116: 2258, 1995 Hedlund, P., Alm, P. and Andersson, K-E.: NO synthase in cholinergic nerves and NO-induced relaxation in the rat isolated corpus cavernosum. Br J Pharmacol, 127: 349, 1999 Jen, P. Y. P., Dixon, J. S., Gearhart, J. P. and Gosling, J. A.: Nitric oxide synthase and tyrosine hydroxylase are colocalized in nerves supplying the postnatal human genitourinary organs. J Urol, 155: 1117, 1996 Tamura, M., Kagawa, S., Kimura, K., Kawanishi, Y., Tsuruo, Y. and Ishimura, K.: Coexistence of nitric oxide synthase, tyrosin hydroxylase and vasoactive intestinal polypeptide in human
22.
23. 24. 25.
26.
27.
28.
875
penile tissue - a triple histochemical and immunohistochemical study. J Urol, 153: 530, 1995 Schirar, A., Chang, C. and Rousseau, J. P.: Localization of androgen receptor in nitric oxide synthase- and vasoactive intestinal peptide-containing neurons of the major pelivic ganglion innervating the rat penis. J Neuroendocrinol, 9: 141, 1997 Schmidt, H. H. H. W.: NO, CO and OH. Endogenous soluble guanylyl cyclase-activating factors. FEBS Lett, 307: 102, 1992 Maines, M. D.: The heme oxygenase system; a regulator of second messenger gases. Ann Rev Pharmacol Toxicol, 37: 517, 1997 Farrugia, G., Irons, W. A., Rae, J. L., Sarr, M. G. and Szurszewski, J. H.: Activation of whole cell currents in isolated human jejunal circular smooth muscle cells by carbon monoxide. Am J Physiol, 264: G1184, 1993 Werkstro¨m, W., Ny, L., Persson, K. and Andersson, K-E.: Carbon monoxide induced relaxation and distribution of haem oxygenase in the pig urethra and oesophagogastric junction. Br J Pharmacol, 120: 312, 1997 Kim, Y. C., Davies, M. G., Marson, L., Hagen, P-O. and Carson, C. C.: Lack of effect of carbon monoxide inhibitor on relaxation induced by electrical field stimulation in corpus cavernosum. Urol Res, 22: 291, 1994 Grundemar, L. and Ny, L.: Pitfalls using metalloporphyrins in carbon monoxide research. Trends Pharmacol Sci, 18: 193, 1997
Vol. 164, 868 – 875, September 2000 Printed in U.S.A.
CHOLINERGIC NERVES IN HUMAN CORPUS CAVERNOSUM AND SPONGIOSUM CONTAIN NITRIC OXIDE SYNTHASE AND HEME OXYGENASE PETTER HEDLUND, LARS NY, PER ALM
AND
KARL-ERIK ANDERSSON*
From the Departments of Clinical Pharmacology and Pathology, Lund University Hospital, Lund, Sweden
ABSTRACT
Purpose: To characterize the distribution of cholinergic nerves in the human corpus cavernosum (CC) and spongiosum (CS) using antibodies to the vesicular acetylcholine transporter (VAChT), and to compare this distribution to those of other transmitters/mediators or transmitter/mediator generating enzymes (heme oxygenases: HO-1 and HO-2; neuronal and endothelial NO synthases: nNOS and eNOS; vasoactive intestinal polypeptide: VIP; and tyrosine hydroxylase: TH), and to investigate NO- and carbon monoxide (CO)-mediated effects. Materials and Methods: Immunocytochemistry, confocal laser scanning microscopy, radioimmunoassay, and functional in vitro studies. Results: Along strands of smooth muscle in the CC and CS, rich numbers of VAChT-, nNOS-, VIP-, TH-, and very few HO-1-immunoreactive (-IR) nerve fibers were observed. Immunoreactivities for VAChT and nNOS, VAChT and VIP, and nNOS and VIP, were generally found in the same varicose nerve terminals. TH-IR nerve fibers or terminals did not contain immunoreactivities for VAChT, NOS or VIP. In the endothelium lining penile arteries, immunoreactivities for eNOS, HO-1, and HO-2 were detected. Single endothelial cells, lining the sinusoidal walls of the CC and CS, were found also to contain eNOS and HO-immunoreactivities. Noradrenaline (NA)-contracted preparations of CC and CS were relaxed by NO, CO, carbachol and by electrical stimulation of nerves. Inhibition of NO synthesis abolished electrically- and carbachol-induced relaxation. In NA-activated strips, relaxation induced by exogenously applied NO, but not those by CO, were accompanied by increases in intracellular levels of cyclic GMP. Conclusions: VAChT, NOS and VIP are found in the same nerve terminals within the human CC and CS, suggesting that these terminals comprise a distinct population of parasympathetic, cholinergic nerves. Endothelially derived NO and the HO/CO system may have a complementary role in penile erection. KEY WORDS: penile erection, innervation, co-localization, confocal microscopy
Pelvic parasympathetic nerves provide a pathway for the vasodilator responses in the penile vasculature, and for the relaxation of corpus cavernosum (CC) smooth muscle necessary for erection. Acetylcholine (ACh), released from cholinergic nerves, may via stimulation of muscarinic receptors on adrenergic nerves decrease the release of noradrenaline (NA), and also release vasoactive agents from the vascular endothelium. In addition, stimulation of cholinergic nerves can lead to release of other relaxation-producing transmitters.1 Numerous nerves in the erectile tissue contain ACh esterase (AChE). However, this enzyme is not confined to cholinergic nerves, and the presence of AChE is not considered sufficient for identification of cholinergic nerves.2 ACh is transferred into vesicles in the nerves by the vesicular acetylcholine transporter (VAChT), and immunohistochemical demonstration of VAChT is considered to be a specific approach for visualization of cholinergic nerves in both the central and peripheral nervous system.3, 4 Nitric oxide synthase (NOS) has been localized to nerves in penile erectile tissues from several species, including humans, and effects mediated by NO have been extensively
studied in isolated CC tissue and in several animal models in vivo.1 In the rat, the presence of NOS in the sinusoidal endothelium (eNOS) is disputed,5 but endothelial NO-dependent relaxations have been demonstrated functionally in human CC preparations.1 However, immunohistochemical demonstration of eNOS in human sinusoidal endothelium has so far not been convincing. Heme oxygenase (HO), a carbon monoxide (CO) producing enzyme, has been demonstrated in neurons and in endothelial and smooth muscle cells in various peripheral organs.6 –9 CO has been shown to induce relaxation of vascular and gastrointestinal smooth muscle preparations.7–10 If CO is involved in smooth muscle regulatory functions in the human penis, this has, to the best of our knowledge, not been clarified. In the present study, we wanted to study and compare in isolated human CC and CS tissue: 1) the distribution of VAChT-immunoreactive (-IR) nerves, and relate it to the total distribution of nerves as well as to populations of nerves containing nNOS, HO, vasoactive intestinal polypeptide (VIP), and tyrosine hydroxylase (TH); 2) the distribution of eNOS immunoreactivity; and 3) the functional effects of NO and CO in isolated preparations of the two tissues.
Accepted for publication April 4, 2000. * Requests for reprints: Department of Clinical Pharmacology, Lund University Hospital, S-221 85 Lund, Sweden. Supported by the Swedish Medical Research Council (Grants no. MATERIALS AND METHODS 6837 and 11205), the Royal Physiographic Society, the Foundations Preparations. Specimens of penile erectile tissue were obof Crafoord, Magnus Bergvall, and Anna-Lisa and Sven-Erik Luntained from 12 men, aged 21 to 71 years (mean 55 years). Ten dgren, and the Medical Faculty, University of Lund, Sweden. 868
869
CHOLINERGIC NERVES CONTAIN NO AND HO
(with normal erectile capacity) underwent cystourethrectomy (no preoperative external radiation) because of bladder malignancy. Cavernous tissue was taken by biopsy peroperatively after consent from the patients had been obtained. One patient underwent a gender reassignment operation. Spongious tissue was also obtained from one donor patient with brain damage after an intracranial hemorrhage. The procedure for obtaining CC and CS tissue was approved by the Ethics Committee, Lund University. Immunocytochemistry. Tissue specimens were fixed and further processed for immunocytochemistry as previously described.11 Cryostat sections were incubated overnight in the presence of a primary antiserum. The primary antisera used (table 1) were rabbit antisera to protein gene product 9.5 (PGP), inducible heme oxygenase (HO-1/Hsp32), constitutive heme oxygenase (HO-2), endothelial NOS (eNOS), and a C-terminal dodecapeptide to cloned human vesicular acetylcholine transporter (VAChT), a sheep antiserum to neuronal NOS (nNOS), a mouse monoclonal tyrosine hydroxylase (TH) antiserum, and a guinea pig antiserum to vasoactive intestinal peptide (VIP). After rinsing, the sections were incubated for 90 minutes in the presence of an appropriate secondary antiserum (see table 2), whereupon, the sections were rinsed and mounted. For the simultaneous demonstration of two antigens,12 some sections were incubated overnight in rabbit antisera to HO-1, or HO-2, rinsed, and then incubated overnight with sheep antiserum to nNOS, or mouse monoclonal THantiserum, or guinea pig antiserum to VIP. Some sections were incubated overnight in guinea-pig VIP or sheep nNOS antisera, rinsed, and then incubated overnight in rabbit VAChT- antiserum. Some sections were incubated overnight in in sheep nNOS antiserum, rinsed, and then incubated overnight in guinea-pig VIP antiserum, or mouse monoclonal TH-antiserum. After rinsing, the sections were incubated in appropriate combinations of secondary antisera, whereupon they were rinsed and mounted. The primary and secondary antisera used, and their combinations, are described in tables 1-4. The characteristics of the rabbit VAChT and sheep nNOS antisera have previously been described.13, 14 In control experiments, no immunoreactivity could be detected in sections incubated in the absence of primary antisera, or incubated with HO-1, HO-2, VAChT or VIP antisera absorbed with excess of the respective antigens (100 g.ml.⫺1). No absorption tests were performed with the TH, eNOS, or nNOS antisera, as antigenic substances were not available. The immunoreactive nerve structures were independently evaluated by the authors with respect to type (nerve trunks and nerve terminals) and relative frequency (arbitrarily graded as very numerous, numerous, moderate number, few and absent.15 In the evaluation of the latter parameter, immunoreactivity to PGP was used as an internal standard as a marker for all populations of nerve fibers.15, 16 In the present study, approximately 250 glasses, each with 4 to 6
sections, were used for the immunohistochemical characterization of the investigated structures. The structures related are referred to as PGP-, HO-1-, HO-2-, nNOS-, eNOS-, TH-, VAChT- and VIP-IR, as cross-reactions with other antigens sharing similar amino acid sequences cannot be completely excluded. The sections were examined in an Olympus BX 40 fluorescence microscope with epi-illumination and filter settings for Fluorescein 5⬘-IsoThioCyanate (FITC)- and Texas Red (TR)- immunofluorescence.17 Confocal laser scanning microscopy. To evaluate if two immunoreactivities were located in the same nerve structures, sections were analyzed in a Bio-Rad MRC-1024 confocal laser scanning equipment (Bio-Rad Laboratories, UK) attached to a Nikon Eclipse E800 upright microscope (Nikon, Japan). The sections were scanned with a 60 x (1.40) oil immersion plan apochromat lens. Sequential scanning using excitation wavelengths of 488 and 568 nm from a KryptonArgon laser. FITC immunofluorescence was detected with a 522 ⫾ 30 nm band pass filter and TR immunofluorescence was detected with a 605 ⫾ 30 nm band pass filter. Functional studies. Strip preparations (1 ⫻ 2 x 5 mm.) were prepared and then suspended in aerated, thermostatically controlled 5 ml. tissue baths (37C, pH 7.4) containing Krebs solution, which was routinely replaced with fresh Krebs solution every 30 minutes. Isometric tension was recorded by means of a Grass Instruments FTO3C force-transducer connected to a Grass 7D polygraph (Grass Instruments Co., USA). An equilibration period of 45 minutes was allowed until mean final stable tensions of 3.8 ⫾ 0.4 mN (n ⫽ 6, n ⫽ 10) for CC, and 3.0 ⫾ 0.3 mN (n ⫽ 12, n ⫽ 30) for CS were obtained. The viability of the preparations was verified by addition of a high K⫹ solution (124 mM) to the organ baths, and contractile responses amounting to 12.4 ⫾ 3.8 mN for CC, and 8.3 ⫾ 1.0 mN for CS, were recorded. The l-noradrenaline (NA) concentration used corresponded to the approximate EC50 value for NA in this tissue (CC, 10⫺6 M; CS 3 ⫻ 10⫺6 M). Electrical field stimulation (EFS) was performed and inhibitory frequency-response relationships were investigated as previously described.11 The effects of NO, CO, carbachol, and EFS in NA-contracted preparations were studied, and the degree of relaxation was expressed as a percentage of the NA-induced contraction. Some preparations were pretreated for 20 minutes with the NO-synthase inhibitor NG-nitro-L-arginine (L-NNA; 10⫺4 M). Radioimmunoassays. Guanosine and adenosine 3⬘: 5⬘cyclic monophosphate (cGMP and cAMP) were analyzed in NA-contracted preparations of erectile tissue (controls) and in NA-contracted preparations exposed to NO and CO. When the effects of the agents reached a maximal and stable level, approximately after 1 minute, the strips were immediately frozen in liquid nitrogen and then stored in 0.5 ml. of 10% trichloracetic acid (TCA) at ⫺70C. The tissue was then homogenized, further processed and the amounts of cGMP and cAMP were quantified as previously described.18 Drugs and solutions. A Krebs solution of the following
TABLE 1. Primary antisera Anti-Serum
Host
Working Dilution
Code
Source
HO-1 (Hsp32) (ra) 1:500 SPA895 StressGen, Victoria, Canada HO-1 (ra) 1:500 OSA100 StressGen, Victoria, Canada HO-2 (ra) 1:500 OSA200 StressGen, Victoria, Canada eNOS (ra) 1:500 Sancta Cruz Biotech, Heidelberg, Germany nNOS (sh) 1:6000 P. Emson, Babraham Institute, Cambridge, England PGP (ra) 1:2000 UltraClone, Wellow, Isle of Wight, England TH (mo) 1:200 Incstar, Stillwater, MO, USA VAChT (ra) 1:1600 E. Weihe, Marburg, Germany VIP (gp) 1:640 Euro-Diagnostica, Malmo¨, Sweden (ra) ⫽ rabbit, (go) ⫽ goat, (sh) ⫽ sheep, (gp) ⫽ guinea pig, (mo) ⫽ mouse. HO(Hsp32) ⫽ Heme Oxygenase, eNOS ⫽ endothelial NO Synthase, nNOS ⫽ neuronal NO Synthase, PGP ⫽ Protein Gene Product 9.5, TH ⫽ Tyrosine Hydroxylase, VAChT ⫽ Vesicular Acetylcholine Transporter, VIP ⫽ Vasoactive Intestinal Polypeptide
870
CHOLINERGIC NERVES CONTAIN NO AND HO TABLE 2. Secondary antisera
Anti-Serum
Host
1.FITC swine anti-rabbit 2.FITC goat anti-guinea pig 3.FITC donkey anti-sheep 4.TR donkey anti-mouse 5.TR donkey anti-rabbit 6.TR donkey anti-sheep FITC ⫽ Fluorescein 5⬘-IsoThioCyanate, TR ⫽ Texas Red
Working Dilution
Code
Source
1:80 1:80 1:80 1:125 1:125 1:125
F 0205 F 6261 T 5393 715-076-151 711-075-152 713-075-147
Dakopatts, Stockholm, Sweden Sigma, St Louis, MO, USA Sigma, St Louis, MO, USA Jackson ImmunoResearch, West Grove, PA, USA Jackson ImmunoResearch, West Grove, PA, USA Jackson ImmunoResearch, West Grove, PA, USA
composition was used (mM): NaCl 119, KCl 4.6, CaCl2 1.5, MgCl2 1.2, NaHCO3 15, NaH2PO4 1.2, glucose 11. An isotonic K⫹ solution (124 mM) was used, in which the NaCl in the normal Krebs solution was replaced by equimolar amounts of KCl. The following drugs were used: NA (Aldrich-Chemie GmbH & Co., Germany), carbachol, tetrodotoxin (TTX) and L-NNA (Sigma Chemical Co., USA). NO and CO were freshly prepared at each experiment. Air-tight glass beakers containing 20 ml. of distilled water were deoxygenated for 1 hour with helium gas. The beakers were then aerated with medical NO or CO gas (purity ⬎99.5%) for 15 minutes until saturated solutions were obtained (NO 3 ⫻ 10⫺3 M, CO 10⫺3 M). Calculations. Student’s paired and unpaired two-tailed t tests were used for statistical comparison of two means. For multiple comparisons, analysis of variance with Bonferroni correction was used. A probability of p ⬍0.05 was accepted as significant. Results are given as mean values ⫾ standard error of the mean (SEM). N denotes the number of individuals, and n denotes the number of strip preparations. All statistical comparisons were made using N. RESULTS
Immunocytochemical findings. In the trabecular smooth muscle tissue, and around arteries in the CC and CS, no nerve trunks displaying VAChT-IR nerve fibers could be observed. However, nerve trunks with immunoreactivities for PGP, nNOS, HO-1, or HO-2 were found. nNOS-IR nerve trunks also contained immunoreactivity for VIP, and single nerve fibers within the trunks contained both nNOS and VIP. TH-IR nerve fibers were also seen in nerve trunks which contained nNOS-IR fibers. Within these nerve trunks, single nNOS- and TH-IR nerve fibers were running close together. As seen in double filter settings and in confocal sections, immunoreactivities for nNOS and TH were found in separate nerve fibers. Nerve trunks with HO-1- and HO-2-IR fibers also contained nNOS- VIP-, or TH-immunoreactivity. Within these nerve trunks, the HO-1- or HO-2-IR fibers were found to exhibit profiles which were similar but not identical to corresponding nNOS-, VIP-, or TH-IR fibers. TABLE 3. Mono-immunolabelling Primary Antiserum (see Table 1)
Secondary Antiserum (see table 2)
HO-1 (Hsp32) (ra) 5 HO-1 (ra) 5 HO-2 (ra) 5 eNOS (ra) 5 nNOS (sh) 3 PGP (ra) 5 TH (mo) 4 VAChT (ra) 5 VIP (gp) 2 After incubation with the two primary antisera, the sections were rinsed, and then incubated for 90 min with first secondary antiserum, rinsed, and then incubated for 90 min with other secondary antiserum. After rinsing, the sections were rinsed and mounted. (ra) ⫽ rabbit, (go) ⫽ goat, (sh) ⫽ sheep, (gp) ⫽ guinea pig, (mo) ⫽ mouse. HO(Hsp32) ⫽ Heme Oxygenase, eNOS ⫽ endothelial NO Synthase, nNOS ⫽ neuronal NO Synthase, PGP ⫽ Protein Gene Product 9.5, TH ⫽ Tyrosine Hydroxylase, VAChT ⫽ Vesicular Acetylcholine Transporter, VIP ⫽ Vasoactive Intestinal Polypeptide.
TABLE 4. Double immunolabelling Combinations of Primary Antisera (see Table 1)
Combinations of Secondary Antisera (see Table 2)
HO-1/Hsp32 (ra) ⫹ nNOS (sh) 3⫹5 HO-1/Hsp32 (ra) ⫹ TH (mo) 1⫹4 HO-1/Hsp32 (ra) ⫹ VIP (gp) 2⫹5 HO-2 (ra) ⫹ nNOS (sh) 3⫹5 HO-2 (ra) ⫹ TH (mo) 1⫹4 HO-2 (ra) ⫹ VIP (gp) 2⫹5 nNOS (sh) ⫹ TH (mo) 3⫹4 nNOS (sh) ⫹ VAChT (ra) 3⫹5 VIP (gp) ⫹ nNOS (sh) 2⫹6 VIP (gp) ⫹ VAChT (ra) 2⫹5 (ra) ⫽ rabbit, (go) ⫽ goat, (sh) ⫽ sheep, (gp) ⫽ guinea pig, (mo) ⫽ mouse. HO(Hsp32) ⫽ Heme Oxygenase, nNOS ⫽ neuronal NO Synthase, PGP ⫽ Protein Gene Product 9.5, TH ⫽ Tyrosine Hydroxylase, VAChT ⫽ Vesicular Acetylcholine Transporter, VIP ⫽ Vasoactive Intestinal Polypeptide
Gracile nerve fibers and varicose nerve terminals containing VAChT-, nNOS- and VIP-immunoreactivities were found in numerous numbers interspersed among trabecular smooth muscle cells. Single fibers containing VAChT-, nNOS-, or VIP-immunoreactivities were observed. In comparison, smooth muscle related PGP-IR varicose nerve fibers were very numerous. The distribution pattern and amount of the different immunoreactivities were similar in specimens from CC and CS (fig. 1, A-D). Very few HO-1-IR, but no HO-2-IR, nerve fibers were seen in the trabecular smooth muscle. In their adventitial layers, arteries were encircled by plexa of VAChT-, nNOS-, VIP-, and TH-IR terminals. Double immunolabeling and confocal microscopy disclosed that the majority of nNOS-IR varicose nerve terminals also contained VAChT- and VIP-immunoreactivity, and that VAChT-IR terminals also contained VIP-immunoreactivity (figs. 2 and 3). nNOS-and TH-IR smooth muscle-related nerve terminals were similarly distributed, but the profiles of the respective nerves terminals were not identical (fig. 3, A). In the CC, the cavernous artery was lined by eNOS-IR endothelial cells, as were small arteries and veins in the CC and CS. The sinusoidal spaces in both tissues were lined by a thin and sometimes disrupted layer of eNOS-IR endothelial cells. The eNOS-IR endothelial cells contained a central nucleus that protruded toward the vascular lumen, and was surrounded by cytoplasm, which extended sideways (fig. 4, A). Immunoreactivities for HO-1 (both HO-1 antisera used) and HO-2 were expressed in endothelial cells of different vessels, including the cavernous artery, and the endothelium lining the walls of the sinusoidal spaces of the CC and CS (fig. 4, B) also contained immunoreactivity for HO-1 (both antisera used) and HO-2. Functional effects. The NA-induced contraction (10⫺6 M for CC and 3 ⫻ 10⫺6 M for CS) amounted to 8.9 ⫾ 1.2 mN (n ⫽ 5, n ⫽ 10) and 6.6 ⫾ 0.8 (n ⫽ 12, n ⫽ 29) in CC and CS, respectively. Addition of NO to precontracted strips, produced concentration-dependent relaxations in both CC and CS (fig. 5). The -logIC50 values for NO in CC and CS were 6.1 ⫾ 0.1 (n ⫽ 5, n ⫽ 5) and 5.6 ⫾ 0.2 (n ⫽ 6, n ⫽ 6; ns). At the same concentrations as used for NO, CO also relaxed the preparations, but not as potently and not with the same efficacy as NO (fig. 5). The corresponding -logIC50 values for CO in CC and CS were 4.3 ⫾ 0.1 (n ⫽ 5, n ⫽ 5) and 4.6 ⫾ 0.1
871
CHOLINERGIC NERVES CONTAIN NO AND HO
FIG. 1. Protein Gene Product (PGP)-IR terminals along smooth muscle bundles in A, corpus cavernosum and B, corpus spongiosum, Texas Red (TR) immunofluorescence. Vesicular Acetylcholine Transporter (VAChT)-IR terminals along smooth muscle bundles in C, corpus cavernosum and D, corpus spongiosum, TR immunofluorescence. Bars ⫽ 100 m.
(n ⫽ 6, n ⫽ 6; ns). The -logIC50 values for NO and CO in the respective erectile tissue were significantly different (p ⬍0.001). In control experiments, addition of vehicle did not produce any effects. Carbachol 10⫺5 M produced relaxant responses in NAcontracted preparations amounting to 58 ⫾ 8% (n ⫽ 5, n ⫽ 4) for CC and 56 ⫾ 7% (n ⫽ 8, n ⫽ 5) for CS. After pretreatment with L-NNA, the carbachol-induced responses were reduced (p ⬍0.01) to 22 ⫾ 6% and 25 ⫾ 5%, respectively. In NA-contracted preparations, EFS produced frequencydependent relaxations, which were abolished by TT ⫻ 10⫺6 M Mean maximal relaxations were obtained at 18 Hz in both CC and CS and amounted to 47 ⫾ 5% (n ⫽ 5, n ⫽ 5) and 53 n˜ 8% (n ⫽ 7, n ⫽ 6; ns). The EFS-induced relaxant responses were at any of the investigated frequencies abolished by pretreatment with L-NNA in both CC and CS preparations (fig. 6). Cyclic nucleotides. The relaxant responses to exogenously applied NO in NA-contracted preparations of CC and CS, were accompanied by increases in the tissue concentration of cGMP. The mean tissue concentrations of cGMP in controls were 0.8 ⫾ 0.1 pmol./mg. protein⫺1 (n ⫽ 5, n ⫽ 5) for CC, and 1.1 ⫾ 0.2 pmol./mg. protein⫺1 (n ⫽ 6, n ⫽ 6; ns) for CS. After administration of NO (10⫺4 M), corresponding values were 4.7 ⫾ 1.0 pmol./mg. protein⫺1 (n ⫽ 5, n ⫽ 5; p ⬍0.01) and 5.8 ⫾ 1.1 pmol./mg. protein⫺1 (n ⫽ 6, n ⫽ 6; p ⬍0.001), respectively. CO, in the same concentration as NO, had no significant effect on the levels of cGMP (0.7 ⫾ 0.1 pmol./mg. protein⫺1 in CC; n ⫽ 5, n ⫽ 5, and 1.3 ⫾ 0.2 pmol./ mg. protein⫺1 in CS; n ⫽ 6, n ⫽ 6). In control specimens, the
mean concentrations of cAMP in CC and CS (25.5 ⫾ 2.1 and 23.7 ⫾ 4.0 pmol./mg. protein⫺1), were not significantly altered after administration of NO or CO. DISCUSSION
The erectile tissues of the corpus cavernosum (CC) and corpus spongiosum (CS) are morphologically similar with endothelium-lined trabecular smooth muscle, supported by fibroconnective tissue, nerves and vessels.1 As seen in the present study, both the CC and CS receive a dense innervation by varicose nerve terminals, gracile nerve fibers and numerous nerve trunks. There was no obvious difference in the amounts or distribution patterns of neuronal structures between the two tissues. Dail et al5 suggested that most neurons derived from the pelvic plexus and destined for the penis comprise a uniform population, demonstrating cholinergic characteristics, and capable of synthesizing NO. Their conclusion was partly based on the distribution of AChE staining. However, as mentioned previously, AChE staining is not specific for cholinergic nerves. We have used antibodies to VAChT, since VAChT is specific for cholinergic nerves. We can conclusively confirm that both human CC and CS receive a rich cholinergic innervation. As previously demonstrated by Weihe et al,3 VAChT immunoreactivity was most prominent in nerve terminals, consistent with the clustering of VAChTcontaining synaptic vesicles at the nerve terminal. In double immunolabeled sections of CC and CS, we found that VAChT- and NOS-, VAChT- and VIP-, and NOS- and VIP-IR
872
CHOLINERGIC NERVES CONTAIN NO AND HO
FIG. 2. Corpus cavernosum with varicose terminals accompanying smooth muscle bundles. A, Vesicular Acetylcholine Transporter (VAChT)-IR terminals, Texas Red (TR) immunofluorescence. B, same section as in A. Vasoactive Intestinal Polypeptide (VIP)-IR terminals with similar profiles as VAChT-IR terminals in A. Fluorescein 5⬘-IsoThioCyanate (FITC) immunofluorescence. C, Nitric Oxide Synthase (NOS)-IR terminals, TR immunofluorescence. D, same section as in C. VIP-IR terminals with similar profiles as NOS-IR terminals in C. FITC immunofluorescence. Bars ⫽ 100 m.
terminals exhibited coinciding profiles, and co-localization was confirmed by confocal microscopy. Thus, nerves containing not only VAChT, but also VIP and NOS, seem to comprise a distinct nerve population in human CC and CS. This is in agreement with what has previously been found in rat CC,19 underlining the similarity in CC innervation between the two species. The reason why NOS and VIP are located to the same terminals is unclear. Since erection and reproduction are basal functions, more than one transmitter system may be required to secure proper function. Single nerve fibers, containing VAChT-, NOS-, or VIP-immunoreactivities, may represent populations of nerves with possible afferent function. Jen et al20 reported that NOS and TH were co-localized in nerves supplying the postnatal human penis. Tamura et al21 reported that in the adult human penis, NOS could be found in nerves containing TH, suggesting that NO may be generated by adrenergic nerves. We found by double immunolabeling that VAChT- and NOS-IR terminals, VAChT- and VIP-IR terminals, and NOS- and VIP-IR terminals generally showed coinciding profiles, whereas profiles of VAChT- and TH-, and NOS- and TH-IR terminals were running closely together, although not coinciding. As verified with confocal laser scanning microscopy, VAChT- and TH-, nNOS- and TH-, and VIPand TH- immunoreactivities were found in separate nerve fibers and terminals. The described morphological arrangement would make possible the cholinergic modulation of NArelease from adjacent adrenergic nerves, as well as simulta-
neous direct smooth muscle relaxant action mediated by the cGMP and cAMP pathways. In the CC, a veno-occlusive mechanism produces a “closed” compartment which generates the intracavernosal pressure needed for penile erection. With a free drainage to penile veins, the CS is a “flow-through” system which is responsible for the erection of the glans.1 Both events depend upon trabecular smooth muscle relaxation, which in the CC is mainly mediated by the NO/cGMP pathway. In the dog CS, NOmediated effects were found to be similar to those seen in the CC, and a functional role for NO in the control of smooth muscle function in the CS was proposed.11 This is in agreement with the present findings, which demonstrate similar NO-mediated relaxant effects in the human CS and CC, and further support a role for NO in the functional control of CS smooth muscle. The endothelium may provide a complementary source of NO during penile erection. ACh-induced relaxation of erectile tissue has been shown to involve the release of NO from endothelial cells.1 No immunoreactivity for eNOS could be demonstrated in endothelial cells of the sinusoidal endothelium of the rat CC.5, 19, 22 Furthermore, carbachol induced only a small relaxant effect in NA-contracted rat CC preparations,19 which suggests that NO released from the endothelium lining the cavernous sinuses is not of major importance for erection in the rat. In contrast, the present study showed that the sinusoidal spaces in both CC and CS from humans were lined by a thin, and sometimes disrupted layer
CHOLINERGIC NERVES CONTAIN NO AND HO
873
FIG. 3. A, confocal microscopy. Corpus cavernosum with terminals showing Tyrosine Hydroxylase (TH)-IR (red) and Nitric Oxide Synthase (NOS)-IR (green) varicosities, which do not coincide. Bar ⫽ 10 m. B, corpus cavernosum with Vesicular Acetylcholine Transporter (VAChT)-IR terminals along smooth muscle bundles. Texas Red immunofluorescence. C, same section as in B. NOS-IR terminals with similar profiles as VAChT-IR terminals in A. Fluorescein 5⬘-IsoThioCyanate (FITC) immunofluorescence. Bars ⫽ 100 m.
FIG. 4. Endothelial cells along sinusoid in corpus cavernosum expressing A, endothelial Nitric Oxide Synthase (NOS)-immunoreactivity and B, Hemoxygenase-1 (HO-1)-immunoreactivity (B). Bars ⫽ 30 m (A) and 50 m (B).
of eNOS-IR endothelial cells. This is also in agreement with the functional findings, showing that carbachol had a good relaxant effect in the preparations, and supports the view that NO produced by eNOS in the sinusoids contributes to erection. CO has been shown to produce relaxation of smooth muscle by stimulating guanylate cyclase,23 and both CO and NO have been implicated as messenger molecules in the central
and peripheral nervous system.24 The present study revealed that relaxations produced by CO were similar in appearance in NA-contracted preparations of human CC and CS. In comparison with the effects obtained with NO in the same tissues, the maximal CO-induced relaxation was about 40% lower than that produced by NO. The tissue-levels of cyclic nucleotides were unaffected by CO, suggesting that other mechanisms than stimulation of guanylate cyclase (or ade-
874
CHOLINERGIC NERVES CONTAIN NO AND HO
of exogenously applied CO, suggest that CO may be involved as a messenger molecule in the regulation of smooth muscle tone in the penis. Conclusion. This study demonstrates that the human CC and CS receive a dense cholinergic innervation. VAChT, NOS and VIP are found in the same nerve terminals within the human CC and CS, suggesting that these nerves comprise a distinct population of parasympathetic, cholinergic nerves. In both CC and CS, HO-can be demonstrated in nerves and in endothelium, and exogenous CO produces relaxation. It seems likely that both the NOS/NO and CO/HO-systems have functional roles in the control of smooth muscle tone of CC and CS. We thank Dr. E. Weihe and Dr. P. Emson for a generous supply of antibodies, and Dr. Anders Håkansson for expert help with the confocal microscopy. REFERENCES
FIG. 5. Addition of NO (squares) or CO (circles) to noradrenaline (l-NA)-contracted strips produced concentration-dependent relaxation in both corpus cavernosum (CC: filled symbols) and corpus spongiosum (CS: open symbols).
FIG. 6. In noradrenaline (l-NA)-contracted preparations, electrical field stimulation produced frequency-dependent relaxation in both corpus cavernosum (CC: open squares) and corpus spongiosum (CS: open circles) preparations, which was effectively suppressed by pretreatment with NG-nitro-L-arginine (10⫺4 M) in both CC (filled squares) and CS (filled circles).
nylate cyclase) are responsible for the relaxant activity of CO. Involvement of K⫹ channels in the action of CO on smooth muscle preparations from the gastrointestinal tract has been proposed.25, 26 If this is the case also in human penile erectile tissues cannot be decided on the basis of the present results. In CC tissue from the rabbit, Kim et al27 found zinc deuteroporphyrin to be without effect on electrically-induced inhibitory responses, and suggested that CO was not involved in neuronally mediated relaxations of cavernous erectile tissue. However, non-specific effects, other than inhibition of HO, have been described for the metalloporphyrins,28 which questions the use of these compounds for the study of the activity of the CO/HO-system. The presently found immunoreactivity for HO isozymes in nerve structures and endothelium of human CC and CS, and the functional effects
1. Andersson, K-E. and Wagner, G.: Physiology of penile erection. Physiol Rev, 75: 191, 1995 2. Lincoln, J. and Burnstock, G.: Autonomic innervation of the urinary bladder and urethra. In: The Autonomic Nervous System. Vol. 6, Chapter 8, Nervous Control of the Urogenital System, edited by C. A. Maggi. Harwood Academic Publishers: London, p. 33, 1993 3. Weihe, E., Tao-Cheng, J-H., Scha¨fer, M. K-H., Erickson, J. D. and Eiden, L. E.: Visualization of the vesicular acetylcholine transporter in cholinergic nerve terminals and its targeting to a specific population of small synaptic vesicles. Proc Natl Acad Sci USA, 93: 3547, 1996 4. Arvidsson, U., Reidl, M., Elde, R. and Meister, B.: Vesicular acetylcholine transporter (VAChT) protein: A novel and unique marker for cholinergic neurons in the central and peripheral nervous systems. J Comp Neurol, 378: 454, 1997 5. Dail, W. G., Barba, V., Leyba, L. and Galindo, R.: Neuronal and endothelial nitric oxide synthase activity in rat penile erectile tissue. Cell Tissue Res, 282: 109, 1995 6. Cook, M. N., Nakatsu, K., Marks, G. S., McLaughlin, B. E., Vreman, H. J., Stevensen, D. K. and Brien, J. F.: Heme oxygenase activity in the adult rat aorta and liver as measured by carbon monoxide formation. Can J Physiol Pharmacol, 73: 515, 1995 7. Morita, T., Perella, M. A., Lee, M-E. and Kourembanas, S.: Smooth muscle cell derived carbon monoxide is a regulator of vascular cGMP. Proc Natl Acad Sci USA, 92: 1475, 1995 8. Ny, L., Alm, P., Ekstro¨m, P., Larsson, B., Grundemar, L. and Andersson, K-E.: Localization and activity of haem oxygenase and functional effects of carbon monoxide in the feline lower oesophageal sphincter. Br J Pharmacol, 118: 392, 1996 9. Zakhary, R., Gaine, S. P., Dinerman, J. L., Ruat, M., Flavahan, N. A. and Snyder, S. H.: Heme oxygenase-2: Endothelial and neuronal localization and role in endothelium-dependent relaxation. Proc Natl Acad Sci USA, 93: 795, 1996 10. Rattan, S. and Chadker, S.: Inhibitory effect of CO on internal anal sphincter: heme oxygenase inhibitor inhibits NANC relaxation. Am J Physiol, 265: G799, 1993 11. Hedlund, P., Larsson, B., Alm, P. and Andersson K-E.: Distribution and function of nitric oxide-containing nerves in canine corpus cavernosum and spongiosum. Acta Physiol Scand, 155: 445, 1995 12. Wessendorf, M. W. and Elde, R. P.: Characterization of an immunofluorescence technique for the demonstration of coexisting neurotransmitters within nerve fibres and terminals. J Histochem Cytochem, 33: 984, 1985 13. Scha¨fer, K-H., Weihe, E., Erickson, J. D. and Eiden, L. E.: Human and monkey cholinergic neurons visualized in paraffin-embedded tissues by immunoreactivity for VAChT, the vesicular acetylcholine transporter. J Mol Neurosci, 6: 225, 1996 14. Herbison, A. E., Simonian, S. X., Norris, P. J. and Emson, P. C.: Relationship of neuronal nitric oxide synthase immunoreactivity to GnRH neurons in the ovariectomized and intact female rat. J Neuroendocrinol, 8: 73, 1996 15. Lundberg, L. M., Alm, P., Wharton, J. and Polak, J. M.: Protein gene product 9.5 (PGP 9.5) A new neuronal marker visualizing
CHOLINERGIC NERVES CONTAIN NO AND HO
16.
17.
18.
19. 20.
21.
the whole uterine innervation and pregnancy-induced and developmental changes in the guinea-pig. Histochemistry, 90: 9, 1988 Gulbenkian, S., Wharton, J. and Polak, J. M.: The visualization of cardiovascular innervation in the guinea pig using an antiserum to protein gene product 9.5 (PGP 9.5). J Auton Nerv Syst, 18: 235, 1987 Ny, L., Alm, P., Ekstro¨m, P., Hannibal, J., Larsson, B. and Andersson, K-E.: Nitric oxide synthase-containing, peptidecontaining, and acetylcholinesterase-positive nerves in the cat lower oesophagus. Histochem J, 26: 721, 1994 Hedlund, P., Alm, P., Ekstro¨m, P., Fahrenkrug, J., Hannibal, J., Hedlund, H., Larsson, B. and Andersson, K-E.: Pituitary adenylate cyclase-activating polypeptide, helospectin, and vasoactive intestinal polypeptide in the human corpus cavernosum. Br J Pharmacol, 116: 2258, 1995 Hedlund, P., Alm, P. and Andersson, K-E.: NO synthase in cholinergic nerves and NO-induced relaxation in the rat isolated corpus cavernosum. Br J Pharmacol, 127: 349, 1999 Jen, P. Y. P., Dixon, J. S., Gearhart, J. P. and Gosling, J. A.: Nitric oxide synthase and tyrosine hydroxylase are colocalized in nerves supplying the postnatal human genitourinary organs. J Urol, 155: 1117, 1996 Tamura, M., Kagawa, S., Kimura, K., Kawanishi, Y., Tsuruo, Y. and Ishimura, K.: Coexistence of nitric oxide synthase, tyrosin hydroxylase and vasoactive intestinal polypeptide in human
22.
23. 24. 25.
26.
27.
28.
875
penile tissue - a triple histochemical and immunohistochemical study. J Urol, 153: 530, 1995 Schirar, A., Chang, C. and Rousseau, J. P.: Localization of androgen receptor in nitric oxide synthase- and vasoactive intestinal peptide-containing neurons of the major pelivic ganglion innervating the rat penis. J Neuroendocrinol, 9: 141, 1997 Schmidt, H. H. H. W.: NO, CO and OH. Endogenous soluble guanylyl cyclase-activating factors. FEBS Lett, 307: 102, 1992 Maines, M. D.: The heme oxygenase system; a regulator of second messenger gases. Ann Rev Pharmacol Toxicol, 37: 517, 1997 Farrugia, G., Irons, W. A., Rae, J. L., Sarr, M. G. and Szurszewski, J. H.: Activation of whole cell currents in isolated human jejunal circular smooth muscle cells by carbon monoxide. Am J Physiol, 264: G1184, 1993 Werkstro¨m, W., Ny, L., Persson, K. and Andersson, K-E.: Carbon monoxide induced relaxation and distribution of haem oxygenase in the pig urethra and oesophagogastric junction. Br J Pharmacol, 120: 312, 1997 Kim, Y. C., Davies, M. G., Marson, L., Hagen, P-O. and Carson, C. C.: Lack of effect of carbon monoxide inhibitor on relaxation induced by electrical field stimulation in corpus cavernosum. Urol Res, 22: 291, 1994 Grundemar, L. and Ny, L.: Pitfalls using metalloporphyrins in carbon monoxide research. Trends Pharmacol Sci, 18: 193, 1997