Penicillamine administration reverses the inhibitory effect of hyperhomocysteinaemia on endothelium-dependent relaxation in the corpus cavernosum in the rabbit


Anthony J. Koupparis, Urology, Bristol Royal Infirmary, Bristol, UK. e-mail:



To elucidate the role of copper in mediating the impact of homocysteine on vasculogenic erectile dysfunction (VED), by investigating the effect of dietary supplementation with the copper-chelator penicillamine to rabbits rendered hyperhomocysteinaemic (HHC) with a methionine-rich diet, as a raised plasma level of homocysteine might be a risk factor for VED.


Homocysteine inhibits the nitric oxide (NO)-dependent relaxation of the corpus cavernosum (CC), an effect which appears to be mediated via the generation of superoxide (O2·-), and H2O2. Copper is a catalyst for the generation of H2O2 in the presence of homocysteine and in the presence of copper, H2O2 undergoes reactions resulting in the generation of O2·-, which reacts with NO to produce peroxynitrite (ONOO-), thereby reducing the bioavailability of NO and impairing NO-mediated relaxation of CC. Smooth muscle strips from CC were obtained from two groups of adult New Zealand White rabbits, one rendered HHC with a diet supplemented with methionine (group 1) and another HHC group that had additional dietary supplementation with penicillamine (group 2). Tissue O2·- levels were measured in each group. After pre-contraction with phenylephrine, relaxation responses of CC strips to carbachol were also assessed in both groups.


Methionine supplementation led to profound HHC in all rabbits. Penicillamine in group 2 reduced the total plasma Cu2+ compared to group 1. There was a markedly lower carbachol-stimulated relaxation of CC from HHC rabbits in group 1, with a mean (sem) maximum relaxation of 37 (4)% (six samples), than in group 2, at 58 (6)%.


These data show that elevated levels in vivo of homocysteine in the rabbit markedly impair NO-dependent relaxation of the CC. Furthermore, this effect appears to be augmented by copper. Further clinical studies on homocysteine and copper status in patients with VED are warranted.


corpus cavernosum


nitric oxide


nonadrenergic, noncholinergic


(vasculogenic) erectile dysfunction


cardiovascular disease




superoxide dismutase


Krebs’ Ringer bicarbonate buffer


Penile erection is initiated by relaxation of the corpus cavernosum (CC), which is mediated by an increase in nitric oxide (NO) formation, released principally by nonadrenergic, noncholinergic nerve (NANC) fibres [1] and the intracavernosal endothelium [2]. NO relaxes cavernosal smooth muscle by activating guanylyl cyclase, which generates cGMP [3]; this then activates protein kinase G, which suppresses calcium mobilization, resulting in an erectile response [2,3]. Therefore, any process that reduces NO formation impairs normal erectile function.

Major risk factors for vasculogenic erectile dysfunction (VED) are diabetes mellitus, smoking and dyslipidaemia [4–6]. These risk factors are associated with an increase in superoxide (O2·-) formation in the vasculature [5–8]; this entity reacts with NO to form peroxynitrite (ONOO-) and other reactive nitrogen species [9]. As a result, the bioavailability of NO is decreased, thereby impairing NO-dependent smooth muscle relaxation and the normal erectile response.

An elevated plasma level of homocysteine is also an independent risk factor for cardiovascular disease (CVD) [10]. In a recent study, we showed that the induction of hyperhomocysteinaemia (HHC) with a methionine-rich diet caused a marked reduction of NO-dependent relaxation (an effect mediated by an increase in O2·- formation) consolidating the view that homocysteine might also be a risk factor for ED [11]. It was advocated that homocysteine-mediated vasculopathy is principally through the auto-oxidation of homocysteine to O2·-[12]. However, homocysteine alone at concentrations found in mild HHC had no effect on in vitro NO-dependent relaxation of the rat aorta [13] or the rabbit CC [14]. By contrast, in the presence of physiological levels of copper, homocysteine inhibited acetylcholine- and carbachol-stimulated relaxation and cGMP formation in these tissues [12,13]. These effects were reversed by superoxide dismutase (SOD), suggesting that copper might augment the generation of O2·- from homocysteine. It was therefore proposed that the relative concentrations of copper and homocysteine, rather than homocysteine alone, determine the vasculopathic impact of HHC [12,13]. To test this possibility in vivo, the effect of the administration of a copper chelator, penicillamine [14–16], on erectile function was investigated in the methionine-fed rabbit model.


New Zealand White rabbits (3 kg initial weight) were fed a standard rabbit food (SDS Ltd, UK) supplemented with 20 g methionine/kg food, the metabolic precursor for homocysteine. Rabbits were allowed free access to the food and water. Two groups of rabbits were assessed; group 1 had the methionine-rich diet and group 2 were also given the copper chelator penicillamine (10 mg/kg/day) in their drinking water. The dose of penicillamine used for this study was equivalent to that used to treat Wilson’s disease, systemic sclerosis, cysteine stones and intractable rheumatoid arthritis [14–16]. Blood samples were taken at weekly intervals by venepuncture of the ear vein, to measure total plasma homocysteine and copper, as previously described [13].

After a month, the rabbits were killed by an i.v. injection with barbiturate; the penises were excised and placed in cold oxygenated Krebs’ Ringer bicarbonate buffer (KRB). Epidermal tissue was removed, the tunica albuginea opened and the CC dissected out, cut into strips of ≈ 1 × 3 mm and placed in oxygenated KRB. The size and weights of the tissues were the same in both groups of rabbits. Strips were placed in oxygenated KRB, and mounted vertically in 1.5 mL capacity organ baths, equipped with two parallel platinum electrodes containing KRB maintained at 37 °C by a thermoregulated circuit. The KRB solution was bubbled with a mixture of 95% O2/5% CO2 maintained at pH 7.4. An initial tension of 2 g was applied to the suspended tissue strips. The tension was recorded on a polygraph (model 7D, Grass Inst., Quincy, MA, USA). All strips were equilibrated for ≥ 30 min, after which they were challenged with KCl (124 mm). Two reproducible contractions varying in magnitude by <10% were consistently obtained. After washout and equilibration, tissues were pre-contracted with phenylephrine and relaxed with cumulative doses of the acetylcholine analogue, carbachol (0.01–10 µm).

To measure tissue O2·-, after dissection the cavernosal segments were transferred to individual wells of a 24-well plate. After transfer, tissue segments were equilibrated in Dulbecco’s modified Eagle’s medium with no phenol red for 10 min at 37 °C in a 95% air-5% CO2 incubator (Heraeus, Hera Cell, Kandro Laboratory Products, Germany). Then 20 µm of horseradish cytochrome c (Sigma Chemical Co., Poole, UK) with or without 500 U/mL of copper-zinc SOD (Sigma) was added and incubated at 37 °C in a 95% air-5% CO2 incubator for 1 h. The reaction medium was removed and the reduction of cytochrome c determined at 550 nm in a spectrometer (Anthos Lucy 1 Laboratory-tech Int., Ringmer, East Sussex, UK) and converted to nmoles of O2·-, using ÄE550 nm = 21.1 mm−−1 as the extinction coefficient. The reduction of cytochrome c that was inhibitable with SOD reflected actual O2·- release. Segments were rinsed in PBS and weighed. The results were expressed as mol of O2·-/mg tissue.

The results are expressed as the mean (sem) of six samples, and analysed using anova for multiple comparisons. Paired comparisons between the groups were assessed using the paired Student’s t-test where anova indicated significance for the multiple comparisons. Statistical significance was accepted when P < 0.01


The methionine-rich diet of 20 g/kg food led to profound HHC in all rabbits; plasma homocysteine levels were significantly higher at 1 and 2 weeks after dietary supplementation with methionine. At 4 weeks the mean plasma levels of homocysteine were 140 (19) mmol/L compared to levels before dosing of 13.1 (1.2) mmol/L (Fig. 1).

Figure 1.

Plasma homocysteine concentrations in rabbits fed a diet of 20 g methionine/kg food over 1 month. Each point is the mean (sem) of nine samples; *P < 0.001 vs pre-dose levels.

Penicillamine had no effect on homocysteine levels but at 4 weeks significantly reduced the total plasma copper, at 6.3 (0.5) µmol/L, compared to group 1, at 9.4 (0.99) µmol/L (Fig. 2). The release of O2·- was significantly greater in group 1, at 0.16 (0.01) nmol/mg tissue/h, than in group 2, at 0.07 (0.01) nmol/mg tissue/h (Fig. 3).

Figure 2.

Plasma copper levels in all rabbits. Each point is the mean (sem) of six samples; *P < 0.01 vs group 1.

Figure 3.

Superoxide release from HHC (group 1) cavernosal segments and from group 2 (penicillamine) segments. Each point is the mean (sem) of six samples; *P < 0.01 vs HHC cavernosal segments.

There was a marked and statistically significantly lower carbachol-stimulated relaxation of the CC from the rabbits in group 1, with a maximum relaxation of 37 (4)%, than in group 2, at 58 (6)% (Fig. 4a). SOD and catalase had no significant effect on the relaxation responses in group 2 (Fig. 4b).

Figure 4.

Concentration-response curves for: a, endothelium dependent carbachol-mediated relaxation of CC strips from group 1 and 2 (P < 0.01 vs group 2); b, carbachol-mediated relaxation of CC in group 2 rabbits, in the presence of SOD (900 U/mL) and catalase (300 U/mL). Each point is the mean (sem) of six samples.


In the present study, the administration of the copper chelator penicillamine reversed the inhibitory effect of HHC on NO-mediated relaxation of the CC. At 1 month after giving penicillamine, when the analysis was undertaken, the concentrations of plasma copper had fallen to almost half of those in group 1, suggesting that this effect might be mediated by a reduction of plasma copper. As relaxation of the CC in response to sodium nitroprusside (that activates guanylyl cyclase) was unaffected by HHC, this indicates that the present effects are not mediated through inhibition of this enzyme. We also previously showed that HHC has no effect on relaxation of the CC when elicited with electrical field stimulation, which promotes relaxation of the CC through the activation of NANC fibres that release dilator peptides, and as such was not tested in this study.

We showed previously that HHC impairs erectile function through the negation of NO release from endothelial NO synthase by the endogenous overproduction of O2·-[11]. By contrast, homocysteine alone is relatively ineffective at inhibiting endothelium-dependent relaxation of the rabbit CC [13]. However, copper appears to augment the inhibitory effect of homocysteine on endothelium-dependent relaxation of the CC by augmenting the formation of O2·-[13]. It was proposed therefore that the possible erectopathic effect of homocysteine in vivo might be mediated by an interaction with endogenous copper [13]. In turn, it was suggested that HHC and copper, both independent risk factors for CVD [10,17], might also be risk factors for ED [11,13]. Further evidence for this is provided by the present study.

Mechanistically, copper augments the formation of O2·- through transition metal interactions with the thiol moiety of homocysteine [18–20]. Other studies showed that copper and homocysteine interact to promote a vasculopathic aetiology, including the oxidation of low-density lipoprotein and impairment of NO-mediated arterial relaxation. Notably, those studies were ex vivo and homocysteine and copper were not present, indicating that the HHC elicits an up-regulation of pro-oxidative enzymes, such as NADPH oxidase. Indeed, we showed that risk factors for ED up-regulate the expression of NADPH oxidase in cavernosal tissue and suggested that this is axiomatic in ED [21–23]. Under normal circumstances, copper is tightly bound to plasma proteins, in particular ceruloplasmin [24]. Several of the copper forms bound to ceruloplasmin are exchangeable (chelatable) which is crucial, as copper is transported by ceruloplasmin and taken up by tissues through a complex chaperone system [25]. However, not only products of oxidative stress, including O2·-, ONOO- and hydrogen peroxide, but also homocysteine and other thiols, dissociate copper from its protein binding sites on ceruloplasmin and fragment the protein [26–34], which might augment the pro-oxidant effects of copper. Such events could explain why a copper chelator prevents the impairment of NO-dependent relaxation, as reported here. Copper bound to ceruloplasmin (chelatable copper) can also catalyse oxidative reactions [35]. It was also shown that copper, at submicromolar concentrations, markedly augments the oxidative destruction of NO in aortae from diabetic but not control rats [36].

In conclusion, HHC appears to elicit an in vivo oxidative impact on erectile function through a copper-mediated mechanism that in turn promotes endogenous production of O2·- and therefore negation of NO-mediated cavernosal relaxation. To investigate this area further, in vivo studies examining the biological effect of different doses of penicillamine might be useful. From a therapeutic perspective, penicillamine is used to treat Wilson’s disease, systemic sclerosis, cysteine stones and intractable rheumatoid arthritis [14–16], but due to side-effects, it is unlikely that penicillamine use would be advisable in patients with ED. Nevertheless, another copper chelator, thiomolybdate, has been effective in reducing inflammation and purportedly has fewer side-effects [37]. An alternative strategy is to reduce plasma homocysteine levels; they can be readily reduced by giving folic acid [38]. We found that giving folic acid to diabetic rabbits reverses the inhibitory effect of experimental diabetes mellitus on cavernosal relaxation (unpublished data). As it was shown that this regimen improves endothelium-dependent relaxation in patients with HHC, it is reasonable to suggest that it would improve erectile function in patients with ED. More clinical studies are required to test this proposal.


None declared. Source of funding: Shackman Trust.