Insulin Resistance and Adrenal Androgen Synthesis Viewed Through a Monogenic Lens

Despite its organ-centric name, the elevated circulating androgen concentrations that drive manifestations of polycystic ovary syndrome (PCOS) reflect cooperation among different organs. These includes the adrenal glands, which produce mostly weak androgens that are activated elsewhere, especially in ovaries, liver, kidneys, and adipose tissue (1). Pharmacological suppression of ovarian and/or adrenal androgen production suggests that serum testosterone in health and PCOS depends on ovaries and adrenal glands about equally (2).

Beyond classical androgens, adrenal production of 11-oxygenated C19 steroids, now studied using liquid chromatography and mass spectroscopy (LC-MS/MS), has also attracted interest (1). Such steroids include 11-hydroxytestosterone, which is converted by 11β-hydroxysteroid dehydrogenase type 2 (11-βHSD2), mostly in the kidneys, to 11-ketotestosterone (11-KT), and 11-hydroxyandrostenedione (11-OHA4). Significant amounts of 11-KT are also generated from 11-OHA4, but this requires 2 steps. First, action of renal 11-βHSD2 generates 11-ketoandrostenedione (11KA4), and this is reduced to 11-KT by AKR1C3, notably in adipose tissue (1, 3). Importantly, the androgenic potency and efficacy of 11-KT are similar to those of testosterone (1). Serum concentrations of adrenal 11-oxygenated androgens are elevated in “common” PCOS and correlate with indices of insulin resistance (IR), leading to the suggestion that they play an underappreciated role in PCOS, and potentially mediate the commonly associated IR (4). This is supported by the decreased insulin sensitivity of untreated nonclassical congenital adrenal hyperplasia and of healthy women on androgen exposure (refs in (5)).

Conversely, primary severe IR (SIR) is commonly associated with increased serum testosterone concentration, which may be extreme (5). This can be normalized by gonadotropin-releasing hormone (GnRH) agonist-induced ovarian suppression, or it resolves upon disappearance of anti–insulin receptor (INSR) antibodies in acquired cases (5). This proves that SIR can cause severe hyperandrogenism and suggests that the causal relationship between hyperandrogenemia and IR is bidirectional. While GnRH agonists produce large improvements in SIR-related clinical hyperandrogenism, the degree of any residual clinical hyperandrogenism is not well documented, and reports of adrenal androgen profiles in primary SIR have been cursory.

A recent study in this Journal from Walzer and colleagues seeks to address gaps in knowledge, using LC-MS/MS to quantify adrenal androgens in 18 women with primary SIR and 23 healthy controls (6). SIR was attributable to acquired or congenital defects in INSR action (7 participants; “SIR-INSR”), or to lipodystrophy (11 participants; “SIR-LD”). The effect of GnRH agonist on 2 SIR-INSR participants was also assessed. As expected, testosterone and androstenedione were elevated in both SIR subgroups. The novel finding was that 11-oxygenated androgens were also elevated in SIR-LD but not in SIR-INSR. The 11-oxygenated androgens comprised 60% of androgens in controls, including around 10% 11-KT, compared with 10% testosterone. In SIR-LD, 38% of circulating androgens were 11-oxygenated, with proportions of 11-KT and testosterone again similar (10%-12%). In the SIR-INSR group, in contrast, 11-oxygenated androgens comprised only 11% of circulating androgens, due to a marked increase in testosterone (54%) without a significant absolute increase in 11-KT (4%). Classic pathway androgens, but not 11-oxygenated androgens, were sharply decreased on GnRH agonist treatment in 2 SIR-INSR women. The androstenedione to 11-OHA4 ratio, used as an index of CYP11B1 activity, was similar in LD-SIR and controls, but markedly decreased in INSR-SIR. CYP11B1 is responsible for adrenal 11β-hydroxylation of C19 steroids. The ratio was unchanged by GnRH agonist in the 2 SIR-INSR participants treated.

Using monogenic disease to interrogate causal relationships between associated phenomena is a powerful approach. Large effect sizes often permit conclusions to be drawn even from small numbers of participants; nevertheless, while results of this study are strongly suggestive, the small numbers of participants still should be noted, especially given within- and between-group heterogeneity. For example, the degree of hyperinsulinemia was higher in the SIR-INSR than the SIR-LD group, while the latter had lower serum leptin concentrations. Moreover, the exclusively Black ethnicity of the SIR-INSR group contrasted with the predominantly White composition of the SIR-LD (55%) and control groups (83%).

Assuming the finding of increased adrenal 11-oxygenated androgen production in SIR-LD but not SIR-INSR is robust, then what is its significance? First, it suggests that in SIR-LD, but not SIR-INSR, adrenal 11-oxygenated androgen production may contribute to clinical hyperandrogenism. If so, then GnRH agonists may be more effective in SIR-INSR than in SIR-LD. This deserves study. Second, there are broader implications for understanding adrenal insulin action. The discordance in 11-oxygenated steroids between SIR-LD and SIR-INSR is reminiscent of the lack of fatty liver or dyslipidemia in SIR-INSR, unlike SIR-LD and common IR (7). Importantly, systemic IR may not mean commensurate IR in all insulin-responsive tissues. Plasma insulin concentration is set by the insulin sensitivity of the tissues responsible for glucose disposal and release, and any tissue that retains greater insulin sensitivity than those will “see” greater insulin action in IR, due to compensatory hyperinsulinemia. This is less likely for INSR defects, which universally impair INSR-mediated insulin action.

The new findings argue that the adrenal should be added to the list of “bystander” tissues exposed to increased insulin action in SIR-LD and common IR, with adverse consequences (8). This is also true of the ovary, with the key difference that ovarian consequences of SIR are seen also in SIR-INSR (5). In this case, extremely elevated insulin is plausibly hypothesized to act via the IGF1 receptor. This study counterbalances evidence that adrenal hyperandrogenism causes IR, demonstrating that primary IR can drive adrenal as well as ovarian androgen production. While the adrenal hyperandrogenemia-IR relationship may well be bidirectional, however, current evidence suggests that the effect of IR on hyperandrogenemia is greater than the converse.

The lack of a zona reticularis in the adrenals of most experimental mouse strains, and the failure of mice to recapitulate the severe hyperandrogenemia of human SIR, mean that human studies are key to dissecting the complex interplay among metabolism and reproductive function. A full account of the influence of insulin action on adrenal/gonadal maturation will ultimately also need to take into account leptin concentrations, and insulin-driven amplification of gonadal gonadotrophin action. More detailed androgen profiling in primary SIR, including both insulin receptoropathy and generalized lipodystrophy, is now much to be desired, ideally allied to systematic documentation of the timing of pubarche, adrenarche, and puberty. All of these have been reported to be premature in some cases of SIR. Such studies may yield insights of both clinical and scientific value both about adrenal actions of insulin, and about the role of adipose tissue in setting “androgen tone”. The opportunity is clear, and the gauntlet down.

Abbreviations

11-βHSD2

11β-hydroxysteroid dehydrogenase type 2

11-KT

11-ketotestosterone

11-OHA4

11-hydroxyandrostenedione

GnRH

gonadotropin-releasing hormone

INSR

insulin receptor

IR

insulin resistance

LC-MS/MS

liquid chromatography–tandem mass spectrometry

PCOS

polycystic ovary syndrome

SIR

severe insulin resistance

Funding

Wellcome Trust, grant 210752/Z/18/Z

Disclosure Summary

The author has no conflicts to declare

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.