Polycystic Ovary Syndrome: A Model of Combinatorial Endocrinology?

Abstract

The clinical research study reported herein by Blank et al. (1) identified insulin as the factor that explained the variable sensitivity to progesterone-induced slowing of GnRH-LH drive previously described in Tanner stage 3–5 hyperandrogenic adolescent girls (2). Specifically, the current analysis revealed that resistance of LH pulsatility to progesterone-induced slowing was proportional to fasting insulin in that group. The authors posited that increased GnRH drive in adolescent girls is a form fruste or precursor of polycystic ovary syndrome (PCOS). This line of investigation buttresses the notion that understanding the neuromodulation of GnRH is critical to explicating the pathobiology of PCOS (3). By providing an explicit example of a critical concept in what could be termed combinatorial endocrinology, namely, that the action of a hormone depends on its context, the study also provides another lesson that is not specific to understanding PCOS. Using this latest insight into the brain phenotype of PCOS as an example, we can infer that the final phenotype of other complex syndromes that present with variation in key features likely represents the interaction of several relevant parameters. Context in endocrinology would include not only age and sex but also time of day and current and possibly past endocrine and metabolic status among other variables. The authors found that LH sensitivity to suppression by progesterone was not due to insulin resistance or androgen excess, a notion that has consumed the field for decades, but rather insulin resistance and androgen excess. That both insulin resistance and increased androgen action confer the full brain phenotype in PCOS may explain the group’s earlier results showing that dampening of androgen action with the anti-androgen flutamide restored GnRH-LH sensitivity to progesterone suppression in PCOS (4). An interaction between insulin resistance and androgen excess appears to explain the brain phenotype of PCOS better than either parameter alone. Argument by analogy would suggest that PCOS is unlikely to be the only endocrine syndrome attributable to a complex interaction of multiple parameters. Dichotomization is a fundamental analytic approach used to isolate causality and the foundation of scientific inquiry and design, but it has limits. Unless explicitly built into the study design, dichotomization into an either/or question does not permit a test of interaction. But physiology and pathophysiology reflect systems biology and combinatorial dynamics. Failure to account for interaction likely explains some of why cellular studies often do not a whole organism describe; conversely, understanding context-specific hormone action may well help to put into perspective the physiological relevance of cellular studies. For instance, ovarian steroids have been shown to alter potassium channel functioning on hypothalamic GnRH neurons and to modulate GnRH neuronal output via this mechanism (5, 6). Not surprisingly, K(ATP) channels also mediated responsiveness to glucose and metabolic inhibition (7). The cellular and rodent models help us to understand the mechanisms that might underlie the current pathophysiological results in the uniquely human condition of female adolescent hyperandrogenism and, vice versa, the human results provide a firm rationale for conducting rodent and cellular studies. It is but a small leap to now imagine that alterations of this type in GnRH cellular physiology would confer differential sensitivity to subsequent insulin exposure and that both the past and current hormone soup might determine the functional characteristics of the GnRH pulse generator. Because synergism exists between insulin and LH in theca cells and because that insight helped to explain the ovarian phenotype found in PCOS, it should come as no surprise that the neuroendocrine apparatus would operate similarly. Indeed, when attempting to discern whether stress-induced suppression of GnRH drive was due to psychogenic or energy challenge, we found profound synergism between psychogenic and metabolic challenge (8). Likewise, Gibbs et al. found that estradiol restored learning in aged female rats, but only in the presence of a cholinesterase inhibitor (9). Given these considerations, the astute study designer is advised to consider whether the investigative approach will permit elucidating “and” as well as “or.” In complex disease, adaptations, and aging, the whole really is greater than the sum of the parts. It is time to remember this dictum not only in the clinic when we are treating patients but also in the experimental setting