Soluble Guanylate Cyclase and Cgmp-Dependent Protein Kinase I Expression in the Human Corpus Cavernosum

Soluble guanylate cyclase and cGMP-dependent protein kinase I expression in the human corpus cavernosum

T Klotz1*, W Bloch2, J Zimmermann1, P Ruth3, U Engelmann1 and K Addicks2

1Department of Urology, University of Cologne; 2Institute I of Anatomy, University of Cologne; and 3Institute of Pharmacology, TU University of Munich, Germany

Nitric oxide (NO) as a mediator in smooth muscle cells causes rapid and robust increases in cGMP levels. The cGMP-dependent protein kinase I has emerged as an important signal transduction mediator for smooth muscle relaxation. The purpose of this study was to examine the existence and distribution of two key enzymes of the NO=cGMP pathway, the cGMP-dependent kinase I (cGK I) and the soluble guanylate cyclase (sGC) in human cavernosal tissue. The expression of the enzymes were examined in corpus cavernosum specimens of 23 patients. Eleven potent patients suffered from penile deviations and were treated via Nesbit's surgical method. Nine long-term impotent patients underwent implantation of ¯exible hydraulic prothesis. Three potent patients underwent trans-sexual operations. Expression of the sGC and cGK I were examined immunohistochemically using speci®c antibodies. In all specimens of cavernosal tissue a distinct immunoreactivity was observed in different parts and structures. We found a high expression of sGC and cGK I in smooth muscle cells of vessels and in the ®bromuscular stroma. The endothelium of the cavernosal sinus, of the cavernosal arteries, and the cavernosal nerve ®bers showed an immunoreactivity against sGC. The distribution analysis of cGK I revealed a predominately vesicular localization in smooth muscle cells. The examination of the endothelium showed no clear immunoreactivity against cGK I. There was no distinct difference in immunoreactivity and cellular distribution between potent and impotent patients. cGMP-dependent kinase I is predominately expressed in cavernosal smooth muscle cells. sGC is expressed in smooth muscle cells, endothelium and nerve ®bers of the corpus cavernosum. In this study, there is no correlation between the distribution of soluble guanylate cyclase and cGMP dependent kinase I in potent or impotent patients. International Journal of Impotence Research (2000) 12, 157±164.

Keywords: corpus cavernosum; cGMP-dependent kinase I; impotence; soluble guanylate cyclase

Introduction regulated phosphodiesterases, cGMP-gated ion channels and cGMP dependent protein kinases.4 cGMP-dependent protein kinases have emerged as 0 Cyclic guanosine 5 -monophosphate (cGMP) medi- important signal transduction mediators and may ates as a second messenger the physiological action regulate the function of numerous proteins.5±7 of nitric oxide (NO), which is known to be one of the Recently it has been shown that activation of main mediators of smooth muscle relaxation in the cGMP-dependent kinase I relaxes smooth muscle corpus cavernosum.1,2 NO, which can diffuse freely cells by decreasing the cytosolic calcium concentra- across cell membranes, acts on the soluble isoform tion.6,7 Because of the fact that smooth muscle of guanylate cyclase (sGC) binding to the heme relaxation is essential for an erection, it is necessary moiety of the protein and thereby activating this to investigate the cGMP-dependent protein kinases enzyme.3 The resultant increase of cGMP is thought in cavernosal tissue. Two distinct types of cGMP- to underlie the physiological action of NO eg in dependent protein kinases have been identi®ed. The smooth muscle cells. There appear to be three main type I kinase (cGK I) occurs in two isoforms, arising classes of intracellular targets for cGMP, cGMP- from alternative mRNA splicing.4,7,8 cGK I is present at high levels in smooth muscle, lung and platelets.4 A second form, type II cGMP-dependent kinase (cGK II) was cloned from mouse brain and rat *Correspondence: T Klotz, Department of Urology, University of Cologne, Joseph-Stelzmannstr. 9, D-50924 KoÈln, Germany. intestine and is expressed in widespread areas of the 9±12 Received 8 September 1999; revised 22 October 1999; brain. Recently we have shown that the endo- accepted 14 January 2000 thelial NO-synthase is one of the main sources of sGC and cGMP-dependent protein kinase I expression TKlotzet al 158 13 nitric oxide in the cavernosal tissue. Nitrinergic 3% H2O2 and 60% methanol PBS for 30 min, then innervation and eNOS expression have shown a permeabilized with 0.2% Triton-X 100 in 0.1 M PBS. broad heterogenicity and no correlation between The sections were then treated with 5% normal goat NOS expression and erectile function was observed. serum (NGS) and 5% bovine serum albumin (BSA) The cGMP kinase I lies `downstream' of the NO solution on PBS. Before each step, the sections were mediated effects, thus it may be possible that there rinsed three times in PBS buffer. Incubation with the are differences between potent and impotent pati- primary antibody occured in a PBS-based solution ents. A loss of cGK I abolishes nitric oxide=cGMP of 0.8% BSA and 20 mM NaN3 for 12 h at 4 C. The dependent relaxation of smooth muscle, at least in anti-sGC antibody was applied in a dilution of the vascular and intestinal system.14 Up to now, 1:1000 and the anti-cGMP-kinase I antibody at cavernosal tissue has not been examined for the 1:800. After rinsing with PBS the sections were cGMP dependent kinase I enzyme. The aim of this incubated with the corresponding secondary bio- study was to investigate the distribution of the tinylated antibody for 1 h at room temperature. A guanylate cylcase and the cGMP dependent protein streptavidin-horseradish peroxidase complex was kinase I in human corpus cavernosum. In this then applied as a detection system (1:200 dilution) context, we examined cavernosal tissues from for 1 h. Finally the staining was developed for 3 ± potent and impotent patients in order to elucidate 5 min with 3,3-diaminobenzidine tetrahydrochlor- possible differences between these groups. ide (DAB) in 0.05 M Tris-HCI buffer and 0.1% H2O2. Counterstaining was performed using methyl green. Negative control sections were incubated without Patients and methods the primary antibody. Patients and collection of tissue Immun¯uorescence Twenty-three human corpus cavernosum tissue specimens were obtained with informed consent Sections of 20 mm, pretreated with BSA and NGS from patients subjected to penile surgery. The were incubated with primary antibody (PECAM specimens were immediately ®xed in 4% parafor- 1:300 or undiluted a-actin smooth muscle solution maldehyde for 4 h and then rinsed in 0.1 M supplied by the manufacturer) and sGC (1:1000) or phosphate-buffered saline (PBS) for 24 h. The tissue cGMP-kinase I (1:800) antibody over night at 4C. was stored for 12 consecutive hours in a PBS After washing three times in PBS the slices were solution containing 18% sucrose for cryoprotection incubated at room temperature for 1 h with anti and then frozen to 780C. mouse Cy2 (1:100) and biotinylated anti rabbit Eleven of the patients (age range 16 ± 63 y; mean (1:400) antibody. Following washing with PBS the age 42 y) with normal erectile functions suffered slices were treated with extravidin Cy3. After from penile deviations. All patients were treated via dehydration in a series of graded ethanol the Nesbit's surgical method. Normal erectile function specimens were embedded in entelan. Analysis was ascertained by anamnestic evaluation. was performed using a Zeiss Axiophot ¯uorescence Nine patients (age range 27 ± 66 y; mean age 46 y) microscope. had shown complete erectile dysfunction for more than 2 y. This group underwent implantation of ¯exible hydraulic penile protheses. One of the Material patients suffered from Klinefelter's syndrome, four patients from severe venous leakage, one patient from impotence due to radical surgery of a bladder The primary anti-GC antibody were from Alexis carcinoma. Two patients had diabetes and a com- (San Diego, CA, USA). The monoclonal primary bined arterial-venous insuf®ciency of the corpus mouse antibodies PECAM-1 and a-actin smooth carvernosum. In one patient, the impotence was of muscle were from Biogenex (San Ramon, CA, unknown origin. USA). The polyclonal cGMP-kinase I antibody was Three patients underwent trans-sexual operations obtained from Professor Peter Ruth from the In- (male ± female). These patients had a long-term stitute of Pharmocology in Munich. The antibody hormonal pretreatment, but reported normal production method was as described previously.4 erections. The secondary biotinylated goat anti-rabbit and goat anti-mouse antibodies were from Vector Labora- Immunohistochemistry tories (Burlingame, USA), mouse Cy2 (Dianova, Hamburg, Germany), biotinylated rabbit anti-mouse (Dako, Denmark), extravidin Cy3 (Sigma, Deisenho- Prior to immunohistochemical examination tissue fen, Germany). BSA was purchased from Sigma sections (20 mm) were placed in a bathing solution of (Deisenhofen, Germany), which also provided the

International Journal of Impotence Research sGC and cGMP-dependent protein kinase I expression T Klotz et al 159 normal goat serum as well as the necessary chemi- and tunica albuginea vessels, without a speci®c cals required for staining with the streptavidin- subcellular localization (Figure 2a ± f). The endo- biotin-peroxidase complex. thelium of the cavernosal sinus and of cavernosal arteries and veins also showed an immunoreactivity against sGC (Figure 2c). The cavernosal tissue Results revealed a known dense nerve network in the ®bromuscular stroma surrounding cavernosal ves- In all specimens of cavernosal tissue, a distinct sels.10 These nerve ®bers were partially immuno- immunoreactivity was observed in different parts positive against sGC (Figure 2b). This observation and structures. We found a high expression of demonstrates the localization of sGC not only in soluble guanylate cyclase and cGMP-kinase I in smooth muscle cells, but also in the endothelium smooth muscle cells of the corpus carvernosum and nerve ®bers (Figure 2a ± d). (Figure 1a ± d). All specimens from impotent pati- The immunohistochemical detection of cGK I in ents with different primary diseases and trans- smooth muscle cells shows a distinct pattern (Figure sexual patients showed an expression of sGC and 3a ± d). A vesicular-like distribution of cGK I in the cGMP-kinase I. There was no distinct difference in smooth muscle cells was found. No immunostaining immunoreactivity and distribution between potent against cGK I was found in the endothelium of and impotent patients (Figure 1a ± d). corpus cavernosum veins and in the endothelium of Concerning the soluble guanylate cylase (sGC) we tunica albuginea vessels. The smooth muscle layer could describe a distinct pattern of distribution in of vessels is cGK I positive (Figure 3c, d). the cavernosal tissue. sGC was predominately found Immunostaining using a polyclonal cGMP-depen- in smooth muscle cells of the ®bromuscular stroma dent protein kinase I antibody revealed that caver- and the vascular smooth muscle layer of cavernosal nosal smooth muscle cells and vascular smooth

Figure 1 A representative comparison of the soluble guanylate-cyclase (sGC) (a, b) and cGMP-kinase I (c, d) expression in corpus cavernosum of patients with (b, d) and without (a, c) erectile dysfunction. At this magni®cation smooth muscle (SM) and venous sinus (S) are recognizable. A distinct staining of sGC and cGK I are observable in the smooth muscle of potent (a,c) and impotent (b, d) patients without recognizable differences. (Cause of impotence in this sections was venous leakage.) Bar a, b 150 mm, c, d 80 mm.

International Journal of Impotence Research sGC and cGMP-dependent protein kinase I expression TKlotzet al 160

Figure 2 Immunohistochemical detection of soluble guanylate cyclase (sGC) of human corpus cavernosum tissue (a ± f) from potent patients: The smooth muscles are distinct immunostained (a). Nerve ®bers (arrow heads) innervating the ®bromuscular stroma are positive for sGC (b). The endothelium (arrows) of corpus cavernosum veins also contains sGC (c). At higher magni®cation a sGC expression in the endothelial cells (arrows) of the sinus cavernosum is demonstrated (e). A sGC immunoreactivity is also found in endothelial cells and smooth muscle of tunica albuginea vessels (V) (d). Without the ®rst antibody no immunoreaction is found (f) Bar a±d 50 mm, e 15 mm, f 75 mm.

muscle layer contained this enzyme (Figures 1 and Discussion 4). We found a double labeling of sGC and PECAM (Figure 4b, d). In the double immunostaining of the Recently we have shown that the endothelial NO- human corpus cavernosum with antibodies for sGC synthase is one of the important sources of NO in or cGK I and PECAM or a-smooth muscle actin. cavernosal smooth muscle cells.13,15 Furthermore Furthermore we can also reveal a double labeling we have demonstrated that erectile dysfunction with cGK I and a-smooth muscle actin (Figure 4e, f). cannot solely be reduced to pathological ®ndings

International Journal of Impotence Research sGC and cGMP-dependent protein kinase I expression T Klotz et al 161

Figure 3 Distribution of cGMP-kinase I in human corpus cavernosum: the immunohistochemical detection of cGMP-kinase I in smooth muscle cells shows a distinct pattern. In cross section of the smooth muscle cells, a vesicular like distribution (arrow heads) of cGMP- kinase I is found (a). In longitudinal sectioned smooth muscle, a cross striation of the immunoreaction products of cGMP-kinase I staining is seen (b). No immunostaining is found in the endothelium (arrows) of tunica albuginea vessels (c), while the smooth muscle layer (SM) of the vessel (V) is cGMP-kinase I positive (c, d). Bar a ± d 25 mm.

in penile NO-synthases, because there has been no phorylation is responsible for smooth relaxation and correlation between the expression of the NO- reduction in vascular tone (Figure 4).17 Smooth synthases and the erectile function in cavernosal muscle relaxation is one of the most important tissue of potent and impotent patients. NO has the effects of the NO=cGMP pathway concerning erectile capacity to activate guanylate cyclase. Submicro- function. In intact smooth muscle cells, cGMP molar concentrations of NO cause rapid and robust dependent kinase I phosphorylates the inositol 16 17 increases in cGMP levels in target cells. Changes of 1,4,5-trisphosphate receptor (IP3R). This phos- the cellular cGMP level as a second messenger lead phorylation may result either in inhibition of Ca2 to physiological responses in these target cells. release or in increased compartmentalization. Thus There are some targets of cGMP action such as cGMP-dependent kinase I is one of most relevant cGMP-gated channels, cGMP-regulated posphodies- trigger enzymes for smooth muscle relaxation. For terases and cGMP-dependent protein kinases. Many the understanding of the complete signal transduc- of these responses are mediated by the activation tion in the corpus cavernosum it is necessary to of the cGMP-dependent kinase I.5,14,17 The cGMP- elucidate the whole NO=cGMP pathway with its dependent protein kinase I has emerged as an `up-and downstream' target proteins. important signal transduction mediator.4,18,19 Sub- To our knowledge this is the ®rst study that sequent cGMP-mediated phosphorylation of pro- deals with the expression of the guanlyate cyclase teins appears to be involved in intracellular Ca2 (sGC) and GMP-dependent kinase I (cGK I) in human sequestration or removal. The removal of cytosolic corpus cavernosum. sGC was found in smooth muscle Ca2 in response to cGMP mediated protein phos- cells without speci®c subcellular localization.

International Journal of Impotence Research sGC and cGMP-dependent protein kinase I expression TKlotzet al 162

Figure 4 Double immunostaining of human corpus cavernosum with antibodies for sGC or cGMP-kinase I and PECAM or a-smooth muscle actin: (a, b) The sGC-immun¯ourescence (b) is more expanded as the immuno¯uorescence for a-smooth muscle actin (a). The comparison of immuno¯uorescence labeling with the endothelial speci®c marker PECAM and the sGC labeling shows a double labeling of sGC and PECAM (arrows). The ¯uorescence staining for cGMP-kinase I (f) is restricted to the a-smooth muscle actin positive areas (e) Bar 25 mm.

The endothelium of cavernosal sinus, of cavernosal lium speci®c PECAM. Concerning the cGMP- arteries and cavernosal nerves also showed an dependent kinase I we found that cavernosal immunoreactivity. This observation demonstrates smooth muscle cells contain this enzyme. This the localization of sGC not only in smooth muscle results were also con®rmed by double immuno- cells, but also in the endothelium. The results were staining with speci®c a-smooth muscle actin anti- con®rmed by double immunostaining with endothe- body.

International Journal of Impotence Research sGC and cGMP-dependent protein kinase I expression T Klotz et al 163 The distribution analysis of cGK I showed a References vesicular localization in smooth muscle cells. In our study, the examination of the endothelium and 1 Burnett AL et al. Nitric oxide: a physiologic mediator of the nerve ®bers revealed no clear expression of cGK I, penile erection. Science 1992; 257: 401 ± 403. although cGK I expression was found in the nerve 2 Rajfer J et al. Nitric oxide as a mediator of relaxation of the ®bers and endothelium of other organs.20,21 Further- corpus cavernosum in response to nonadrenergic, noncholi- more we have not found in our study any distinct nergic neurotransmission. New Engl J Med 1992; 326: 90 ± 94. differences between potent and impotent patients 3 Ignarro LJ. Heme-dependent activation of guanylate cyclase by nitric oxide: a novel signal transduction mechanism. Blood in regards to the distribution of sGC and cGMP Vessels 1991; 28: 67 ± 73. dependent kinase I. We have to stress the limita- 4 Pfeifer A et al. Structure and function of cGMP-dependent tions of our study regarding the small number of protein kinases. Rev Physiol Biochem Pharmacol 1999; 135: patients which did not allow a statistical analysis. 105 ± 149. 5 Lincoln TM, Corbin JD. Characterization and biological role of For example, immunohistochemistry is not a quan- the cGMP-dependent protein kinase. Adv Cyclic Nucleotide titative technique and another assay such as Res 1983; 15: 139 ± 192. immuno-blotting with more patients might reveal 6 Cornwell TL, Lincoln TM. Regulation of intracellular quantitiative changes in the enzymes between potent Ca2 levels in cultured vascular smooth muscle cells. Reduc- and impotent patients. Thus our results have a tion of Ca2 by atriopeptin and 8-bromo-cyclic GMP is mediated by cyclic GMP-dependent protein kinase. J Biol premature character regarding differences between Chem 1989; 264: 1146 ± 1155. potent and impotent patients. 7 Felbel J et al. Regulation of cytosolic calcium by cAMP and As for the NO-synthases erectile dysfunction cGMP in freshly isolated smooth muscle cells from bovine cannot be reduced to pathological ®ndings in sGC trachea. J Biol Chem 1988; 263: 16764 ± 16771. 13,15 8 Wernet W, Flockerzi V, Hofmann F. The cDNA of the two and cGK I. It should be stressed that our study isoforms of bovine cGMP-dependent protein kinase. FEBS Lett shows only the expression or distribution of the 1989; 251: 191 ± 196. enzymes but not the functional status. In the case 9 Uhler MD. Cloning and expression of a novel cyclic GMP- of the cGMP-dependent kinases, this problem is dependent protein kinase from mouse brain. J Biol Chem 1993; alleviated by recent developments for the analysis of 268: 13586 ± 13591. 23 10 Jarchau T et al. Cloning, expression, and in situ localization of the function of cGMP dependent kinases. rat intestinal cGMP-dependent protein kinase II. Proc Natl Because of the selective expression of cGK I in Acad Sci USA 1994; 91: 9426 ± 9430. smooth muscle cells of the corpus cavernosum, it 11 EI-Husseini AE et al. Nitric oxide regulates cylic GMP- seems possible that cGMP kinase could be thera- dependent protein kinase phosphorylation in rat brain. J Neurochem 1998; 71: 676 ± 683. peutically activated or inhibited by speci®c sub- 12 Sandberg M et al. Molecular cloning and predicted full-length 23,24 stances. Gene knockout approaches may amino acid sequence of the type Ib isoenzyme of cGMP- promise new insights into major functions con- dependent protein kinase from human placental. FEBS Lett. trolled by cGK I and justify a more detailed 1989; 255: 321 ± 329. investigation into the underlying mechanisms. The 13 Bloch W et al. Evidence for the involvement of endothelial nitric oxide synthase from smooth muscle cells in the erectile function challenge remains to de®ne the speci®c effects of of the human corpus cavrnosum. Urol Res 1998; 26: 129 ± 135. agonists and cGMP that are dependent on cGK I as a 14 Pfeifer A et al. Defective smooth muscle regulation in cGMP key enzyme for smooth muscle relaxation. kinase I-de®cient mice. EMBO J 1998; 17: 3045 ± 3051. 15 Klotz T et al. Die verschiedenen quellen der stickstoffoxid- freisetzung im menschlichen corpus cavemosum. Akt Urol 1998; 29: 246 ± 251. 16 Kim YC et al. Modulation of vasoactive intestinal polypeptide Conclusions (VIP)-mediated relaxation by nitric oxide and prostanoids in the rabbit corpus cavernosum. J Urol 1995; 153: 807 ± 810. 17 Komalavilas P, Boerth NJ, MacMillan-Crow LA, Cornwell TL. This study indicates that cGK I is present in the cGMP signaling through cAMP-and cGMP-dependent protein corpus cavernosum. It remains to be determined kinases. Adv Pharmacol 1995; 34: 305 ± 322. whether the enzyme is important for erectile 18 Lohmann SM et al. Distinct and speci®c functions of cGMP- dependent protein kinases. Trends Biochem Sci 1997; 22: dysfunction. The distribution of the cGK I correlates 307 ± 312. with the distribution of the soluble guanylate 19 MoÈnks D et al. Expression of cGMP-dependent protein kinase cylcase (sGC) in smooth muscle cells. cGK I is I and its substrate VASP in neointimal cells of the injured rat a key enzyme for smooth muscle relaxation and carotid artery. Eurp J Clin Invest 1998; 28: 416 ± 423. 20 Draijer R et al. Expression of cGMP-dependent protein kinase I the development of speci®c agonists may have and phosphorylation of its substrate, vasodilator-stimulated therapeutical consequences. phosphoprotein, in human endothelial cells of different origin. Circ Res 1995; 77: 897 ± 905. 21 Huber A et al. Protein kinase G expression in the small intestine and functional importance for smooth muscle relaxation. Am J Physiol 1998; 275: G629 ± G637. Acknowledgements 22 Ebel J et al. Neuronal colocalization of neuropeptide Y, nitric oxide synthase, tyrosine hydroxylase and vasoactive intestinal peptide and morphometry of innervation density of the corpus This study was supported by a ®nancial grant of cavernosum penis from patients with erectile dyfunction. Cell P®zer GmbH Karlsruhe, Germany. Vision 1997; 4: 271 ± 279.

International Journal of Impotence Research sGC and cGMP-dependent protein kinase I expression TKlotzet al 164 23 Smolenski A et al. Functional analysis of cGMP-dependent 24 Lohse JL, FoÈrstermann U, Schmidt HHHW. Pharmacology of protein kinases I and II as mediators of NO=cGMP effects. NO: cGMP signal transduction. Naunyn Schmiedebergs Arch Naunyn Schmiedebergs Arch Pharmacol 1998; 358: 134 ± 139. Pharmacol 1998; 358: 111 ± 112.

International Journal of Impotence Research