madman
Super Moderator
Androgens and male sexual function (2022)
Giovanni Corona, MD, Ph.D., Consultant Endocrinologist, Giulia Rastrelli, MD, Ph.D., Associate professor of Endocrinology, Linda Vignozzi, MD, Ph.D., Associate professor of Endocrinology, Mario Maggi, MD, Ph.D., Full professor of Endocrinology
Sexual symptoms are the most specific determinants of low testosterone (T) observed during adulthood. In this narrative review, we summarize the most important evidence supporting the positive relationships between endogenous T levels and sexual activity in the adult male, by using preclinical and clinical observations. In addition, we also report an update of our previous meta-analysis evaluating the effects of T treatment (TRT) on sexual functioning in subjects with T deficiency. Available data indicate that TRT of symptomatic hypogonadal men can improve several aspects of sexual life, including erection. However, the effect is rather modest and lower in subjects with associated metabolic conditions. The specific observed effects are similar to those derived from lifestyle intervention. Since TRT might result in body composition improvement, it is reasonable to suppose that initial treatment with T can improve the willingness of hypogonadal subjects to perform physical exercise and to adhere to healthier behavior. Similar data were derived from animal models. However, it should be important to recognize that lifestyle modifications should be the first step to promote weight reduction. TRT can be combined with lifestyle interventions only in symptomatic hypogonadal subjects especially in the presence of comorbid metabolic conditions.
Introduction
Testosterone (T) is a gender-specific sex hormone, being, during adulthood, at least 20-fold higher in men than in women. In the male, T is essentially produced by the testis, as the adrenal contribution to the circulating hormone levels is negligible. Testicular production of T is characterized by three specific surges: one during fetal life, another in the early neonatal life, and the last one during puberty remaining essentially stable thereafter [1-3]. Fetal production of T is driven by chorionic gonadotropin (hCG), while the other surges are driven by the pituitary luteinizing hormone (LH). Fetal T production plays an essential role in shaping the male internal and external genitalia and, most probably, in programming effects on later brain development, e.g. driving core gender identity and sexual orientation [4,5]. The neonatal surge of T happens during the first 3-6 months of life and it is often referred to as mini-puberty. Mini-puberty remains a mysterious phenomenon, but it is hypothesized that the T rise in the male newborn might affect several psycho-biological aspects, including body and penile growth, number of Leydig cells within the testis, and brain organization [1,4,5]. During puberty, T rise allows definitive development of secondary sexual characteristics, growth acceleration, and modifications in body composition, including an increase in bone and muscle mass, along with a decrease in fat mass [1]. In addition, several psychological modifications occurring during male puberty may be regulated by T, as, for example, aggressive behavior and spatial ability [6,7].
In contrast to the aforementioned, well-recognized, differentiating, and organizing activities of T - spanning from the early fetal life to the entire pubertal process - its role in adult male life is more related to a non-crucial supporting role in several biological activities. In particular, during adulthood, all the effects of T are related to maintaining and sustaining the psycho-biological modifications initiated during the pubertal transition including body composition, spermatogenesis, erythropoiesis, male attitudes and behaviors, bone health, and sexual functions. However, none of these T effects is essential for survival; in fact, the life expectancy of eunuchs is similar to that of the general population [8]. Low T during adulthood has been associated with physical issues (osteoporosis, reduced mobility, fatigue, anemia), psycho-neurological symptoms (depression, decreased energy, cognitive deficits), and sexual symptoms (decreased spontaneous and sex-related erection, low libido, reduced orgasmic pleasure, delayed ejaculation, decreased ejaculatory volume) [9e13]. Late-onset hypogonadism (LOH) is an umbrella term that syndromically associates, in the aging male, some or all of the aforementioned symptoms or signs of low T (below 12 nmol/L) [14,15]. In a seminal collection of recommendations [15] released by several Andrology Societies (i.e. International Society of Andrology, International Society for the Study of Aging Male, European Association of Urology, European Academy of Andrology, and American Society of Andrology), it was decided that low T is not enough to define a true hypogonadal state; in fact, also low T-related symptoms should be present. A few years later, a large epidemiological study (European Male Aging Study, EMAS) - enrolling 3369 aged men in eight European centers - indicated that, in both the training and validation sample, only three sexual symptoms (reduced sexual desire and spontaneous and sexual-related erections) - but not other physical or psychological symptoms - clustered with low total T (<11 nmol/L) and calculated free T (cfT ¼<220 pmol/L) [16]. Although statistically significant, all the associations with sexual symptoms were relatively weak, suggesting that the same sexual symptoms can be generated in overall eugonadal individuals by other medical conditions [16]. Later on, the EMAS biochemical thresholds were essentially validated in a setting of men consulting for sexual dysfunction [17]. In fact, the assessment of Youden index showed that the best thresholds for detecting men with androgen deficiency-related sexual symptoms were 10.4 nmol/L for total T and ranges from 225 to 260 pmol/L for cfT [17]. However, association does not mean causation. In fact, the same sexual symptoms, often viewed as a mirror of T deficiency, can also be explained by several morbidities afflicting the aging male [18]. The Chronic Disease Score (CDS) is a widely accepted measure of comorbidities, based on the aggregate number of prescribed medications for several chronic conditions [19]. In the aforementioned clinical setting of men complaining of sexual dysfunction, we reported that the weighted effect of CDS versus total T is 2-3 times higher in explaining a decreased basal and dynamic peak systolic velocity (DPSV) and acceleration at penile color Doppler ultrasound (PCDU). Even the relative risk of having an impaired erection, either spontaneous or sex-related, was more dependent on high CDS than on low total T [20]. In that study [20], the most genuine symptom of T deficiency was reduced sexual desire. Accordingly, in the longitudinal extension of the EMAS study, the development of secondary hypogonadism has the strongest correlation with decreased sexual desire, rather than other sexual symptoms [21].
*In this review, we will summarize the relationships between endogenous T levels and sexual activity in the adult male, by using preclinical and clinical observations. In addition, we also reviewed, using a meta-analytic method, the effect of T treatments in subjects with T deficiency, focusing on the effect on sexual symptoms.
*Regulation of penile erection by endogenous testosterone
Penile erection is a neuro-hemodynamic phenomenon causing engorgement of the two corpora cavernosa (CC) by arterial blood, finally leading to rigidity and erection for mechanical compression of the emissary venous vessels, thereby decreasing blood outflow [28]. Cell-to-cell interactions between the endothelial layer and the surrounding smooth muscle cells (SMC) are the main regulators of arterial inflow into the CC. Activation of penile erection is controlled by the major pelvic ganglia that contain, within the pelvic plexus, both parasympathetic and sympathetic neurons.
Sympathetic signals (i.e. noradrenaline), along with other substances (e.g. endothelin-1), maintain the SMC of the penile arteries and of the cavernous spaces in a contractile status by increasing intracellular calcium concentrations. Both noradrenaline and endothelin-1 also activate a calcium sensitizing pathway involving activation of the small GTPase protein RhoA and of its downstream kinase ROCK. The final effect is the activation of myosin light chain kinase (MLCK) and SMC contraction, which limits arterial inflow and causes a flaccid state [28].
During an erection, parasympathetic and non-adrenergic, non-cholinergic (NANC) nerve activity causes and sustains the release from nerve endings and from endothelial cells of nitric oxide (NO), through neuronal and endothelial NO synthase (nNOS and eNOS, respectively). NO easily diffuses into the SMC increasing its downstream effectors (cGMP, guanylate cyclase, PKG), causing calcium sequestration, and calcium efflux. This allows SMC to relax and blood flow to increase [28]. The crucial role of NO in penile erection has been demonstrated in pioneering studies with genetic manipulation of the NO-forming enzymes eNOS and nNOS [29e31]. Returning to the basal, contracted state is facilitated by cGMP catabolism through a series of phosphodiesterases (PDEs). In the penis, among different PDE, PDE5 is the most represented and biologically active in metabolizing cGMP [32]. Other neurotransmitters, such as vasoactive intestinal peptide (VIP), prostaglandins, and adenosine (ADO), facilitate SMC relaxation by activating adenylate cyclase and calcium mobilization. ADO is a purinergic neuromodulator that exerts a profound effect on penile erection acting through a series of receptors, with ADORA2B being crucial, because, in a mouse model, its genetic deletion is associated with loss of erection and its excess to priapism [33]. Considering the crucial role of the endothelial and SMC in the dynamic process of rigidity/detumescence, it is obvious that any alteration in their representation could profoundly alter penile physiology.
T exerts a crucial role in facilitating an increase in penile blood flow (Fig. 1). Fig. 2 shows the relationship between quintiles of endogenous T and measured maximal, prostaglandins E1 (PGE1)- induced, DPSV at PCDU in a large series (n ¼ 2751) of men consulting our outpatient clinic for sexual dysfunction [12,25,34]. In an age- and CDS-adjusted ANCOVA model, the lowest T quintile - which corresponds to the EMAS-defined hypogonadal state (i.e. low T and sexual symptoms) - is characterized by a significant decrease in maximal penile blood flow. In contrast, in the eugonadal range (total T > 10.4 nmol/L), further increases in penile blood flow were marginal or insignificant. This is tantamount to say that deficiency of T in the hypogonadal range, but not in the eugonadal one, is independently associated with an impaired penile blood flow.
However, studies in humans have not helped in discerning the mechanism of T action in penile erection. Preclinical studies in animal models could, therefore, be helpful in understanding the role of T in favoring penile erection. In the last 12 years, we set up a model of metabolic syndrome (MetS) by feeding rabbits a westernized, cholesterol-rich, diet (HFD [21]). Beside MetS, HFD induced hypogonadotropic hypogonadism and erectile dysfunction (ED) [22]. We, later on, demonstrated that training rabbits to perform endurance exercise reversed the hypogonadal condition and erectile dysfunction. By selecting data from these HFD rabbits performing or not physical exercise and from those fed a regular diet (RD) [22-24], we now report associations between gene expression of the AR and the aforementioned families of factors that drive penile erection/detumescence. Results are shown in Table 1. Expression of AR is closely associated with markers of smooth muscle function (SM22, aSMA, RhoA, ROCK1, ROCK2, MYH11) and with the entire NO signaling pathway, ranging from NO formation (DDHA1, eNOS, nNOS), its downstream action (GCSa1, GCSb1, PKG) to its degradation (PDE5). In addition, AR is associated with ADO action (ADORA1, ADORA2A, ADORA2B, ADORA3) and degradation (ADA, AMPD2). Similar results were previously published concerning rabbits treated with a gonadotropin-releasing hormone (GnRH) analog (triptorelin) and supplemented or not with T [35]. PDE5 expression in the penis is an important prerequisite for PDE5 inhibitors (PDE5i) activity [22,36]. Therefore, investigating its relationship with androgen signaling is an important point for understanding penile physiology. Fig. 3, panel A, shows gene expression of PDE5 in the penile extract as a function of increasing quintiles of AR expression. Data are derived from the aforementioned rabbit groups [22-24]. KruskaleWallis analysis indicates significant differences among groups (p < 0.0001). In particular, post-hoc ManneWhitney U test confirmed that there is a stepwise increase in the expression of PDE5 as a function of increasing quintiles of androgen receptor (AR) expression. However, when the expression of PDE5 in the penis was evaluated as a function of quintiles of circulating total T, it was found that only the lowest quintile of T was associated with a reduction of PDE5 expression (Fig. 3 panel B). This lowest quintile represents hypogonadal rabbits, with all total T below one standard deviation of the normal RD value (6.56 ± 4.62 nmol/L). For higher T quintiles there was no further increase in PDE5 expression.
*In conclusion, results from our preclinical and clinical studies, here summarized, suggest that penile erection is also an androgen-dependent phenomenon (see also Fig. 1). However, the relationship between endogenous circulating T levels and erection is evident only in the case of T deficiency and not in eugonadal conditions, suggesting a saturation effect, the molecular basis of which is still to be determined.
*Regulation of sexual interest by endogenous testosterone
Sexual desire is the subjective feeling of wanting to initiate a sexual experience. Hence, it is an important prerequisite for sexual excitation and arousal, which may include, in men, penile erection. Sexual desire is an essential component of the human sexual response, with a particular valence for species perpetuation and for couple enjoyment. Its reduction in men, i.e., male hypoactive sexual desire disorder (MHSDD), has been conceptualized by several editions of the Diagnostic and Statistical Manual of Mental Disorders, including the latest version (DSM-5 [37]), as a long-lasting (more than six months) persistent or recurrently deficient sexual or erotic thoughts, fantasies, and desire for sexual activity that causes marked distress or interpersonal difficulty, with the disease not being a result of medical illness, another psychological disorder, or the effects of a drug.
In a survey on 3714 men consulting for sexual dysfunction at the University of Florence, Florence, Italy, reduced libido was found in 36.4% of the sample, often associated with other sexual dysfunctions such as ED and reported premature or delayed ejaculation, as it was isolated only in 5% of the subjects [38]. Reduced libido was not associated with other risk factors in 33% of the sample (primary reduced libido), while it was often associated with other medical or psychological conditions (secondary reduced libido). In fact, among subjects with hyperprolactinemia or psychopathology, 84%, and 48%, respectively, reported reduced libido (see also Fig. 1). In subjects with low T (i.e. total T < 12 nmol/L), 40% of cases reported any reduction in sexual desire. The main characteristics of low T-associated reduced libido were a severe ED, a decrease in sexual activity, a perceived reduction in ejaculate volume and in sleep-related erection, along with a higher scoring in free-floating anxiety and depressive symptomatology [38]. As a result, reduced libido in subjects with low T could impair several aspects of a couple's sexual behavior, resulting in a source of deterioration of the couple's relationship (as measured by Structure Interview on Erectile Dysfunction Scale 2, i.e. marital domain [38])
An age-adjusted association between low T and reduced sexual desire was reported in some [16,39,40], but not all [41,42], epidemiological studies. In the EMAS study, a low frequency of sexual thoughts was present in 27.5% of the population and was associated with both cFT (p < 0.0001) and total T (p ¼ 0.048) [16]. In particular, a reduction of each 1 nmol/L T below 8 nmol/L was associated with a 48% (20-83%) increased risk of low desire. Hence, a threshold of 8 nmol/L was proposed for the association between low T and sexual desire. In the same study, a threshold for cFT was also found, corresponding to 220 pmol/L. In a previous study conducted in a German series of 434 consecutive male patients aged 50-86 years. Consulting for andrological problems - a higher threshold was proposed for explaining the association between low T and [43,44] decreased libido (i.e. 15 nmol/L [45]). In a more recent survey, conducted at the University of Florence, Italy, in 4890 men attending an outpatient clinic for sexual dysfunction, it was found that the best thresholds to verify the association between reduced T and severe sexual symptoms, including low libido, impaired morning erection, and ED, was total T < 10.4 nmol/L or cfT<225 pmol/L [17]. As mentioned before, similar results were also derived from the EMAS study, because the association between low T and syndromic clustering of sexual symptoms, including low libido, was evident for total T < 11 nmol/L and cfT<220 pmol/L) [16].
The human brain is endowed by AR. AR is distributed in discrete areas of the brain, all involved in the regulation of the male sexual response, including temporal, preoptic, hypothalamus, amygdala, midbrain, frontal and prefrontal cortex areas, and cingulate gyrus (see for review in [46]). Acute testosterone exposure increases activation of most of them, as detected by functional magnetic resonance imaging (fMRI, see for review in [47]). Similarly, Park et al., [48], using fMRI showed that, in healthy volunteers, erotic visual stimulation was able to activate the inferior frontal lobe, cingulate gyrus, insula gyrus, corpus callosum, thalamus, caudate nucleus, globus pallidus, and inferior temporal lobe. The latter results were blunted in hypogonadal men and restored by T administration [48]. In the pioneering study of Stoleru et al. [49] with positron emission tomography (PET) it was found that sexually explicit visual stimulation induced a T rise and activates several paralimbic areas as a function of circulating T levels. Hence, it is conceivable that the AR mediates T action on sexual desire (see also Fig. 1). However, the possibility that other potential targets of T within the brain, such as estrogen receptors, are involved in the T-action on the sexual drive cannot be ruled out [50].
Accordingly, a previous double-blind placebo-controlled RCT, performed by experimentally inducing hypogonadism with GnRH-analogue in 400 healthy men, showed that both estrogen and T deficiency contributes to the decline of sexual desire [50]. Similarly, a sexual desire improvement has been reported in men with aromatase deficiency after transdermal estradiol treatment in subjects affected by (a rare condition impairing estrogen formation), suggesting a possible role of estrogen in regulating male sexual desire [51]. Conversely, adrenal hormones including dehydroepiandrosterone (DHEA) and its sulfate (DHEAS), as well as cortisol and aldosterone are not involved in the regulation of male sexual desire [52].
*Regulation of male orgasm by endogenous testosterone
Ejaculation is a spinal reflex that can be classically divided into three main phases: emission, expulsion, and orgasm. During the first, reversible part of the ejaculation (emission), the semen is propelled throughout the male genital tract until the posterior portion of the urethra. Seminal vesicles secretion, and to a lesser extent prostatic secretion, contributes substantially to the ejaculatory bolus. Closure of the bladder neck prevents the retrograde spillover of the ejaculate bolus into the bladder. During ejaculation an irreversible forceful series of contractions of the muscle of the pelvic floor allows the ejection of ejaculatory bolus, containing the semen, through the urethral meatus. The latter phase is usually associated with orgasm, a complex experience that may occur as sexual pleasure reaches a climax [53]. However, in prepubertal children, orgasm can occur without ejaculation [53]. The orgasmic reaction is controlled by a complex network of sympathetic and parasympathetic nerves located in the thoracolumbar and sacral spinal cord, respectively, with a somatic component controlled by Onuf's nucleus, along with a lumbar spinal generator of ejaculation (SGE). The entire male genital tract, including prostate and seminal vesicles, development, and function, is under androgen control, as it is the neural control of the process (see for review [53,54]; Fig. 1). Hence, it is rather obvious that derangements in androgen action are associated with several ejaculatory dysfunctions, including reduced ejaculatory volume and premature and delayed ejaculation.
*Effects of testosterone supplementation on sexual function according to an updated meta-analysis of RCTs
Out of 357 retrieved articles, 18 were included in the study ([55e72]; Table 2). Available RCTs enrolled 1810 subjects in the active arm and 1762 in the placebo group respectively. The mean follow-up was 40.1 weeks and the mean age of the patients included was 60.2 years. The trials considered differed in baseline total T levels and TRT was administered in different doses and formulations (Table 2).
Data on the effect of TRT on erectile function, including IIEF-erectile function domain (EFD) or IIEF-5 score (erectile function, EF component), were available in 13 and 5 studies, respectively (Table 2). When the combined data for IIEF-EFD and IIEF-5 scores were considered I2 was 75.97 (p < 0.0001). The Beggadjusted rank correlation test (Kendall tau 0.36; p ¼ 0.053), based on TRT vs. placebo on EF component, suggested no major publication bias. Overall TRT provided a significant improvement of the EF component, as compared to placebo (Standardized mean ¼ 0.40 [0.24; 0.56]; p < 0.0001; see also Fig. 4). Meta-regression analysis showed that the effect of TRT was lower in those studies with a higher prevalence of diabetes mellitus (DM, Fig. 5, panel A). In addition, a trend toward a negative association with higher body mass index (BMI) was also observed (S ¼ 0.025 [0.055; 0.003]; p ¼ 0.08 and I ¼ 1.096 [0.175; 2.018]; p < 0.05). Conversely, no association with trial duration, mean age, mean baseline total T, and mean baseline IIEF score at baseline was observed (not shown).
The negative association between overall TRT-induced EF improvement and diabetes was confirmed in a multivariate regression model, after adjusting for age, trial duration, and mean T at enrolment (adj. r ¼ 0.093, p < 0.0001). Similar results were observed when BMI was introduced in the same regression model (not shown).
Data including I2 calculated when only trials using IIEF-EFD as the outcome was considered was 35.4 (p ¼ 0.104). The Begg-adjusted rank correlation test (Kendall tau 0.35; p < 0.11), considering only this sub-group, suggested no major publication bias. A sensitivity analysis restricted to the latter trials confirmed the positive effect of TRT in improving erectile function as compared to placebo (Fig. 6, and Supplementary Figure 1 Panel A). After excluding those trials enrolling subjects with mean baseline T levels above 12 nmol/L from the analysis [58,71], meta-regression analysis showed an inverse relationship between baseline T levels and final outcomes when TRT was compared to placebo (Fig. 5 Panel B). The latter was confirmed even when adjusted for age, trial duration and diabetes mellitus (adj.r ¼ - 0.243; p < 0.0001). The introduction in the same regression model of BMI did not modify the association (not shown).
*Other sexual function components
Data on the effect of TRT on IIEF-libido, intercourse satisfaction, and overall sexual satisfaction domains were available in 10 studies, including 2528 individuals, respectively. Overall, TRT resulted in an improvement of IIEF-desire, intercourse, and overall sexual satisfaction domains (Fig. 6, and Supplementary Figure 1, Panels B-D).
Data on the effect of TRT on the IIEF-orgasmic domain were available in 11 studies, including 2604 individuals. Even in this group, TRT determined a significant improvement over placebo (Fig. 6, and Supplementary Figure 1, panel E). The data were confirmed even when one study [66] that included a mixed population of subjects with premature and delayed ejaculation was excluded from the analysis (Fig. 5, and Supplementary Figure 1 Panel E).
Conclusions
T supplementation to symptomatic subjects with T deficiency is associated with a statistical improvement in several aspects of sexual life, including erection. However, the effect is modest and less apparent in subjects with established metabolic conditions, such as T2DM. In terms of amelioration of erectile function, the effect of TRT is similar to those obtained with lifestyle interventions, such as dieting and physical exercise, which, by the way, often improves the underlying metabolic disturbance. However, it should be important to recognize that lifestyle modifications should be the first step to promote weight reduction. TRT can be combined with lifestyle intervention only in symptomatic hypogonadal subjects especially in the presence of comorbid metabolic conditions. Accordingly, some preliminary results support better metabolic and body composition outcomes when lifestyle modifications are combined with TRT [85]. No data on erectile function are available. Considering that T supplementation not only improves sexual life but also body composition and muscle mass, it is conceivable that initial treatment with T will improve the willingness of hypogonadal subjects to perform physical exercise and to adhere to healthier behavior. Later on, TRT can be interrupted as soon the effect of the healthier lifestyle has orchestrated its effect. In fact, in an animal model of MetS, T administration to hypogonadal rabbits improved their ability to perform the physical exercise by distinct effects on muscle fiber diameter and composition [23] and by reducing hypothalamic metaflammation [86]. Hence, a combined treatment with lifestyle intervention and T supplementation should be suggested as an initial approach only in symptomatic hypogonadal subjects with metabolic conditions not wanting fatherhood in the near future. Further, studies are advisable to better evaluate the long-term effects, in particular on metabolic and sexual outcomes, off the latter approach.
Giovanni Corona, MD, Ph.D., Consultant Endocrinologist, Giulia Rastrelli, MD, Ph.D., Associate professor of Endocrinology, Linda Vignozzi, MD, Ph.D., Associate professor of Endocrinology, Mario Maggi, MD, Ph.D., Full professor of Endocrinology
Sexual symptoms are the most specific determinants of low testosterone (T) observed during adulthood. In this narrative review, we summarize the most important evidence supporting the positive relationships between endogenous T levels and sexual activity in the adult male, by using preclinical and clinical observations. In addition, we also report an update of our previous meta-analysis evaluating the effects of T treatment (TRT) on sexual functioning in subjects with T deficiency. Available data indicate that TRT of symptomatic hypogonadal men can improve several aspects of sexual life, including erection. However, the effect is rather modest and lower in subjects with associated metabolic conditions. The specific observed effects are similar to those derived from lifestyle intervention. Since TRT might result in body composition improvement, it is reasonable to suppose that initial treatment with T can improve the willingness of hypogonadal subjects to perform physical exercise and to adhere to healthier behavior. Similar data were derived from animal models. However, it should be important to recognize that lifestyle modifications should be the first step to promote weight reduction. TRT can be combined with lifestyle interventions only in symptomatic hypogonadal subjects especially in the presence of comorbid metabolic conditions.
Introduction
Testosterone (T) is a gender-specific sex hormone, being, during adulthood, at least 20-fold higher in men than in women. In the male, T is essentially produced by the testis, as the adrenal contribution to the circulating hormone levels is negligible. Testicular production of T is characterized by three specific surges: one during fetal life, another in the early neonatal life, and the last one during puberty remaining essentially stable thereafter [1-3]. Fetal production of T is driven by chorionic gonadotropin (hCG), while the other surges are driven by the pituitary luteinizing hormone (LH). Fetal T production plays an essential role in shaping the male internal and external genitalia and, most probably, in programming effects on later brain development, e.g. driving core gender identity and sexual orientation [4,5]. The neonatal surge of T happens during the first 3-6 months of life and it is often referred to as mini-puberty. Mini-puberty remains a mysterious phenomenon, but it is hypothesized that the T rise in the male newborn might affect several psycho-biological aspects, including body and penile growth, number of Leydig cells within the testis, and brain organization [1,4,5]. During puberty, T rise allows definitive development of secondary sexual characteristics, growth acceleration, and modifications in body composition, including an increase in bone and muscle mass, along with a decrease in fat mass [1]. In addition, several psychological modifications occurring during male puberty may be regulated by T, as, for example, aggressive behavior and spatial ability [6,7].
In contrast to the aforementioned, well-recognized, differentiating, and organizing activities of T - spanning from the early fetal life to the entire pubertal process - its role in adult male life is more related to a non-crucial supporting role in several biological activities. In particular, during adulthood, all the effects of T are related to maintaining and sustaining the psycho-biological modifications initiated during the pubertal transition including body composition, spermatogenesis, erythropoiesis, male attitudes and behaviors, bone health, and sexual functions. However, none of these T effects is essential for survival; in fact, the life expectancy of eunuchs is similar to that of the general population [8]. Low T during adulthood has been associated with physical issues (osteoporosis, reduced mobility, fatigue, anemia), psycho-neurological symptoms (depression, decreased energy, cognitive deficits), and sexual symptoms (decreased spontaneous and sex-related erection, low libido, reduced orgasmic pleasure, delayed ejaculation, decreased ejaculatory volume) [9e13]. Late-onset hypogonadism (LOH) is an umbrella term that syndromically associates, in the aging male, some or all of the aforementioned symptoms or signs of low T (below 12 nmol/L) [14,15]. In a seminal collection of recommendations [15] released by several Andrology Societies (i.e. International Society of Andrology, International Society for the Study of Aging Male, European Association of Urology, European Academy of Andrology, and American Society of Andrology), it was decided that low T is not enough to define a true hypogonadal state; in fact, also low T-related symptoms should be present. A few years later, a large epidemiological study (European Male Aging Study, EMAS) - enrolling 3369 aged men in eight European centers - indicated that, in both the training and validation sample, only three sexual symptoms (reduced sexual desire and spontaneous and sexual-related erections) - but not other physical or psychological symptoms - clustered with low total T (<11 nmol/L) and calculated free T (cfT ¼<220 pmol/L) [16]. Although statistically significant, all the associations with sexual symptoms were relatively weak, suggesting that the same sexual symptoms can be generated in overall eugonadal individuals by other medical conditions [16]. Later on, the EMAS biochemical thresholds were essentially validated in a setting of men consulting for sexual dysfunction [17]. In fact, the assessment of Youden index showed that the best thresholds for detecting men with androgen deficiency-related sexual symptoms were 10.4 nmol/L for total T and ranges from 225 to 260 pmol/L for cfT [17]. However, association does not mean causation. In fact, the same sexual symptoms, often viewed as a mirror of T deficiency, can also be explained by several morbidities afflicting the aging male [18]. The Chronic Disease Score (CDS) is a widely accepted measure of comorbidities, based on the aggregate number of prescribed medications for several chronic conditions [19]. In the aforementioned clinical setting of men complaining of sexual dysfunction, we reported that the weighted effect of CDS versus total T is 2-3 times higher in explaining a decreased basal and dynamic peak systolic velocity (DPSV) and acceleration at penile color Doppler ultrasound (PCDU). Even the relative risk of having an impaired erection, either spontaneous or sex-related, was more dependent on high CDS than on low total T [20]. In that study [20], the most genuine symptom of T deficiency was reduced sexual desire. Accordingly, in the longitudinal extension of the EMAS study, the development of secondary hypogonadism has the strongest correlation with decreased sexual desire, rather than other sexual symptoms [21].
*In this review, we will summarize the relationships between endogenous T levels and sexual activity in the adult male, by using preclinical and clinical observations. In addition, we also reviewed, using a meta-analytic method, the effect of T treatments in subjects with T deficiency, focusing on the effect on sexual symptoms.
*Regulation of penile erection by endogenous testosterone
Penile erection is a neuro-hemodynamic phenomenon causing engorgement of the two corpora cavernosa (CC) by arterial blood, finally leading to rigidity and erection for mechanical compression of the emissary venous vessels, thereby decreasing blood outflow [28]. Cell-to-cell interactions between the endothelial layer and the surrounding smooth muscle cells (SMC) are the main regulators of arterial inflow into the CC. Activation of penile erection is controlled by the major pelvic ganglia that contain, within the pelvic plexus, both parasympathetic and sympathetic neurons.
Sympathetic signals (i.e. noradrenaline), along with other substances (e.g. endothelin-1), maintain the SMC of the penile arteries and of the cavernous spaces in a contractile status by increasing intracellular calcium concentrations. Both noradrenaline and endothelin-1 also activate a calcium sensitizing pathway involving activation of the small GTPase protein RhoA and of its downstream kinase ROCK. The final effect is the activation of myosin light chain kinase (MLCK) and SMC contraction, which limits arterial inflow and causes a flaccid state [28].
During an erection, parasympathetic and non-adrenergic, non-cholinergic (NANC) nerve activity causes and sustains the release from nerve endings and from endothelial cells of nitric oxide (NO), through neuronal and endothelial NO synthase (nNOS and eNOS, respectively). NO easily diffuses into the SMC increasing its downstream effectors (cGMP, guanylate cyclase, PKG), causing calcium sequestration, and calcium efflux. This allows SMC to relax and blood flow to increase [28]. The crucial role of NO in penile erection has been demonstrated in pioneering studies with genetic manipulation of the NO-forming enzymes eNOS and nNOS [29e31]. Returning to the basal, contracted state is facilitated by cGMP catabolism through a series of phosphodiesterases (PDEs). In the penis, among different PDE, PDE5 is the most represented and biologically active in metabolizing cGMP [32]. Other neurotransmitters, such as vasoactive intestinal peptide (VIP), prostaglandins, and adenosine (ADO), facilitate SMC relaxation by activating adenylate cyclase and calcium mobilization. ADO is a purinergic neuromodulator that exerts a profound effect on penile erection acting through a series of receptors, with ADORA2B being crucial, because, in a mouse model, its genetic deletion is associated with loss of erection and its excess to priapism [33]. Considering the crucial role of the endothelial and SMC in the dynamic process of rigidity/detumescence, it is obvious that any alteration in their representation could profoundly alter penile physiology.
T exerts a crucial role in facilitating an increase in penile blood flow (Fig. 1). Fig. 2 shows the relationship between quintiles of endogenous T and measured maximal, prostaglandins E1 (PGE1)- induced, DPSV at PCDU in a large series (n ¼ 2751) of men consulting our outpatient clinic for sexual dysfunction [12,25,34]. In an age- and CDS-adjusted ANCOVA model, the lowest T quintile - which corresponds to the EMAS-defined hypogonadal state (i.e. low T and sexual symptoms) - is characterized by a significant decrease in maximal penile blood flow. In contrast, in the eugonadal range (total T > 10.4 nmol/L), further increases in penile blood flow were marginal or insignificant. This is tantamount to say that deficiency of T in the hypogonadal range, but not in the eugonadal one, is independently associated with an impaired penile blood flow.
However, studies in humans have not helped in discerning the mechanism of T action in penile erection. Preclinical studies in animal models could, therefore, be helpful in understanding the role of T in favoring penile erection. In the last 12 years, we set up a model of metabolic syndrome (MetS) by feeding rabbits a westernized, cholesterol-rich, diet (HFD [21]). Beside MetS, HFD induced hypogonadotropic hypogonadism and erectile dysfunction (ED) [22]. We, later on, demonstrated that training rabbits to perform endurance exercise reversed the hypogonadal condition and erectile dysfunction. By selecting data from these HFD rabbits performing or not physical exercise and from those fed a regular diet (RD) [22-24], we now report associations between gene expression of the AR and the aforementioned families of factors that drive penile erection/detumescence. Results are shown in Table 1. Expression of AR is closely associated with markers of smooth muscle function (SM22, aSMA, RhoA, ROCK1, ROCK2, MYH11) and with the entire NO signaling pathway, ranging from NO formation (DDHA1, eNOS, nNOS), its downstream action (GCSa1, GCSb1, PKG) to its degradation (PDE5). In addition, AR is associated with ADO action (ADORA1, ADORA2A, ADORA2B, ADORA3) and degradation (ADA, AMPD2). Similar results were previously published concerning rabbits treated with a gonadotropin-releasing hormone (GnRH) analog (triptorelin) and supplemented or not with T [35]. PDE5 expression in the penis is an important prerequisite for PDE5 inhibitors (PDE5i) activity [22,36]. Therefore, investigating its relationship with androgen signaling is an important point for understanding penile physiology. Fig. 3, panel A, shows gene expression of PDE5 in the penile extract as a function of increasing quintiles of AR expression. Data are derived from the aforementioned rabbit groups [22-24]. KruskaleWallis analysis indicates significant differences among groups (p < 0.0001). In particular, post-hoc ManneWhitney U test confirmed that there is a stepwise increase in the expression of PDE5 as a function of increasing quintiles of androgen receptor (AR) expression. However, when the expression of PDE5 in the penis was evaluated as a function of quintiles of circulating total T, it was found that only the lowest quintile of T was associated with a reduction of PDE5 expression (Fig. 3 panel B). This lowest quintile represents hypogonadal rabbits, with all total T below one standard deviation of the normal RD value (6.56 ± 4.62 nmol/L). For higher T quintiles there was no further increase in PDE5 expression.
*In conclusion, results from our preclinical and clinical studies, here summarized, suggest that penile erection is also an androgen-dependent phenomenon (see also Fig. 1). However, the relationship between endogenous circulating T levels and erection is evident only in the case of T deficiency and not in eugonadal conditions, suggesting a saturation effect, the molecular basis of which is still to be determined.
*Regulation of sexual interest by endogenous testosterone
Sexual desire is the subjective feeling of wanting to initiate a sexual experience. Hence, it is an important prerequisite for sexual excitation and arousal, which may include, in men, penile erection. Sexual desire is an essential component of the human sexual response, with a particular valence for species perpetuation and for couple enjoyment. Its reduction in men, i.e., male hypoactive sexual desire disorder (MHSDD), has been conceptualized by several editions of the Diagnostic and Statistical Manual of Mental Disorders, including the latest version (DSM-5 [37]), as a long-lasting (more than six months) persistent or recurrently deficient sexual or erotic thoughts, fantasies, and desire for sexual activity that causes marked distress or interpersonal difficulty, with the disease not being a result of medical illness, another psychological disorder, or the effects of a drug.
In a survey on 3714 men consulting for sexual dysfunction at the University of Florence, Florence, Italy, reduced libido was found in 36.4% of the sample, often associated with other sexual dysfunctions such as ED and reported premature or delayed ejaculation, as it was isolated only in 5% of the subjects [38]. Reduced libido was not associated with other risk factors in 33% of the sample (primary reduced libido), while it was often associated with other medical or psychological conditions (secondary reduced libido). In fact, among subjects with hyperprolactinemia or psychopathology, 84%, and 48%, respectively, reported reduced libido (see also Fig. 1). In subjects with low T (i.e. total T < 12 nmol/L), 40% of cases reported any reduction in sexual desire. The main characteristics of low T-associated reduced libido were a severe ED, a decrease in sexual activity, a perceived reduction in ejaculate volume and in sleep-related erection, along with a higher scoring in free-floating anxiety and depressive symptomatology [38]. As a result, reduced libido in subjects with low T could impair several aspects of a couple's sexual behavior, resulting in a source of deterioration of the couple's relationship (as measured by Structure Interview on Erectile Dysfunction Scale 2, i.e. marital domain [38])
An age-adjusted association between low T and reduced sexual desire was reported in some [16,39,40], but not all [41,42], epidemiological studies. In the EMAS study, a low frequency of sexual thoughts was present in 27.5% of the population and was associated with both cFT (p < 0.0001) and total T (p ¼ 0.048) [16]. In particular, a reduction of each 1 nmol/L T below 8 nmol/L was associated with a 48% (20-83%) increased risk of low desire. Hence, a threshold of 8 nmol/L was proposed for the association between low T and sexual desire. In the same study, a threshold for cFT was also found, corresponding to 220 pmol/L. In a previous study conducted in a German series of 434 consecutive male patients aged 50-86 years. Consulting for andrological problems - a higher threshold was proposed for explaining the association between low T and [43,44] decreased libido (i.e. 15 nmol/L [45]). In a more recent survey, conducted at the University of Florence, Italy, in 4890 men attending an outpatient clinic for sexual dysfunction, it was found that the best thresholds to verify the association between reduced T and severe sexual symptoms, including low libido, impaired morning erection, and ED, was total T < 10.4 nmol/L or cfT<225 pmol/L [17]. As mentioned before, similar results were also derived from the EMAS study, because the association between low T and syndromic clustering of sexual symptoms, including low libido, was evident for total T < 11 nmol/L and cfT<220 pmol/L) [16].
The human brain is endowed by AR. AR is distributed in discrete areas of the brain, all involved in the regulation of the male sexual response, including temporal, preoptic, hypothalamus, amygdala, midbrain, frontal and prefrontal cortex areas, and cingulate gyrus (see for review in [46]). Acute testosterone exposure increases activation of most of them, as detected by functional magnetic resonance imaging (fMRI, see for review in [47]). Similarly, Park et al., [48], using fMRI showed that, in healthy volunteers, erotic visual stimulation was able to activate the inferior frontal lobe, cingulate gyrus, insula gyrus, corpus callosum, thalamus, caudate nucleus, globus pallidus, and inferior temporal lobe. The latter results were blunted in hypogonadal men and restored by T administration [48]. In the pioneering study of Stoleru et al. [49] with positron emission tomography (PET) it was found that sexually explicit visual stimulation induced a T rise and activates several paralimbic areas as a function of circulating T levels. Hence, it is conceivable that the AR mediates T action on sexual desire (see also Fig. 1). However, the possibility that other potential targets of T within the brain, such as estrogen receptors, are involved in the T-action on the sexual drive cannot be ruled out [50].
Accordingly, a previous double-blind placebo-controlled RCT, performed by experimentally inducing hypogonadism with GnRH-analogue in 400 healthy men, showed that both estrogen and T deficiency contributes to the decline of sexual desire [50]. Similarly, a sexual desire improvement has been reported in men with aromatase deficiency after transdermal estradiol treatment in subjects affected by (a rare condition impairing estrogen formation), suggesting a possible role of estrogen in regulating male sexual desire [51]. Conversely, adrenal hormones including dehydroepiandrosterone (DHEA) and its sulfate (DHEAS), as well as cortisol and aldosterone are not involved in the regulation of male sexual desire [52].
*Regulation of male orgasm by endogenous testosterone
Ejaculation is a spinal reflex that can be classically divided into three main phases: emission, expulsion, and orgasm. During the first, reversible part of the ejaculation (emission), the semen is propelled throughout the male genital tract until the posterior portion of the urethra. Seminal vesicles secretion, and to a lesser extent prostatic secretion, contributes substantially to the ejaculatory bolus. Closure of the bladder neck prevents the retrograde spillover of the ejaculate bolus into the bladder. During ejaculation an irreversible forceful series of contractions of the muscle of the pelvic floor allows the ejection of ejaculatory bolus, containing the semen, through the urethral meatus. The latter phase is usually associated with orgasm, a complex experience that may occur as sexual pleasure reaches a climax [53]. However, in prepubertal children, orgasm can occur without ejaculation [53]. The orgasmic reaction is controlled by a complex network of sympathetic and parasympathetic nerves located in the thoracolumbar and sacral spinal cord, respectively, with a somatic component controlled by Onuf's nucleus, along with a lumbar spinal generator of ejaculation (SGE). The entire male genital tract, including prostate and seminal vesicles, development, and function, is under androgen control, as it is the neural control of the process (see for review [53,54]; Fig. 1). Hence, it is rather obvious that derangements in androgen action are associated with several ejaculatory dysfunctions, including reduced ejaculatory volume and premature and delayed ejaculation.
*Effects of testosterone supplementation on sexual function according to an updated meta-analysis of RCTs
Out of 357 retrieved articles, 18 were included in the study ([55e72]; Table 2). Available RCTs enrolled 1810 subjects in the active arm and 1762 in the placebo group respectively. The mean follow-up was 40.1 weeks and the mean age of the patients included was 60.2 years. The trials considered differed in baseline total T levels and TRT was administered in different doses and formulations (Table 2).
Data on the effect of TRT on erectile function, including IIEF-erectile function domain (EFD) or IIEF-5 score (erectile function, EF component), were available in 13 and 5 studies, respectively (Table 2). When the combined data for IIEF-EFD and IIEF-5 scores were considered I2 was 75.97 (p < 0.0001). The Beggadjusted rank correlation test (Kendall tau 0.36; p ¼ 0.053), based on TRT vs. placebo on EF component, suggested no major publication bias. Overall TRT provided a significant improvement of the EF component, as compared to placebo (Standardized mean ¼ 0.40 [0.24; 0.56]; p < 0.0001; see also Fig. 4). Meta-regression analysis showed that the effect of TRT was lower in those studies with a higher prevalence of diabetes mellitus (DM, Fig. 5, panel A). In addition, a trend toward a negative association with higher body mass index (BMI) was also observed (S ¼ 0.025 [0.055; 0.003]; p ¼ 0.08 and I ¼ 1.096 [0.175; 2.018]; p < 0.05). Conversely, no association with trial duration, mean age, mean baseline total T, and mean baseline IIEF score at baseline was observed (not shown).
The negative association between overall TRT-induced EF improvement and diabetes was confirmed in a multivariate regression model, after adjusting for age, trial duration, and mean T at enrolment (adj. r ¼ 0.093, p < 0.0001). Similar results were observed when BMI was introduced in the same regression model (not shown).
Data including I2 calculated when only trials using IIEF-EFD as the outcome was considered was 35.4 (p ¼ 0.104). The Begg-adjusted rank correlation test (Kendall tau 0.35; p < 0.11), considering only this sub-group, suggested no major publication bias. A sensitivity analysis restricted to the latter trials confirmed the positive effect of TRT in improving erectile function as compared to placebo (Fig. 6, and Supplementary Figure 1 Panel A). After excluding those trials enrolling subjects with mean baseline T levels above 12 nmol/L from the analysis [58,71], meta-regression analysis showed an inverse relationship between baseline T levels and final outcomes when TRT was compared to placebo (Fig. 5 Panel B). The latter was confirmed even when adjusted for age, trial duration and diabetes mellitus (adj.r ¼ - 0.243; p < 0.0001). The introduction in the same regression model of BMI did not modify the association (not shown).
*Other sexual function components
Data on the effect of TRT on IIEF-libido, intercourse satisfaction, and overall sexual satisfaction domains were available in 10 studies, including 2528 individuals, respectively. Overall, TRT resulted in an improvement of IIEF-desire, intercourse, and overall sexual satisfaction domains (Fig. 6, and Supplementary Figure 1, Panels B-D).
Data on the effect of TRT on the IIEF-orgasmic domain were available in 11 studies, including 2604 individuals. Even in this group, TRT determined a significant improvement over placebo (Fig. 6, and Supplementary Figure 1, panel E). The data were confirmed even when one study [66] that included a mixed population of subjects with premature and delayed ejaculation was excluded from the analysis (Fig. 5, and Supplementary Figure 1 Panel E).
Conclusions
T supplementation to symptomatic subjects with T deficiency is associated with a statistical improvement in several aspects of sexual life, including erection. However, the effect is modest and less apparent in subjects with established metabolic conditions, such as T2DM. In terms of amelioration of erectile function, the effect of TRT is similar to those obtained with lifestyle interventions, such as dieting and physical exercise, which, by the way, often improves the underlying metabolic disturbance. However, it should be important to recognize that lifestyle modifications should be the first step to promote weight reduction. TRT can be combined with lifestyle intervention only in symptomatic hypogonadal subjects especially in the presence of comorbid metabolic conditions. Accordingly, some preliminary results support better metabolic and body composition outcomes when lifestyle modifications are combined with TRT [85]. No data on erectile function are available. Considering that T supplementation not only improves sexual life but also body composition and muscle mass, it is conceivable that initial treatment with T will improve the willingness of hypogonadal subjects to perform physical exercise and to adhere to healthier behavior. Later on, TRT can be interrupted as soon the effect of the healthier lifestyle has orchestrated its effect. In fact, in an animal model of MetS, T administration to hypogonadal rabbits improved their ability to perform the physical exercise by distinct effects on muscle fiber diameter and composition [23] and by reducing hypothalamic metaflammation [86]. Hence, a combined treatment with lifestyle intervention and T supplementation should be suggested as an initial approach only in symptomatic hypogonadal subjects with metabolic conditions not wanting fatherhood in the near future. Further, studies are advisable to better evaluate the long-term effects, in particular on metabolic and sexual outcomes, off the latter approach.
