Corporal reactivity to adenosine and prostaglandin E1 in alloxan-induced diabetic rabbit corpus cavernosum, and the effect of insulin therapy


Professor Dr Y. Sarioglu Cumhuriyet Üniversitesi Tip Fakültesi, Farmakoloji Anabilim Dali, 58140 Sivas, Turkey.



To investigate the changes in corporal reactivity to adenosine and prostaglandin E1 (PGE1 ) in corpus cavernosal strips from alloxan-induced diabetic rabbits and to determine the effects of insulin therapy.

Materials and methods

Twenty-four New Zealand white rabbits were studied in three equal groups: group 1, control; group 2, diabetes induced by the administration of alloxan hydrochloride intravenously; group 3, as group 2 but with insulin administered after the induction of diabetes. At the end of 8 weeks, the reactivity of corpus cavernosal strips from the animals was assessed in organ chambers.


The relaxation responses of corpus cavernosal strips to adenosine were similar in all groups, but the response to PGE1 was impaired in groups 2 and 3 compared with that in controls.


If vasoactive drugs are to be used for the diagnosis or treatment of diabetic erectile impotence, direct-acting vasodilators, e.g. adenosine, must be used. In alloxan-induced diabetes, the corporal reactivity to PGE1 was impaired and insulin therapy did not restore the relaxation responses.


Relaxation of corporal smooth muscle, with expansion of the lacunar spaces, is a prerequisite for penile erection. The relaxation reduces the cavernosal venous outflow by compressing venules under the tunica albuginea, the surrounding fibrous structure. Therefore, the tone of smooth muscle regulates the corporal veno-occlusive mechanism, which plays a key role in the development of erection [1[2]–3].

It has been known for more than a decade that acetylcholine and several other agents can release relaxing factors from the vascular endothelium [4]. One of these factors has been identified as nitric oxide (NO), synthesized from l-arginine [5]. Recent studies have shown that both acetylcholine and neuronally mediated relaxation in rabbit and human corpus cavernosum tissue involves the release of NO or an NO-like substance [6[7][8][9][10]–11].

Diabetes mellitus is a contributing cause of erectile impotence and the prevalence of impotence in diabetic men has been estimated to be 35–75% [12,13]. There is controversy about the aetiology of impotence in diabetic men, but recent studies suggest that diabetes is associated with the development of autonomic neuropathy and endothelial dysfunction in blood vessels. Moreover, studies in diabetic men and animal models of diabetes have shown that diabetic impotence is associated with impaired neurogenic and endothelium-mediated relaxation of trabecular smooth muscle. This impairment in relaxation responses in both mechanisms of diabetes appears to involve an alteration in cGMP/NO pathway [14[15][16]–17].

Adenosine is a metabolite of the putative neurotransmitter ATP and has a potent vasodilatory effect directly on various peripheral vascular beds [18]. Thus adenosine may be useful in the treatment of erectile impotence, because the cavernosal sinus is one such vascular bed. Adenosine has a vasodilatory role in rabbit femoral artery, in canine carotid artery and in canine penile erectile tissue [19,20]; Takahashi et al. [21] showed that adenosine also produced a full erection when administered intracavernosally in dogs. The vasodilatory effect of adenosine on vascular beds is not mediated by endothelium-derived relaxing factor [22,23], whereas adenine nucleotides (ATP and ADP) exert endothelium-dependent relaxation in various vascular smooth muscle [24,25]. It was suggested that adenosine and ATP have a potent relaxant activity on the corpus cavernosum through a mechanism different from the NO/cGMP pathway, and that purinergic transmission may be an important component involved in the initiation and maintenance of penile erection [26,27].

Since the intracavernosal injection of PGE1 was reported to produce penile erection, it has been the subject of clinical trials for the management of impotence [28]. Human penile trabecular tissue has the ability to synthesize various prostanoids [29[30]–31]. It has been suggested that arachidonic acid cascade products may be involved in the control of penile erection and PG dehydrogenase activity has been detected in human corpus cavernosum, which may be why PGE1 when given intracavernosally seldom causes prolonged erection and priapism [31,32].

The present study assessed the effects of 8 weeks of alloxan-induced diabetes on the corporal reactivity to PGE1 and adenosine in a rabbit model. A further aim was to determine the effects of insulin therapy in corporal reactivity to these substances and thus help elucidate the pathophysiology of diabetes.

Materials and methods

Induction of diabetes

Litter-mate and weight-matched New Zealand white rabbits were anaesthetized by an intramuscular injection with xylazine (5 mg/kg) and ketamine (20 mg/kg). Sixteen rabbits in which diabetes mellitus was induced by administering alloxan (150 mg/kg intravenously) formed the experimental groups. The criterion for the induction of diabetes was a blood glucose concentration of ≥ 2.5 g/L at the end of first week of alloxan exposure. Animals were re-awakened by the intravenous administration of yohimbine (0.2 mg/kg). Animals then received 10% glucose in drinking water for the first 12 h, then were fed regular rabbit chow and water for 8 weeks. Eight normal rabbits received no treatment and were maintained as litter-mate controls for 8 weeks. The 16 diabetic rabbits were divided into two groups; eight rabbits remained diabetic and the other eight were treated with 6 U/day NPH insulin, injected subcutaneously.

Preparation and assessment of corpus cavernosal tissue

After 8 weeks, the rabbits were killed by intravenous injection with pentobarbital (50 mg/kg); they were exsanguinated and the penis removed en bloc. A ventral incision was made in the right and left corpora, the tunica dissected and the cavernosal tissue exposed. The corporal tissue was immediately placed into an organ chamber after dissection.

Corpus cavernosal strips (≈2×2×15 mm) were assessed in 20 mL organ chambers for isometric tension measurement. The strips were tied with silk to a force transducer (FT 03, Grass Instruments, Quincy, MA, USA) on one end and fixed with silk ties to a glass support on the other. The organ chambers contained physiological saline solution composed of (mmol/L): NaCl 118, KCl 4.7, CaCl2 2.5, NaHCO3 25, MgSO4 1.2, KHPO4 1.2, glucose 11. The solution was gassed with 95% O2 and 5% CO2 during the study. The pH of the solution was 7.4 and the temperature was maintained at 37°C. After mounting, the preparations were allowed to equilibrate for 2 h, during which the tension adjusted to 2 g.

After equilibration, relaxation was assessed against a background of pre-contraction with 10 μmol/L phenylephrine, and adjusted to 70–80% of the maximal contraction. At the plateau of contraction, relaxation responses to cumulative concentrations of PGE1 and adenosine were evaluated, with the changes in isometric tension recorded on a Grass model 79E polygraph. All corpus cavernosal strips isolated from control and experimental rabbits were contracted with 124 mmol/L KCl and relaxed with papaverine hydrochloride (100 μmol/L) after pre-contraction with phenylephrine to test the soundness of trabecular smooth muscle. All drugs were dissolved in distilled water.

The relaxant effects of PGE1 and adenosine were expressed as a percentage of the pre-contraction with phenylephrine. To evaluate the effects of agonists the maximum response (Emax ) and pD2 values (i.e. the negative logarithm of the concentration for the half-maximal response, ED50 ) were calculated as predicted from the Scatchard equation for drug–receptor interaction. Agonist pD2 values (apparent agonist affinity constants) were calculated from each agonist concentration–response curve by linear regression, as a measure of the sensitivity of the tissues to each agonist. All data are expressed as the mean (sem ). Groups were compared statistically using general linear models of anova followed by Scheffe’s F-test. Values of P<0.05 were considered to indicate statistical significance.


The body weights, blood glucose and insulin concentrations of all groups at the end of 8 weeks are listed in Table 1. The differences in body weights, blood glucose and insulin concentrations in group 2 were significantly different from those of the control group (P<0.05), but the values were similar in groups 1 and 3 (P>0.05).

Table 1.  Characteristics of the animals, and maximum relaxation response (Emax ) and pD2 values after exposure to agonists in strips of corpus cavernosal tissue obtained from the three groups of eight rabbits. Group 1, control; group 2, diabetic; group 3, insulin-treated diabetic Thumbnail image of

In isolated corpus cavernosal strips pre-contracted with phenylephrine (10 μmol/L) the administration of adenosine (1–300 μmol/L) gave concentration-dependent relaxation responses (Fig. 1a). There was no significant difference among the groups for these responses (P>0.05) and between the Emax and pD2 values (Table 1). The administration of PGE1 (1–30 μmol/L) also gave concentration-dependent relaxation responses (Fig. 1b). These responses were significantly impaired in groups 2 and 3 compared with controls (P<0.05), but there was no difference between groups 2 and 3 (P>0.05). The Emax and pD2 values were significantly different in groups 2 and 3 compared with controls (Table 1). The administration of papaverine (100 μmol/L) caused relaxation but there was no significant difference among the groups (P>0.05), and none between the Emax and pD2 values (Table 1). When tissues were contracted with 124 mmol/L KCl, similar tensions were achieved in all groups.

Figure 1.

Concentration–response curves of relaxation in response to (a) adenosine and (b) PGE1 in isolated strips of rabbit corpus cavernosum. Each point is the mean (sem ) percentage of the contraction induced by 10 μmol/L phenylephrine. Eight preparations were used in each case. Group 1 (control), dark green; group 2 (diabetic), light green; group 3 (treated diabetic), light red.


Alloxan-induced diabetes produced a significant decrease in rabbit body weight and serum insulin concentration, and an increase in serum glucose concentration; these were prevented by insulin treatment. It has been established that a relatively brief period of alloxan-induced diabetes can induce changes in the physiological mechanisms that modulate trabecular smooth muscle in rabbits [7,16]. The similarity of these changes to those in tissues from diabetic impotent men make the alloxan-induced diabetic rabbit a suitable model to investigate the effects of diabetes on penile smooth muscle.

Hedlund and Andersson showed that PGF , PGI2 , high concentrations of PGE2 and TX A2 analogues contracted the corpus cavernosum [33]. They found that PGE1 effectively relaxed human trabecular tissue and cavernosal artery segments contracted by noradrenaline and PGF . PGE1 has been shown to inhibit the release of noradrenaline from penile adrenergic nerves and to increase intracellular cAMP levels in corporal smooth muscle; this phenomenon is believed to be essential in the mechanism of erection [3,34].

Studies of blood vessels suggest that diabetes-related impaired relaxation may be a result of increased formation of vasoconstrictor prostanoids [35]. It has been shown that elevated glucose increases the production of arachidonic acid metabolites during stimulation with acetylcholine [36]. Azodzoi et al. [7] reported that the neurogenic and endothelium-dependent relaxation of corporal tissue from diabetic and control groups, after treatment with indomethacin, was significantly greater than before treatment. Moreover, they observed that despite tissue treatment with indomethacin, neurogenic and endothelium-dependent relaxation of the diabetic group was depressed compared with controls, and they concluded that there was no role for prostanoids in mediating impaired relaxation in diabetics.

In the present study, the relaxation reaction to PGE1 was impaired in isolated corpus cavernosal strips from diabetic rabbits. The administration of insulin to alloxan-induced diabetic rabbits did not improve corporal reactivity to PGE1in vitro. On the other hand, the adenosine-induced relaxation responses were similar in all isolated corpus cavernosal strips. Adenosine has been considered a neuromodulator in penile erection because of its erectile effect after intracavernosal administration to dogs [21]. In previous studies it has been shown that adenosine has a potent relaxant effect in various vascular structures and this effect was independent of endothelium [19,21,22]; adenosine was suggested as an ideal intracavernosal agent for the diagnosis and treatment of impotence, as it could repeatedly produce a full erection for a reasonable period with no side-effects [19]. Mantelli et al. [27] reported that adenosine has potent relaxant activity on the corpus cavernosum, acting through a different mechanism from the NO pathway, and that the receptors involved in the effect of adenosine are of the A2a subtype. However, Chiang et al. [37] reported that adenosine receptors involved in the modulation of neurotransmission in rabbit corpus cavernosum appear to be the A2b subtype. They also reported that adenosine induced an increase in human cavernosal arterial velocity and resistive index, measured by colour duplex ultrasonography, and the combination of adenosine and 10 μg PGE1 was more effective in increasing the resistive index and erection grade than was 20 μg PGE1 alone. They concluded that adenosine seems to be an important neuromodulator for penile erection and could be an alternative combination in the treatment of impotence.

There are no published results of the intracavernosal administration of PGE1 in a large series of men with diabetic erectile impotence. In these patients it has been reported that tumescence occurred with intracavernosal administration, but not after administering PGE1 via the internal pudendal artery [38,39]. Ishii et al. [38] concluded that vascular damage was the main cause of the poor response to PGE1 injection and vascular disturbance was clinically doubtful in patients with diabetes mellitus in whom no erection occurred.

In conclusion, if vasoactive drugs are to be used for the diagnosis or treatment of diabetic erectile impotence, direct-acting vasodilators, e.g. adenosine, must be used; additionally, it is necessary to combine the results obtained from large series with in vitro corporal reactivity responses to clarify the effect of PGE1 treatment for diabetic erectile impotence.


This study was supported by a grant from Cumhuriyet University Research Fund, Sivas.