Growth hormone (GH) and insulin-like growth factor I (IGF-I) receptors are present on pituitary gonadotrophs and on testicular Leydig and Sertoli cells. Thus, the GH/IGF-I system may modulate the pituitary-gonadal axis in males. This is a randomized cross-over study. Eight healthy male volunteers (mean age 35, range 29–46 years) were treated with GH for 3 weeks (1st week 0.01, 2nd week 0.02, 3rd week 0.03 mg/day/kg) or a GH receptor antagonist (Pegvisomant) (1st week 10, last 2 weeks 15 mg/day), separated by 8 weeks of washout. Before and after the two treatment periods, concentrations of luteinizing hormone (LH), follicle-stimulating hormone, testosterone, oestradiol, sex hormone–binding globulin, inhibin B and Anti-Müllerian Hormone (AMH) were measured. During GH treatment, IGF-I increased [(median (IQR)] 166 (162–235) vs. 702 (572–875) μg/L, p < 0.001) together with oestradiol [(mean ± SD) 78 ± 23 vs. 111 ± 30 pm, p = 0.019], and the oestradiol/testosterone ratio (p = 0.003). By contrast, AMH (42 ± 14 vs. 32 ± 7 pm, p = 0.018), Inhibin B (211 (146–226) vs. 176 (129–204) ng/L, p = 0.059) and LH (3.8 ± 1.5 vs. 3.2 ± 1.2 U/L, p = 0.096) decreased. During pegvisomant treatment IGF-I (204 (160–290) vs. 106 (97–157) μg/L, p = 0.001) and oestradiol (86 ± 28 vs. 79 ± 25 pm, p = 0.060) decreased. No significant changes or trends in the other reproductive hormones occurred during the two treatment regimens. GH/IGF-I activity was positively associated with serum oestradiol, suggesting that GH/IGF-I stimulates aromatase activity in vivo. As a novel observation, we found that high GH activity was associated with reduced levels of the Sertoli cell marker AMH. Further studies are needed to evaluate possible effects of GH on Sertoli cell function and/or spermatogenesis.
Growth hormone (GH) and its primary downstream mediator insulin-like growth factor I (IGF-I) appear to be involved in different aspects of male reproductive function (Chandrashekar et al., 2004). GH and IGF-I receptors have been identified on pituitary gonadotrophs as well as on Leydig, Sertoli and germ cells, pointing to the possibility that GH action may regulate the pituitary-gonadal axis and be involved in gonadal steroidogenesis and spermatogenesis (Lobie et al., 1990; Harvey et al., 1993; Mertani et al., 1994; Chandrashekar et al., 2004). Clinical data are limited and conflicting (Ovesen et al., 1993; Juul et al., 1998), but there are data supporting that GH action may play a role in male reproductive function (Laron & Sarel, 1970; Laron & Klinger, 1998; Carani et al., 1999). Besides possible direct pituitary or testicular effects GH/IGF-I action may reduce concentrations of sex hormone–binding globulin (SHBG) with subsequent increased concentrations of bioavailable androgens (Giavoli et al., 2006), although controversy exists (Blok et al., 1997). Finally in vitro studies suggest that IGF-I action may increase the conversion of testosterone to oestradiol via induction of aromatase activity (Zhang et al., 2010).
In recent years, measurements of Anti-Müllerian hormone (AMH) have received increasing interest as a marker of Sertoli cell function and of intratesticular androgen activity (Aksglaede et al., 2010). AMH is predominantly produced by immature Sertoli cells and triggers in foetal life the involution of the foetal Müllerian ducts in boys (Grinspon & Rey, 2010). AMH is inhibited by testosterone. During puberty concentrations of AMH decrease dramatically, but throughout adult life serum AMH remains at measurable and stable levels (Aksglaede et al., 2010). However, the physiological function as well as factors regulating AMH expression in adulthood remains widely unexplored, including a possible influence of GH action.
In this study, we examined the possible effects on various circulating markers of male reproductive function of 3 weeks of GH excess and 3 weeks of GH inhibition in healthy male volunteers. Male reproductive function was evaluated by serum concentrations of gonadotropins [follicle-stimulating hormone (FSH), luteinizing hormone (LH)], Leydig cell markers (testosterone, oestradiol, SHBG) and Sertoli cell markers (inhibin B and AMH).
Subjects and methods
This was a randomized cross-over study (Andreassen et al., 2012). Participants were treated for 3 weeks, in random order, with increasing doses of subcutaneous (sc.) injections of GH (genotropin, mini-quick, Pfizer, 1st week 0.01 mg/kg/day, 2nd week 0.02 mg/kg/day, 3rd week 0.03 mg/kg/day), or with sc. injections of a GH receptor antagonist (pegvisomant, Pfizer, 1st week 10 mg/day, last 2 weeks 15 mg/day). During GH treatment, the aim was IGF-I levels within the acromegalic range. The dose was chosen based on experiences from previous experimental studies (Yuen et al., 2002; Ehrnborg et al., 2007; Thum et al., 2007). During GH receptor blockage, we aimed to mimic severe growth hormone deficiency. The 15 mg/day dose was chosen based on experiences with treatment of acromegalic patients with severe GH hypersecretion (Jorgensen et al., 2005). The two treatment regimens were separated by an 8-week wash-out period. Half of the participants had GH as the initial treatment whereas the other half started with Pegvisomant treatment. Before administration of GH /Pegvisomant, participants were thoroughly instructed by a specialized nurse. The study medication was administrated once daily at 11 o'clock pm.
During the two treatment periods, patients were seen once every week, approximately 8 h after injection of medicine. Furthermore, there was a follow-up visit 8 weeks after last injection of study medication. At each visit vital signs were measured and fasting blood samples collected for routine biochemical testing including plasma glucose. Before and after the 3 weeks treatment with GH or Pegvisomant, additional serum samples were drawn and stored at −80 °C for later biochemical analyses of GH-, IGF-I-, IGF-binding protein 3 (IGFBP-3), insulin and reproductive hormones. Insulin measurements were included in the analyses as insulin levels may be a confounding factor influenced by GH activity (Moller & Jorgensen, 2009) and influencing levels of SHBG (Vermeulen, 1996).
Inclusion criterion was age between 22 and 65 years. Exclusion criteria were chronic disease, body mass index (BMI) above 30 kg/m2, daily intake of prescribed medication, previous cancer disease, drug or alcohol abuse. Participants should have a normal physical examination at the screening visit, blood pressure below 140/90 mmHg and concentrations of routine biochemical parameters within normal range.
The original study population consisted of nine healthy male volunteers (Andreassen et al., 2012). One of the participants (49 years old) had signs of impaired testicular function with high baseline serum FSH (29 U/L) and low serum inhibin B (40 ng/L). Therefore, this individual was excluded from the study. The remaining eight individuals were all Caucasians, mean age 35, range 29–46 years, height 184 (175–190) cm, weight 85.8 ± 12.4 kg and BMI 26.2 (23.2–27.6) kg/m2. No one had a history with cryptorchidism, infertility or sexual problems. Three are fathers. A limitation to this study is that examination of external genitalia was not performed and no semen analysis was done. However, the regular physical examination, the reproductive hormone concentrations and the medical history strongly indicated an intact reproductive function in all participants.
Serum IGF-I and IGFBP-3 were determined by immunoassay (IMMULITE 2000, Siemens Healthcare Diagnostics, Los Angeles, CA, USA). Intra- and interassay coefficients of variation (CV) were <4 and <9% (Sorensen et al., 2010). Serum GH was measured by a commercial time-resolved immunoflourometric assay (TR-IFMA) (Delfia; Perkin Elmer Life Sciences, Turku, Finland), following validated modifications as previously described (Orskov et al., 2007). Normally, the assay is a one-step assay, in which serum and an europium-labelled GH detection antibody are incubated in microtiter wells pre-coated with another GH antibody. In contrast to the detection antibody, the coating antibody does not recognize Pegvisomant and accordingly, it is possible to measure GH in the presence of Pegvisomant when employing separate incubation steps for serum samples and the detection antibody. Intra- and interassay CVs are not affected by the modification, and are as stated by the manufacturer (Orskov et al., 2007).
Serum Pegvisomant was measured by an in-house radioimmunoassay (RIA) using diluted serum samples which had been depleted of endogenous GH following pre-incubation in the microtiter wells from the Delfia GH TR-IFMA. Intraassay CVs ranged from 3 to 8%, whereas the interassay CV was 6% (Orskov et al., 2007). Insulin was measured by a chemiluminescence immunoassay (Advia Centaur; Siemens Healthcare Diagnostics, Ballerup, Denmark). Intra- and interassay CVs were both 3%.
Serum concentrations of FSH, LH, SHBG, testosterone and oestradiol were measured by TR-IFMAs (Delfia; Perkin Elmer). Intra- and interassay CVs for both FSH and LH were 3 and 5%. Intra- and interassay CVs for both SHBG and testosterone were <8 and <5%. For oestradiol intra- and interassay CVs were <4 and 8%. Serum inhibin B was measured by a specific two-sided enzyme immunometric assay (Serotec, Kidlington, UK). In our hands, intra- and interassay CVs were 15% and 18%. Free testosterone was calculated from the testosterone and SHBG concentrations using the method by Vermeulen et al. (1999), with the assumption of an average serum albumin concentration of 43 g/L. Serum AMH was measured by a sensitive immunoassay [Immunotech; Beckman Coulter Ltd., Marseilles, France (A16507)] with a detection limit of 2 pm. In our hands, the intraassay CVs were <7.8% and 5.4% at 13 and 123 pm. The interassay CVs were <11.6 and 10.9%, at 19 and 99 pm. Methods for reproductive hormones have previously been validated in our laboratory (Andersson et al., 1997; Aksglaede et al., 2010).
Changes in the different hormones were evaluated by Student's paired t-test. For each variable, two analyses were conducted: pre-GH treatment vs. 3 weeks GH treatment and pre-Pegvisomant vs. 3 weeks Pegvisomant treatment. Data are expressed as mean ± SD for normal distributed variables and as median (IQR) for non-normally distributed variables.
The study was approved by the local committee of ethics in Copenhagen (protocol ID H-3-2012-013). The study was registered at www.clinicaltrials.gov protocol ID NCT00969644.
There was a significant increase in serum IGF-I and IGFBP-3 following 3 weeks of GH treatment and a decrease during GH receptor blockage (Fig. 1 and Table 1). Serum GH measured 8 h after injection increased in parallel to the escalating doses of injected GH, whereas Pegvisomant treatment was accompanied by a small increase in endogenous GH levels (Table 1). Serum concentrations of Pegvisomant after 3 weeks treatment were in the range that has been reported previously in sufficiently treated acromegalic patients with severe hypersecretion of GH (Jorgensen et al., 2005) (Table 1). As expected, GH treatment induced insulin resistance with a significant increase in concentrations of insulin (Table 1 and Fig. 1) and fasting glucose (4.9 ± 0.3 vs. 5.6 ± 0.3 mm, p = 0.036). Pegvisomant treatment did not influence glucose metabolism as judged by changes in insulin concentrations (Fig. 1 and Table 1) or glucose concentrations (p = 0.68). In the two study arms, we observed identical IGF-I baseline concentrations. Thus, there was no evidence of a carry-over effect between the two treatment regimens.
Table 1. Levels of GH, IGF-I, IGFBP-3, insulin, Pegvisomant (Peg) and reproductive hormones measured before and after 3 weeks of GH and Pegvisomant treatment. For all variables except serum Pegvisomant, the following comparisons were made: pre-GH vs. 3 weeks GH (post-GH); pre-Pegvisomant vs. 3 weeks Pegvisomant (post-Peg.)
p-value pre- vs. post-GH
p-value pre- vs. post-Peg
SDS, standard deviation score; Testo, testosterone; Est, oestradiol; Cal., calculated; GH, growth hormone; IGF-I, insulin-like growth factor I; IGFBP-3, IGF-binding protein 3.
−1.67 (−0.50 to –2.02)
19.1 ± 7.1
18.7 ± 6.3
21.0 ± 7.7
20.1 ± 7.7
78 ± 23
111 ± 30
87 ± 28
79 ± 25
0.0044 ± 0.0016
0.0064 ± 0.0022
0.0044 ± 0.0014
0.0042 ± 0.0012
Luteinizing hormone, U/L
3.8 ± 1.5
3.2 ± 1.2
4.3 ± 1.6
4.1 ± 1.5
Sex hormone–binding globulin, nm
35 ± 16
33 ± 12
35 ± 16
34 ± 16
Cal. free T, nm
Inhibin B, ng/L
Anti-Müllerian Hormone, pm
42 ± 14
32 ± 7
43 ± 17
40 ± 14
Follicle-stimulating hormone, U/L
Reproductive hormones, GH treatment
During GH treatment, oestradiol (p = 0.019) and the oestradiol/testosterone molar ratio (p = 0.003) increased without any changes in total serum testosterone (p = 0.70) (Table 1 and Fig. 2). The absolute increase in serum oestradiol was 32.8 (95% CI 7.1–58.4) pm corresponding to an average increase of 42%. The average increase in the oestradiol/testosterone ratio was 45%. Parallel to the increase in oestradiol, serum LH tended to decrease (p = 0.096) (Table 1). Concentrations of SHBG (p = 0.23) and the calculated free fraction of testosterone were unchanged (p = 0.60) (Table 1).
Serum AMH [p = 0.018, ΔAMH 10.0 (2.3–17.7) pm, average decrease 24%] and serum inhibin B [p = 0.059, Δinhibin B 18 (−1-37) ng/L] decreased. Concentrations of FSH were unchanged (p = 0.45) (Table 1 and Fig. 2).
Reproductive hormones, Pegvisomant treatment
Growth hormone receptor blockage was accompanied by a decrease in serum oestradiol [p = 0.060, Δoestradiol 7.5 (−0.4–15.4) pm, average decrease 9%]. The other variables reflecting reproductive function did not show any trends or significant changes (Table 1 and Fig. 2).
All variables returned to baseline levels with no significant changes between the screening visit and the follow-up visit. There were no serious side effects during neither of the two treatment regiments and all participants completed the study. During GH treatment, there was a significant increase in body weight (86.0 ± 12.7 vs. 88.0 ± 13.4 kg, p = 0.001) most likely reflecting water retention which is a well-known effect of GH excess. Three reported joint and muscle pain and four reported mild headache. During Pegvisomant treatment, three reported increased fatigue and two complained about increased hunger which might reflect a transient decrease in blood sugar induced by lack of GH action on glucose metabolism.
When the excluded individual with high concentrations of FSH and low inhibin B was included in the analyses, it only had minor influence on the changes in reproductive hormone concentrations and it did not change the main conclusions.
This study gives further support to a link between GH/IGF-I activity and male reproductive function. We found a strong association between GH/IGF-I activity and circulating concentrations of oestradiol in our controlled study on healthy men. Oestradiol levels increased substantially during 3 weeks of GH treatment and decreased during GH receptor blockage. The changes in oestradiol were observed despite unchanged concentrations of total testosterone and the calculated free fraction of testosterone. Thus, GH treatment was accompanied by a highly significant increase in the oestradiol/testosterone ratio, suggesting an altered aromatase activity. The parallel trend towards a decrease in LH supports increased oestrogen activity with negative feedback at the pituitary gland (Raven et al., 2006). Circulating oestradiol in adult men originates from the conversion of testosterone by aromatase in peripheral tissues (primarily adipose tissue) and directly from the testes where aromatase is highly expressed in Leydig cells (Haverfield et al., 2011). The testes contribute with approximately one fourth of total circulating oestradiol (Kelch et al., 1972). In vitro it has been shown that IGF-I action can induce aromatase activity (Zhang et al., 2010). Furthermore, one previous clinical study reported increased concentrations of oestradiol (with unchanged testosterone) after GH replacement therapy in accordance with our present observation (Juul et al., 1998).
As a possible clinical implication, high concentrations of endogenous IGF-I might contribute to increased oestradiol/testosterone ratio in puberty with subsequent development of pubertal gynaecomastia. It should be noted that in the majority of participants the IGF-I concentrations obtained after the 3 weeks of GH treatment did not exceed those observed in normal pubertal boys (Sorensen et al., 2010). In animal models (Ng et al., 1997) and after GH treatment in adults (Cohn et al., 1993), breast development has been reported, supporting that GH/IGF-I activity may be involved in the pathogenesis of gynaecomastia. Besides effects on peripheral tissues, it has become clear that the balance between androgens and oestrogens is fundamental for normal testicular development and fertility (Haverfield et al., 2011; Carreau et al., 2012). However, the role and regulation of oestrogens are still poorly understood (Haverfield et al., 2011). Based on our results, it seems reasonable to suggest that GH and IGF-I are involved in regulation of intratesticular oestrogen formation.
The increase in concentrations of insulin and IGF-I observed during GH treatment would be expected to decrease levels of SHBG (Vermeulen, 1996; Giavoli et al., 2006). On the other hand, the concomitant increase in oestradiol induced by GH/IGF-I activity would per se tend to increase SHBG concentrations (Pugeat et al., 1996). Thus, the opposing influence of insulin/IGF-I vs. oestradiol on SHBG formation might explain the unchanged levels of SHBG throughout the study.
As another novel result we found that high GH/IGF-I activity suppressed AMH concentrations. As AMH seems to be synthesized exclusively in gonadal tissue (Aksglaede et al., 2010), this result further supports that changes in circulating GH/IGF-I affect intratesticular processes. The mechanisms connecting GH activity with AMH formation are unknown, but it seems plausible to suggest that changes in intratesticular sex steroid syntheses could be involved. As one possibility intratesticular induction of aromatase with local oestrogen formation might reduce production of AMH. Mature human Sertoli cells express both classical oestrogen receptors (Cavaco et al., 2009) and G protein-coupled oestrogen receptors (Rago et al., 2011) leaving the possibility that not only androgens but also oestrogens may influence the synthesis of AMH. This viewpoint is supported by one in vivo animal study with grafting of testicular tissue to female chick embryos. Pre-treatment of the male donors with oestradiol counteracted regression of the Müllerian ducts else wise induced by the implanted testes. Based on their results, the authors suggested that oestradiol down-regulates AMH in males (Stoll et al., 1993). As another possibility reduction in AMH levels during GH treatment could also be caused by higher levels of intratesticular testosterone since both GH and IGF-I seems to potentate LH-induced Leydig cell steroidogenesis (Horikawa et al., 1989). However, we find this possibility less likely as the circulating concentrations of testosterone were unchanged. Furthermore, concentrations of inhibin B, another marker of Sertoli cell function and spermatogenesis would be expected to increase (Jensen et al., 1997; Jorgensen et al., 2010) in response to increased intratesticular testosterone and this was not observed. On the contrary, there was a decrease in serum inhibin B (p = 0.059). During conditions with changes in intratesticular testosterone, levels of AMH and inhibin B usually change in opposite directions. This has been observed when testosterone increases in puberty (Hero et al., 2012), and when testosterone is suppressed as a result of medical castration in prostate cancer treatment (Eldar-Geva et al., 2010). Finally, it cannot be excluded that systemic or locally produced IGF-I could have a direct effect on AMH and inhibin B as IGF-I receptors have been identified on Sertoli cells (Vannelli et al., 1988).
In conclusion, GH/IGF-I action seems to influence the balance between androgens and oestrogens, supporting that GH/IGF-I stimulates aromatase in vivo. High GH/IGF-I activity was accompanied by reduced AMH levels. The underlying mechanisms and clinical significance are unknown and this topic merits further investigations.
The study was supported by an unrestricted grant by Pfizer, Denmark. None of the authors have any conflict of interest to declare regarding this study.