To examine the effects of androgens on erectile response and the expression of nitric oxide synthase (NOS) isoform mRNAs in the penile corpus cavernosum of castrated rats.
To examine the effects of androgens on erectile response and the expression of nitric oxide synthase (NOS) isoform mRNAs in the penile corpus cavernosum of castrated rats.
The study comprised 50 adult male Sprague-Dawley rats in five groups: sham controls; castrated; castrated and receiving testosterone; castrated and receiving dihydrotestosterone (DHT); castrated and receiving testosterone and 5α-reductase inhibitor (finasteride). Androgen replacements were administered via implants of silicone tubing. After 7 days, some animals underwent electrical stimulation of the cavernosal nerves and the remainder were used for further analysis. NOS activity was measured in the soluble fraction of the corpus cavernosum, using the Griess reaction. Total RNA was isolated and nNOS and eNOS mRNA expression examined using semiquantitative reverse-transcriptase polymerase chain reaction.
Castration caused a marked decrease in erectile response and the ratio of maximal intracavernosal pressure (ICPmax) to systemic blood pressure (SBP), although both testosterone and DHT effectively restored the response to normal. NOS activity and the amount of nNOS mRNA were reduced in castrated rats but restored by androgen replacement. Although there was no significant difference in NOS activity between the androgens, nNOS mRNA expression was higher in rats treated with DHT. There were no effects of androgen in rats treated with finasteride, as the ICPmax/SBP ratio, NOS activity and amount of nNOS mRNA decreased. eNOS mRNA expression was independent of androgen.
Androgens enhance nNOS gene expression in the penile corpus cavernosum of rats, suggesting that they play an important role in maintaining NOS activity. Of the two androgens, DHT was more potent.
Although it is generally accepted that erectile response in mammals is regulated by androgens, and androgen receptors have been identified in cavernosal tissue [1,2], the extent of the involvement and precise role of these steroids remains to be established. To date, a variety of animal models, including man, has been used to elucidate the role of androgens in erectile response. It has been shown that castration results in a diminished frequency and duration of penile erection [3–5], a significant decrease in the magnitude of intracavernosal pressure  and a considerable reduction in the erectile response to electrical field stimulation (ES) .
Nitric oxide (NO), a potent vasodilator, has recently been shown to be responsible for the relaxation of smooth muscle cells in the penile corpus cavernosum [8,9]. Activities of NO-producing enzymes, NO synthases (NOSs), have been identified in the rat penile cytosol fraction. Penile NOS activity has been shown to be influenced by castration, which decreased NOS activity of penile tissue; activity was later restored by testosterone replacement to normal levels . Further studies have confirmed that as the administration of l-arginine failed to augment erectile response while supplementation of testosterone in castrated rats resulted in increased nNOS mRNA synthesis, enzyme activity is dependent on the availability of androgens rather than of substrates . It can therefore be postulated that androgens play an essential role in the production and maintenance of erectile response via modulation of NOS activity. However, there is no clear evidence of the extent to which androgens control the NO-dependent erectile mechanism. It is still not clear whether androgens exert their effects on a specific NO-production system by activating a specific NOS isoform gene(s).
Thus the purpose of this study was to investigate the effects of androgens on NOS activity of the penile corpus cavernosum in castrated rats under varying androgenic hormone status. To verify the androgenic effect on NOS gene activation, transcriptional changes in NOS isoforms were also examined. A further goal was to determine whether these effects of testosterone are mediated via its conversion into DHT.
The study comprised 50 adult male Sprague-Dawley rats (250–300 g body weight), maintained under controlled lighting and treated according to local regulations. The rats were castrated under ketamine anaesthesia (35 mg/kg intraperitoneal). The animals were divided into five experimental groups of five each: group 1, intact (sham-operated) controls; group 2, castrated; group 3, castrated with testosterone supplementation via 2.2-cm Silastic™ tubing (Dow Corning, Midland, MI, USA; outside diameter 3.17 mm; inside diameter 1.57 mm); group 4, castrated, with DHT supplementation via 1-cm tubing; and group 5, castrated with testosterone supplementation and daily administration of finasteride (40 mg/kg body weight) orally. Animals in groups 1 and 2 were implanted with empty tubing. After treatment for 7 days, the rats were assessed to determine circulating androgen levels by radioimmunoassay, using blood collected from the heart.
Rats were anaesthetized as described and the erectile response measured as previously described . Briefly, after surgical exposure of the hypogastric and pelvic nerve, the major pelvic ganglion and the cavernosal nerves, the last were stimulated with a glass electric stimulator and a bipolar platinum electrode (1–3 V, 20 Hz for 5 ms). During stimulation, erections were evaluated on the basis of maximal intracavernosal pressure (ICPmax). After the ischiocavernosal muscle was divided, the silver-white tunica albuginea of the penile crus was exposed. A 22 G butterfly needle was inserted into the penile crus and connected to the monitor. For the duration of the experiments, the animals were re-injected every 45 min with 35 mg/kg ketamine. Systemic blood pressure (SBP) was simultaneously monitored by inserting a polyethylene tube (PE-50) into the right femoral artery.
To determine whether androgens modulate the ability of the penis to synthesize NO, NOS activity was measured in the corpus cavernosal tissues. The activity in tissue homogenates was determined in penises removed from animals not subjected to other experiments, as previously described [10,12]. Briefly, rats were anaesthetized and the penile shaft and bulb (excluding skin and glans) excised and stored at −80°C. Each penis was weighed and homogenates prepared in six volumes of cold medium containing 0.32 mol/L sucrose, 20 mmol/L HEPES (pH 7.2), 0.5 mmol/L EDTA, 1 mmol/L dithiothreitol and protease inhibitors (3 μmol/L leupeptin and 1 mmol/L phenylmethylsulphonylfluoride), using a Polytron homogenizer. The cytosolic and particulate fractions were separated by centrifugation at 12 500 g for 60 min. To the cytosol fraction (200 μL), 10 μL each of 2 mmol/L arginine and 2 mmol/L NADPH were added. After incubating the mixture for 1 h at 37°C, 200 μL of Griess reagent (1% sulphanylamide, 0.1% naphthylethylene diamine and 2.5% H3PO4 ) was added and incubated further for 10 min at room temperature. The nitrate concentration was measured by comparison with a standard curve, using linear gradient concentrations of sodium nitrate. To normalize interassay variations of enzyme specific activity against the control value for each series, a complete set of penile cytosol samples per selected group was used for each determination of citrulline synthesis. Specific activity was calculated as μmol/10 min per gram of tissue.
To measure the expression of NOS genes, total RNA was isolated using TRIzol (Gibco BRL, Gathisberg, MD, USA). The reverse transcription reaction was carried out using a commercially available kit (Promega, Madison, WI, USA). Briefly, 1.0 μg of total RNA was primed with 0.5 μg of the oligo (dT)15 and incubated for 30 min at 42°C. For each sample, 1×RT buffer (10 mmol/L Tris-HCl, pH 8.8, 50 mmol/L KCl, and 0.1% Triton X-100), 1.0 mmol/L dNTPs, 5 mmol/L MgCl2 , and 300 units of M-MLV RT in a volume of 20 μL were combined. Oligonucleotide primers for nNOS had the following sequences: sense primer, 5′-ACCTGAAGAGCACACTGGAAAC-3′ and antisense primer, 5′-GATGGCCGACCTGAGATTC-3′; these amplified a 428 bp fragment for the nNOS sequence . Nucleotide sequences of primers for eNOS were: sense primer, 5′-CAGGCTGCCTGTGAAACTTT-3′ and antisense primer, 5′-TTGCTGCTCTGTAGGTTCTC-3′ . For an internal standard, cyclophilin mRNA was also amplified. The primers used for cyclophilin were 5′-TGTTCTTCGACATCACGGC-3′ (sense) and 5′-TTATGGCGTGTGAAGTCACC-3′ (antisense), amplifying a transcript of 216 bp. Semiquantitative RT-PCR of nNOS gene involved the ‘primer dropping’ method ; 4 μL of the cDNA product was used and the reaction was carried out in a final volume of 20 μL containing 2.5 mmol/L MgCl2 , 1×PCR buffer (20 mmol/L Tris-Cl, 50 mmol/L KCl, pH 8.4, Triton X-100), 2.5 units of Taq DNA polymerase, 0.2 mmol/L dNTPs and 25 pmol/L of each primer. A PTC-200 thermal cycler (MJ Research, Inc, Watertown, MA, USA) completed eight cycles consisting of heat denaturation at 94°C for 45 s, annealing at 54.5°C for 45 s and extension at 72°C for 1 min. After eight cycles the primer for cyclophilin (25 pmol/L concentration of sense and antisense) was added and the thermal cycler completed 24 additional cycles under the same conditions. Because there were large differences in the values of the melting temperatures of the primers, it was not possible to co-amplify eNOS and cyclophilin cDNAs. After adding 4 μL of cDNA products to each tube, separate PCR reactions were therefore carried out. The PCR condition of eNOS was: 94°C for 45 s, 51°C for 45 s and 72°C for 1 min, total 30 cycles. Amplification of cyclophilin was completed during 24 cycles under the same conditions. In preliminary experiments of up to 30 and 24 cycles, the PCR and product amplification of eNOS and cyclophilin, respectively, was linear. Using a 2% agarose gel, a 10 μL aliquot of each reaction product was subjected to electrophoresis followed by staining with ethidium bromide. The intensities of ethidium bromide fluorescence were determined using a CCD-camera (Fujifilm, Japan). The density of nNOS and eNOS were normalized to that of cyclophilin and the values expressed as arbitrary units.
Experimental values were expressed as the mean (sem); Student's t-test and the Wilcoxon rank-sum test were used to calculate probabilities when comparing two groups independently from the others, and P<0.05 considered to indicate significance. For comparing all groups with the controls, a one-way anova was computed.
Castration caused a marked decrease in the circulating plasma levels of androgens; in control rats (group 1) the testosterone level was 1.58 (0.50) ng/mL and the DHT level was 61.7 (16.0) pg/mL. Castrated rats (group 2) had an undetectable level of testosterone and a lower DHT level, at 15.8 (6.6) pg/mL, but levels were higher, at 1.43 (0.36) ng/mL and 58.1 (26.9) pg/mL of testosterone and DHT respectively, in rats in group 3. The DHT level in rats in group 5, at 17.9 (5.7) pg/mL, was not significantly different from those of rats in group 2 and finasteride thus effectively inhibited 5α-reductase activity.
To confirm the previously reported effects of androgens on erectile response, ES of the major pelvic ganglion, at the voltage producing maximum pressure, was administered to animals on completion of treatment. The effects of castration and androgen supplementation on the erectile response to ES were expressed as the mean (sem) of the ratio of ICPmax to SBP. Table 1 shows that the erectile response was significantly less in groups 2 and 4 than that in group 1. Castration reduced the ICPmax/SBP ratio by 40% (P<0.005), although both testosterone and DHT replacement restored the magnitude of response to control levels. The response of group 5 determined whether testosterone alone induced a normal erectile response; because there was a 29% decrease in group 5 (P=0.012), it is clear that the effect of testosterone was attenuated by finasteride treatment.
The NOS activity results are presented in Table 1, expressed as a percentage of the normal controls; there was a close relationship between ES response and NOS activity for each of the groups assayed. Castration caused a significant decrease in penile NOS activity, although this was restored by androgen treatment to levels which were not significantly different from the control value. Although both testosterone and DHT did not completely restore NOS activity, the levels reached did not differ significantly, although they were significantly higher than that of group 2. However, the administration of finasteride to rats receiving testosterone resulted in a reduction of NOS activity by 55%.
The effect of androgens on NOS activity was assessed by determining the regulation of NOS mRNA expression in the penis using the RT-PCR method with isoform-specific primers, measuring the availability of nNOS and eNOS mRNAs. Figure 1 shows the results of electrophoretic analysis of the RT-PCR products of nNOS mRNA. The expected size of the PCR products of nNOS (428 bp) and cyclophilin (216 bp) were obtained and the results analysed further by computing the ratio of nNOS to cyclophilin peak areas for each sample (Table 1). Castration caused an 80% reduction in the amount of nNOS mRNA, compared with that in control rats. The amount of nNOS mRNA in cavernosal tissue was significantly greater in rats in group 1 and 5 than in those that were castrated. While DHT replacement restored levels to control values, testosterone was less effective, as nNOS mRNA expression was only 38% of the control value. The administration of finasteride to rats receiving testosterone significantly decreased nNOS mRNA expression. Figure 1B shows the RT-PCR products of eNOS mRNA, with an expected size of 310 bp. The results also show that there was no significant difference in relative expression levels between castrated and androgen-replaced rats (Table 1). As there was no significant difference between castrated and control animals, it is also evident that castration did not alter eNOS mRNA expression.
Despite androgen receptor levels decreasing rapidly in the penis after puberty , the effect of androgen on normal penile erection is supported by several lines of experimental results. The present results confirm observations by others [3,9,16] showing that castration in rats decreases erectile function and androgens restore it. These results showed that androgens modulate smooth muscle relaxation in the penile corporal cavernosum; the ICPmax/SBP ratio was reduced in castrated rats but restored to normal levels by androgen treatment (Table 1). Similar results have previously been reported by others; maximal cavernosal pressure is markedly reduced after castration but returns to normal levels after treatment with testosterone . It has also been shown that androgens act to maintain both the inflow and outflow of blood from cavernosal spaces during erection . In castrated rats with spinal-cord transection, erectile response (ex copula) declines but is restored by testosterone replacement . In the same animals, oestrogen fails to restore penile responsiveness  but restores copulatory behaviour . As reports from this and other laboratories have shown, the extent to which intracavernosal pressure rises during erection depends on the availability of functional androgens.
Convincing evidence has shown that erection in the rat is androgen-dependent and mediated by NO [10,20,21]; androgens may act primarily to stimulate or maintain the activity of NOS for the production of erectile response [3,7,22,23]. NOS activity has been measured as the conversion of arginine to citrulline and was shown to be lower in castrated rats than in those which were intact or supplemented with testosterone . When adult rats are infused with an inhibitor of NO synthesis, cavernosal pressure decreases in both castrated and testosterone-replaced animals . In the present study, total NOS activity decreased in castrated rats, while androgens restored enzyme activity to control levels (Table 1). These results correlate with those relating with the responsiveness to ES (the ICPmax/SBP ratio; Table 1). Both responses indicate that with reconstitution of the androgen milieu, erectile response returns to near-normal values and NOS activity is restored. From an overall perspective it may be suggested that the role of androgens in the production of penile erection is modulated through NOS activity. Modulation of NOS by androgens has also been shown by histological examination showing that the density of NOS-positive nerve fibres decreased in castrated animals but recovered to normal levels when rats were treated with androgens .
It has previously been suggested that in the production of erectile response in castrated animals, DHT is a more potent androgen . The present results also showed that as the 5α-reductase inhibitor (finasteride) administered to rats receiving testosterone resulted in decreased NOS activity, the conversion of testosterone to DHT may be required to exert their effects.
Although cDNA probes for NOS isozymes are currently available, few reports have described the detection of NOS mRNA in the penis. Furthermore, it is still unclear whether the effect of androgens on NOS expression is isoform-specific. As penile erection is dependent on NO [7,15,21], it is not surprising that castration has been shown to decrease NOS activity. However, there are conflicting reports as to whether this reduction is also accompanied by a decrease in nNOS content, as determined by immunoblot detection . On the other hand, it has been proposed that castration inhibits penile NOS activity without affecting NOS levels .
The effects of androgens on penile NOS mRNA in castrated rats appear to be NOS-isoform specific. The present results showed that castration decreased nNOS mRNA expression, whereas androgens reversed it (Table 1). Similar results, that testosterone treatment of castrated animals results in a marked increase of nNOS gene expression, have been reported previously . It is therefore assumed that androgens increase nNOS mRNA levels by enhancing nNOS gene expression or by decreasing mRNA degradation, resulting in more available enzyme for the production of NO in rat penis. The present finding that nNOS mRNA expression is less influenced by testosterone than is DHT also confirms the previous suggestion that DHT induces a greater erectile response .
The present results showed a discrepancy between NOS activity and the amount of nNOS mRNA in rats which received testosterone; the possibility that nNOS expression is regulated at the translational level, as postulated previously by Chamness et al. , may account for this finding. It is also likely that activity of the translational machinery may be limited to the use of only a small portion of nNOS transcripts. Another possibility is that androgenic effects are not of primary importance for the expression of NOS mRNAs. However, these effects on erectile response appear to be species-dependent. It has been postulated that for the maintenance of the erectile mechanism, only a minute level of androgens might be required, and that in men, this level is not affected by physiological reductions in serum androgens . In fact, the administration of finasteride to normal men does not significantly impair erectile function . On the other hand, the blockage of testosterone action by a 5α-reductase inhibitor in castrated rats markedly reduces both erectile response and nNOS expression to castrate levels.
There has been no attempt to identify a specific NOS isoform in penile cytosolic fraction; in fact, cytosolic enzyme activity represents most of the penile NOS [7,15]. The present results show that the amount of eNOS mRNA was not significantly affected by androgens. An oestrogen response element has previously been found within the promoter region of bovine eNOS gene , suggesting that oestrogen may be a physiological regulator of eNOS expression. However, as there was no significant difference between finasteride-and androgen-treated rats, oestrogenic effects were not apparent in the present model. Although oestrogen can be produced in castrated rats by androgen supplementation, the amount may be negligible and insufficient to induce eNOS gene activation.
In conclusion, it is apparent that androgens are indispensable to normal erectile response in rats, although their effects seem to depend partly on the NOS-mediated erectile mechanism. In humans, the most common causes of impotence are ageing and diabetes, which are frequently accompanied by a moderate reduction of free serum testosterone [3,10]. A reduction in erectile capacity and penile NOS activity has been found in both senescent  and congenitally diabetic rats . Both clinical situations therefore suggest an essential role for androgens in the production of normal erection, although it is possible that other hormones also contribute to maintenance of the erectile mechanism and may affect nNOS content in the penis. This is supported by recent findings showing a significant relationship between the adrenal gland and penile erection .