Elevated 11-oxygenated androgens are not a major contributor to HPG-axis disturbances in adults with congenital adrenal hyperplasia due to 21-hydroxylase deficiency

1. Introduction

Congenital adrenal hyperplasia (CAH) encompasses a group of hereditary autosomal recessive disorders that are characterized by mutations in key enzymes of cortisol synthesis. Among these, patients affected by mutations in the 21-alpha hydroxylase gene are the largest entity (1). Depending on the severity of the underlying mutation, patients are either suffering from isolated glucocorticoid-(simply virilizing form= SV) or additional mineralocorticoid-deficiency, when less than 1-2% of enzymatic activity is retained (salt-wasting = SW). Glucocorticoid deficiency in turn results in ACTH overstimulation due to the loss of hypothalamo-pituitary negative feedback which subsequently drives excessive adrenal androgen production (2).

Hypothalamus-pituitary-gonadal (HPG)-axis disturbances are a common phenomenon in CAH patients of both sexes. In particular in women in the fertile age, menstrual disturbances occur in up to 30% (3-5). The most obvious mechanism for this finding is the inherent extra-gonadal androgen and progesterone production in CAH. Correspondingly, menstrual cycle disturbances in female patients with CAH have been associated with poor treatment control as indicated by increased 17-OHP, testosterone, progesterone and androstenedione levels (6). In line, a recent study from the DSD-life cohort demonstrated that an androstenedione/testosterone ratio ≥1, indicating excessive adrenal testosterone production, is a strong predictor of sub-normal testosterone levels in men with CAH (7). Most of these patients correspondingly present with low or inadequately low gonadotropin levels indicating hypogonadotropic hypogonadism, while some patients also show signs of primary testicular damage mainly due to testicular adrenal rest tumor (TART) formation (7-9)

However, HPG-axis disturbances can also be found in patients who are regarded well controlled based on the aforementioned treatment markers (4,10). The exact culprit for HPG disturbances is therefore not always clear, as all major steroids, namely testosterone, estradiol and progesterone are important parts of the HPG feedback loop (11).

In recent years, a further class of adrenal-derived androgens has come into focus in the context of CAH. These are 11-oxygenated 19-carbon androgens, in particular 11-ketotestosterone (11KT) and 11β-hydroxyandrostenedione (11OHA4). Androstenedione can be converted to 11OHA4 by 11β-hydroxylase, then to 11-ketoandrostenedione through the action of 11β-hydroxysteroid dehydrogenase type 2 and finally to 11KT (12,13). It has been demonstrated that 11KT and its derivate 11KT-dihydrotestosterone act at the androgen receptor with equal potency to the classic testosterone and dihydrotestosterone (DHT) (14). Correspondingly, it has been shown that these steroids are significantly elevated in patients with classic CAH (15). While it has been suggested that elevated 11KT levels might contribute to HPG-suppression at the hypothalamic and pituitary level, their clinical relevance has still to be determined (16). Although testosterone exerts direct negative effects at the hypothalamic levels via AR binding (17,18), estradiol seems to be the dominant suppressor in this context (19). As 11KT cannot be directly aromatized questions a dominant role in HPG suppression in this context (14).

In the present study, we were therefore interested if 11-oxygenated androgens, in particular 11-KT would add information on HPG-axis disturbances in adult patients with CAH due to 21-hydroxylase deficiency in addition to established markers of disease control.

In detail, we hypothesized that 11KT would be higher in men with hypogonadotropic hypogonadism than in those with eugonadism, that 11OHA4 would be higher in males with TART than in those without and that 11KT and 11OHA4 would be higher in women with menstrual irregularities in contrast to those with eumenorrhoea. In an exploratory analysis we were further interested if 11KT and 11OHA4 would differ between women with a regular cycle and those on oral contraceptive therapy to estimate if these potential covariates should be accounted for in the clinical context if our main hypotheses would hold true.

2. Material and Methods

3. Results

Hypogonadism in men

Androstenedione levels were significantly higher in men with secondary hypogonadism than in those with eugonadism (4.5x; p = 0.023). A similar pattern could be observed with 17-OHP (14.4x; p = 0.007), 11KT (3.5x; p = 0.020), 11OHA4 (5.7x, p = 0.009) and ACTH (9.3x; p = 0.004) (Figure 1).

• Download figure
• Open in new tab

Figure 1. Differences between men with hypogonadism and eugonadism

Androstenedione levels were significantly higher in men with secondary hypogonadism than in those with eugonadism (4.5x; p = 0.023). Same was true for 17-OHP (14.4x; p = 0.007), 11KT (3.5x; p = 0.020), 11OHA4 (5.7x, p = 0.009) and ACTH (9.3x; p = 0.004). (Boxes represent the first to the third quartile, vertical line indicates medians, crosses indicate outliers).

Amenorrhea in women

ACTH (4.3x; p = 0.003), androstenedione (2.5x; p = 0.019), 17-OHP (8.1x; p = 0.002), 11KT (5.2x; p = 0.002), 11OHA4 (3.7x; p = 0.023) and testosterone (3.1x; p = 0.007) were significantly higher in women with amenorrhea versus regular cycling women (Figure 2).

• Download figure
• Open in new tab

Figure 2. Differences between regular cycling and amenorrhoeic women

ACTH (4.3x; p = 0.003), androstenedione (2.5x; p = 0.019), 17-OHP (8.1x; p = 0.002), 11KT (5.2x; p = 0.002), 11OHA4 (3.7x; p = 0.023) and testosterone (3.1x; p = 0.007) were significantly higher in women with amenorrhea in comparison to those with a regular cycle. (Boxes represent the first to the third quartile, vertical line indicates medians, crosses indicate outliers).

Equivalence dosage of HC was lower (12.7mg/m² vs 16.4mg/m^2; p = 0.019). There was no difference in terms of GC vs HC use or HOMA-IR. There was also no difference in DHEAS levels and the A/T-ratio. There was no significant difference in these parameters between those with irregular and those with a regular cycle (data not shown).

Oral contraceptive use

Androstenedione (p = 0.011), the FAI (p = 0.001) and DHEAS (p = 0.049) were significantly lower in women with a regular menstrual cycle in comparison to women currently taking an oral contraceptive pill, while there was no difference in total testosterone, 17-OHP, 11KT or 11OHA4 (Figure 3).

• Download figure
• Open in new tab

Figure 3. Oral contraception vs regular cycle

When comparing those with a regular menstrual cycle with those currently taking an oral contraceptive pill, androstenedione (p = 0.011), the FAI (p = 0.001) and DHEAS (p = 0.049) were significantly lower, while there was no difference in total testosterone, 17-OHP, 11KT or 11OHA4.). (Boxes represent the first to the third quartile, vertical line indicates medians, crosses indicate outliers).

4. Discussion

We could demonstrate in this relatively large cohort that 11-oxygenated androgens are significantly elevated in patients of both sexes with HPG-axis disturbances. They however do not seem to be the most relevant androgens in this context.

The prevalence of hypogonadism in men and menstrual irregularities in women in our cohort was comparable to that reported in the literature (7,24) and hence emphasizes that HPG-axis disturbances are still a fairly common issue in this patient population. Despite a significant elevation of 11KT and 11OHA4, being in accordance with earlier studies (16), 17-OHP was a better predictor of hypogonadism in male patients. Our data therefore do not indicate a leading role of 11KT in HPG-axis suppression. This may be due to the fact that 11KT cannot be converted into estrogenic steroids and many effects that had originally been attributed to testosterone in men are actually mediated via aromatization to estradiol (25). This also includes sex steroid regulation at the hypothalamus/pituitary level (19).

The significant role of 17OHP may be explained by the fact that progesterone is also a very potent suppressor of the HPG-axis in men (26,27). As 17-OHP is a direct metabolite of progesterone, it is a potent agonist of the PR (28) as also indicated by a strong positive correlation between both steroids in patients with CAH (29).

In contrast to previous work (16) and our initial hypothesis, neither 11KT nor 11OHA4 were elevated in men with TART and most steroid measures, including 11KT and 11OHA4 showed a negative instead of a positive correlation with FSH as a potential marker of gonadal damage. Although our cohort was age-wise comparable to those adult patients reported by Turcu et al., a considerable proportion of their patients with TART were children. While TARTs in younger patients might still respond to HPA-axis suppression, they seem to become less modifiable in the long-term due to fibrotic modifications (30). In line, presence of TART has been associated with poor treatment quality in some (8) but not all studies (9,31). Despite the factor of age, we cannot exclude that there were further differences in non-recorded potential confounders such as long-term disease control distinguishing our patient groups.

11KT and 11OHA4 levels were also higher in women with secondary amenorrhea than in those with a regular cycle, as having been reported before (16). But again, when adding other markers of disease control to our model, testosterone was superior in explaining amenorrhea. Correspondingly, menstrual cycle disturbances in female patients with CAH have been associated with poor androgen control (6). Testosterone modulates GnRH-pulse frequency and therefore interferes with the complex hormonal interplay during the menstrual cycle (32). In addition, it also has direct effects on the endometrium (33).

Of note, there were no significant differences in 11KT or 11OHA4 between those with a regular cycle and those currently taking hormonal contraceptives. This may be due to a lower binding affinity for SHBG that has been suggested but not yet demonstrated for these steroids. This could be of clinical relevance, as this might explain why some patients benefit less from oral contraceptive treatment with regard to hyperandrogenic symptoms such as hirsutism and acne (34). However, prospective studies are needed to substantiate this finding.

There are several limitations of our study. Firstly, due to the cross-sectional design we can only show associations and no causality between dependent and independent variables. Prospective studies with modifying interventions are therefore needed. A further limitation in this context is that, in contrast to the general recommendation, the definition on hypogonadism in our study was based on a single, instead of two testosterone measures in the morning (21). However, diurnal variance of testosterone and LH in CAH do generally differ from the general male population due to the extra-gonadal androgen production and also depend on GC pharmacokinetics (1). A further limitation of our study is that we only used a single steroid measurement in the morning for our analyses. Steroid measurements from blood drawn before the intake of the morning GC dosage might therefore yield other results. However, all patients included were either receiving a morning dosage of HC or prednisolone or an evening dosage of dexamethasone which show, given the wide inter-individual variability, a comparable suppression profile for the chosen time window (35,36).

Lastly, although we in investigated a large group of adult patients with classic CAH, given the rarity of the disease, subgroups were rather small, stepwise regression had to be used for our final models. Nonetheless, given the strong differences in steroid markers between groups we consider our results to be valid.

5. Conclusion

In the present study, 11KT and 11OHA4 did not seem to add additional information to the understanding of HPG-axis disturbances in CAH. Further studies have to show if 11oxC19 steroids are of relevance for explaining for other health issues in CAH. It might be of particular interest, if they could be a superior biomarker of disease control. To our knowledge it has not yet been shown to what extent 11-oxygenated androgens show a diurnal variation, potentially addressing the still problematic issue of timing of blood sampling with regard to different GC treatment regimens. Former approaches to overcome these restrictions like serial salivary (36) or blood sampling (37), or hair steroid measures (52) are on the one hand very time consuming and on the other hand have so far not yet been proven to be superior in terms of predicting clinically relevant long-term outcomes. Lastly, it would have to be demonstrated that 11KT can be independently targeted by treatment-modification, independent of other adrenal androgens.

Funding

This work was supported by the IFCAH grant 2013 to NR and by the Deutsche Forschungsgemeinschaft (Heisenberg Professorship 325768017 to NR and Projektnummer: 314061271-TRR205 to NR). The authors declare no competing interests.