The diabetic information superhighway has been trafficked lately by news of GlaxoSmithKline’s Avandia, and the recent FDA declaration that the once-promising drug may be more dangerous than previously thought, and may substantially increase the risk of heart problems for patients.
Concern about new drugs is hardly novel, but amidst all this recent news, I can’t help but consider the potentially unforeseen side effects of avant-garde therapeutic agents as I read about them.
Take 11β-HSD1 inhibitors, for example. These guys are pretty cool; a number of pharmaceutical companies, including Merck, Pfizer, Amgen, and Incyte, have formulations of these 11β-hydroxysteroid dehydrogenase type 1 inhibitors in phase I and II clinical trials, with the hope of adding a new, salable tool to the type 2 diabetic’s toolbox.
Here’s how it works: throughout the course of the day, the adrenal gland releases the hormone cortisol, either due to normal circadian (daily) and ultradian (hourly) rhythms, or in reaction to stress. Cortisol is an important stress hormone, but many of its effects, while good in a fight-or-flight scenario, are bad over time in terms of diabetes and obesity: cortisol increases glucose production and release in the liver; increases appetite in anticipation of any necessary reaction to the stress-causing situation; and alters metabolic processing such that fat is preferentially stored in the central abdominal region. So, for diabetics, this means stored glucose directly increases blood sugar, new calories are ingested and increase both blood sugar and weight, and, to top it all off, the new weight tends to be the “bad,” visceral, disease-causing weight.
There are a couple of therapeutically important pit-stops for cortisol throughout this process, though; the active hormone cortisol can be converted in the kidney, colon, and some other tissues into the inactive molecule cortisone. The enzyme 11β-HSD2 aids this conversion, and allows cortisol to be stored out of the bloodstream.
The other side of the equation is controlled by 11β-HSD1, which converts the stored cortisone back into cortisol in the liver, fat tissue, the brain, and other tissues. This conversion and release happens in time with the adrenal gland and natural rhythms, and therefore contributes to the disadvantageous effects of circulating cortisol.
11β-HSD1 inhibitors, then, aim to disrupt this pathway; restrict the ability of 11β-HSD1 to convert cortisone into cortisol and you decrease the amount of circulating cortisol, thereby reducing the amount of glucose released in the liver and slowing the increase of appetite and unfortunate storage of visceral fat.
Or so the theory goes. In light of the recent Avandia news, I give not one but two ears to those who are inclined to raise concerns, as Erika Harno and Anne White do in their paper in Trends in Endocrinology and Metabolism, “Will treating diabetes with 11β-HSD1 inhibitors affect the HPA-axis?” Their particular concern: if you artificially reduce the level of circulating cortisol by means of the 11β-HSD1 gateway, will the adrenal gland just overcompensate, activating the hypothalamic-pituitary-adrenal (HPA) axis, and producing extra cortisol for release, thereby defeating the whole point of the inhibitors? Plus, if the HPA-axis all starts working overtime to compensate for the lower detected levels of cortisol, cortisol wouldn’t be the only hormone overproduced. THe adrenal gland might also simultaneously produce excess adrenal androgens, which, at high levels, have been known to cause osteoporosis, hypertension, poly cystic ovarian syndrome, hirsutism (think female beards!), and a number of other nasty-sounding complications.
In their review, Harno and White analyze a number of studies that have been done in mice and humans, hoping to determine the likelihood that 11β-HSD1 inhibitors would cause such adverse complications. They first thing they found was that measuring and adequately evaluating cortisol levels is extremely difficult, given the natural hourly fluctuation, the fact that the stress on subjects during observation often changes the patterning of cortisol release, and the deficiency in rodents of the adrenal androgens that are also of interest.
Despite these difficulties, Harno and White were able to gain some insight from studies of mice which were engineered to lack the enzyme 11β-HSD1 entirely. THe complete deletion of 11β-HSD1 in mice had, overall, inconclusively minimal effects on adrenal action. Mice of a certain genetic background saw a 70% increase in the weight of the adrenal gland, implying increased adrenal activity, but mice of another genetic background saw only a 20% increase, which, given the 46% change in adrenal weight throughout the course of a normal mouse day, is not too jarring.
Looking at the amount of cortisol secreted in the 11β-HSD1 knockout mice, researchers found again inconclusiveness; mice from the first genetic type had a higher base level of cortisol, but peak levels remained mostly unchanged. These knockout mice also took longer to return to base levels of cortisol after stress. However, the 11β-HSD1 knockout mice from the other genetic typing remained largely unchanged, retaining normal levels and rates of cortisol secretion.
Points are added in to the “Pro” column of 11β-HSD1 inhibitors by studies of the inhibitor compounds themselves. In a Merck study of the new drug’s effects on mice and rats, circulating levels of the precursor and co-marker of cortisol were not increased, implying the hypothalamus was not inducing overaction of the adrenal gland. Similarly, the human trials done thus far by drug companies indicate that the hormonal precursors to adrenal cortisol production are not upregulated, and the HPA axis remains unperturbed.
In the end, then, Harno and White conclude that there is a pretty good chance that 11β-HSD1 inhibitors will not increase the release of cortisol from thim adrenal gland. In other words, 11β-HSD1 inhibitors are probably good additions to the diabetic toolbox.
But, as we saw with Avandia and as we see with all the “likely”s and “maybe”s that come with the manipulation of molecular pathways in the body, this is a risky endeavor. Harno and White propose some potential ways to reduce the risk of upregulating the adrenal gland– using tissue-specific 11β-HSD1 inhibitors, for example, to focus on tissues that are shown to have less of an effect on the feedback loop up to the adrenal gland; or reducing cortisol levels only during certain times of the day, at advantageous places within the natural circadian rhythm– but they also point out that more study of the cortisol pathway is needed, especially in long-term clinical trials on human patients. It seems clear to me, too, that variation in patient responses should be analyzed in terms of genetic profiles, given the differing responses in genetically disparate mouse models.
And all this is possible. And all this is quite potentially very good. But we should not forget that we are pioneers in a New World, full of both promise (think John Winthrop’s City on a Hill) and danger (think Roanoke).