Anatomy, Physiology, and Pathophysiology of Erectile Dysfunction

Dopamine

Central dopaminergic neurons project to the MPOA and the paraventricular nucleus [17]. Furthermore, dopaminergic neurons have been identified that travel from the caudal hypothalamus to innervate the autonomic and somatic nuclei in the lumbosacral spinal cord [18,19]. Thus, dopamine can be expected to participate in the regulation of both the autonomic and somatic components of the penile reflexes.

Both the two major families of dopamine receptors, D1-like (D1, D5) and D2-like (D2, D3, D4) receptors [20], have been associated with central erectile functions; however, the D2-like receptor subtype seems to have the predominating effect. The nonselective dopamine receptor agonist, apomorphine, when administered systemically to male rats, was found to induce penile erection [21], simultaneously producing yawning and seminal emission. Similarly, low-dose systemic administration of other dopamine agonists initiates erection [1]. The effects of these agonists can be attenuated by centrally but not peripherally acting dopamine receptor antagonists.

Oxytocin

Oxytocinergic spinal projections from the supraoptic and paraventricular nuclei of the hypothalamus are likely to influence the sacral autonomic outflow more than the somatic outflow [22,23]. The finding that immunoreactive oxytocin-containing spinal neurons associate with sacral preganglionic neurons supports the idea that oxytocin has an important role in the autonomic spinal circuitry that mediates penile erection [24,25]. Oxytocin is a potent inducer of penile erection when injected into the lateral cerebral ventricle, the paraventricular nucleus, or the hippocampus of laboratory animals; intrathecal oxytocin can also initiate an erection. These erections can be blocked by the administration of oxytocin antagonists given intracerebroventricularly (i.c.v.) or intrathecally, or by electrolytic lesion of the paraventricular nucleus. Additionally, noncontact erections can be reduced by a selective oxytocin receptor antagonist administered into the lateral ventricles, which supports the view that oxytocin mediates this response [26].

Adrenocorticotropic Hormones (ACTH) and Related Peptides

Administered i.c.v., the ACTH and α-melanocyte-stimulating hormones (α-MSH) are able to induce penile erection, along with grooming, stretching, and yawning [1,2,27]. These effects are most probably mediated via stimulation of melanocortin (MC) receptors, of which five different subtypes have been cloned and characterized [28,29]. Alpha-MSH/ACTH seem to act in the hypothalamic periventricular region, and grooming, stretching, and yawning, but not penile erection, was reported to be mediated by MC₄ receptors [27,30]. It is unclear, however, what MC receptor subtype(s) can be linked to the erectile responses. For example, the MC₃ receptor is found in high density in the hypothalamus and limbic systems [31], regions known to be important for erectile functions. The site and mechanism of action responsible of α-MSH/ACTH seem to be different from those involving dopamine or oxytocin [15].

Martin et al. [32] concluded that current evidence indicates that the MC₄ receptor subtype contributes to the pro-erectile effects observed with MC pan-receptor agonists. However, the putative receptor subtypes, pathways, and mechanisms implicated in mediating the pro-erectile effects of MCs remain to be fully elucidated. Melanotan II, a synthetic analogue of α-MSH, when given subcutaneously, was shown to have pro-erectile effects in men with psychogenic impotence [33]. Still, the therapeutic potential of α-MSH analogues remains to be established [34–36].

Nitric Oxide (NO)

Several investigators have shown that within the central nervous system, NO can modulate sexual behavior and penile erection [37–41]. NO may act in several discrete brain regions, e.g., in the MPOA [40,41] and the paraventricular nucleus [5,30]. NO production increases in the paraventricular nucleus of male rats during noncontact penile erections and copulation, confirming that NO is a physiological mediator of penile erection at the level of the paraventricular nucleus [39].

As mentioned previously, injection of NO synthase (NOS) inhibitors i.c.v. or in the paraventricular nucleus prevents penile erectile responses induced by dopamine agonists, oxytocin, and N-methyl-D-aspartate (NMDA) in rats. NO may also mediate the actions of ACTH/α-MSH and 5-HT_2C agonists, which elicit erections when injected into the intracerebroventricular system, according to mechanisms unrelated to oxytocinergic neurotransmission [38]. The inhibitory effect of NOS inhibitors was not observed when these compounds were injected concomitantly with L-arginine, the substrate for NO [38].

Excitatory Amino Acids

Microinjections of L-glutamate into the MPOA elicits an increase in intracavernous pressure [4], and behavioral studies have shown that NMDA increases the number of penile erections when injected into the paraventricular nucleus [42–44]. Furthermore, NMDA, amino-3-hydroxy-5-methyl-isoxazole-4-propionic acid, or trans-1-amino-1,3-cyclo-pentadicarboxylic acid increases intracavernosal pressures (ICPs) when injected into the paraventricular nucleus [45]. The effect of NMDA was prevented by intracerebroventricular administration of an oxytocin antagonist [42]. The NOS signal transduction pathway is considered to mediate the effect of NMDA. Injection of the amino acid leads to an increased concentration of NO metabolites in the paraventricular nucleus [46], and the administration of NOS inhibitors into the paraventricular nucleus and i.c.v. blocked the NMDA effect [42,47].

Serotonin

Neurons containing serotonin (5-hydroxytryptamine, 5-HT) can be found in the medullary raphe nuclei and ventral medulla reticular formation, including the rostral nucleus paragigantocellularis, and bulbospinal neurons containing 5-HT project to the lumbar spinal cord in the rat and cat [1]. Some serotonergic fibers occur in close apposition with sacral preganglionic neurons and motoneurons, and synapses were demonstrated at the ultrastructural level [25]. These morphological findings support the involvement of 5-HT in both the supraspinal and spinal pharmacology of erection, with participation in both the sympathetic and parasympathetic outflow mechanisms.

In animals, 5-HT seems to exert a general inhibitory effect on male sexual behavior [48], although the amine may be inhibitory or facilitatory depending upon its action at different sites and at different 5-HT receptors within the central nervous system [49,50]. This may explain conflicting reports of 5-HT agonists either enhancing or depressing sexual function. Yonezawa et al. [51] found that p-chloroamphetamine, an indirect serotonin (5-HT) agonist, elicited both penile erection and ejaculation simultaneously in anesthetized rats. It was suggested that these effects were mainly produced by the release of 5-HT as limited to the lower spinal cord and/or peripheral sites. The use of selective 5-HT receptor agonists and antagonists can reveal different components of male copulatory behavior [9].

Gamma-Aminobutyric Acid (GABA)

Cumulative data resulting from investigations on the role of GABA in penile erection indicate that this neurotransmitter may function as an inhibitory modulator in the autonomic and somatic reflex pathways involved in penile erection [52]. Activation of GABA_A receptors in the paraventricular nucleus of male rats reduced penile erection and yawning in response to apomorphine, N-methyl-D-aspartate (NMDA), and oxytocin [52]. Dorfman et al. examined age-related changes of the GABA_B receptor in the lumbar spinal cord of Sprague Dawley rats of different ages using quantitative autoradiography [53]. GABA_B receptor affinity showed significant age-dependent and regional increases. The GABA_B receptor decrease in aged rats did not seem, however, to be related to the inhibitory function in penile erection.

Cannabinoids

Administration of endogenous and exogenous cannabinoids was shown to be associated with changes in penile erection and modulation of male sexual behavior [54,55]. The cannabinoid CB1 receptor antagonist SR 141716A was shown to potentiate the penile erection responses to apomorphine in rats [56]. Also, it was shown that cannabinoid CB1 receptors present in the paraventricular nucleus may influence erectile function and sexual activity possibly by modulating paraventricular oxytocinergic neurons mediating erectile function [57]. It was also demonstrated that SR 141716 induced penile erection by a mechanism involving excitatory amino acid neurotransmission causing activation of neuronal NOS (nNOS) in paraventricular oxytocinergic neurons [58].

Opioid Peptides

Available information supports the hypothesis that opioid µ-receptor stimulation centrally prevents penile erection by inhibiting mechanisms that converge upon central oxytocinergic neurotransmission [1]. In rats, morphine injected into the paraventricular nucleus prevents noncontact penile erections (i.e., when penile erection is induced in the male by the presence of an inaccessible receptive female) and impaired copulation. These morphine effects are apparently mediated by prevention of the increased NO production that occurs in the paraventricular nucleus during sexual activity [59]. Morphine also prevents apomorphine-, oxytocin-, NMDA- and noncontact-induced penile erection and yawning by inhibiting NOS activity in the paraventricular nucleus [60–62].

Prolactin

Long-term hyperprolactinemia can depress sexual behavior, reduce sexual potency in men, and depress genital reflexes in rats [50,63]. Acute and chronic central prolactin treatment in rats, however, may have stimulatory and inhibitory effects on male sexual behavior, respectively [64]. Correspondingly, striatal dopaminergic activity was shown to be increased and decreased by acute and 5-day central prolactin treatment [64], supporting the view that the effects of prolactin are associated with changes in striatal dopaminergic activity. Prolactin has been shown to inhibit the dopaminergic incerto-hypothalamic pathway to the MPOA [65]. In humans, it is still unclear whether the negative effects of hyperprolactinemia on erectile function are mediated centrally by way of reduction in sexual interest and sex drive [66], or through a direct effect of prolactin on corpus cavernosum smooth muscle contractility. In dogs, a direct effect on the corpus cavernosum was suggested [67]. In any case, the effect seems independent of circulating testosterone levels and gonadal axis function [68].

Sexual Hormones

Androgens, particularly testosterone, are necessary (although not sufficient) for sexual desire in men. They are essential in the maintenance of libido and have an important role in regulating erectile capacity [69–73]. In men with normal gonadal function, however, there is no correlation between circulating testosterone levels and measures of sexual interest, activity, or erectile function [74]. Following castration in the male (which may reduce plasma testosterone levels by 90% [75]), or other causes leading to a reduction in androgen levels, there is generally a decline in libido, and sometimes in erectile and ejaculatory functions. Testosterone administration restores sexual interest and associated sexual activity in hypogonadal or castrated adult men [76–78]. The testosterone dose–response relationships for sexual function and visuospatial cognition differ in older and young men, higher testosterone doses needed in the elderly for normal sexual functioning [72]. El-Sakka et al. [79] assessed the pattern of age-related testosterone depletion in patients with erectile dysfunction (ED). They found a significant decrease in testosterone level throughout the 4-year follow-up in patients with ED. Patients with decreasing testosterone were older than patients with a steady testosterone level. When castration has been performed in humans, the resultant sexual function may range from a complete loss of libido to continued normal sexual activity. Thus, the role of androgens in erectile function is complex, and androgen deprivation may not always cause erectile impotence, either in man [80], or in rats [81].

Inositol 1,4,5-trisphosphate (IP₃), 1,2-diacylglycerol (DAG), and protein kinase C (PKC)

The binding of vasoconstrictor agonists, such as norepinephrine (NE), endothelin-1 (ET-1), angiotensin II (AT-II), prostaglandin (PG) F_2α and thromboxane (Tx) A₂, to their respective receptors stimulates phospholipase C beta (PLC-β). This membrane-bound enzyme hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP₂) to liberate IP₃ and DAG. IP₃ binds to specific receptors (IP₃R) on the smooth endoplasmic reticulum (SER) to stimulate the release of Ca²⁺ from intracellular stores. IP₃Rs act as Ca²⁺-activated Ca²⁺ channels. Binding of IP₃ to these receptors not only activates the channel, but also increases the sensitivity of the IP₃R to Ca²⁺ and facilitates Ca²⁺-induced release of Ca²⁺[102]. Upon dissociation of agonists from their receptors, free Ca²⁺ is recycled back into the SER by the SER Ca²⁺-ATPase pump. It restores intracellular Ca²⁺ levels to the basal state and thereby is an important mechanism of signal termination.

DAG is also an important intracellular second messenger generated by PLC. DAG directly activates PKC. With regard to smooth muscle tone, PKC can regulate ion channels or phosphorylate multiple substrates to facilitate contraction [93,103]. PKC may also mediate Ca²⁺-independent contraction, since several of the PKC isoforms are insensitive to Ca²⁺ while still being activated by DAG. In VSMCs, termination of DAG/PKC signaling is accomplished predominantly by hydrolysis of DAG by lipases to yield free fatty acids and glycerol [103].

Cyclic Nucleotides

Generation of cyclic nucleotides (cyclic guanosine monophosphate [cGMP] and cyclic adenosine monophosphate [cAMP]) by guanylyl and adenylyl cyclases is a primary mode of mediating penile vascular and nonvascular smooth muscle relaxation. Vasodilators such as vasoactive intestinal polypeptide (VIP) and PGs E and D activate G-protein (Gs)-coupled receptors that can stimulate plasma membrane-associated adenylyl cyclase, whereas soluble guanylyl cyclase (sGC) can be directly activated by NO or carbon monoxide (CO) by binding to the heme moiety of the enzyme. Increased levels of intracellular cAMP and cGMP cause the activation of cAMP-dependent and cGMP-dependent protein kinases (cAK and cGK) [104–106]. Thus, in addition to specific activation, there is potential cross-activation of cAK (also called protein kinase A) and cGK (also called protein kinase G), which may be a mechanistic basis for signal cross talk. However, it has been postulated that activation of cGK by cGMP is the main mechanism to mediate relaxation of penile erectile smooth muscle. cGK are derived from two different genes that encode type I (cGKI) and type II (cGKII). In smooth muscle, only cGKI is expressed and exists as two splice variants (cGKI alpha and cGKI beta). While specific roles of the two different cGKI isoforms are an area of continuing investigation, there is evidence that both isoforms differ considerably in their functional properties [107–109]. Immunoprecipitation studies indicate that in smooth muscle cells, cGKI is associated with IP₃R and a protein known as IP₃R-associated cGK substrate (IRAG), both of which act as substrates for the kinase [110]. Phosphorylation of IP₃R and IRAG decreases agonist-induced Ca²⁺ release from the SER [111]. In addition, cGKI is known to phosphorylate phospholamban, a small membrane-associated protein that constitutively inhibits the SER Ca²⁺-ATPase. Phosphorylation of phospholamban inactivates its inhibitory control of ATPase activity and increases Ca²⁺ reuptake into the SER, where Ca²⁺ is bound to proteins such as calsequestrin. Protein kinase A can also phosphorylate phospholamban to increase Ca²⁺ reuptake [112]. Thus, the combined actions of cGKI and cAK can inhibit Ca²⁺ release from intracellular stores or stimulate Ca²⁺ re-uptake.

Additional and perhaps equally important pharmacomechanical mechanisms by which cGMP and/or cGKI may cause relaxation also involve inhibition of Rho kinase and stimulation of MLCP [113]. Some studies suggest that Rho kinase and MLCP can both be phosphorylated by cGKI to antagonize Rho kinase activity and stimulate MLCP. In the penis, immunostaining for cGKI alpha and cGKI beta has been observed within the smooth musculature and the endothelium of cavernous arteries and sinusoids. Double-staining protocols revealed the colocalization of alpha-actin, cGMP, endothelial NOS (eNOS), and cGKI isoforms [114]. Findings from in vitro functional studies are also in support of a significance of the cGKI in the control of human penile erectile tissue [115,116].

PDEs

One of the main mechanisms by which cyclic nucleotide signaling is terminated is by the action of PDEs, a heterogenous group of hydrolytic enzymes. PDEs are classified according to their preference or affinity for cAMP and/or cGMP, kinetic parameters of cyclic nucleotide hydrolysis, relative sensitivity to inhibition by various compounds, allosteric regulation by other molecules, and chromatographic behavior on anion exchange columns. Out of 11 families of PDEs consisting of more than 50 isoenzymes identified to date [117–119], six (PDE 1, 2, 3, 4, 5, and 11) have been proven to be of pharmacological importance. Since the distribution and functional significance of PDE isoenzymes varies in different tissues, isoenzyme-selective inhibitors have the potential to exert specific effects on the target tissue. Preferential expression of PDE5 in the corpus cavernosum and the cGMP-mediated relaxation of the cavernous smooth muscle during sexual stimulation have made inhibition of this enzyme by sildenafil, vardenafil, or tadalafil a clinical benefit in the management of ED. The purified protein is a homodimer of 99.6 kDa subunits and binds two zinc atoms per monomer which are necessary for catalysis.

Since PDEs form a biochemically and structurally diverse family of proteins, there might be more than one PDE isoenzyme or isogene serving as potential drug target in the treatment of ED. In the 1990s, the presence of PDE isoenzymes 2, 3, 4, and 5 was reported in cytosolic supernatants of human erectile tissue [120]. In addition, the expression of mRNA transcripts specifically encoding for 14 different human PDE isoenzymes and isoforms in human cavernous tissue was shown by means of real-time polymerase chain reaction (RT-PCR) and Northern blot analysis: PDE1A, PDE1B, PDE1C, PDE2A, and PDE10A, which hydrolyze both cAMP and cGMP; the cAMP-specific PDE isoenzymes PDE3A, PDE4 (A-D), PDE7A, and PDE8A, and the cGMP-specific PDEs PDE5A and PDE9A [121]. The intracellular level of cAMP in human erectile tissue is mainly regulated by the cAMP-degrading PDE isoenzymes 3 and 4. Results obtained in vitro suggest that PDE3 and PDE4 might be the predominant isoenzymes in the human corpus cavernosum [122]. Interestingly, it has also been shown that the reversion of tension mediated by an alpha₁-adrenoceptor of isolated human corpus cavernosum induced by sildenafil and tadalafil was reversed by the cAK inhibitor Rp-8-CPT-cAMPS, suggesting an involvement of cAMP-mediated mechanisms in the action of PDE5 inhibitors [115]. On the basis of these observations, an important complementary role might be considered for the adenylyl cyclase/cAMP/cAK pathway in the control of cavernous smooth muscle tone.

NO

NO is the primary mediator of nonadrenergic, noncholinergic (NANC) parasympathetic input and endothelium-dependent relaxation in the corpus cavernosum [1]. NO can regulate a wide array of physiological functions in mammals. It is synthesized on demand from the amino acid L-arginine and molecular oxygen by a family of enzymes known as NOS. Three distinct isoforms of NOS have been identified which were originally named after the tissues in which they were first described. nNOS (NOS type I) and eNOS (or NOS type III) are Ca²⁺/calmodulin-dependent, constitutive isoforms. Inducible NOS (iNOS or NOS type II) is a Ca²⁺-independent isoform that is mainly expressed in macrophages and tissues following an immunological stimulus [130]. NO can readily cross plasma membranes to enter target cells and promote the synthesis and accumulation of cGMP by the activation of the sGC. sGC is a heterodimeric protein that contains a prosthetic heme attached to a histidine residue of the β-subunit, which is required for the activation of the enzyme. Although the binding of NO occurs in the β-subunit, both subunits are required for the stimulation of enzyme activity [131,132].

PGs

PGs are produced by the action of cyclooxygenases on the common precursor arachidonic acid [133]. Human corpus cavernosum smooth muscle cells in culture have been shown to produce prostaglandin E (PGE₂) and PGF_2α. It has been demonstrated that prostanoids can induce both relaxation and contraction in penile corpus cavernosum. PGE is the only endogenous PG that appears to elicit relaxation of human trabecular smooth muscle, the others causing constriction or having no effect on smooth muscle tone. Exogenous PGE₁ and PGE₂ relax isolated cavernosal tissue at submicromolar concentrations, while PGE₂ causes contraction at concentrations of 10 µM or greater. Prostaglandin E receptor (EP) receptors, which mediate the response to PGE, have been the most extensively studied and are categorized into four pharmacologic subclasses (EP₁–EP₄) [134,135]. Several different isoforms of the EP₃ receptor have been identified and arise from alternative splicing of a single gene product. In general, the EP₁, EP_3I, and EP_3III receptors mediate smooth muscle contraction by stimulating phosphatidylinositol hydrolysis or inhibiting adenylyl cyclase, while EP₂, EP_3II, and EP₄ receptors mediate smooth muscle relaxation by coupling to Gs protein and stimulating adenylyl cyclase to increase intracellular cAMP [136].

Peptide Regulators—VIP and ET

The density and distribution of VIPergic nerves within the penis has led many to postulate that VIP, in addition to NO, is an important NANC neurotransmitter regulating penile erection. VIP-immunoreactive nerves are widely distributed throughout the male genitourinary system, and VIP and NOS containing nerves are often colocalized in penile tissue [1]. Isolated corpus cavernosum tissue from various species, including human, exhibits relaxant responses to VIP. These effects are accompanied by an increase in tissue levels of cAMP. The role of endogenous VIP in mediating penile erection is further supported by the observations that anti-VIP antibodies and VIP receptor antagonists inhibit nerve-mediated relaxation in isolated cavernosal tissue strips [1]. However, the role of VIP as a modulator of trabecular smooth muscle tone remains inconclusive since the effects of the peptide in vivo are not necessarily consistent with ex vivo or in vitro findings. Intracavernosal administration of VIP in humans has yielded varying results, ranging from no effect to partial tumescence to full erection. Thus, while descriptive studies are supportive of VIP's potential role in mediating or enhancing the onset and maintenance of penile erection, the mechanisms underlying its regulation and action have yet to be completely elucidated.

ET-1 is one of the most potent vasoconstrictors yet described [137,138]. This peptide has also been shown to have growth factor activity, stimulating mitogenesis in fibroblasts, smooth muscle, and endothelial cells. Similar to NO, ET release from the intimal lining of blood vessels can be induced by shear stress. In human corpus cavernosum, ET-1 is synthesized by the endothelium and elicits strong, sustained contractions of corpus cavernosum smooth muscle [139,140]. Both ET receptor subtypes (ET_A and ET_B) have been identified in penile corpus cavernosum. They are distributed on both the endothelium and the smooth muscle and are distinguished by their binding affinity for ET-3 [141]. Exogenous ET-1 or ET-2 cause equipotent contraction in isolated cavernosal tissue strips, whereas ET-3 induces much weaker contraction in corpus cavernosum. While the receptor-binding affinity of ET-1 and ET-2 is not necessarily greater than that of other contractile factors, the rate of dissociation is significantly slower than many ligands [141]. This may account for the unique ability of ETs to maintain long-lasting, sustained contraction in corpus cavernosum smooth muscle. It has been suggested that ET may also exert vasodilatory effects at low concentrations through a “super-high” affinity form of the ET_B receptor. Although this vasodilatory role of ET in penile erection remains unclear, it has been demonstrated in the rat that ET-3 and submaximal doses of ET-1 increase ICP, potentially by stimulating NO production [142].

Diabetes Mellitus

Diabetes mellitus is a common chronic disease, affecting 0.5–2% worldwide. The prevalence of diabetes as a comorbidity has remained at 20–25%, irrespective of whether treating clinics are endocrine based or andrology based [149,150]. ED in diabetics is more common than retinopathy or nephropathy [151]. The Massachusetts Male Aging Study reported that up to 75% of men with diabetes have a lifetime risk of developing ED, much higher rates than 52% (40–70 years of age) [152–154]. The onset of ED occurs in the earlier age for those with diabetes, presenting within 10 years of the diabetic onset in more than 50% of patients with any type of diabetes [155].

MetS

MetS, which is also called insulin resistance syndrome or syndrome X, includes glucose intolerance, insulin resistance, obesity, dyslipidemia, and hypertension. Several epidemiological data have identified MetS as potential risk factors of ED. Grover et al. [156] evaluated the effect of various cardiovascular risk factors on ED in a primary care setting. ED was found in 49.4% according to a score of less than 26 on the International Index of Erectile Function-erectile function (IIEF-EF) domain in a cross-sectional survey of 3,921 Canadian men. The presence of diabetes (odds ratio 3.13), undiagnosed hyperglycemia (odds ratio 1.46), impaired fasting glucose (odds ratio 1.26), and MetS (odds ratio 1.45) were identified as independent risk factors for ED.

Diabetes Mellitus

In 12% of type 1 diabetic men, ED was the first symptom of diabetes [157]. The prevalence of coronary artery disease (CAD) (20%) and peripheral vascular disease (5%) among men with diabetes is far higher than in the general population. Pathologic changes in the cavernous arteries [158], ultrastructural changes in the cavernous smooth muscle [159], and impaired endothelium-dependent relaxation of the corporeal smooth muscle [160] have also been noted in penile specimens from diabetic men with ED.

The presence of ED in diabetic patients could be the harbinger of fatal cardiovascular disease. Gazzaruso et al. [161] demonstrated a higher prevalence of ED in diabetic patients with silent CAD than those without any evidence of myocardial ischemia. ED was associated with more than 14 times higher risk for silent CAD in diabetic men. In a subsequent study, ED was associated with higher major cardiovascular morbidity and mortality in diabetic patients with silent CAD [161].

Hemoglobin (Hb)A1c levels have been shown to increase with the severity of ED [162,163] and was found to be an independent predictor of the EF score in 78 men with type 2 diabetes [164]. However, there have been conflicting results about the beneficial effects of strict glycemic control on erectile function. Some studies reported improved erectile function following the reduction of HbA1c, while others reported no significant change despite the aggressive blood sugar control [165,166].

Hypogonadism is often associated with diabetic ED patients. Corona et al. found hypogonadism in 24.5% of men with diabetes and ED vs. 12.6% in the rest of the men with ED [167]. Testosterone supplementation in human diabetics with ED receiving pharmacological treatment might be advantageous in the diabetic men in whom PDE5 inhibitors given for ED do not work [168].

MetS

MetS and increased waist-to-hip ratio have been associated with a higher proportion of moderate-to-severe ED in men older than 50 years [169]. Conversely, ED may be predictive of MetS presence in men with a body mass index (BMI) of <25 kg/m [145,170]. These interesting findings suggest that ED may be a warning sign for MetS in men otherwise considered at low cardiovascular risk.

Studies indicated the possible role of inflammation and endothelial dysfunction in the development of ED for patients with MetS. Men with MetS had an increased prevalence of ED, reduced endothelial function score, and higher circulating concentrations of high sensitivity C-reactive protein (CRP) compared with men without metabolic disorders [171]. This and other studies clearly showed the relationship between MetS, the inflammatory endothelial activation, and the prevalence of ED [171,172].

As mentioned earlier, low circulating androgen levels are clearly a risk factor for MetS and the reverse relationship is true as well. ED might occur as a possible consequence of hypogonadism and MetS. A recent study by Zhody et al. [173] elegantly related hypogonadism to ED and MetS by analyzing BMI measurements in 158 obese men. With increasing BMI, the frequency of hypogonadism and ED increased, while total serum T showed a strong negative correlation. To assess the effect of BMI on vasculogenic ED, the authors examined this relationship in the absence of other risk factors and found that for a BMI <25, three out of 13 men (23.1%) had vasculogenic ED as compared with 32 out of 54 men (59.3%) with a BMI ≥25.

Currently, no direct pharmacologic therapy for MetS is available. Esposito et al. [174] assessed the effect of weight loss and increased physical activity on men with ED associated with obesity. BMI decreased significantly in the intervention group, and was associated with a decrease in serum concentrations of interleukin-6 and CRP. Erectile function scores improved significantly with lifestyle intervention but remained stable in the control group. In multivariate analyses, changes in BMI, physical activity, and CRP were independently associated with erectile function improvement. Thus, lifestyle changes are associated with improved sexual function and lowered inflammation in obese men with ED.

Nitrergic Dysfunction

Erection is activated by NO release from nNOS at NANC nerve terminals. Maintenance of cavernosal vasodilation is thought to occur through the activation of eNOS in endothelial cells, presumably in response to shear stress. Impaired vasodilator signaling often results from NANC nerve dysfunction and/or endothelial dysfunction, leading to ED. Numerous studies have demonstrated type I diabetic animals to have impaired cavernosal relaxation to electrical field stimulation as well as decreased ICP following electrical cavernosal nerve stimulus, indicative of nitrergic dysfunction [175–181]. Decreased penile nNOS content was detected in various rodent models of type 2 diabetes [182–184]. Impaired nitrergic-mediated relaxation in type 2 diabetic mice has also been reported; however, the extent of the impaired relaxant response was modest, leading the authors to question the true pathophysiologic relevance of this finding to the ED phenotype [185].

Endothelial Dysfunction

Endothelial dysfunction is characterized by lowered NO bioavailability resulting from decreased eNOS expression or activity, or increased NO scavenging. It is clear that an attenuation of endothelium-dependent vasodilation of cavernosal tissue is present in several animal models of type 1 and type 2 diabetes [144,147,184–186]. The activation of eNOS can occur by hemodynamic stimuli, such as shear stress, as well as through protein signaling, such as by VEGF, leading to eNOS phosphorylation on serine 1177 [147,187]. In addition to phosphorylation events, eNOS activity and subsequent NO production are regulated by substrate concentration, cofactor availability, and enzyme coupling. Relevance of dysfunctional eNOS enzyme regulation remains speculative in regard to diabetic ED (see citation [147] for review).

Oxidative Stress

Chronic hyperglycemia induces free radical production through formation of advance glycation end-products (AGE), lipid peroxidation, polyol pathway activation, superoxide production, and activation of PKC [188]. Increased penile and serum AGE and reactive oxygen species (ROS) levels have been detected in type I diabetic rodents [189,190]. Impairments in NO-mediated cavernosal relaxation in these rodents are prevented with superoxide dismutase or a peroxynitrite decomposition catalyst, supporting a delirious role of ROS in type I diabetic ED [191–193].

Studies examining oxidative stress in animal models of type 2 diabetes or MetS are scant. Decreased antioxidant levels, such as glutathione (GSH), may result in elevated ROS/oxidative stress in type 2 diabetic men. Kovanecz et al. found prolonged treatment with pioglitazone, a peroxisome proliferator-activated receptor (PPAR)-γ agonist said to have anti-inflammatory effects and to improve the glutathione/glutathione disulphide (GSH/GSSH) ratio [194], suggesting that glycemic-stabilizing agents may also have benefit in decreasing damaging ROS.

Cavernosal Hypercontractility

Increased contractile function of the cavernosum can result from heightened sympathetic activation or potentiated intracellular contractile signaling of smooth muscle cells. Many animal models of diabetic ED have pointed to cavernosal hypercontractility as a pertinent mechanism underlying the disease phenotype. A recent review extensively outlines potential pro-signaling pathways in the penile smooth muscle cell that may contribute to diabetic ED [148].

Studies by Carneiro et al. and Luttrell et al. have suggested the presence of increased contraction in response to sympathetic activation in type II diabetic ED [185,195]. Wingard et al. found the heightened contractile signaling in the type 2 diabetic rodent in response to phenylephrine and ET-1 to be mediated by overactivity of PKC and Rho kinase, two primary kinases mediating smooth muscle cell tone [186].

Veno-occlusive Dysfunction

The limiting of blood outflow through mechanical compression of the emissary veins against the tunica albuginea is essential for the maintenance of elevated corporal pressures and a rigid erection. Studies in animal models of type 2 diabetes have suggested that a veno-occlusive disorder may underlie the ED phenotype. Kovanecz et al. found Zucker diabetic fat rats to have an inability to sustain adequate intracorporal pressure after the cessation of penile saline infusion, suggesting the presence of a veno-occlusive disorder [194]. These studies have recently been validated in the db/db mouse model of type 2 diabetes [185].

Increased Vasoconstriction

RhoA/Rho kinase pathway plays an important role in calcium sensitization and tonic contraction of smooth muscle. Cardiovascular diseases are associated with an enhancement of RhoA/Rho kinase activity [245]. RhoA and Rho kinase expression is elevated in the cavernosal tissue from hypertensive rats, possibly contributing to reduced erectile function in these rats [246]. ET-1 levels are elevated in plasma from patients with hypertension and hypercholesterolemia [247,248]. Patients with organic ED also show higher venous and cavernosal ET-1 levels [249]. In cavernosal tissue from DOCA-salt hypertensive rats (specific animal model), contractile responses to ET-1 were increased and relaxation caused by ET_B activation was reduced [250]. A decrease in ET_B receptors in cavernosal tissue from hypercholesterolemic rabbits was also detected [251]. AT-II has been detected in endothelial and smooth muscle cells of human corpus cavernosum [252] and caused its contraction [253]. AT-II converting enzyme expression is upregulated in rats with arteriogenic ED [254]. In fact, administration of an AT₁ antagonist has been shown to improve erectile function in hypertensive patients [255].

Impaired Neurogenic Vasodilatation

Impaired neurogenic relaxation of cavernosal tissue has been observed in a rabbit model of cavernosal ischemia [209], in SHR, affecting both NO neurotransmission and CO neurotransmission [256] and after cigarette extract administration [241] while was unaffected in a rabbit model of hypercholesterolemia [207].

Endothelial Dysfunction

Coronary endothelial dysfunction has been associated with the presence of ED [257]. Endothelium-dependent flow-mediated dilation (FMD) of the brachial artery was significantly reduced in ED patients, although nitroglycerine-induced dilation was also impaired [258,259]. The impairment of FMD correlates to the severity of ED [260]. Endothelium-dependent penile blood flow increases generated by reactive hyperemia were reduced in patients with ED. At the same time, endothelial function in the forearm vasculature was not significantly altered in ED patients [261].

Penile Structural Alterations

Objective reduction of smooth muscle cells has been demonstrated in patients with organic ED [262]. A decrease in cavernous trabecular smooth muscle and an increase in connective tissue are correlated with diffuse venous leakage and a failure of the veno-occlusive mechanism, hence resulting in ED [263,264]. SHR show hyperproliferation of smooth muscle in cavernosal tissue and penile vasculature, which correlates with blood pressure values. Increased fibrosis was also observed [265]. Hypertension-induced alterations were improved after antagonism of type 1 AT-II receptors (AT₁) [266].

Diuretics

This class of drug has been extensively studied following early trials which showed a high prevalence of self-reported ED. Possible mechanisms include decreased vascular resistance and lowered zinc levels leading to reduced androgen production. Appropriate controlled studies with ED as an end point give consistent results despite trends toward lower-dosage schedules [225]. Similar findings were documented from the Treatment of Mild Hypertension Study (TOMHS) where the prevalence of ED at 2 years in men taking a low-dose thiazide was twice that of both the placebo group and those on alternative agents [272]. Interestingly, after 4 years of treatment, prevalence of ED in the placebo group approached that of the thiazide group, a finding not fully explained by dropouts. It may be that thiazide therapy unmasks latent ED at an earlier stage rather then being directly causal. Thus, it is likely that thiazide diuretics are associated with ED in men with hypertension, although this may represent unmasking of an existing problem and the effect can be reduced by lifestyle changes.

β- and α-Adrenoceptor Antagonists

In penile tissue, activation of β-adrenoceptors leads to corporal relaxation [1] and vasodilation of penile arteries [273]. This response is attenuated in vitro by nonselective drugs such as propanolol, possibly by blocking postjunctional β₂-adrenoceptors [273,274], but not by cardiac selective agents such as practolol and atenolol. Depending on the lipophilicity of the β-adrenoceptor antagonists (e.g., propranolol is hydrophobic, while atenolol is hydrophilic), they may also exert an inhibitory effect within the central nervous system, perhaps leading to lowered sex hormone levels [275]. The effect of β-adrenoceptor antagonists on erectile function to a large degree can be explained by their mechanisms of action (e.g., they are either general β-adrenoceptor antagonists, selective β₁-adrenoceptor antagonists, or also possess vasodilatatory properties) (Figure 2). Nonselective drugs such as propanolol were associated with higher prevalence of ED compared with patients treated with placebo or ACE inhibitors [276,277]. Later trials using agents with higher selectivity for the β₁-adrenoceptor such as acebutolol have shown a substantial reduction in ED as a side effect with no difference being found against the placebo and ACE inhibitor groups [272]. This also applies to the use of selective β₁-adrenoceptor antagonists in the prophylaxis of angina [278]. The general β-adrenoceptor antagonists which also cause vasodilation by blocking α₁-adrenoceptors, e.g., carvedilol, have, in a crossover study, been reported to be associated with worsening in sexual function [279]. Some of the recently introduced β₁-adrenoceptor antagonists such as nebivolol have also vasodilatatory effects mediated by release of NO. In crossover studies, nebivolol, in contrast to the selective β₁-adrenoceptor antagonists, metoprolol, and atenolol, did not decrease sexual intercourse activity in hypertensive men, and may even have positive effects on erectile function [280,281].

In clinical trials, α₁-adrenoceptor antagonists (e.g., doxazosin) used to treat hypertension [272] or lower urinary tract symptoms [282] are not associated with complaints of ED and indeed had lower rates than placebo groups. Drugs stimulatory to the α₂-adrenoceptors such as clonidine result in diminished erectile function both clinically and experimentally by peripheral and central mechanisms [274,283].

ACE Inhibitors and AT₁ Receptor Antagonists

In addition to circulating AT-II, ACE and chymase are expressed in erectile tissue, and functional as well as binding studies suggest that AT-II induces contraction by activation of AT₁ receptors [270]. Moreover, AT-II increases during the detumescence phase in man [284]. The ACE inhibitor captopril does not cause any significant adverse effect on sexual function in awake rats [274], and enalapril may even improve erectile function in SHR [285]. This is also supported by clinical studies comparing an ACE inhibitor with other agents and placebo. All three studies found either no difference compared with placebo or improved sexual function from baseline compared with other antihypertensive drugs [272,275,276,286].

In studies of hypertensive animals, AT₁ receptor antagonists (e.g., losartan, valsartan, and candesartan) reverse structural changes in the penile vasculature and appear to conserve erectile function [285,287–289]. Moreover, in clinical cross-sectional studies, AT₁ receptor antagonists, in contrast to other antihypertensive drugs, even tend to improve erectile function [225]. In direct comparison with the β-adrenoceptor antagonist carvedilol, valsartan has a beneficial effect on existing sexual dysfunction at baseline and has no adverse sexual effects during 12 months of treatment [279]. In the case of losartan, 3-month treatment was also reported to improve sexual function [290].

Calcium Channel Blockers

Smooth muscle contraction requires increased cytosolic calcium derived from internal stores and extracellular fluid. It would, therefore, be anticipated that calcium channel blockers would have a permissive effect on penile erection but might inhibit bulbospongiosal contraction during ejaculation. This is supported by findings of in vitro studies which demonstrated a modest relaxant effect on isolated cavernosal smooth muscle [283] and penile arteries [291]. Clinical studies have demonstrated no adverse effect on erection and ejaculatory complaints seem short-lived [275]. In the TOMHS, there was no significant excess risk of ED in the amlodipine group compared with placebo-treated patients [272]. Another study also showed no increase in the prevalence of ED when hypertension was treated with diltiazem alone or in combination with an ACE inhibitor [292]. A comparative study of two calcium channel antagonists showed that neither had any significant effect on sexual function, although two patients withdrew from the nifedipine arm because of reduced libido [293].

Lipid-Lowering Drugs (Fibrates, Statins)

ED was reported to be a frequent side effect of treatment of hyperlipidemic subjects using clofibrate [298] or gemfibrozil [299]. In patients referred to a clinic for primary hyperlipidemia, an increased risk of ED was also observed in patients treated with fibrates [300]. Fibrates interact with PPARs [301], which in the liver stimulates microsomal esterification of estradiol and testosterone [302]. That may explain the increased prevalence of ED reported in patients treated with fibrates.

In patients with hyperlipidemia and treated with statins such as simvastatin and pravastatin and referred to a clinic for primary hyperlipidemia, an increased risk for ED was reported [300,303]. It is controversial whether simvastatin is associated with ED [304–306], and so far, in patients treated with simvastatin, underlying diseased vasculature rather than the drug appears to be the cause of ED. In contrast to simvastatin, positive effects on erectile function was reported for atorvastatin on erectile function [307–310]. These reported positive effects of atorvastatin, in contrast to simvastatin, on erectile function suggest that the effect of statins is not a class effect of statins. Moreover, in an observational prospective study, it was reported that differences in dose, relative efficacy, or relative lipophilicity of statin did not show correlation with changes in IIEF score over a 6-month period [311]. However, statins are structurally a heterogenous group of compounds and, therefore, other mechanisms than the lipid-lowering effects may have significance for the effect of a statin on erection [246,312,313].

Aldosterone Receptor Antagonists

Treatment with spironolactone and eplerenone increased survival in systolic chronic heart failure. In addition to being an aldosterone receptor antagonist, spironolactone also blocks androgen receptors (ARs) and that may explain why it is associated with sexual dysfunction. In contrast, the newer aldosterone receptor antagonist, eplerenone, is devoid of effects on sex hormone receptors.

Antipsychotics

It is difficult to separate disease from drug effect, as well as to obtain reliable information from patients with psychotic illness. There is a paucity of controlled studies. Mainly, older antipsychotic drugs result in decreased erection and anorgasmia, while newer antipsychotics appear to have lower incidence of sexual dysfunction. Among newer antipsychotics, risperidone appears to have highest rate of sexual dysfunction, while there are insufficient data on aripiprazole and ziprasidone.

Antidepressive and Anxiolytic Drugs

Depression by itself makes it difficult to separate effect of illness from additive effect of drugs, as mood disorders may lead to lack of interest and emotional withdrawal from the sexual partner. Antidepressants can have numerous effects on sexual function including altered sexual desire, erection difficulties, and orgasm problems. Selective serotonin reuptake inhibitors (SSRIs) and venlafaxine can negatively affect all the steps of the male sexual response cycle (desire–arousal–excitement–orgasm). Bupropion, nefazodone, and mirtazapine have lower rates of sexual dysfunction than SSRIs. The evidence has been summarized in a recent Cochrane review of 15 randomized studies [314]. The main conclusion regarding comedication to correct ED was an effect of sildenafil in three of the clinical trials, while the available evidence was insufficient regarding other strategies for correction of ED. Anxiolytic agents such as bupropion, acting mainly by inhibiting dopamine reuptake, and buspirone, which acts on 5-HT_1A receptors, are not associated with sexual side effects in placebo-controlled trials [315] and can be used to alleviate sexual symptoms caused by other antidepressant medication [316]. There are insufficient randomized data assessing effect of dose reduction, e.g., drug holidays, on sexual function in patients treated with antidepressants.

Opiates

Long-term intrathecal administration of opiates results in hypogonadotropic hypogonadism and associated sexual dysfunction that can be restored with appropriate supplementation [317]. Administration of opioid antagonists to older men with ED, however, did not improve erectile function measured objectively by nocturnal penile tumescence monitoring [318]. Opioids do have a generalized depressant effect on sexual function when directly administered to the MPOA in rat brain, but treatment with the opioid receptor antagonist, naloxone, had no sexual effect on healthy male volunteers [283].